The F-35 Lightning II:

From Concept to
Cockpit

Edited by
Jeffrey W. Hamstra
Lockheed Martin Corporation

P R O G R E S S I N A S T R O N A U T I C S A N D A E R O N A U T I C S

Timothy C. Lieuwen, Editor-in-Chief

Volume 257
The F-35 Lightning II: From
Concept to Cockpit
The F-35 Lightning II: From
Concept to Cockpit

Edited by

Jeffrey W. Hamstra, Lockheed Martin Corporation

Volume 257
Progress in Astronautics and Aeronautics

Timothy C. Lieuwen, Editor-in-Chief

Georgia Institute of Technology
Atlanta, Georgia

Published by
American Institute of Aeronautics and Astronautics, Inc.
12700 Sunrise Valley Drive, Suite 200, Reston, VA 20191-5807
American Institute of Aeronautics and Astronautics, Inc., Reston, VA.

Copyright # 2019 by Lockheed Martin Corporation. Published by the American Institute of

Aeronautics and Astronautics, Inc., with permission.
ISBN 978-1-62410-566-1
PROGRESS IN ASTRONAUTICS AND AERONAUTICS

EDITOR-IN-CHIEF

Timothy C. Lieuwen
Georgia Institute of Technology

EDITORIAL BOARD

Sam Adhikari Christopher H. M. Jenkins

Sysoft Corporation Montana State University

Paul M. Bevilaqua Eswar Josyula

Lockheed Martin (Ret.) U.S. Air Force Research Laboratory
Steven A. Brandt
Mark J. Lewis
U.S. Air Force Academy
Institute for Defense Analyses

José A. Camberos Dimitri N. Mavris

U.S. Air Force Research Laboratory
Georgia Institute of Technology

Richard Christiansen Alexander J. Smits

Sierra Lobo, Inc. Princeton University

Richard Curran Karen Thole

Delft University of Technology The Pennsylvania State University

Simon Hook Oleg A. Yakimenko

Jet Propulsion Laboratory U.S. Naval Postgraduate School
TABLE OF CONTENTS

Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvii

Chapter 1 F-35 Program History: From JAST to IOC . . . . . . . . . . . . . 1

Arthur E. Sheridan, Lockheed Martin Aeronautics Company, Fort Worth, TX;
Robert Burnes, F-35 Lightning II Joint Program Office, Arlington, VA
I. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
II. Joint Advanced Strike Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
III. Concept Demonstration Phase: Joint Strike Fighter . . . . . . . . . . . . . . 7
IV. System Development and Demonstration . . . . . . . . . . . . . . . . . . . . 15
V. International Participation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
VI. Transition to Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
VII. Operations: Road to Initial Operational Capabilities . . . . . . . . . . . . . 65
VIII. Upcoming Plans: Future Developments . . . . . . . . . . . . . . . . . . . . . . 70
IX. Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
Appendix: Acronym List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

Chapter 2 F-35 Air Vehicle Configuration Development . . . . . . . . . 77

Mark A. Counts, Brian A. Kiger, John E. Hoffschwelle and Adam M. Houtman, Lockheed
Martin Aeronautics Company, Fort Worth, TX; Greg Henderson, Lockheed Martin
Aeronautics Company (retired), Beulah, WY
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
II. Initial Development: Technologies for Tactical Aircraft . . . . . . . . . . . 78
III. Concept Demonstration Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
IV. Concept Demonstration Aircraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
V. Preferred Weapon System Concept . . . . . . . . . . . . . . . . . . . . . . . . . 93
VI. System Development and Demonstration . . . . . . . . . . . . . . . . . . . . 97
VII. STOVL Weight Attack Team . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
VIII. Weight Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
IX. Test and Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

vii
viii TABLE OF CONTENTS

X. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118

Chapter 3 F-35 Air Vehicle Technology Overview . . . . . . . . . . . . . 121

Chris Wiegand, Bruce A. Bullick, Jeffrey A. Catt, Jeffrey W. Hamstra,
Greg P. Walker and Steve Wurth, Lockheed Martin Aeronautics Company,
Fort Worth, TX
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
II. Integrated Air Vehicle Subsystems . . . . . . . . . . . . . . . . . . . . . . . . . 124
III. Propulsion Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
IV. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158

Chapter 4 F-35 Digital Thread and Advanced Manufacturing . . . . 161

Don A. Kinard, Lockheed Martin Aeronautics Company, Fort Worth, TX
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
II. F-35 Lean Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
III. F-35 Materials and Structures: Supportable Low Observables . . . . 166
IV. The Digital Thread . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
V. The Future of Advanced Manufacturing . . . . . . . . . . . . . . . . . . . . . 176
VI. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181

Chapter 5 F-35 Weapons Design Integration . . . . . . . . . . . . . . . . 183

Douglas M. Hayward and Andrew K. Duff, Lockheed Martin Aeronautics
Company, Fort Worth, TX; Charles Wagner, F-35 Joint Program Office,
Arlington, VA
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
II. F-35 Weapons Suite and General Arrangement . . . . . . . . . . . . . . . 184
III. Internal Weapon Bay Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
IV. External Weapon Carriage Design . . . . . . . . . . . . . . . . . . . . . . . . . 195
V. Suspension and Release Equipment Design . . . . . . . . . . . . . . . . . . 198
VI. Gun System Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
VII. Stores Management System Design . . . . . . . . . . . . . . . . . . . . . . . . 206
VIII. Future Weapons Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
IX. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
TABLE OF CONTENTS ix

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212

Chapter 6 F-35 System Development and Demonstration

Flight Testing at Edwards Air Force Base and Naval Air Station
Patuxent River . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
Mary L. Hudson and Michael L. Glass Lockheed Martin Aeronautics
Company, Edwards Air Force Base, CA; Lt Col Tucker “Cinco” Hamilton,
United States Air Force, Edwards Air Force Base, CA; C. Eric Somers
and Robert C. Caldwell, Lockheed Martin Aeronautics Company,
Naval Air Station Patuxent River, MD
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 213
II. Edwards Air Force Base . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
III. Naval Air Station Patuxent River . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
IV. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251

Chapter 7 F-35 Structural Design, Development,

and Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Robert M. Ellis, Philip C. Gross, Joseph B. Yates and John R. Casement,
Lockheed Martin Aeronautics Company, Fort Worth, TX; R.H.L. (Tad) Chichester,
Lockheed Martin Aeronautics Company (retired), Fort Worth, TX; Kathryn
Nesmith, F-35 Joint Program Office, Arlington, VA
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
II. TASK I: Design Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
III. TASK II: Design Analysis and Development Testing . . . . . . . . . . . . 262
IV. TASK III: Full-Scale Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
V. TASK IV: Certification and Force Management Development . . . . . 282
VI. TASK V: Force Management Execution . . . . . . . . . . . . . . . . . . . . . . 284
VII. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285

Chapter 8 F-35 Flight Control Law Design, Development,

and Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
Jeffrey J. Harris and James Richard Stanford, Lockheed Martin Aeronautics
Company, Fort Worth, TX
I. Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
II. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
III. Control Law Design Methodology . . . . . . . . . . . . . . . . . . . . . . . . . 290
x TABLE OF CONTENTS

IV. Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

V. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311

Chapter 9 F-35 Propulsion System Integration, Development,

and Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
Steven P. Wurth and Mark S. Smith, Lockheed Martin Aeronautics
Company, Fort Worth, TX
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 313
II. F-35 Propulsion System Overview . . . . . . . . . . . . . . . . . . . . . . . . . 314
III. Conceptual Design, Test, and Verification for Concept
Feasibility: ASTOVL/LSPM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
IV. Preliminary Design, Test, and Verification of Functional
Feasibility: CDP/X-35 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
V. Detailed Design, Test, and Verification for Operational
Deployment: SDD/F-35 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
VI. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363

Chapter 10 F-35 Subsystems Design, Development, and

Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Drew Robbins, John Bobalik, David De Stena, Ken Plag, Keith Rail and Ken Wall,
Lockheed Martin Aeronautics Company, Fort Worth, TX
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
II. Systems Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
III. Requirements Derivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 375
IV. Key Technology Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . 378
V. Integration Qualification Effort/Challenges . . . . . . . . . . . . . . . . . . . 385
VI. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398

Chapter 11 F-35 Mission Systems Design, Development, and

Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
Greg Lemons, Karen Carrington, Thomas Frey and John Ledyard,
Lockheed Martin Aeronautics Company, Fort Worth, TX
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 399
II. The Vision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
TABLE OF CONTENTS xi

III. F-35 Sensor Suite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 403

IV. Fusing the Data into Information . . . . . . . . . . . . . . . . . . . . . . . . . . 413
V. Next-Generation Cockpit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
VI. Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
VII. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 419
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 420

Chapter 12 F-35 Information Fusion . . . . . . . . . . . . . . . . . . . . . . . 421

Thomas L. Frey, J. Chris Aguilar, Kent R. Engebretson, David K. Faulk and
Layne G. Lenning, Lockheed Martin Aeronautics Company, Fort Worth, TX
I. Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
II. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
III. Fusion Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
IV. The F-35 Information Fusion Approach . . . . . . . . . . . . . . . . . . . . . 427
V. Information Tiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428
VI. Evidence-Based Combat Identification . . . . . . . . . . . . . . . . . . . . . . 430
VII. Autonomous Sensor Management . . . . . . . . . . . . . . . . . . . . . . . . . 433
VIII. Cooperative Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 435
IX. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437

Chapter 13 F-35 Carrier Suitability Testing . . . . . . . . . . . . . . . . . . 441

Tony Wilson, Lockheed Martin Aeronautics Company, Fort Worth, TX
I. Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 441
II. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
III. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
IV. Shore-Based Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
V. Ship-Based Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 485
VI. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 500
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 502

Chapter 14 F-35 Aerodynamic Performance Verification . . . . . . . . 503

David G. Parsons, Austin G. Eckstein and Jeff J. Azevedo, Lockheed Martin
Aeronautics Company, Fort Worth, TX
I. Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
II. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
III. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 504
IV. Performance Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 508
xii TABLE OF CONTENTS

V. Flight Test Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511

VI. Flight Test Analysis Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516
VII. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523

Chapter 15 F-35 High Angle of Attack Flight Control

Development and Flight Test Results . . . . . . . . . . . . . . . . . . . . . . . 525
Daniel G. Canin, Jeffrey K. McConnell and Paul W. James, Lockheed
Martin Aeronautics Company, Fort Worth, TX
I. Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
II. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
III. Aircraft Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527
IV. Control Law Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534
V. Flight Test Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545
VI. Flight Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
VII. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 570
VIII. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 572
Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 572
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 573

Chapter 16 F-35 Weapons Separation Test

and Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
Christopher F. Hetreed, Matthew D. Carroll and Joe E. Collard, Lockheed Martin
Aeronautics Company, Fort Worth, TX; Richard C. Snyder,
Naval Air Warfare Center, Naval Air Station Patuxent River, MD
I. Nomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
II. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 576
III. F-35 Lightning II and Various SDD Weapons . . . . . . . . . . . . . . . . . 579
IV. Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
V. Ground Tests: Validation of the Dynamic System Model . . . . . . . . 599
VI. Typical Store Separation Flight Test Instrumentation . . . . . . . . . . . 602
VII. F-35 Store Separation Flight Testing: Validating Store
Aerodynamics Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 606
VIII. F-35 Store Separation Flight Test Vignettes . . . . . . . . . . . . . . . . . . 627
IX. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 637
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 638
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 639
TABLE OF CONTENTS xiii

Chapter 17 F-35 STOVL Performance Requirements

Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 641
David G. Parsons and Daniel E. Levin, Lockheed Martin Aeronautics Company,
Fort Worth, TX; David J. Panteny and Peter N. Wilson, BAE Systems,
Warton, United Kingdom; Michael R. Rask, Cummings Aerospace, Niceville,
FL; Brad L. Morris, Naval Air Systems Command, Naval Air Station
Patuxent River, MD
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 642
II. Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 643
III. F-35 STOVL Performance Requirements Verification . . . . . . . . . . . . 645
IV. Short Takeoff Performance Verification . . . . . . . . . . . . . . . . . . . . . 646
V. Vertical Landing Bring-Back Performance Verification . . . . . . . . . . 659
VI. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 679
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 680

Chapter 18 F-35 Climatic Chamber Testing and

System Verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681
Victorio J. Rodriguez and Billie Flynn, Lockheed Martin Aeronautics Company,
Fort Worth, TX; Marc G. Thompson, BAE Systems, Warton,
United Kingdom; Steven Brelage, Naval Air Warfare Center Aircraft
Division, Naval Air Station Patuxent River, MD
I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681
II. F-35 vs Previous Aircraft Climatic Testing . . . . . . . . . . . . . . . . . . . . 682
III. Mckinley Climatic Laboratory Test Facility . . . . . . . . . . . . . . . . . . . 684
IV. Challenges and Decisions in Defining and Scheduling
Climatic Test Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 684
V. Aircraft Modification, Facility, and Test Plan Preparation . . . . . . . . 687
VI. Entry into Test Facility and Test Execution . . . . . . . . . . . . . . . . . . . 694
VII. F-35 Icing Cloud Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 696
VIII. Thermal Management System Testing . . . . . . . . . . . . . . . . . . . . . . 704
IX. F-35 Climatic Testing Results Summary and Lessons Learned . . . . 705
X. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 708
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 709

About the Authors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 711

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 721

Supporting Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 743

FOREWORD

My love for aviation—specifically high-performing aircraft—went a mirror

opposite route than the title of this book: I went from cockpit to concept. As a
Naval aviator I had the privilege to fly 38 different types of military aircraft and
logged more than 3,000 flight hours.
After my active duty flying days, I moved to the industry side of aviation,
working on several military aircraft programs for Lockheed Martin, including
leading the team building the F-22 Raptor, followed by leading the F-35 Joint
Strike Fighter program through its pre-competition concept development phase
and competitive contract selection.
The F-35 program, at the highest level, encompasses a complex problem state-
ment addressing much more than developing the next fighter.
For the last several decades, the operational model for long term combat and
peace keeping operations revolved around tri-service, coalition participation for
both operational and financial burden sharing reasons. No single country can
go it alone any longer.
The reality of this concept, however, has historically been complex and unrea-
listic due to underlying technology, lack of platform interoperability and political
differences among and between participants. This was the basis for the complex
undertaking called the Joint Strike Fighter.
As you read the history of the F-35 Joint Strike Fighter team, think about the
challenges they faced from the start:
1. Meet a very complex operational requirement: Design, develop, test, field
and sustain a family of highly common first line fighters designed to fully
incorporate the unique basing requirements of conventional air force oper-
ations, expeditionary operations off small ships and unprepared fields and
projecting power from the sea from large deck aircraft carriers.
2. Make that platform stealthy, supersonic and embed the most sophisticated
suite of multi-spectral sensors ever employed in a fighter.
3. Incorporate high volume production capabilities that can capture economies
of commonality and scale.
4. Ensure our closest allies can participate with the goal of true interoperability
and burden sharing in future combat and peace keeping operations.
5. Revitalize a global industrial base that has atrophied over the years by imple-
menting revolutionary new manufacturing capability based on very high pre-
cision, automation and motion-based production systems, and transferring
technology that enables the industries of our allied partner nations to
compete on a global world stage with the best of the best.
6. Assemble a world class prime contractor partnership and global supply chain
that can perform to the unprecedented production and sustainment require-
ments of the F-35 Joint Strike Fighter program objectives.

xv
xvi FOREWORD

The F-35 Lightning II: From Concept to Cockpit is the seminal archive for the
journey that the F-35 team experienced. It is a remarkable collection of lessons
learned on a very unique venture to change the acquisition paradigm for tri-
service and coalition allied procurement of front-line combat capability. I was pri-
vileged to lead this transformational team through the challenges of the first
decade and I salute the men and women that delivered this incredible capability.

Capt. Charles Thomas (Tom) Burbage, U.S. Navy, Ret.

F-35 Joint Strike Fighter Program
Executive Vice President and General Manager, Ret.
PREFACE

In the years since the F-35 Joint Strike Fighter development program was awarded
to Lockheed Martin, countless articles have appeared in the popular media dis-
cussing the merits and challenges of the program. F-35 information has also
occasionally appeared in the scientific literature through AIAA or other technical
venues. However, an integrated story of the F-35’s technical development has
heretofore never been presented in a single work.
The purpose of The F-35 Lightning II: From Concept to Cockpit is to tell that
story—a full-spectrum history of the design, development and verification of the
F-35 Lightning II as described by the engineers, scientists and managers intimately
involved throughout the development program. The journey to achieve the F-35’s
extraordinary capability was long with untold challenges encountered on the path.
The reader will find many of the key technical challenges, and the innovative sol-
utions that resulted, discussed herein.
The reader should realize this work is not intended as an overall assessment of
the F-35 or the Joint Strike Fighter program. Rather, it is an engineering develop-
ment story from the perspective of insiders, many of whom dedicated the best part
of their career to F-35 and are justifiably very proud of their work.
The book is based on 18 technical papers presented in a two-day F-35 track at
the 2018 AIAA AVIATION Forum. The “From Concept to Cockpit” subtitle
appropriately summarizes the contents, with the chapters grouped into three
major sections:

† F-35 Program Overview, which begins with an overall history of the program
and further presents discussion of aircraft configuration design, weapons inte-
gration, key technologies brought into the program, and overviews of aircraft
manufacturing and the flight test program.
† F-35 Air Vehicle Design, which presents discussion of key aircraft systems,
including airframe structure, flight controls, propulsion, subsystems and
mission systems.
† F-35 Test and Verification, which presents discussion of the flight and ground
test programs undertaken to verify the F-35’s airworthiness, flight perform-
ance and mission capability.
The chapters are in a logical progression but are not interdependent, and so one
may read the contents in any order without confusion.
Although much information is included in this work, there are several impor-
tant topics that have been left out. Given the sensitive nature of military products,
not every technology or system can be discussed.
I wish to express appreciation to the staff at AIAA for their invaluable help and
guidance regarding this publication as well as for their encouragement and leader-
ship in bringing the F-35 program to an AIAA event.

xvii
xviii PREFACE

Special thanks is also due to several Lockheed Martin individuals for their out-
standing contributions in championing, leading, and supporting the Concept
to Cockpit project, from initial discussions through the presentations at the
AVIATION Forum and preparation of this work:

† Ron Bessire and Lee Sampson, Engineering & Technology

† Russel Baldonado, Daniel Buck, Alyssa Campbell, Karen Hagar, Kathleen
Hearons, Ken Ross and Dennis Soultaire, Communications
† Rick Mange, J.D. McFarlan and Joey Sikorski, F-35 Program
Our greatest thanks must go to the 60 plus individuals from Lockheed Martin, our
F-35 partner companies and suppliers, and the F-35 Joint Program Office
who contributed the 18 chapters of this work. The authors brought to bear
their knowledge, experience and significant personal time to develop the
content. Most of all, they brought a passion to tell the integrated F-35 engineering
story and help advance the state of the aerospace profession by doing so. Without
their expertise and dedication, this work would not have been possible.

Jeffrey W. Hamstra
Sr. Fellow, Lockheed Martin Corporation
 CHAPTER 1

F-35 Program History: From

JAST to IOC
Arthur E. Sheridan
Lockheed Martin Aeronautics Company, Fort Worth, TX
Robert Burnes†
F-35 Lightning II Joint Program Office, Arlington, VA

The Joint Strike Fighter program leading to the Lockheed Martin family
of F-35 aircraft has been unprecedented in terms of scope and challenge.
This paper reviews the background and need for the air system. It sum-
marizes the environment, objectives, approach, and results of each of
three distinct development phases, and highlights some of the most sig-
nificant challenges encountered and solutions achieved. It also covers
initial production and sustainment achievements in parallel. Despite
the ambitious goals and numerous challenges, the development program
is drawing to a close, and a system is now being produced and sustained
that meets its customers’ warfighting requirements.

I. BACKGROUND
The origins of the Joint Strike Fighter (JSF) program can be traced to the long-
standing commitment of the U.S. Marine Corps (USMC) and United Kingdom
(UK) Royal Air Force (RAF) and Royal Navy (RN) to develop a Short Takeoff
and Vertical Landing (STOVL) strike fighter, and to the end of the Cold War.
Drastic defense budget reductions after the Cold War, together with aging fleets
of fighter aircraft in the United States and across the west, demanded a new
level of cooperation in development and production. The U.S. Department of
Defense (DoD) Bottom-Up Review in 1993 cancelled previously separate
fighter/attack development plans of the U.S. Air Force (USAF), U.S. Navy
(USN), and USMC that aimed at replenishing U.S. fleets but became viewed as
unaffordable. The need for new aircraft procurement was compelling, however,
due to the end of production of legacy fighters (Fig. 1). Furthermore, the large
number of aircraft types in use by the United States and its allies could not be

 Program Management Principal, F-35 Customer Programs, AIAA Associate Fellow.

†
Director, Program Operations and Director, Program Management.

1
2 A. E. SHERIDAN AND R. BURNES

Fig. 1 Historical and projected U.S. fighter procurement profile (circa 2001).

affordably maintained (Fig. 2). Trends toward joint operations and coalition
warfare required significant improvements in interoperability. In this environ-
ment, service leaders in the United States and United Kingdom agreed to
develop a single program to address the next generation of affordable strike plat-
forms. Additional affordability strategies contributing to the environment were
acquisition reform initiatives advocating performance-based specifications and
concurrent development, as well as the desire to exploit the digital revolution
with simulation-based acquisition, digital design, and paperless commerce.
Existing U.S. service strike fighter requirements were widely disparate, ranging
from the U.S./U.K. Advanced STOVL (ASTOVL) (a small supersonic STOVL

Fig. 2 Legacy fighter types expected to be replaced by JSFs.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 3

airplane for the USMC and United Kingdom with a maximum empty weight of
24,000 lb) to the Navy’s A-12 (a stealthy carrier-based, twin-engine, long-range
medium bomber), to a low-cost fleet-structure fighter to succeed the USAF’s
F-16. References [1] and [2] provide summaries of U.S. precursor programs
and their sequence, as well as the early development of STOVL propulsion con-
cepts that together form the genesis of what is now the F-35 program. At the
time, the industry had great doubt that a single aircraft could be designed to
satisfy the needs of all services. For this reason, and lacking common air-vehicle
requirements, DoD did not approve the creation of an aircraft acquisition
program; rather, the initial program mandate was to invest jointly in technologies
that could be applied irrespective of a specific aircraft configuration, and to
perform configuration studies to determine whether a common family of aircraft
could meet service needs.

II. JOINT ADVANCED STRIKE TECHNOLOGY

The services formed a JSF Program Office (JSFPO) drawing from the Naval Air
Systems Command (NAVAIR), USAF Aeronautical Systems Command (ASC),
and U.K. Ministry of Defence (MoD), and created the Joint Advanced Strike
Technology (JAST) program in 1994. The JSFPO was (and remains) located in
Arlington, Virginia, where top personnel from each systems command are
co-located. The leadership structure was established with roles of the Senior
Acquisition Executive (SAE), Program Executive Officer (PEO), and deputy
PEO positions alternating among the departments of the USN and USAF, and
rotating at nominally two-year intervals; that is, a USAF SAE would be served
by a Department of the Navy PEO (USN or USMC officer) and a USAF
Deputy PEO. The pattern would then reverse upon the change of command on
a nominally two-year basis.

A. TECHNOLOGY DEMONSTRATIONS
The new program issued initial contracts to industry under the JAST banner, with
separate contracts for individual technology maturation and demonstration
efforts. Contracts were issued to all eventual JSF competitors, but they were to col-
laborate in planning the efforts and were required to share the results across the
industry. JSF Integrated Subsystems Technology (J/IST) is a prime example of
such an effort, and within that, the More-Electric Actuation program is an
example that was led by Lockheed Martin and eventually incorporated into the
F-35 configurations. Other proprietary technologies were pursued separately by
competitors, a Lockheed Martin example being the diverter-less supersonic
inlet that performs as a mixed-compression inlet that avoids boundary layer inges-
tion through the use of innovative shaping with no moving parts, providing
a lightweight and smooth configuration with favorable signature integration.
4 A. E. SHERIDAN AND R. BURNES

Both the J/IST electric power and actuation concept and the diverter-less inlet
were demonstrated on separate F-16 platforms [3, 4].

B. CONCEPT DEFINITION AND DESIGN RESEARCH

In parallel with these technology programs, the JAST program also issued
Concept Definition and Design Research (CDDR) contracts. These efforts were
begun to conceive of specific aircraft configurations and specific performance
requirements. In general terms, the requirements were to provide configuration
variants to serve the services’ differing basing needs: Conventional Takeoff and
Landing (CTOL) for USAF, STOVL for the USMC and RAF/RN, and a Carrier
Variant (CV) for USN operations with catapults and arresting gear. The
up-and-away performance requirements for range/payload and maneuverability
were to be common among the variants and approximately equivalent to
combat-equipped current F-16s and F/A-18s. Signature levels, mission systems,
and weapons capabilities were to be studied over a wide trade space. An afford-
ability target of 28 million dollars in 1994 base-year economics was established
for the CTOL variant—roughly equivalent to the price of an F-16 Block 50,
including targeting and electronic warfare pods and external fuel tanks. The
different variants were to be as common as possible to take advantage of econom-
ies of scale and interoperability.
In light of budget constraints, the U.S. services recognized the value of com-
monality and saw that overly specific requirements could negate those benefits.
So, they re-examined basic assumptions embedded in their previous development
projects and adopted some important changes. Most notably, for the first time
since the A-7, the USN accepted the single-seat and single-engine requirement,
which was essential to achieve a practical STOVL configuration for the USMC.
The result of the CDDR phase was that the configuration concepts and corre-
sponding requirements trade studies instilled sufficient confidence in the JSFPO
and the participating services to proceed to the next phase to develop a family
of air systems to be known as the JSF, designed to satisfy a Joint Operational
Requirements Document (JORD) that would also be developed in the next
phase: the Concept Demonstration Phase (CDP).

C. INDUSTRY COMPETITORS
Three industry competitors participated in the program, each with configura-
tion families based on different STOVL propulsion concepts. Reference [2]
gives an overview of the evolution of STOVL concepts preceding JAST. The Lock-
heed Martin CDDR designs were all based on the Shaft-Driven Lift Fan (SDLF)
described below.
The Boeing concept was based on direct lift, which relied on diverting the
majority of engine exhaust to Harrier-like swiveling nozzles at the center of
gravity for hover. Exhaust flow was abruptly switched to and from an aft-mounted
F-35 PROGRAM HISTORY: FROM JAST TO IOC 5

vectoring nozzle during wingborne/jetborne transitions. A series of remote nozzles

provided attitude control, and a jet screen was located near the main inlet to reduce
hot-gas ingestion. At this stage in the program, the Boeing configuration was a
delta-wing arrangement that was essentially common for all three variants.
The McDonnell Douglas (later acquired by Boeing)/Northrop Grumman/
British Aerospace (now BAE Systems) concept was a conventional wing-tail
arrangement with conventional propulsion for the CTOL and CV variants. The
STOVL variant employed the lift-plus-lift/cruise propulsion system similar in
arrangement to the Russian YAK-38 and YAK-141 aircraft. It was to have a
single lift engine mounted forward and a combination of swivel nozzles and aft
conventional nozzle for main-engine exhaust, similar to but much shorter than
the Boeing exhaust system.

D. THE LOCKHEED MARTIN AIR VEHICLE CONCEPT

The Lockheed Martin air vehicle concept centered on the SDLF propulsion system
[1, 4] illustrated in Fig. 3, which was key to both STOVL capability and family
commonality among variants. Together with a vectoring lift-fan nozzle and a con-
tinuously vectoring three-bearing swivel nozzle (3BSD) for the engine exhaust, the
system inherently addressed the major challenges and typical causes of failure for
previous supersonic STOVL concepts.

Fig. 3 Comparison of the Lockheed Martin JSF CTOL and STOVL propulsion systems.
6 A. E. SHERIDAN AND R. BURNES

1. VERTICAL THRUST-WEIGHT MARGIN

The lift fan augments thrust relative to the basic nonafterburning engine by
approximately 40%, achieving thrust more efficiently with higher mass flow and
lower exhaust velocity.

2. HOVER BALANCE/TRIM
The SDLF arrangement creates natural vertical-thrust posts around the aircraft
center of gravity, so pitch and roll control is achieved simply through shifting
of upward vertical thrust among the four inherent nozzles. Importantly, the
high thrust capability of the forward-placed lift fan allows the aft nozzle and
engine to be placed at the aft end of the fuselage, permitting conventional arrange-
ments of aerodynamic configuration, structure, and systems. This is key to the
efficient conventional configuration that is well-suited to all three variants.

3. CONTINUOUS TRANSITION
Continuously vectoring lift-fan and engine exhaust nozzles permit a smooth tran-
sition (wingborne to/from jetborne flight) without requiring a propulsion-system
mode change during transition. This simplifies transition and reduces risk. The
propulsion system converts to STOVL mode prior to downward transition and
converts out of STOVL mode after upward transition.

4. INDUCED EXTERNAL ENVIRONMENT

The forward lift fan exhaust is generated solely by the compression of air by the
fan. Although the exhaust is still energetic, it is much cooler and slower than
engine exhaust. In the aft, engine exhaust is more benign due to the extraction
of energy to drive the lift fan by the low-pressure turbine. Exhaust from the roll
nozzles is similar to that from the lift fan and is produced by the main engine
fan. The result is an acceptable induced thermal, acoustic, and flow environment
imposed on the aircraft, deck, or ground surface, as well as surrounding personnel,
aircraft, or equipment. Furthermore, the much cooler lift-fan exhaust blocks the
warmer engine exhaust from reaching the forward parts of the aircraft in hover,
minimizing hot-gas ingestion in the inlets.
At Lockheed Martin, two series of configurations were developed [5]. The 100
series was derived from the Defense Advanced Research Projects Agency’s
ASTOVL and Common Affordable Lightweight Fighter (CALF) programs.
These were canard-configured designs with an SDLF for the STOVL variant,
like that tested under the ASTOVL program in the National Aeronautics and
Space Administration Ames Research Center’s full-scale wind tunnel [6]. The
commonality strategy for the CTOL and CV variants was simply to remove the
SDLF system and replace the aft vectoring nozzle with a two-dimensional vector-
ing nozzle, as implemented on the F-22. The 200 series of configurations used the
F-35 PROGRAM HISTORY: FROM JAST TO IOC 7

same propulsion system with a conventional wing-tail arrangement. The common-

ality approach was to use identical aerodynamic configurations for the CTOL and
STOVL variants. The CV variant was to be common as well, but at this stage had
greater wing area due only to enlarged wing leading- and trailing-edge flaps and
an extended wing tip, so the airfoil shape was thinner and the profile somewhat
compromised to facilitate a common wing box with the other variants.
A conventional wing-tail arrangement was selected for two primary reasons.
First, extensive research by Lockheed Martin leading to the F-22 design concluded
that the conventional arrangement produced the best transonic turning maneuver
performance and maximum lift for desired longitudinal stability levels, with com-
petitive supersonic drag characteristics [7]. Second, although in this phase there
was no explicit requirement, the need for low Carrier Powered-Approach Speed
(VPA) indicated much lower risk for a wing-tail design compared to a delta-wing
or canard configuration, both from a lift perspective (maximum lift and at
approach Angle of Attack [AOA]), and from a low-speed control perspective
(control-power and adverse control coupling). The canard configuration carried
through the ASTOVL and CALF programs was designed for CTOL/STOVL capa-
bility without regard for carrier compatibility with a CV variant.
At this stage, commonality among the variants was very high within the con-
figuration family, including airframe structure and Vehicle Systems (VSs). It was
expected that the benefits of extensive commonality would outweigh the costs of
improvements in performance that might be gained by optimizing the structure
and systems to specific service requirements. In later stages of the program, the
airframe structure and some VSs did evolve away from commonality at the
detail level, as described later in the paper. Mission Systems (MSs) remained
nearly 100% common throughout.

III. CONCEPT DEMONSTRATION PHASE: JOINT STRIKE FIGHTER

The JAST program to date and CDDR results were deemed promising enough to
proceed to the next phase of an air system acquisition program. The priorities, or
pillars, of the program were established as lethality, survivability, supportability,
and affordability.
The CDP program had ambitious objectives. First, the Preferred Weapons-
System Concept (PWSC), representing the future production air system, would
be developed and matured. Second, the requirements set would be refined itera-
tively with the PWSC. Third, key technologies would be matured and demon-
strated, including flight test of a full-scale Concept Demonstrator Aircraft (CDA).

A. THE COMPETITION
In 1996 the three competitors (Lockheed Martin, the McDonnell Douglas/British
Aerospace/Northrop Grumman team, and Boeing) submitted proposals for the
8 A. E. SHERIDAN AND R. BURNES

CDP phase. By this time the high stakes of the program were clear, so the govern-
ment maintained a fair competition by placing limits on spending for the program
to keep contractors from attempting to buy the competition. In early 1997
Lockheed Martin and Boeing were selected for the CDP program.
The CDP contract values were each just more than $1 billion, including the
CDAs and engine development. After McDonnell Douglas’s loss, both Northrop
Grumman and British Aerospace were still interested in participating in the
program, and both clearly had valuable technical capability, so there was a court-
ship period during which both companies considered and were being considered
for joining Boeing or Lockheed Martin as teammates. Ultimately both companies
teamed with Lockheed Martin. Soon after the CDP down-select, McDonnell
Douglas merged into the Boeing Company, making the McDonnell Douglas JSF
resources available to the Boeing CDP effort.
On the Lockheed Martin team, both new teammates were full-fledged
aircraft prime contractors, and each brought unique strengths to the team.
Northrop Grumman had extensive experience with low observables and a long
legacy in carrier suitability, and British Aerospace had a legacy and unique
capabilities relative to STOVL aircraft and extensive capabilities in precision fab-
rication. Teaming agreements were established outlining the teammates’ respon-
sibilities and work share, as well as provisions for sharing intellectual property
within the program. During this phase, the development team functioned as a
single unified team, largely co-located in Fort Worth, Texas, and numerous key
leadership positions were filled by personnel from Northrop Grumman or
British Aerospace.

B. REQUIREMENTS DEVELOPMENT
Establishing requirements was a key objective of the CDP program, resulting
in the JORD, a JSF Model Specification (JMS), and Key Performance
Parameters (KPPs). The JORD was preceded by a series of JSF Interim Require-
ments Documents (JIRDs) that were released on a roughly annual basis. Require-
ments maturation was closely overseen by service representatives, with frequent
reviews by the Operational Advisory Group (OAG) and Senior Warfighters
Group (SWG).
As the trivariant configurations matured, requirement trade studies were
conducted in parallel (separately by both competitors) to determine which com-
binations of capabilities were achievable and affordable. Indeed, requirements
management was the principal affordability lever applied during this phase.
The basic aircraft sizing was determined through several iterations of Cost and
Operational Performance Trades (COPTs), addressing aircraft performance
(e.g., mission, maneuver, basing). These results fed JIRD-I and JIRD-II between
1995 and 1997, establishing initial aerodynamic performance, low-observability
requirements, and overall supportability and avionics targets. The COPTs
were followed by formal Cost-as-Independent-Variable (CAIV) studies using
F-35 PROGRAM HISTORY: FROM JAST TO IOC 9

Fig. 4 Typical JSF COPT and CAIV trade study data.

campaign-level operational analysis measures of merit to determine the most

cost-effective combinations of sensors, weapons, signature, and maneuver/flight
envelope capabilities. These fed JIRD-III in 1998 and the draft JORD in April
1999 with more detailed avionics and supportability requirements (Fig. 4).
JSF was the first program to apply CAIV in a quantifiable way. Dozens of
trade studies of individual capabilities were conducted to quantify operational
benefit and impact on remaining life-cycle cost. These trade results were prior-
itized and plotted as the CAIV curve (Fig. 4). The joint requirements authors
understood that affordability would restrain imagination when it came to capa-
bilities, and the CAIV curve facilitated reconciliation of personal biases and
warfighting value. Some capabilities are difficult to quantify in terms of cam-
paign analysis, such as the gun. These cases were settled using consensus tech-
niques within the OAG, SWG, and JSFPO, and among contractor operational
analysts.
Although, on the surface, performance requirements did not vary significantly
through the phase, there were subtle ground rule changes and a few added par-
ameters that drove commonality among variants apart, particularly in the air-
frame. For example, revisions to the USAF design-mission penetration altitude
and Mach number drove required fuel volume in the CTOL variant, and increas-
ing the vertical load factor requirement to 9g for the CTOL caused most structural
members and actuators to diverge from the corresponding STOVL parts. The
addition of the VPA requirement as a KPP for the CV variant directly resulted
in increased CV wing area and movement to a noncommon wing-box planform
and cascading changes into other systems, such as actuators.
Within the industry team, a rigorous systems engineering process was main-
tained to decompose the top-level performance-based specification requirements
to lower tiers. Decomposition was documented in requirements work packages
and tracked via IBM Rational DOORS software. Affordability requirements
were imposed as design-to-life-cycle-cost targets for each design area.
10 A. E. SHERIDAN AND R. BURNES

C. PREFERRED WEAPONS SYSTEM CONCEPT

Lockheed Martin PWSC configurations were designated as the 230 series. The first
released configuration family, 230-1, was designed to satisfy JIRD-I requirements
and retained a common wing-box planform for all three variants. This became the
point of departure for the CDAs, later designated X35A, B, and C, so the external
lines and other design data were transferred to the CDA design team in Palmdale,
California. Subsequent design iterations through 230-5 were driven by wing, pro-
pulsion, and internal arrangement changes [5]. In general, the design team strove
to make the configurations as compact as possible. Overall fuselage dimensions
were heavily constrained by the propulsion system and weapons bays. Wing
area and span for CTOL and STOVL were being driven up by subsonic maneuver
requirements, but the design team was constrained by hover weight. One major
breakthrough was the adoption of a Variable-Area Vane-Box Nozzle (VAVBN)
for the lift-fan exhaust. Previously, the lift fan exhausted through a Telescoping
Vectoring Exhaust Nozzle (TVEN), referred to as the pram hood by British Aero-
space personnel. This nozzle, when sized for Short-Takeoff (STO) thrust at aft
vector angles, had too great an effective area at the vertical hover condition,
limiting available hover thrust [4]. The variable area feature of the VAVBN,
however, allowed thrust optimization at all vector angles, improving Vertical
Landing (VL) capability, which in turn permitted a larger wing area for combat
turn performance.
The external configuration matured to balance subsonic vs supersonic vs
basing performance, stability and control, signature characteristics, and sensor
locations and fields of regard. This was supported by extensive wind-tunnel and
radar-range testing in addition to extensive Computational Fluid Dynamics
(CFD) and computational electromagnetics analyses. As the configurations
evolved, sometimes subtle changes created significant effects. In one example
just before the 230-1/X-35 lines were to be frozen, a slight increase in wing inci-
dence was incorporated to improve STO performance that also required reshap-
ing the upper fuselage over the inlet duct. This change showed the desired increase
in lift in the low-speed wind tunnel, but in transonic tests showed an unexpected
and unacceptable reduction in directional stability. This created an extra design
iteration just before the X-35 configuration could be frozen. Perhaps the most sig-
nificant evolution of the external configuration was in the CV wing size (and tails)
driven by the addition of the VPA requirement as a KPP. Iterations of the CV
design included successively larger wing areas in order to maintain sufficient
risk margin to the requirement. This caused the CV variant to abandon the
common wing-box geometry with the other variants, although the substructure
locations were retained to allow for common assembly tooling.
The internal arrangement, which directly affected the external lines as well,
matured significantly as internal systems matured. The size of subsystem com-
ponents often increased with each design iteration, but the biggest challenge
was in integration, accounting for mounting provisions, connectors and
F-35 PROGRAM HISTORY: FROM JAST TO IOC 11

couplings, routing of tubes and harnesses, bend radii, separation requirements,

and maintainer access. Major geometric integration challenges included carriage
and clearance of the long list of internal weapons, landing gear retraction and
stowage, expendable countermeasures dispensers, Electro-Hydrostatic Actuation
System (EHAS) actuators and electronic units, engine and accessories removal
and installation, and inflight opening doors. Thermal management was also a
continuing challenge. As would be expected of a 5th-Generation-class fighter,
system heat loads are absorbed internally as much as possible, but the STOVL
lift fan also rejects substantial heat, adding to the challenge. Also, each variant
had unique center-of-gravity requirements, which again were most demanding
for STOVL. The result is that, although many fuel-system components were
common among the variants, the integrated fuel system arrangements became
unique. Unique center-of-gravity requirements also significantly affected weapons
suspension and release equipment design [8].
The need for hover (and takeoff) thrust-weight margin placed demanding
requirements on nearly all aspects of the design family, but most strongly on
STOVL. Refinement of the STOVL propulsion system to improve both thrust
and aeromechanical operability continued throughout PWSC development, as
well as weight reduction. Early in the phase, the lift-fan diameter was increased
by 2 in. for added thrust. Significant changes were made to the primary, lift fan,
and auxiliary air inlets for improved pressure recovery and distortion, the most
visible being adoption of the aft-hinge lift-fan inlet door [4]. The main inlet
was shortened by approximately 40 in. as a weight-saving measure, and the four-
edge main-inlet aperture was replaced by the three-edge configuration to improve
high-AOA distortion.
The PWSC manufacturing approaches for affordability also had a direct
impact on the configurations. One approach was targeted at rapid assembly
(five months) with minimal tooling. Major subassembly components were to be
fully stuffed with subsystem components and joined during final assembly
using discrete quick-mate joints with relatively few large fasteners accompanied
by fluid and harness couplings at every joint. The final assembly mate planes
were selected to minimize complex rigging tasks after mate. The goal was to
avoid mate planes that would be crossed by the canopy, inlet duct, or weapons-bay
doors, but because the inlet duct and canopy overlapped in fuselage station, a mate
joint at Fuselage Station 270 was accepted that crossed the canopy.
Another goal was to maintain common tooling among all variants. Even if web
or flange thicknesses varied between/among variants, pad-ups were envisioned to
maintain the same interfaces with tooling. Very tight outer mold line tolerances
at seams would be maintained by precision machining inner mold lines of the
composite skins and the assembled substructure.
As the PWSC product definition and production system requirements firmed
up during the CDP program, the teaming agreements between Lockheed Martin
and teammates Northrop Grumman and BAE Systems were amended to define
specific work content responsibilities for the teammates in coming detailed
12 A. E. SHERIDAN AND R. BURNES

design and the future production and sustainment phases. The team organization,
personnel assignments, and business practices migrated to more discrete sub-
contractor relationships. In development, both teammates were to have broad
participation in numerous areas. In the production program, Northrop
Grumman would have full responsibility for the center fuselage and inflight
opening doors, notably the large weapons-bay doors, as well as the weapons-
bay-door drive; arresting-gear, fire-protection, inertial-navigation, and global-
positioning systems; landing-aid antennas; and MS common components. BAE
Systems would be responsible for the aft fuselage, horizontal- and vertical-tail
boxes, CV outboard wing boxes, and the crew-escape, fuel, life-support, and
ice-detection systems.
The final PWSC configuration, 230-5, became the basis for the Engineering
and Manufacturing Development (EMD) proposal in early 2001.

D. CONCEPT DEMONSTRATOR AIRCRAFT: THE X-35S

As mentioned previously, the CDA or X-35 aircraft were derived from the 230-1
PWSC configurations. These aircraft were crucial to demonstrate the SDLF
system’s capability and verify that variants with a high degree of commonality
could satisfy the diverse requirements of each service. To that end, Lockheed
Martin produced two aircraft with one basic design. Unique features for each
variant could be installed on either aircraft. The Boeing X-32A and Lockheed
Martin X-35A aircraft are shown in Fig. 5.
The Lockheed Martin demonstrators did not reflect the PWSC’s airframe
structure, nor most of the subsystems. Risk reduction for those elements was
accomplished through other unique demonstrations. Rather, the X-35s rep-
resented the aerodynamic configuration, the full-scale propulsion system, and
the flight control laws (not hardware). The airframes were built-up airframes as
one-off prototypes. They were built as single assemblies, not from mated subas-
semblies. As much as possible, systems components were taken from existing

Fig. 5 Boeing X-32A and Lockheed Martin X-35A CDAs.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 13

aircraft. The landing gear, for instance, were taken from the A-6 for the main gear
and the F-15E for the nose gear.
Four key differences between the X-35B STOVL variant and the proposed
PWSC were: 1) the X-35B retained the TVEN pram-hood lift-fan nozzle rather
than the later VAVBN; 2) the side-hinged lift-fan inlet doors on the X-35B
were replaced with a single aft-hinged door; 3) the center-hinge arrangement of
the auxiliary air inlet doors was replaced by side-hinge doors; and 4) the X-35B
retained the use of three-handle STOVL inceptors much like the Harrier arrange-
ment, rather than the two-handle unified-control scheme adopted for the
proposed PWSC.
The first airplane (Ship 1) first flew in October 2000 as a CTOL variant at
Palmdale. It immediately transitioned to Edwards Air Force Base (AFB), Califor-
nia, to conduct up-and-away flight tests representative of both the CTOL and
STOVL variants. After 28 flights in five weeks, the airplane returned to Palmdale
for conversion to the STOVL variant by installing the SDLF, 3BSD, and roll-
control ducts. During its brief test program, the X-35A was flown by Lockheed
Martin, USAF, USMC, and U.K. pilots. It achieved 5g, 20-deg AOA, supersonic
speeds, and aerial refueling. Quantitative results for specific range and maneuver-
ing performance closely matched predictions. Solid Level 1 handling qualities
were observed by all six pilots throughout the test envelope, and trouble-free
propulsion performance demonstrated engine/inlet compatibility.
Ship 2 first flew in December 2000. It was configured as the X-35C with larger
wing leading- and trailing-edge flaps and larger tail surfaces. It was flown by Lock-
heed Martin, USN, USAF, and U.K. pilots for approximately 40 flights at Edwards
AFB, including Field Carrier Landing Practice (FCLP), aerial refueling, and super-
sonic flight, before ferrying across the country in February 2001 to Naval Air
Station (NAS) Patuxent River, Maryland, for an additional 30 flights of aggressive
field carrier landing tests over four weeks, accomplishing 258 FCLPs. Again,
up-and-away tests supported prediction models, and excellent flying qualities
were noted. Eight government and Lockheed Martin pilots flew the X-35C con-
figuration (Fig. 6).
Ship 1 completed its conversion to the X-35B configuration in February 2001
and entered a series of ground tests prior to first flight. In parallel, the STOVL
propulsion system completed accelerated mission testing, and the flight-control
software underwent final regression tests. The first ground test involved mounting
the airplane on struts just above a grated hover pit for restrained measurements
of net forces produced by the installed propulsion system, and to demonstrate
the functionality of the integrated flight and propulsion control system. Then
the grate was replaced by a solid ground plane to determine the effects of close
ground proximity on net aeropropulsive forces. Also, ground environment acous-
tic, thermal, and flow measurements were made on the aircraft’s lower surface, on
the ground plane, and at a 50-ft radius from the airplane.
Initial STOVL-mode flights were made in late June 2001 over the grated hover
pit to establish hover performance limits and clear the bottom of the jetborne
14 A. E. SHERIDAN AND R. BURNES

flight envelope. This established known conditions for the end point of sub-
sequent decelerating transitions. After demonstrating conversions to/from
STOVL mode at altitude, systematic build-downs to hover were performed,
including semi-jetborne landings, leading to the first VL on the AM-2 landing
pad at Edwards AFB. Successively slower STOs were made down to a rotation
speed of 60 kt. Level-1 flying qualities were consistently seen during the tests.
The engine was free of hot-gas ingestion, and infrared images confirmed that
the forward lift-fan flow effectively blocked engine exhaust from the engine
inlets and forward portions of the aircraft.
Aerodynamic, propulsion, and environment measurements from these
ground and flight tests were better than predictions in nearly all cases, as conser-
vative margins in prediction models were found to be unnecessary. During the
first flight, a press-up (vertical takeoff from the grated hover pit), British Aero-
space test pilot Simon Hargreaves pressed up to a stable hover 20 ft high when
he only expected to barely break contact from the grated pit. In the lead-up to
STOVL flight tests, the test team contemplated plans to follow Ship 2 to NAS
Patuxent River in order to do hover tests at favorable conditions at sea-level
elevation and cooler temperatures. In reality, the system produced robust vertical
performance in the high desert at 2300-ft elevation and in the middle of summer,
so the entire test program was completed there.
On 20 July 2001 the X-35B became the first aircraft in history to achieve both
supersonic and hovering flight in a single sortie (Fig. 7). For more than 40 years
prior to the X-35, U.S. and European aerospace enterprises had pursued Vertical
and/or Short Takeoff and Landing (V/STOL) fighter technology, yielding only
the subsonic Harrier as a truly operational system. Besides the X-35B, only
three research vehicles had previously demonstrated both hovering and superso-
nic capability with the same configuration: the German VJ-101C, French Mirage

Fig. 6 X-35C carrier-suitability flight test accomplishments.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 15

Fig. 7 X-35B performance of Mission X: STO, supersonic dash, and VL.

III-V, and Soviet Yak-141. However, none of these aircraft represented oper-
ational weapons platforms, and their limited performance prevented them from
accomplishing hover and supersonic flight in a single mission. With that back-
ground, the X-35B team set out to execute Mission X, consisting of an STO, a
supersonic dash, and a VL to conclusively demonstrate that the SDLF propulsion
system would overcome the fundamental incompatibilities of supersonic and
STOVL flight and enable the common configuration family.

IV. SYSTEM DEVELOPMENT AND DEMONSTRATION

The Lockheed Martin proposal for what was then called the EMD phase was
submitted in February 2001, while the X-35 flight tests were still underway. At
that time, only the X-35A CTOL had completed its testing and was still being con-
verted to the X-35B STOVL. The X-35C CV was preparing to ferry from Edwards
AFB to NAS Patuxent River to complete carrier suitability tests. Supplements to
the proposal covering successful X-35 flight test results from all three variants
were later submitted in April and August. Formal questions and answers were
exchanged over the summer via evaluation notices and face-to-face discussions,
and final proposal revisions were due in mid-September.
The proposal evaluation criteria included three equally weighted factors:
affordability, EMD (each component of which received both proposal ratings
and risk ratings), and past performance. The affordability factor addressed the
proposed products, including subfactors for the air vehicle, autonomic logistics
(sustainment system), and remaining life-cycle cost (procurement, operations,
and support). The EMD factor addressed the proposed program, including sub-
factors for technical plans, management, and EMD cost.
16 A. E. SHERIDAN AND R. BURNES

The entire proposal evaluation process remained on schedule despite delays in

planned STOVL flight tests experienced by both competitors and the 9/11 terror-
ist attacks on the Pentagon. The Defense Acquisition Board met as planned on 24
October to confirm that the JSF was ready to enter the next phase of development,
now termed System Development and Demonstration (SDD). The dramatic
public announcement of the winning contractor was also made on schedule on
26 October 2001 [9].
The announcement that Lockheed Martin had won the contract was made by:
1. The source selection authority, Secretary of the Air Force James Roche
2. Under Secretary of Defense for Acquisition, Technology and Logistics
Edward C. “Pete” Aldridge
3. Secretary of the Navy Gordon England
4. U.K. Parliamentary Under-Secretary (Ministry of Defence) (Procurement)
Lord William Bach
5. Maj. Gen. Michael Hough, USMC, JSF PEO
Secretary Roche stated that both competitors’ proposals were “very strong,” but
that the Lockheed Martin team “emerged continuously as the clear winner” on
a best-value basis, considering strengths, weaknesses, and relative risks [9]. Inter-
estingly, the F-35 designation seemed to be determined and announced on the
spot in response to a reporter’s question, based on the X-35 designation of the
Lockheed Martin demonstrator. In keeping with the demonstrator designations,
the CTOL variant was designated F-35A, the STOVL variant was designated
F-35B, and the CV variant was designated F-35C.
No time was wasted in executing the contract; the model contract submitted
with the proposal was signed by the government the same day on which the down-
select was announced.

A. THE PROGRAM
The objectives of the SDD program were to: 1) develop an affordable family of air
systems (air vehicles plus autonomic logistic systems) that meet service require-
ments and significantly reduce life-cycle cost; 2) develop a life-cycle plan that sup-
ports production, fielding and operational support, and eventual disposal; and 3)
demonstrate and implement affordability initiatives. In the interest of affordability
and rapid development, the program was aligned with several acquisition-reform
initiatives, including integrated product and process development, Performance-
Based Specifications (PBSs) and contractor Total System Performance Responsi-
bility (TSPR), Simulation-Based Acquisition (SBA), concurrent development, and
Performance-Based Logistics (PBL). A JSFPO policy minimized the use of
government-furnished property, except for full-scale ground-test and flight-test
facilities and, notably, the propulsion system (development and hardware).
F-35 PROGRAM HISTORY: FROM JAST TO IOC 17

Fig. 8 Development and LRIP schedule originally specified in the EMD CFI.

The overall program plan was specified by the government in the EMD Call
for Improvements (CFI), the title given to the request for proposals, as reproduced
in Fig. 8. The overall program period of performance was to be an aggressive 126
months, with first flights of the CTOL, STOVL, and CV variants planned for 48,
53, and 62 months after go-ahead, respectively. The plan also specified certifica-
tion of three blocks of progressive MS capabilities to execute mission vignettes
with particular weapons loadouts.
Though not part of the SDD program, the Low-Rate Initial Production (LRIP)
program was planned to be highly concurrent with development. Long-lead
funding for the first LRIP lot was planned to be authorized before the first
flight of the CTOL variant, with full funding to be authorized just after the
STOVL first flight. Six annual lots were envisioned, with production quantities
ramping from 10 Block-1 aircraft in the first lot to 168 Block-3 aircraft in the sixth.
Lockheed Martin was awarded a $19-billion cost-plus-award-fee contract.
Pratt & Whitney was separately awarded a $4-billion contract. The two companies
established an associate contractor agreement to govern the coordination and
integration of the propulsion system in the aircraft. In the Lockheed Martin con-
tract, the majority of the available award fee was tied to the customer’s assess-
ments of contract performance during each six-month period covering the
affordability, developmental cost control, management, and technical categories,
as well as an overall comprehensive rating. However, a significant portion of
the contract fee, known as the Schedule B award fee, was to be determined
by an objective comparison of actual LRIP production costs with specified
18 A. E. SHERIDAN AND R. BURNES

Affordability Improvement Curves (AICs) for each variant. The AICs corre-
sponded to the cost trend, when averaged over the projected production
program, that would correspond to the original $28 million (fiscal year 1994
dollars) CTOL Unit Recurring Flyaway (URF) cost target, when adjusted for econ-
omic escalation, production quantity and rate, variant configuration, and scope
changes. This device was intended to provide a direct incentive for achieving
affordability goals but was very difficult to implement. As the program progressed
and setbacks to performance were encountered, the fee structure was ultimately
renegotiated.
The Lockheed Martin SDD program plan implementation through first flights
of all variants aligned with the CFI, as shown in Fig. 9. In Lockheed Martin’s plan,
a single air system Preliminary Design Review (PDR) covering all three variants
was set for 17 months after receiving Authority to Proceed (ATP), rather than
only the 12 months specified in the government’s CFI, compressing the timespan
available to design, manufacture, and check out the CTOL variant after the PDR.
Lockheed Martin’s plan also allowed more time for each variant to produce the
detailed designs, with separate air system Critical Design Reviews (CDRs) for
each variant at 31, 37, and 44 months, respectively, rather than the specified
21 months for all variants in the CFI plan, greatly reducing the time available
for manufacturing and/or requiring a greater degree of concurrency between
design and manufacturing.

Fig. 9 Excerpt from Lockheed Martin’s initial SDD schedule.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 19

In order to accomplish the trivariant PDR event, lines-freeze was to be

completed only nine months after ATP, barely enough time to conduct a single
wind-tunnel test cycle (design, build, test, and analyze). A second milestone,
lines validation, was established recognizing that additional testing may be
required to finalize the geometry. However, structural layouts would have to be
completed based on the initial lines freeze. The other pacing element to begin
detailed structural design was aerodynamic loads. The time available was not
enough to conduct a full loads wind-tunnel program, so the initial structural
sizing was done using data from the X-35s, and adjustments were made based
on varying levels of CFD analysis.
In summary, the early stages of the SDD plan depended on a high level of con-
currency among aerodynamic lines and loads, structural design, and manufactur-
ing. The final PWSC configuration from the CDP phase was well defined at the at
the outset of SDD, but wind-tunnel testing of the specific configurations had not
been completed in the CDP due to the need to focus resources on X-35 flight tests.
The design team expected a few weeks to pass between the down-select and ATP,
during which advance work would mitigate some of the concurrency risk, but as
mentioned earlier, the two events actually occurred on the same day. In fact, the
first critical challenge the team had to overcome was to ramp up personnel very
rapidly, increasing the staff from 400 to 4000 people in the first eight months.
Therefore, the early SDD schedule was very challenging.

B. SYSTEM REQUIREMENTS DEPLOYMENT

The SDD phase began with an operational requirements document taken from
the draft JORD produced by the JSFPO and U.S. and U.K. services at the end
of the CDP program. This established KPPs for the system, defining values
for combat radius, CV recovery (VPA), STOVL performance (flat-deck and
ski-jump STO distance, and VL bring-back), interoperability, Radio-Frequency
(RF) signature, mission reliability, sortie generation rate, and logistics footprint.
The JSF Contract Specification (JCS) aligned with the JORD and was included
in the Lockheed Martin prime contract. It was an entirely capabilities-based PBS.
Roles and missions for each of the U.S. services and the United Kingdom were
defined along with mission usage. Performance, interface, and environment
requirements were defined for lethality, survivability, basing, security, safety,
reliability, supportability and training, situational awareness, and handling qual-
ities, along with their corresponding definitions.
One of the initial tasks was to formalize requirements allocations to create per-
formance and functional requirements at all tiers of the product hierarchy and
organization. This was done through mission decomposition and requirements
work packages. Requirements allocations were documented and managed
through the use of a commercial requirements management tool tailored to the
F-35’s needs. PBSs for system suppliers could be generated directly by the tool.
20 A. E. SHERIDAN AND R. BURNES

C. KEY PROGRAM STRATEGIES

The development team recognized that the complexity and scope of the SDD
program, combined with the aggressive schedule and affordability goals, would
require extraordinary team performance. In addition to the broad acquisition
reform initiatives, a number of engineering and management strategies were
ingrained in the organization from the outset, some widely applied and some
narrowly focused. The following are examples of the strategies used.

1. JSF FIRST AND JSF ENTERPRISE GUIDING PRINCIPLES

Program leadership continued to reinforce JSF First! as the program motto,
reflecting the high stakes in the success of the program to U.S. and allied militaries,
the prime teammates’ and critical subcontractors’ businesses, and individual
careers. In other words, whatever selfish advantage one organization might gain
within the program paled in comparison to the benefits of the highly functioning
overall team. The Lockheed Martin program general manager also established 10
enterprise-guiding principles governing the way people worked together, categor-
ized in support of five objectives:
1. Inspire excellence.
2. Expect the exceptional.
3. Seek to connect.
4. Foster trust and respect.
5. Value the individual.

2. AFFORDABILITY AND BEST VALUE

The affordability pillar permeated all aspects of the F-35 program. In SDD, the
emphasis shifted from COPT and CAIV top-level requirements tradeoffs to the
allocation of design-to-cost targets to the system, subsystem, and airframe com-
ponent levels, and to the execution of affordability initiatives to progressively
achieve the targets (Fig. 10). Integrated Product Teams (IPTs) were accountable
to report progress against burndown plans on a monthly basis, and suppliers
formally reported statuses quarterly on projected recurring cost and their own
affordability initiatives.

3. TOTAL SYSTEM PERFORMANCE RESPONSIBILITY AND PERFORMANCE-BASED

SPECIFICATIONS
TSPR was a DoD acquisition reform doctrine designed to save costs by eliminat-
ing redundant or unnecessary work between the government and prime contrac-
tor. In this approach, Lockheed Martin accepted TSPR and was free to make use of
F-35 PROGRAM HISTORY: FROM JAST TO IOC 21

Fig. 10 Typical monthly design-to-cost data summary.

MIL-SPEC and MIL-STD standards but was not bound by them if there
were industry or commercial equivalents that resulted in meeting overall require-
ments. In this approach, the government customer was afforded insight into the
program without traditional government oversight. The JCS was a PBS dictating
the required system performance capabilities (the what) but not specifying the
methods by which the system was developed to achieve the required performance
(the how). Lockheed Martin, in turn, employed PBSs in most of the major MS and
VS subcontracts.

4. SIMULATION-BASED ACQUISITION
SBA was a DoD initiative to exploit advances in information technologies, specifi-
cally modeling and simulation, to enable better, faster, and cheaper weapons-
systems acquisition. To oversimplify, the objective was to verify system behavior
and performance vs requirements in a digital virtual environment. To these ends,
Lockheed Martin and the JSFPO invested heavily in a robust system of labora-
tories dedicated to F-35 development and verification. The main laboratory facili-
ties are depicted in Fig. 11. These produced high-fidelity simulations used to verify
requirements. Extensive physical testing in various domains served not to verify
requirements directly but rather to validate the models that comprised the
virtual verification environment. The dedicated F-35 laboratory environment
included the following:
. VS Integration Facility
. VS Processor/Flight Controls Integration Facility
. MS Integration Laboratories (including systems integration stations with
simulated, RF-stimulated, and open-air [actual hardware in physical aircraft
model] inputs)
. Air System Integration Facility
. Verification Simulator
22 A. E. SHERIDAN AND R. BURNES

Fig. 11 Integrated, co-located high-fidelity laboratory complex.

Lessons learned from legacy programs drove the laboratory requirements and
design. First, co-location in one geographic area of all major test and verification
laboratory venues was determined to be essential, as opposed to the dispersed
facilities of the F-22 program in Seattle, Washington; Fort Worth; and Marietta,
Georgia. The more integrated and synergistic F-35 environment was available
to the development team on a 24/7 basis. Second, in alignment with DoD SBA
policy statements, reuse of software and data was found to save costs. Reuse
was a strategic initiative from the very early stages of laboratory development.
Reuse allowed the team to avoid duplicative efforts to develop the middleware/
executive layers of software and aircraft/sensor models across all of the laboratory
venues. The F-35 laboratories had two to three times greater capability than that of
any preceding program by any measure, with unprecedented connectivity. These
capabilities were key elements in the SDD plan, shaping the makeup of the flight
test aircraft fleet and supporting a target of only 5000 flights. These flights were
planned to provide a top-level validation of the integrated models, not to verify
lower-level requirements.

5. COLLABORATIVE ENVIRONMENT AND THE DIGITAL THREAD

Another strategy to exploit advanced information technologies involved sharing
needed data among a wide range of participants and among phases of the
program life cycle. Considerable investment was made early in the program in
tools to facilitate nearly instant access to data across the team, while at the
same time reliably controlling access to restricted data for ITAR or other
F-35 PROGRAM HISTORY: FROM JAST TO IOC 23

reasons. The JSF Data Library was established as a program-wide repository for
collaborative file exchange and vaulting of program records of all types. For
detailed designs, the product data manager, later known as the product lifecycle
manager, facilitated near-real-time collaboration among nine primary design-
release organizations spanning 17 time zones, as well as approximately 50 suppli-
ers with design authority for their systems, and more than 100 build-to-print
suppliers across the globe [10, 11]. Design information was comprehensive for
each part and contained in Build-to Packages (BTPs), including CATIA solid
models, drawings, analysis documents, specifications, manufacturing planning,
and tooling information, all organized into the product data structure, including
changes. The system ensures that consistent data are used in the as-designed,
as-planned, as-built, and as-maintained configurations.

6. CONCURRENT ENGINEERING AND INTEGRATED PRODUCT TEAMS

The F-35 air vehicle, production system, and support system were developed sim-
ultaneously. Even the earliest flight test aircraft were built on production hard
tooling at Lockheed Martin and, to varying degrees, at teammates’ and suppliers’
sites. Also, the manufacturing system makes use of the same support equipment
developed or selected for use in the field. Design-for-manufacture and the sup-
portability pillar were among the design team’s primary goals from the outset.
Multidiscipline IPTs were the norm, with representation from manufacturing
and sustainment disciplines, and virtual simulations of assembly and maintenance
tasks were done in parallel with the parts/assemblies definition [10].

7. SYSTEMS ENGINEERING AND REQUIREMENTS MANAGEMENT

Top-level requirements, either directly from the JCS or derived from mission
decomposition, were allocated down to all tiers of the design team to be verified
up to form the systems engineering V. Requirements management is addressed
later in the paper.

8. RISK MANAGEMENT AND KEY SYSTEM DEVELOPMENT AND INTEGRATIONS

The program established and maintained a robust risk management process fol-
lowing the classical ISO 31000 approach, identifying and assessing risks and
executing mitigation plans for moderate to high risks. This process was
applied at each organizational level. Most often, risks were identified at lower
tiers and evaluated using consequence thresholds (i.e., impacts on cost, schedule,
or performance) tailored to that tier. Moderate or low risks were managed at
that tier, but high risks were elevated for assessment at the next tier, and so
on. In this way, for instance, a risk that could be managed within a Tier IV
IPT’s budget, that did not impact the critical-path schedule, or that only affected
internally imposed requirements might be handled at Tier IV, whereas a risk
with high cost impacts, affecting critical milestones or top-level performance
24 A. E. SHERIDAN AND R. BURNES

requirements (e.g., a KPP) would be elevated to the program level. Risk reviews
were held monthly or as needed, and new risks were processed. Existing risks,
once mitigated, were retired. At the outset of SDD, the program-level risks
included:

Risk 01: Software Executability Risk 07:

Production URF
Risk 02: Lift System Hardware Risk 08:
Process & Tool Performance
Risk 03: Trained Manpower Risk 09:
Air Vehicle Weight Control
Availability Risk 10:
Electro-Hydrostatic Actuator
Risk 04: Supplier/Partner Development
Management Risk 11: Canopy Bird Strike Compatibility
Risk 05: MS Fusion Algorithms Risk 12: Requirements Control
Risk 06: Deck Crew Environment under PB

For the most part, mature technologies were chosen for inclusion in the F-35
air system design, produced by the prior JAST and CDP phases or other prior
programs. However, it was understood that successful development was highly
dependent on complex cross-IPT integration of multiple technologies to an extra-
ordinary extent, even among high-performance fighters. Integration challenges
could be physical, functional, or both. An evaluation of integration challenges
was done at the outset of SDD by mapping candidates’ complexity vs potential
program impacts to arrive at the Key System Developments and Integrations
(KSDIs). Teams were created around each KSDI with a leader responsible to
the program manager for planning and executing the cross-IPT effort, particularly
in the early stages of SDD as the system requirements were finalized and the
system became fully defined. The list of 22 KSDIs is presented here:
. Interoperability Integration
. Prognostics-and-Health-Management Development & Integration
. Outer-Mold-Line Definition
. Lean Manufacturing
. Low-Observables Aperture/Edge/Sensor Integration
. Virtual Weapons System/Simulation-Based Acquisition
. Subsystems—Airframe Integration
. Cockpit Integration
. Integrated-Core-Processor Development & Integration with Sensors
. Mission-Systems Software Development and Domain Integration
. Vehicle-Systems Software Development and Domain Integration
. Integrated Flight-Propulsion Control
F-35 PROGRAM HISTORY: FROM JAST TO IOC 25

. Integrated Subsystems Development and Integration

. STOVL Propulsion System
. Integrated First Flight Checkout
. Shipboard Suitability Integration
. Joint Distributed Information System Development & Integration
. Integrated Avionics
. Design for Reliability, Maintainability, & Supportability
. Factory-to-Field Support Equipment Commonality
. Air Vehicle/Training Software Commonality
. Air Vehicle Service Life

D. KEY CHALLENGES: WEIGHT GROWTH

One of the most significant challenges to the success of the F-35 program was that
of air vehicle weight growth and control. Weight control is critical to any flying
machine, but the STO and VL requirements for the F-35B are particularly unfor-
giving. During the first two years of SDD, many factors contributed to both real
weight growth in all variants, attributed to meeting evolving requirements, as well
as the realization of higher weights as estimating accuracy matured. The com-
bined effects resulted in a projected failure to meet both STO and VL KPP
thresholds that threatened the success of the entire program. Figure 12 presents
current estimates of STOVL-variant empty weight through early SDD, compared
to various target measures.
At the entry to the SDD phase, the three variants were at a preliminary design
level of maturity. At that stage, weight estimates, particularly for the airframe
structure, relied on Lockheed Martin’s parametric prediction tools. With these,
estimates for each component were based on regressions of statistical data from
past aircraft for like components, based on the size, shape, material mix, and
structural requirements (e.g., Mach number and load factor). Features of the
F-35 configurations that were not present in the historical aircraft, such as low
observables, variant commonality, integrated subsystems, and rapid-assembly fea-
tures, were accounted for with incremental weight allowances added to the para-
metric predictions. In addition to the current estimate, a growth allowance to
Initial Operational Capability (IOC) was also added for use in performance pre-
dictions and structural sizing, based on prior programs’ history.
Throughout the CDP and early SDD phases (prior to PDR), JSFPO mass
properties engineers produced independent estimates using similar tools, and
the two organizations performed frequent weight reconciliations as the air
vehicle configurations evolved. The two estimates remained separate until PDR,
with the government’s estimate always being higher, but the reconciliation
26 A. E. SHERIDAN AND R. BURNES

Fig. 12 Current weight-empty estimates as of year-end 2003 compared to plan and critical
KPP requirements.

process facilitated a mutual understanding of each team’s analyses. A unified

JSFPO and Lockheed Martin weight-empty status was an entry criterion for PDR.
Completion of PDR in late March 2003, 17 months into the program, kicked
off detailed design of the CTOL variant, with the first aircraft designated
2AA:0001 (abbreviated as AA-1). At that time, the adjusted parametric weight
estimates (including estimated growth to IOC), together with aerodynamics and
propulsion estimates, supported achievement of all KPP thresholds, and these
estimates were used to allocate weight budgets to airframe components and sub-
systems for detailed design.
As the detailed design and analysis processes matured the AA-1’s structure
from preliminary design to sized layouts and eventually to released BTPs, periodic
weight estimates based on a detailed accounting of parts became available, known
as Bottom-Up Weights (BUWs). Unsurprisingly, the earliest BUW totals were
significantly greater than both Lockheed Martin’s and the JSFPO’s parametric
estimates, because the part details were not yet optimized. However, at PDR
it was recognized that the roughly 4800-lb gap was too great to expect part
optimization alone to close. More work was needed to improve the efficiency of
the overall structural arrangement, and some concessions to commonality and
rapid assembly had to be reconsidered due to significantly more weight impact
than what was allowed. Furthermore, PDR identified a number of unresolved
design integration issues primarily related to weapons clearance and loading in
the internal weapons bay, integration of the internal gun on the CTOL variant,
and clearance to parts of the STOVL propulsion system.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 27

1. THE BLUE RIBBON ACTION TEAM

It was clear that rapid closure of these issues was needed to progress detailed
design on the very aggressive release schedule. AA-1 CTOL first flight was targeted
just 31 months after PDR, as well as the other variants that were to follow closely.
To address the issues, a Blue Ribbon Action Team (BRAT) was assembled from
technical leaders of the contractor team, the JSFPO, and experts from NAVAIR
and ASC. The objective was to resolve open PDR weight and integration issues
while preserving the baseline SDD plan as much as possible. The BRAT estab-
lished a time-phased weight reduction curve with targets for June and September
2003, and closure to the CDR target weight by December 2003. They conducted
detailed reviews of every major airframe component, every part weighing 5 lb
or more, and a bottom-up estimate of smaller parts: tubes, harnesses, fasteners,
brackets, and sealants.
Initial results from the BRAT reviews were good. Design decisions resolved
the key integration issues, and the June 2003 weight estimates improved the
June BRAT targets for all variants. The most significant configuration changes
involved increasing the cross-sectional area to improve structural load paths
and internal equipment volume (which penalized supersonic performance) and
eliminating quick-mate joints between major assemblies (which nearly doubled
the planned final assembly timespan). Work resumed on the AA-1 BTPs, but
the combination of delayed BTP starts and added final assembly timespan resulted
in a forecasted delay in AA-1 first flight by approximately seven months, into the
second quarter of 2006.
By September 2003, however, although the optimization of large parts contin-
ued to show small improvements, a more complete accounting for small airframe
parts and other parts revealed a large increase over the June BUW. Primary con-
tributors to the other weight category changes from June were tubing and instal-
lation (clamps, brackets), systems installation, shims and spacers, and fasteners.
The total CTOL airframe weight estimate increased about 800 lb from June to
September. During the fall of 2003, calculated weights for released AA-1 BTPs
also continued a slow but steady increase, as did the government-furnished
propulsion system weight. At that time, the BUW estimates replaced parametric
estimates as the current estimates of record.

2. THE STOVL WEIGHT ATTACK TEAM

The actions taken during the BRAT period saved significant weight and resulted in
a CTOL configuration that, although not optimum, would have satisfied its driving
mission-radius KPP. However, when the weights were projected onto the much
more weight-sensitive STOVL variant, it failed to meet all of its KPP thresholds,
and significantly would have essentially no usable VL capability. As this reality
became quantified through the fall of 2003, program leadership on both contractor
and customer sides decided that more fundamental changes in the configurations
28 A. E. SHERIDAN AND R. BURNES

and the program plan were necessary. The first priority was to define a viable
STOVL configuration with robust margins to the KPP requirements that still
aligned with the program pillars of lethality, survivability, supportability, and
affordability. It was decided to largely pause the program until confidence in the
STOVL variant could be restored. Once that was accomplished, it was recognized
that a major redesign, affecting all variants, would be required, as well as rebaselin-
ing the SDD program. It was also decided that the AA-1 detail design and pro-
duction process was providing extremely valuable data and experience, so it
would continue in a parallel effort producing a de facto prototype of the aircraft.
In December 2003, program leadership directed that a dedicated team, to be
known as the STOVL Weight Attack Team (SWAT), be assembled to reestablish
the configuration family. During the first quarter of 2004, leadership, member-
ship, funding, and operating principles of the team were decided upon, and in
early March the plan was approved by the program leadership and the DoD
SAE. In mid-March the dedicated SWAT team crowded into unused office
space at Lockheed Martin’s Fort Worth headquarters; at teammate Northrop
Grumman’s El Segundo, California, site; and at teammate BAE Systems’ Samles-
bury, Lancashire (United Kingdom) facilities. The team was made up of roughly
550 full-time equivalent personnel. A kickoff meeting for the whole team in
mid-March laid out the full extent of the problem, the stakes at risk, and the
overall team approach. In addition to the dedicated SWAT, the entire program
team was engaged, starting with a weight stand-down day and a program-wide
weight incentive program for weight-saving ideas.
Operating principles for the team were designed to accelerate decision making
but maintain systems engineering rigor and configuration control, so these
disciplines were well represented on the team. Affordability and supportability
remained prime objectives as well, with dollars-per-pound thresholds enforced
and affordability, reliability, and maintainability impacts reported for every pro-
posed change. A simple two-tier decision process was set up with a weekly
tempo. The SWAT board had delegated authority and funding resources within
certain constraints to enact changes, and a higher-level multiboard met the next
day for decisions outside of those constraints. The multiboard consisted of the
entire executive leadership team, including engineering, manufacturing, sustain-
ment, business management and contracts, and others, whose presence was
required each week.
Recognizing that weight reduction actions could compromise various other
objectives, means were put in place to prevent hijacking of a weight-saving
change by advocates for disciplines adversely affected by changes before they
were aired at the SWAT board or multiboard. All inputs from the customer com-
munity were filtered through only two JSFPO senior leaders.
SWAT activities were organized into several thrusts.

1. Optimize zero-fuel weight. The highest volume of trades involved exhaustive

optimization of existing airframe, MS, and VS components by examining
F-35 PROGRAM HISTORY: FROM JAST TO IOC 29

each component, assembly, or part for excess margins, nonessential redun-

dancies or growth capability, inefficiencies, opportunities for low-risk tech-
nology insertion, or commonality penalties.
2. Leave no stone unturned. A review was made of past design or trade study
decisions affecting weight in light of the now better-understood weight situ-
ation. Multiple previously disjointed weight savings were reviewed, updated,
and consolidated. A team was set up to review new weight savings ideas from
the employee weight incentive program. Potential weight increases were
treated with the same rigor as weight savings. Design studies were pursued
to avoid or mitigate the increases, if possible; if not, weight increases were
accounted for exactly as weight reductions.
3. Reduce fuel weight. At the outset of the SWAT effort, STO performance was
the driving KPP requirement, and of course mission fuel was the largest single
element of takeoff gross weight. The most extensive CFD-based aerodynamic
drag optimization to date on the program, accompanied by wind tunnel ver-
ification, was used to optimize mission performance and reduce the KPP
mission fuel required.
4. Improve STOVL performance. STO maneuver trajectory and control effector
usage were optimized to maximize gross weight capability, and VL control
allowances were optimized to maximize available vertical thrust. Improve-
ments were made in inlet and nozzle performance to increase available
installed STOVL-mode thrust, but increases in core engine temperatures
and uninstalled thrust were not considered.
5. Challenge requirements. Requirements at all levels were examined for areas
that were overly specific. The primary focus was on internal self-imposed
requirements by the contractor team, followed by conservative interpretations
of JCS requirements. JCS requirements were prioritized in terms of value to
the warfighter based on operational analyses. KPP threshold values were
held to be sacrosanct, but interpretations of the definitions, ground rules,
and assumptions underpinning them were challenged. Although require-
ments studies were conducted throughout the SWAT phase, relaxation of
requirements was reserved until all other changes were implemented.
6. Quash weight increases. New processes for approving detailed design releases
were instituted to ensure that each part was designed to the minimum
acceptable weight.

Intensive activity in the SWAT phase lasted approximately seven months after the
kickoff. More than 600 design changes were approved in the period, highlights of
which are described in Ref. [5]. Most of the changes, amounting to about 2600 lb
of weight reduction and 600 lb of installed thrust improvement, were made within
the discipline teams, airframe, VS, MS, or propulsion. However, many key trade
studies required the integration of the overall aircraft configurations, particularly
30 A. E. SHERIDAN AND R. BURNES

requirements studies. These utilized the CAIV approach to identify capabilities

that drove aircraft weight significantly but had relatively little operational
benefit to the warfighter. Scout configurations were designed covering a broad
range of weapons capabilities, combat flight performance, signature-control
features, and MS functions and performance. Analysis of these configurations
quantified weight impacts as functions of mission capabilities.
Meanwhile, mission-effectiveness evaluations across the design space were
done in conjunction with the OAG. The most promising capability combi-
nations were evaluated in a July 2004 manned simulation exercise known as
Trial Buccaneer at the U.K. MoD’s Cutlass simulation facility at Abbey Wood
using USAF, USMC, USN, RN, and RAF pilots. The exercise concluded that
the selected capability combination could effectively execute the stressing
JORD missions. These combat requirements tradeoffs resulted in: 1) reducing
the STOVL weapons bay capacity to a 1000-lb weapon (plus missile) required
for the STOVL variant (previously, a common bay with the CTOL/CV was
incorporated as a growth provision); 2) reducing outboard wing store stations
to 1000-lb class; 3) limiting dual external air-to-ground loads and external
tanks to subsonic; 4) removing unused wiring (growth provision) from air-to-air
stations; 5) reducing structural limit speed to actual achievable airspeed; and 6)
slightly reducing signature treatment. Together these changes contributed about
500 lb of weight reduction. Some of these changes affected ORD requirements,
so they had to be approved by the SWG and Joint Requirements Oversight
Council.
The other dimension of requirements relief used to finally close the thrust-
weight gap was in definitions of the STO and VL requirements in conjunction
with the USMC. STO mission fuel requirements were reduced by optimizing
the ingress/egress altitude. Similarly, required fuel reserves for VL were
reduced by revising the prescribed wave-off/go-around pattern, control allowan-
ces, and on-deck fuel reserves to better match the operations of the legacy Harrier
fleet. These revisions increased allowable STO/VL gross-weight capability by
approximately 700 lb, so the net gap closure attributable to requirements revisions
totaled 1200 lb.
Figure 13 depicts weight-empty status through the SWAT phase as it was
monitored weekly and reported to the team and to external stakeholders at
all levels. The green curve shows the SWAT burndown plan derived from
the probability-weighted list of planned design trades, their potential weight
savings, and expected completion dates. This was the plan against which
SWAT progress was measured. Black diamonds indicate the current estimate,
including all approved changes to date. In some weeks, more weight increases
were recognized than savings. Also, maturating AA-1 and engine designs contin-
ued to increase in weight, as indicated by red diamonds. These increases were also
included in the current estimates as they were projected onto the STOVL variant.
Each week a weight-empty projection was reported based on the in-work and
remaining trade studies.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 31

Fig. 13 Progress and forecast of weight-empty achieved and required during SWAT effort.

The red solid line in the figure represents the weight-empty Not-to-Exceed
(NTE) target to satisfy the STO and VL KPP thresholds. Steps in the line reflect
changes in propulsion, aerodynamic, or control characteristics, or changes in
requirement ground rules. For example, the upward step in early August 2004
reflects the adoption of installed thrust improvements [5]. The difference
between the forecast and the NTE target represents the expected gap once
planned trade studies were complete. In late August 2004, the forecast and
NTE target lines converge to eliminate the gap because the decision was made
at the program level to adopt the proposed STOVL configuration as a baseline
and to accept the requirements revisions described previously. Completing
trade studies continued into October 2004, and ratification of the configuration,
requirement adjustments, and approval to recommence the detailed design
phase was obtained through a series of DoD decision boards, culminating in
approval by the SWG, Configuration Steering Board, and finally the Defense
Acquisition Board, also in October 2004.

3. POST-SWAT
Closure of the SWAT effort resulting in a viable STOVL configuration was a major
success, overcoming the most significant technical challenge to date on the
program, but it was also a major disruption of the program plan. Indeed, a com-
plete replan of the SDD program was required, as described later in this paper.
One of the most significant changes was in the sequence of variant-detailed
design efforts. Fabrication and assembly operations were already underway on
AA-1, but issues identified at PDR and addressed by the BRAT, together with
32 A. E. SHERIDAN AND R. BURNES

slower than planned ramp up of resources, meant that first flight, originally
planned for the fourth quarter of 2005, was then planned for the third quarter
of 2006, and eventually slipped to 15 December 2006. Following the SWAT
phase, it was recognized that a new detailed design was needed for the CTOL
variant based on the new STOVL configuration. However, the previous
crawl-walk-run approach was replaced with the principle, Do the hardest one
first, and the others will benefit. Therefore, detailed design of the STOVL configur-
ation became primary, with the new CTOL design effort planned to be only
slightly staggered and in parallel with the STOVL. In the new plan, the first
flight of the STOVL variant was delayed 18 months from its original date in
the first quarter of 2006 to the third quarter of 2007.
As the program exited the SWAT phase, a new target approach was developed
to protect KPP threshold performance at IOC, in light of expected weight growth
and uncertainties remaining at the time. Figure 14 shows the STOVL NTE
line established in conjunction with the JSFPO and Pratt & Whitney that included
a 3% growth allowance for the Lockheed Martin –responsible aircraft as well as an
engine weight-growth allowance. In addition, it was recognized that uncertainties
and variability in weight estimation, propulsion performance, and aerodynamic
effects combined to create substantial uncertainties in STOVL performance capa-
bility. Therefore, a Monte Carlo uncertainty analysis was conducted to determine
additional margin to be imposed on weight-empty requirements. As the program
matured through the detail design, manufacturing, and flight test phases, these

Fig. 14 Post-SWAT weight-empty management plan with design margins.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 33

uncertainties were expected to lessen. Therefore, the uncertainty margin was com-
puted as a function of time (Fig. 14). The combined growth and uncertainty
margins established the Weight Tripwire line in the figure. As the program pro-
gressed, a weight status that exceeded the tripwire compelled additional weight
reduction actions to be undertaken to offset the overage.
A predictable result of the weight challenge and program-wide effort to
resolve it was a strong emphasis on weight management going forward, at both
organizational and individual levels. A weight czar was appointed as part of the
chief engineer’s office, the mass-properties engineering staff was substantially
increased, and more rigorous weight reviews were implemented in the BTP
release and change management processes. In addition, the entire design staff
became highly sensitized to the need for weight optimization. Designers were
required to demonstrate that all parts were at or near the minimum acceptable
weight as part of the BTP release reviews. Similarly, all change-board reviews
began with an update of current weight-empty compared to the tripwire line
(accounting for growth plus uncertainty) and a review of potential weight
reductions that could be implemented if necessary to offset any weight increases
predicted for changes under consideration that day. These measures were very
successful in enforcing a zero-weight-growth policy; the STOVL weight-empty
at CDR was only 90 lb greater than the final SWAT configuration (0.4%
growth). However, this vigilance did increase pressure on BTP release schedules,
affecting SDD’s overall schedule performance.
The SWAT effort projected a reduction in URF cost, due mainly to the sim-
plification and elimination of systems components, of about $700,000 per aircraft,
relative to the SWAT starting point, and a substantial reduction in operating and
support costs when accounting for fuel savings due to weight reduction. Support-
ability KPPs were virtually unaffected. STOVL weight-empty was reduced by
more than 3000 lb (and CTOL and CV by approximately 2400 and 1900 lb,
respectively), but the resultant current estimates were still roughly 2000 lb
heavier than those at the outset of SDD. The post-SWAT configurations were
measurably heavier and more complex than was recognized at the outset of the
contract. Furthermore, commonality among the variants was reduced by an esti-
mated 7% overall. These impacts on affordability created challenges in other
aspects of the program.

E. KEY CHALLENGES: MAJOR SDD PROGRAM REPLANS—OVER TARGET BASELINE

Obviously, the extensive redesign effort to solve weight growth issues and reestab-
lish KPP-compliant configurations had significant impacts on the overall program
cost and schedule. The review and approval of the new 240-4 family of configur-
ations was tantamount to a new PDR, roughly 18 months after the original event.
A complete replan of the program was implemented immediately following
the SWAT effort with a new variant firing order, new systems engineering,
design engineering, and program management policies, as well as new program
34 A. E. SHERIDAN AND R. BURNES

leadership and a new organizational structure. In order to distinguish pre- and

post-SWAT designs, a new type-version designation system was established:
2AF for CTOL, 2BF for STOVL, and 2CF for CV. Therefore, the first STOVL air-
craft would be 2BF:0001, abbreviated as BF-1. As previously stated, AA-1 was con-
tinued at its own pace; its CDR (now called a design integration maturity review)
was held in April 2004, just as the SWAT was beginning, and first flight occurred
in December 2006. The AA-1 effort provided valuable data and experience to the
design team as a de facto prototype, but this was a major element of scope not
included in the SDD program at contract award. Furthermore, most of the MS
and VS suppliers had completed or begun CDRs of their own based on weight
allocations that turned out to be too generous from the air system PDR. These
were heavily influenced by the CTOL requirements because that was the most
urgent schedule. Many of these systems were similarly set back to a PDR state
of maturity. All these factors combined to require the first of what would even-
tually become three program replans (Fig. 15). This would come to be referred
to as Over Target Baseline – 1 (OTB-1). After receiving the government’s agree-
ment and approval, the SDD contract was renegotiated, adding 18 months to the
schedule and more than $5 billion to the cost.

1. POST-SWAT RESET
The post-SWAT replan still called for the production of the original 14 test air-
craft and 7 full-scale ground-test articles built to the new design, in addition to
AA-1, which was built to the pre-SWAT configuration. The flight test program
was to include five STOVL, five CTOL, and four CV aircraft. Three of the

Fig. 15 Progression of SDD replans and milestone schedules.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 35

CTOL and one each of the STOVL and CV aircraft were to be fully equipped with
avionic systems and low observable treatments and dedicated to MS and signature
flight tests. The remaining flight test aircraft were referred to as Flight Sciences
aircraft and were equipped with instrumentation for performance, flying qualities,
and structural tests. Static and durability test airframes were planned for each
variant, as well as a dedicated test airframe for CV drop and barrier tests. Original
SDD production plans included a factory-built CTOL airframe for Radar Cross-
Section (RCS) pole-model tests. However, at this time it was determined that this
was impractical because the integration of the rotator would negate any fidelity
benefit of using the factory-built airframe. Instead, a more conventional model
structure would be utilized for the pole tests, but the model would include
actual production components for key features, such as the canopy, flight-
operable doors, and flight control surfaces.

2. MID-COURSE UPDATE
By the fall of 2006, a great deal of progress was made in the SDD program with
KPPs intact, but adverse cost and schedule trends were building that would
result in a second program replan. As cost and schedule pressures increased
over the next two years, program leadership instituted a series of intensive
efforts to contain costs and achieve key milestones that supported the on-schedule
completion of SDD and the beginning of production.
Technical progress was steady during the period. CDRs were completed for all
three variants. Beginning with a first flight in December 2006, AA-1 flight tests
were validating modeling, simulation, and analysis tools. Laboratories and
flying testbeds were verifying the performance of system components, as well as
their integration. Many of the ground-test and flight-test aircraft were in assembly
with fit and quality that confirmed the effectiveness of the digital thread from
design to manufacturing. Flight and laboratory software were demonstrating
better-than-legacy stability. KSDIs were successfully achieving milestones in
cross-discipline integration. All but two of the initial program-level risks were
retired, and those that were being elevated to the program level were tied more
to cost/schedule than technical performance.
This technical progress, however, was proving to take more effort and more
time than allotted by the SDD plan. Cost and schedule threats and pressures
emerged in many elements in the program.
In the airframe design release process, weight optimization required more
design analysis iterations and resulted in less commonality among the variants
than had originally been assumed. The structural arrangements and external
lines remained highly common among the three optimized variants, but they
ended up sharing only very few specific airframe part numbers. In order to
obtain enough resources to develop and release airframe BTPs, engineering per-
sonnel from across the team were employed, in addition to subcontracted
resources. Airframe design locations spanned 17 time zones, from Australia to
36 A. E. SHERIDAN AND R. BURNES

Europe to California. This effort was successful due to the common collaborative
digital thread design environment and toolsets, but the rigors of the digital
thread and the complexity of overseeing the worldwide operation contributed
to cost and schedule pressures. The results of these factors were that more
BTPs were required than originally planned, and each BTP required more
design and analysis effort than originally planned.
Manufacturing of the SDD test aircraft, both fabrication and assembly,
became the largest cost and schedule driver during this period. Significant
changes were made to the initial tooling system used for AA-1, and they were
still in progress as assembly of the weight-optimized aircraft was underway.
A number of design and tooling features aimed at easing producibility were
eliminated in the weight-optimization process. Prolonged BTP release schedules,
combined with longer-than-planned turnaround at fabrication suppliers, resulted
in inefficient assembly due to out-of-sequence and out-of-station work, and
sometimes work stoppage while awaiting parts. Suppliers’ initial fabrication
efforts were hampered by complexities associated with weight-optimized parts,
both in machine programming and in actual fabrication times. The parts out-
sourced included some of the most challenging parts ever produced by industry;
for example, the wing carry-through/landing gear attachment bulkhead on the
CTOL and CV variants are machined from the largest titanium forgings
yet produced.
The volume of engineering change traffic per BTP was well predicted in
the program plan, but the number of BTPs was increased, and the cost of pro-
cessing and implementing changes affected both engineering design and
manufacturing.
Systems suppliers also contributed significantly to the threats and pressures.
Overall, a large number of system suppliers required management reserve
funding to cover overruns to their development contracts, creating the second
largest cost driver. For example, the EHAS, originally highly common, became
largely unique for each variant. The EHAS was pioneered and demonstrated
in the preceding J/IST program using an F-16, but requirements for the F-35
had diverged substantially, and major technical challenges had to be overcome
related to motor design/regenerative power, thermal management, seals, pumps,
weight, and high-voltage separation. Similarly, the integrated CNI system scope
was initially underestimated, ultimately requiring development of 22 hardware
items and approximately 1.4 million software lines of code. Some communica-
tions, navigation, and identification functions required new technology invention,
for example, the Multifunction Advanced Data Link.
Throughout this timeframe, managers at all tiers, together with the Affordabil-
ity team, worked to identify, quantify, and mitigate these pressures within their
own spans of control, but the trends at the program level indicated that the
current program plan could not be executed within the OTB-1 budget or schedule.
However, clear direction was received from the JSFPO that no additional funding
would be available for the program.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 37

In a manner reminiscent of the SWAT effort, Lockheed Martin and JSFPO

program managers jointly chartered a special team to reexamine original premises
underlying the SDD program plan, in light of progress to date and remaining
risks. The objective was to realign remaining funds with essential tasks to
enable the successful completion of SDD within the OTB-1 budget and schedule.
Timing for action was urgent because the budget was being consumed at a high
spend rate, but the budget profile was declining steeply. Therefore, decisions
had to be made quickly to protect the remaining budget.
This effort was known initially as the Mid-Course Risk Reduction, and later as
simply the SDD Completion Plan. The team was composed of major IPT leads or
senior members of their staffs from both Lockheed Martin and the JSFPO. The
charter was to ensure that essential program objectives were accomplished,
while still achieving affordability and weight targets, and establishing the training
center, international production, and test sites. The essentials included:

1. Complete the design definition (drawings and software).

2. Build required ground-test and flight-test aircraft.
3. Validate and certify the design (developmental/operational tests and systems
qualification).
4. Verify JSF contract specification requirements.

Like the SWAT effort, the team took a multipronged approach. First, scope was
defined, and realistic cost and schedule estimates were made for all known cost
threats. These, together with all remaining baseline tasks, were comprehensively
reviewed by the joint leadership team to determine which tasks were essential
and to what extent. Recommendations to add, reduce, or eliminate tasks were
made by the entire multi-IPT team in order to preserve balance. Although this
continuing effort identified on the order of $1 billion in task reductions, the cumu-
lative increases from cost threats more than negated the savings in the end.
Second, the team developed a set of candidate cross-cutting initiatives that
challenged the underlying premises, ground rules, and assumptions built into
the original SDD plan. Candidate initiatives ranged widely, from a management
organizational structure to certification practices to manufacturing quantities.
The majority of remaining resources at the time were devoted to produce and
test the last of the flight-test aircraft and structural test airframes, so premises
and roles for those efforts were scrutinized. This effort resulted in the most pro-
minent changes that were recommended and approved, namely the elimination of
two flight-test aircraft: AF-5 and CF-4.
Finally, the third prong involved a return to the CAIV process, this time
focused on the remaining development costs. The air system designs were
nearly complete or very well defined in this period, so design changes would
only increase costs. Therefore, the CAIV tradeoffs involved primarily the extent
to which the designed-in capabilities would be certified in the SDD program;
38 A. E. SHERIDAN AND R. BURNES

that is, the military utility of specific capabilities (weapons loadouts, employment
flight envelopes, and corresponding MS software functions) were ranked against
the cost, timespan, and number of test flights required to certify the capability.
Although some options defined along that spectrum did yield substantial cost
savings and schedule improvements, no significant capability deferrals were
accepted in the recommended mid-course update to the SDD completion plan.
Despite the intensity and priority of these efforts, they did not identify enough
cost savings to fully offset realistic estimates of known cost pressures. Further-
more, because the SDD test aircraft were manufactured on the same assembly
lines as the LRIP aircraft, schedule delays for the test aircraft would result in
similar delays for production. This, in turn, would delay operational testing, so
that a schedule extension for SDD was going to be required. Finally, a new 2008
program baseline was approved, OTB-2, which recognized a 12-month extension
to the SDD schedule and more than $1 billion in additional cost (Fig. 15).

3. NUNN-MCCURDY BREACH AND TECHNICAL BASELINE REVIEW

The need to re-baseline and add significant funds to the SDD program for a
second time, combined with the funding required to begin production, resulted
in increased scrutiny of the program by the government beyond the JSFPO. In
the aftermath of OTB-2, continuing schedule challenges with both manufacturing
and flight tests, and evolving policies and priorities within the government, even-
tually led to a critical breach of Nunn-McCurdy cost-growth thresholds and a
third major OTB.
The F-35 PEO initiated an Independent Manufacturing Review Team (IMRT)
in 2009 led by former F-35 PEO RADM (Ret.) Craig Steidle. This team consisted
of government and industry experts and was chartered to evaluate current
program plans, funding, manning, and facilities and to assess whether they
were adequate to achieve planned production ramp-up and sustain the predicted
maximum production rates. The IMRT review was to encompass program man-
agement, product definition, parts fabrication, assembly and test, supply chain
management, and global sustainment. The IMRT conducted reviews in 2009
and 2010, and their recommendations included a more formal production inte-
grated master plan, increased emphasis on affordability, more timely funding,
and improved management metrics. Overall, the IMRT affirmed the production
system and plan but concluded that production quantities should be limited to
no more than a 50% increase year to year.
As the OTB-2 program plan was being formally approved and implemented
as the new baseline, several DoD agencies were developing independent cost and
schedule estimates. In order to improve clarity of the results and minimize
resources drawn from the JSFPO and industry teams, the PEO insisted that these
agencies join forces to form a single independent review team with a single
outcome. An acquisition decision memorandum was issued naming the Office of
the Secretary of Defense (OSD) Cost Analysis Improvement Group (CAIG) as
F-35 PROGRAM HISTORY: FROM JAST TO IOC 39

the lead of a joint team that included other OSD offices, NAVAIR, the Air Force
Cost Accounting Agency, and ASC. The joint team, which became known as the
Joint Estimate Team (JET), was initially chartered to estimate program costs and
requirements for fiscal years 2010 through 2015 in the Future Years Defense
Plan and the president’s FY2010 budget, but later the charter was extended to
cover the costs of the entire program, and the team eventually had an important
role in the 2010 Nunn-McCurdy recertification. The team consisted of approxi-
mately 25 personnel with expertise in cost estimating, scheduling, and various tech-
nical disciplines. A similar joint assessment team was separately chartered in 2009
to review the F135 engine program at Pratt & Whitney and Rolls-Royce.
Throughout 2008 and 2009, the JET conducted a series of contractor visits to
collect data for its independent cost and schedule estimates. Detailed reviews at
Lockheed Martin (with teammates in attendance) were typically two to three
days in duration and occurred at three- to six-month intervals, with extensive
data exchange and numerous side meetings between reviews. The JET also
made site visits to the seven highest-value system suppliers, both engine manufac-
turers and the Edwards AFB test site. As the JET estimates evolved, feedback was
given to Lockheed Martin with the objective of confirming the JET’s understand-
ing of the facts presented. The data reviews were focused on development at this
time, and were quite comprehensive, covering technical risk, engineering staffing,
drawing productivity and change volume, software productivity and growth,
laboratory capacity, supplier staffing, flight-test productivity, schedule risk,
test-aircraft manufacturing productivity and timespan, labor rates, and material
and systems procurement costs.
Although JET SDD estimates were not public, they were significantly and
unsurprisingly greater than those built into the joint Lockheed Martin/JSFPO
OTB2 baseline due to the differing bases for the estimates. The JSF program
was initially founded on a number of acquisition reform practices designed to
separate it from historical programs—this out of necessity for a rapid, affordable
recapitalization of the western fighter fleet. In contrast, the JET estimates were
firmly based on those same historical programs.
However, by October 2009 cost and schedule pressures against the OTB-2
baseline were again building, and it was evident within the program that more
software and flight test resources were needed. Again, detailed joint reviews
were held, but this time with a different mandate. In the efforts preceding
OTB-2, the mandate had been to not provide more money, but the new
program priority was the low-risk completion of the remaining program. Joint
teams worked through the summer of 2010 to recommend a new baseline that
included an additional software test line, an additional dedicated flight test air-
craft, and the temporary use of several production aircraft for flight tests.
Although the focus of the JET reviews and data gathering had been on the
SDD program, the CAIG also maintained independent estimates of production
URF costs on a routine basis using its own methodology. These estimates were
used within the OSD and were typically conservative (i.e., high) compared to
40 A. E. SHERIDAN AND R. BURNES

Fig. 16 Cost growth history and breakdown.

program estimates reported in the annual F-35 Selected Acquisition Report

(SAR). However, when Congress enacted the Weapons System Acquisition
Reform Act of 2009, the Cost Analysis and Program Evaluation (CAPE) organiz-
ation was established. This organization absorbed the former CAIG and assumed
the cost-reporting responsibility for major defense acquisition programs like
the F-35. As a result, the 2009 SAR cost estimates, based on the JET SDD estimate
and the CAIG production estimate, increased significantly compared to the prior
year’s estimate, despite the fact that F-35 aircraft configurations had actually
remained stable since 2005.
Figure 16 depicts the trend of reported F-35 Program Acquisition Unit Cost
(PAUC) over the period of the SDD program to date, compared to the original
acquisition program baseline. PAUC includes development, production (URF),
and support system costs on a per-unit basis. The trend of increasing costs consists
of three distinct periods. In the early program, air vehicle weight and complexity
had increased until the SWAT effort arrested and partially reversed the trend,
resulting in a net increase in URF that drove a steep increase in PAUC. In the
middle period, SDD costs rose significantly (OTB-1 and OTB-2), causing a
steady but shallower increase in PAUC. Finally, more conservative estimating
methodology in the CAPE estimates for both SDD and production resulted in a dis-
crete upward step in PAUC reflected in the 2009 SAR, driven primarily by URF
estimates rather than SDD. In April 2010, this step, along with other similar cri-
teria, triggered the formal declaration of a critical breach of cost-growth limits
defined by the Nunn-McCurdy Act and resulted in the third program re-baseline.
A Nunn-McCurdy critical breach has major implications for a program [12],
subjecting it to a detailed review for potential termination. If not terminated, the
DoD was required to certify to Congress that: 1) the program was essential to
F-35 PROGRAM HISTORY: FROM JAST TO IOC 41

national security; 2) there were no lower-cost alternatives; 3) the CAPE had deter-
mined new cost estimates to be reasonable; 4) the program had higher priority
than others from which funding would be taken; and 5) management was able
to control additional cost growth. In addition to other requirements, the statute
required that the program be restructured in a way that addressed the root
cause of the cost growth.
Immediately following the critical breach, the DoD established IPTs to address
each of the five recertification requirements, chartered a sixth team to conduct a
comprehensive Technical Baseline Review (TBR), and initiated a root-cause
analysis. The recertification requirements were satisfied, and on 2 June 2010 the
DoD issued an acquisition decision memorandum certifying the F-35 program.
Immediately afterward the TBR began to engage, with more than 100 government
personnel organized into five teams: Air Platform, Mission Systems, Test and
Evaluation, Service Integration, and System Acceptance.
The TBR’s stated objectives were:

1. Assess the planning baseline to ensure that cost and schedule planning
reflected the technical scope of SDD and was adequate to execute the
program.
2. Assess the technical planning for gaps (i.e., risks, issues, or other concern
areas) to ensure that resolution or mitigation was covered on a technical basis.
3. Provide a final assessment three to four weeks prior to the November 2010
Defense Acquisition Board (DAB) meeting.

The team addressed the Lockheed Martin air system, the Pratt & Whitney propul-
sion system, and other government costs for SDD. TBR products for each team’s
domain were three-point cost estimates (best case, most likely, and worst case),
a recommended schedule and risk assessment, and technical findings regarding
gaps and risks. Although the TBR objectives were aligned in general with the
ongoing replanning activity already well underway within the program, the com-
bined result constituted the largest (and last) replan of the program (OTB-3). It
stretched the completion by an additional 36 months for tasks within the SDD
contract scope. The TBR also identified gaps in the SDD scope and added
tasks, most notably a third lifetime of structural durability testing. This resulted
in the addition of a further 21 months to the SDD period. The overall increase
for the Lockheed Martin contract cost was greater than $6 billion (Fig. 15). Fol-
lowing this third replan, implemented in 2011 as the program baseline, perform-
ance tracked well to the plan.

F. OTHER SELECTED KEY CHALLENGES

A development program of the magnitude and complexity of SDD encounters and
must overcome more challenges than can be covered in a paper such as this.
42 A. E. SHERIDAN AND R. BURNES

However, the following are brief summaries of some issues encountered that are
representative of the types of problems overcome and that had significant effects
on the course of SDD or high-profile impacts on other program elements.

1. STOVL PROBATION
As program personnel were working through the implementation of the OTB-3
baseline in early 2011, the CTOL and CV test aircraft were exceeding planned
flight rates. However, the STOVL fleet experienced low rates due to a combination
of unrelated development issues associated with the STOVL-unique propulsion
system. Around the same time, the STOVL structural durability test airframe
experienced cracking in the main wing carry-through bulkhead within the first
lifetime of testing. Figure 17 locates components affected by these issues in the air-
craft. Concerns about these issues themselves, as well as potentially unacceptable
consequences to weight/performance and maintenance time/cost, prompted the
government to impose a two-year probation on the STOVL variant. In January
2011, then –Secretary of Defense Robert Gates announced the probation, or
“period of increased scrutiny,” and stated that, if the issues could not be resolved
in that time, the STOVL variant should be cancelled [13]. As a result, the CTOL
and CV test programs would no longer depend on the STOVL as the lead variant

Fig. 17 F-35B components with issues triggering STOVL probation.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 43

in development. In addition, the production quantity for STOVL in the current

LRIP-5 contract was cut from 13 aircraft to only 3, a reduction of 16 aircraft
from the previous contract.
Rapid progress in resolving the identified issues and improving the STOVL
flight rate, together with a highly successful first deployment of two STOVLs
to an L-class ship, resulted in lifting of the probation after just one year. On
20 January 2011, Gates’s successor Leon Panetta, at the NAS Patuxent River test
site, stated, “The STOVL variant is demonstrating the kind of performance and
maturity that is in line with the other two variants of the JSF.” Panetta said, “The
STOVL variant has made sufficient progress so that as of today, I am lifting the
STOVL probation” [14]. Summaries of the key development issues that led to
the probation and their resolutions follow.
The STOVL durability test article developed a significant crack on the main
wing carry-through bulkhead early in the first life of testing. Prior to the start
of durability testing, Lockheed Martin predicted the number of findings expected
during the test program based on legacy aircraft of each variant type. The actions
called out in the government’s probation letter were already part of the durability
test plan. All findings in the test were evaluated and correlated to determine the
life of the part against the design spectra. Parts with life deficiencies were rede-
signed to full life and incorporated into the first available production lot, following
the standard configuration and weight management procedures. Any required
modification designs for aircraft already fielded were created and programmed
for implementation prior to the effective flight hours being reached in the field.
In this case, the finding was correlated to the test spectra to calculate an
updated service life for the component. A design change involving sculpting the
area where the crack initiated to reduce stress concentration was incorporated
into production at LRIP-4. The change had a negligible effect on weight. For
the 13 production aircraft already built, a modification was retrofitted. This
involved similar blending of material at the stress concentration point, as well
as the addition of external straps to reduce gross stresses. These added approxi-
mately 70 lb to weight-empty. The retrofit was scheduled for each affected aircraft
prior to reaching 577 flight hours. Modification on the SDD test aircraft required
only the local blending to provide adequate life extension to finish the flight
test program.
In 2010 the test team identified a vibration issue with the Auxiliary Air Inlet
(AAI) door during high-speed semi-jetborne testing. A redesign of the AAI door
quickly ensued. While the AAI door was being redesigned in 2011, the NAS
Patuxent River team continued to expand the STOVL mode flight envelope in
areas not impacted by the door vibration issue. In fact, the team was able to
expand sufficient semi-jetborne and jetborne flight envelope that year to allow
the initial developmental sea trial aboard the U.S.S. Wasp in October 2011 with
two test aircraft [15]. After completing the successful sea trial, a redesigned
AAI door system was installed on BF-1 later that same year. The redesign was suc-
cessfully verified through regression flight testing.
44 A. E. SHERIDAN AND R. BURNES

During flight testing, also in 2010, an analysis of thermal data projected that
the roll-post-nozzle actuators would exceed their maximum temperature capa-
bility when the aircraft was subjected to the required 1% hot day (1208F) con-
ditions. The additional heat was attributed to higher-than-expected leakage
around the roll-post nozzle. A series of parallel actions was taken to protect the
aircraft during flight test and alleviate operational limitations:

1. Establish lower temperature thresholds on flight test monitored aircraft.

2. Establish operating time and ambient temperature limits on unmonitored
flight test and fleet aircraft.
3. Add insulation to the actuator housing.
4. Redesign the roll-post actuator for higher temperature capability.

The insulation was introduced to production aircraft in 2011 and subsequently

removed in 2015 when Rolls-Royce introduced the increased-capability actuator
to production. Despite extensive operations conducted in one of the hottest
areas of the United States (Yuma, Arizona), there have been no temperature-
related failures associated with the roll-post-nozzle actuators.
Throughout the development of the SDLF concept, driveshaft linear and
angular deflection was a driving design consideration. The flex couplings at both
ends of the driveshaft are designed to flex as the aircraft maneuvers and thermally
expand and contract during a mission. The combination of higher-than-expected
thermal growth and manufacturing variation resulted in conditions under which
the driveshaft flex couplings were predicted to stretch or compress beyond the
designed axial deflection limits. For the SDD flight test aircraft, until a design
fix could be put in place, a gap measurement between the engine flange and dri-
veshaft flange was performed to establish a manufacturing variation datum for
axial deflection. Using various models, prediction tools, and real-time aircraft
and engine flight test instrumentation, an axial tension and compression
margin estimate was made available for real-time monitoring in the flight test
control room. This monitoring typically constrained operations when the aircraft
was hot (high fuel temperatures) and the engine was at a low power setting. An
interim design improvement was introduced in 2011 in which a classed spacer
was placed between the engine and driveshaft flanges to optimize the gap.
Although this solution was relatively simple and quick to implement, it intro-
duced undesirable logistics and maintenance complexity. In 2016 a new, increased
axial capability flex coupling was introduced that accommodated deflections
throughout the full F-35B operating range and alleviated the additional mainten-
ance burden of the classed spacers.
During F-35B conventional mode flight testing as early as 2008, elevated
clutch housing temperatures were observed that correlated with uncommanded
lift fan rotation. A Lockheed Martin, Pratt & Whitney, and Rolls-Royce team
conducted a root-cause/corrective-action activity and concluded that the
F-35 PROGRAM HISTORY: FROM JAST TO IOC 45

uncommanded lift fan rotation had been caused by tight tolerances between
clutch plates on newly built clutches. The resultant friction between the plates
caused the clutch-case temperature to increase throughout the flight and poten-
tially exceed design limits. With no immediate solution available to keep the
clutch plates separated, Lockheed Martin developed a passive cooling modifi-
cation to the existing active cooling system. In 2011, a clutch thermal monitoring
system was also added to provide pilot awareness of clutch thermal state during
both the uncommanded lift fan rotation and pilot-commanded STOVL conver-
sions. In parallel with the cooling modifications, Rolls-Royce initiated an effort
to thin the clutch plates to reduce clutch drag. This change was incorporated
into production clutch deliveries beginning in 2014. Because this change
reduced clutch life, Rolls-Royce implemented a more durable clutch plate material
in 2016 that restored the maximum number of clutch engagements to exceed the
specification level. Since these design changes were implemented, there have been
no reported up-and-away clutch heating events.

2. CV ARRESTING HOOK REDESIGN

One of the fundamental features of the CV variant allowing arrested landings on a
carrier is the Arresting Hook System (AHS). The F-35C AHS was among the most
challenging systems to design and one of the last to mature sufficiently to com-
plete CDR. The requirement to completely enclose the system when retracted,
combined with the relatively forward location of the engine nozzle, resulted in
a configuration with the hook point longitudinal position considerably closer to
the main landing gear than in any legacy aircraft. When roll-in testing of the
system began at the USN’s Lakehurst, New Jersey, facility during 2011, and follow-
ing initial fly-in tests, results demonstrated an unacceptably poor engagement
rate. The low rate was due to the hook point bouncing over the arresting wire,
caused by the unique physical dynamics of the system. Nose and main landing
gear tires excited a wave motion in the wire that resulted in the center of the
wire being on or close to the deck just as the hook point passed. Together with
a relatively low hold-down force in the system that allowed the hook to bounce
upward, and a blunt hook point, the hook often failed to catch the wire.
This problem led to a major redesign of the AHS [16], but fortunately did not
have a significant effect on the surrounding airframe, systems installations, or
door arrangement. The changes aimed chiefly to flatten and sharpen the
leading edge of the hook-point shoe to improve the pickup of the wire and signifi-
cantly increase the hold-down force on the hook to reduce bounce-up. A new
CDR was conducted for the system, and testing resumed in early 2014, leading
to ship trials later that year with excellent engagement rate results. By the end
of 2014, 16 CV airplanes had already been delivered. Those, and an additional
two aircraft, would ultimately be delivered before the change was implemented
on the production line, requiring retrofits of a total of 18 aircraft.
46 A. E. SHERIDAN AND R. BURNES

3. FUEL TANK INERTING REDESIGN

Challenges related to the Onboard Inert Gas Generation System (OBIGGS) also
began to emerge in 2012 that eventually affected SDD flight tests, LRIP pro-
duction, and retrofit modifications that paced USMC and USAF IOCs.
F-35 air vehicles rely on the OBIGGS for lightning protection in order to avoid
weight penalties associated with passive systems. Unlike aircraft with aluminum
skins that create a Faraday cage, keeping currents induced by lightning strikes
on the exterior of the aircraft, a composite-skinned aircraft like the F-35 experi-
ences those currents on the external surface as well as within the aircraft structure.
This is primarily due to the fact that lightning-level currents will attach to the
nearest metal fastener and penetrate the skin. This was well-understood early in
the aircraft design, and features were included to protect against direct effects
due to lightning strikes. Indirect effects were also mitigated to ensure that com-
ponents withstood currents induced in the aircraft wiring system. The remaining
lightning-strike risk is the ignition of fuel system ullage (vapor above the fuel level)
if arcing occurs as a result of a lightning strike. Available passive measures to
prevent arcing would have resulted in unacceptable weight, so requirements
were placed on the OBIGGS to maintain low enough oxygen concentrations in
the fuel tank ullage to prevent ignition. These requirements were more demanding
than other vulnerability requirements, so flow rates were adjusted to ensure that
the ullage remained below the oxygen concentration requirement. However, at a
special CDR for the system in 2012, it was determined that the system did not
meet the requirement at some locations for short-duration transients. Therefore,
several modifications to the design, already in production, would be needed to
satisfy the program’s safety hazard risk-index requirements for all variants.
The required changes involved redesigns of valves and orifices and the
addition of wash lines for nitrogen-enhanced air to certain locations in fuel
tanks. Although changes to the system architecture were modest, some of the
new components were installed in locations that were difficult to access once
the aircraft was assembled. Furthermore, some of the components were entirely
new, requiring a complete design-qualification cycle in addition to procurement
and manufacturing spans before the new hardware would be available. Further-
more, fuel-system software changes were needed. Therefore, the change
implementation was divided into two phases. The first phase covered the com-
ponents that were the most difficult to access and was incorporated at the
soonest possible effectivity in LRIP-6 production, while the remaining hardware
was targeted at LRIP-7 effectivity. A retrofit plan was developed for all
prior aircraft.
However, in 2014, a new issue related to the OBIGGS emerged. It was deter-
mined that, at certain fuel states and during high-g maneuvers, the arrangement of
the OBIGGS and fuel siphon tanks could result in fuel tank pressures that
exceeded design limits. The immediate effect of this discovery was the imposition
of maneuver and weather limitations on the fleet. Fuel-system software could be
F-35 PROGRAM HISTORY: FROM JAST TO IOC 47

modified to reduce tank pressures, but at the expense of inerting performance.

For the F-35B STOVL aircraft, such a software change resulted in small pockets
with oxygen concentration above the limit. However, ignition laboratory tests
were performed at Wright-Patterson AFB, Ohio, that produced pressure data
confirming that the structure could withstand such ignition events. This solution
was verified using the fuel-system simulator and flight tests, but production
implementation was delayed until LRIP-8. The verification of the solution, acqui-
sition of parts, and modifications of 10 F-35Bs were just in time for USMC IOC
declaration in July 2015.
For the F-35A CTOL configuration, the tank overpressure condition was actu-
ally more severe than on the STOVL due to differences in the fuel system arrange-
ment and maneuver requirements between the variants, and its schedule was only
slightly less pressing. The CTOL benefitted from the work done on STOVL, but
the CTOL solution required the incorporation of an all-new software-controlled
pressure relief valve, and a new wash line routed from the outboard wing tank
all the way to near the center of the fuselage. This change would not be incorpor-
ated in production until midway through LRIP-9 production. Like the STOVL,
depot modifications to the required number of F-35A aircraft were only com-
pleted just before the USAF IOC declaration in August 2016.
The F-35C CV fuel system configuration is similar to the F-35B, but the ulti-
mate OBIGGS modification was somewhat different than either the STOVL or
CTOL. Production effectivity for the change is LRIP-10, and as with the other
variants, retrofits will be installed in previously delivered F-35Cs to relieve
flight restrictions.
As a result, the F-35 is now fielded with no inflight restrictions related to
lightning.

4. HELMET AND SEAT REDESIGNS

The F-35 Helmet-Mounted Display (HMD) system [17] and US16E ejection seat
[18] each has unprecedented capabilities (Fig. 18). Each has overcome significant
development challenges in its own right, and the progress of each development
effort has been intertwined with the other.
The HMD serves as the pilot’s primary display, providing integrated flight
reference information, tactical and navigational display integration, and digital
imagery, including night vision. The system development during SDD has
involved several iterations. Early versions faced functional performance issues
with night vision acuity, tracking/boresight alignment and latency, display
jitter, green glow, and obscuration by the canopy bow frame. In addition, in
tests of the ejection seat, issues arose with the security of the visor attachment
and yawing moments exerted on the pilot’s head due to the oxygen hose.
Finally, qualification tests identified the need to modify the helmet transmitter
unit and helmet/vehicle interface cable. Based on the Gen II version of the
system, the system was deemed deficient to requirements and the 2010 TBR
48 A. E. SHERIDAN AND R. BURNES

Fig. 18 HMD and US16E ejection seat.

found the system not suitable to complete SDD or fleet operation. Work contin-
ued on the system to develop a Gen III version to address these issues, but as a risk
mitigation fallback, Lockheed Martin pursued an alternate system using separate
night vision goggles.
Numerous changes incorporated into the Gen III design successfully
addressed most of the issues, and the fallback option was eventually dropped.
Night vision capability was improved with a new camera and software changes,
helmet display unit hardware improvements, and the addition of a fixed-camera
assembly. Tracking improvements resulted from the addition of a boresight reticle
unit to allow the pilot to see alignment status, optical trackers to the HMD and
fixed camera, and an inertial measurement unit to the helmet. In addition,
several other hardware and software improvements were made. The visor attach-
ment was strengthened, and an oxygen hose attachment was added to minimize
head turning. Finally, the helmet transmitter unit and helmet/vehicle interface
cable were strengthened. Although these changes together addressed the func-
tional performance shortfalls of the Gen II design, they also added head-borne
mass to the helmet and shifted the center of mass forward.
The US16E ejection seat was also designed to meet an unprecedented level of
requirements. The system is designed to function safely over a wide range of flight
conditions, from static hover very near the ground to high altitudes and very high
equivalent airspeeds, and over virtually any attitude. Moreover, the system was
F-35 PROGRAM HISTORY: FROM JAST TO IOC 49

designed to accommodate a very wide range of pilot shapes and sizes, from a
103-lb female to a 245-lb male. Simply stated, the key design challenge for the
seat is that high forces are required to eject and then decelerate heavy pilots at
extreme conditions, but these forces impose the risk of neck injury on light
pilots. These risks are increased with increased mass of the pilot’s helmet and
center of mass that is not aligned with the seat forces. High neck loads can
occur during the ejection and parachute opening phases.
Prior to mid-2015, the ejection seat had been qualified with a Gen II helmet,
but a repeat test at low speed with a light pilot, performed in support of redesign-
ing the seat sequencer, showed that neck injury criteria were exceeded, in contrast
to earlier tests. A review of the earlier test data revealed that the test mannequin’s
head was being supported by the parachute riser at the critical load condition,
giving misleadingly low measured neck loads. In late August 2015, U.S. services
imposed a minimum weight limit of 136 lb for F-35 pilots. In response to this,
two improvements were developed to the ejection seat. First, a lightweight
aircrew switch was added to the seat. Selecting the lightweight pilot position
adjusts the timing of the drogue chute sequence, lowering the speed of main
chute deployment. Second, a fabric head support panel was added between the
parachute risers to prevent hyperextension of the neck (whiplash).
Both of these changes mitigate neck loads in the parachute opening phase of
seat operation but do not affect the catapult stage. To address the catapult stage, a
weight limit was established for the HMD that would satisfy neck-injury criteria
for light pilots. In May 2015 the JSFPO directed that a Gen III Light version of the
HMD/helmet be developed. The primary weight reduction is achieved by the
introduction of a missionized visor. This removes the external tinted visor and
introduces a two-visor system with one clear display visor and one tinted
display visor. The pilot will be able to swap the visor on his or her helmet while
in flight to adjust to changing environments.
The production effectivity for the revised ejection seat is LRIP Lot 10, which
commenced aircraft deliveries in January 2018, and the Gen III Light HMD began
production deliveries.

G. KEY RESULTS: NOTABLE FULL-SCALE TEST ACCOMPLISHMENTS

Although the ambitious SDD program endured numerous significant technical,
schedule, and cost challenges, the F-35’s equally ambitious 5th-Generation
system capabilities have been successfully achieved, based on extensive full-scale
test results.

1. FLIGHT TESTS
Reference [15] provides a complete summary of F-35 flight tests. The test program
was conducted by the F-35 Integrated Test Force (ITF) composed of engineering,
flight operations, maintenance, and management personnel from Lockheed
50 A. E. SHERIDAN AND R. BURNES

Martin, Pratt & Whitney, USAF, USN/USMC, international partners, and suppli-
ers, as needed, in a single integrated organization. Two primary test sites provided
extensive base and test-range infrastructure: Edwards AFB and NAS Patuxent
River. Numerous other test locations provided specialized test capabilities, includ-
ing L-class amphibious assault ships and CVN-class carriers. Figure 19 highlights
a few significant flight test milestones.
In April 2018 the final SDD test flight was completed. Overall, more than 9000
test flights accomplished more than 65,000 test points in more than 17,000 hours
of testing over nearly 10 years. In recognizing the joint government/industry
team, Lockheed Martin’s program manager stated that the F-35 flight test
program represented the most comprehensive, rigorous, and safest developmental
flight test program in aviation history. SDD flight testing highlights included full
flight envelope performance and flying qualities, high AOA, STOVL development
testing, ship trials, 183 weapon separation tests, 42 weapons delivery accuracy
tests, and 33 mission effectiveness tests, which included numerous multiship mis-
sions of up to eight F-35s against advanced threats.

2. FULL-SCALE GROUND TESTS

In addition to flight testing, numerous ground-based tests using full-scale hard-
ware contributed to validating models and verifying requirements. A sample of
ground tests performed is illustrated in Fig. 20.
For each variant, two full-scale airframes were produced for structural tests on
the same assembly line as that used for flight test and production aircraft. One of

Fig. 19 Notable flight test milestones: AA-1 first takeoff, BF-1 first vertical landing, X-35B
bomb drop, and successful F-35C arresting wire engagement.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 51

Fig. 20 Notable full-scale ground tests: F-35A gun-fire, antenna model, F-35C drop, F-35A
live fire, F-35B climatic, and RCS model.

each variant’s airframes was used for static testing to confirm the strength
and stability of the structures. The other was dedicated to durability testing,
which was extended to three lifetimes as a scope addition included in OTB-3.
The CTOL tests were performed at BAE Systems’ Brough, United Kingdom,
facility, as was the CV durability test. The STOVL tests were done at Lockheed
Martin’s facility in Fort Worth, as was the CV static test. The CV static test
article was also used for drop testing to verify structural integrity for extreme-
sink-rate carrier landings at Vought Aircraft Industries in Grand Prairie,
Texas [11].
Full-scale live-fire vulnerability tests were performed at the Naval Air Warfare
Center Weapons Division at NAS China Lake, California, using the fully equipped
AA-1 CTOL aircraft, the CV structural static-test airframe with an engine
installed, and the STOVL structural static-test airframe. Ballistic testing of the
STOVL propulsion system was also conducted.
Extensive climatic tests were performed in 2014 and 2015 on a fully equipped
STOVL aircraft at the McKinley Climatic Laboratory located at the Eglin AFB,
Florida [19]. These tests covered a wide range of climatic conditions, but also a
wide variety of F-35 flight conditions, including simulated hover for the
STOVL variant.
Although they did not include a representative airframe structure, other
full-scale tests involving salient flight hardware were accomplished using pole
models for signature and antenna aperture testing. The full-scale signature
model included flight-representative versions of all salient features and was
tested at the Lockheed Martin Helendale, California, facility. Full-scale integrated
antenna tests were conducted at the USAF Research Laboratory Rome Research
Site in Rome, New York.
52 A. E. SHERIDAN AND R. BURNES

H. ALTERNATE ENGINE PROGRAM

In addition to the lead engine, the Pratt & Whitney F135, in SDD the JSFPO con-
tracted Lockheed Martin to integrate into the F-35 air system an alternate General
Electric/Rolls-Royce F136 engine. The rationale was that competitive engine
offerings would spur continuous development and reduce price as the respective
engine manufacturers competed for work share, as has been exhibited with F-16
engine competitions. To those ends, the F-35 was designed not only to be compa-
tible with the F136, but also to allow the physical interchange of the F135 and F136
engines in the field, operating in such a way that propulsion operations were
transparent to the pilot regardless of engine type.
Therefore, the F-35 inlet was sized to support either engine, with growth
margin. Engine interfaces were closely managed to ensure compatibility with
either engine. Lift-system hardware (lift fan, driveshaft, and roll posts) and even
the CTOL and STOVL engine exhaust ducts and nozzles were intentionally estab-
lished as common hardware, provided by Pratt & Whitney/Rolls-Royce for both
the F135 and F136. Specifically, the F135 and F136 shared a number of identical
interfaces, such as the front engine mounts, hydraulic mounts, and Engine
Starter/Generator (ES/G) pumps. Some flexibility in F-35 interfaces was designed
in to accommodate minor location differences in the mounting locations of the
rear engine side mount and power feeders to the ES/G. Portions of the tradition-
ally airframe-owned main fuel line and electrical panels were provided to the pro-
pulsion system contractors to be engine-mounted to allow freedom in the
terminal routing of those lines.
To provide Level 1 flying qualities, both propulsion system contractors
were responsible for meeting nontraditional control requirements, especially
for STOVL operations. For example, when operating in STOVL mode the pilot
commands forward and vertical acceleration with stick and throttle inputs.
F-35B control laws command the engine system to provide a required thrust
level, thrust split between the main-engine nozzle and lift fan and the roll-post
command. In many respects the propulsion system is now treated like any
other aerodynamic effector. Accuracy, bandwidth, rate capability, and system
response to failure are managed by the propulsion system. For example, although
different engines may operate at different rpm at a common control point, the
effector output is a closed loop managed by the F-35 propulsion system to
provide predictable performance to the pilot.
Success in these endeavors was established at the JSF program’s August 2004
alternate engine readiness review.
Under separate JSFPO contracting, the General Electric/Rolls-Royce Fighter
Engine Team took a pre-SDD F136 engine into conventional testing in July
2004, followed by initial STOVL engine testing in 2005. After transitioning
to SDD, the F136 completed its PDR in 2006 and CDR in February 2008. The
first SDD F136 began ground testing in February 2009, followed by STOVL
ground testing in November 2010. By the end of 2010, the F136 had accrued
F-35 PROGRAM HISTORY: FROM JAST TO IOC 53

more than 1000 test hours, with flight test engines planned for delivery in 2011
and first flights planned for each of the three variants soon thereafter.
Though cancelled in 2011 after a long and highly publicized government
funding debate, technologies and lessons learned from the F136 program live
on in the joint government/industry Versatile Affordable Advanced Turbine
Engines program.

V. INTERNATIONAL PARTICIPATION
The scope and complexity of international participation in the F-35 program has
been both an asset and a challenge. From the earliest roots of the joint U.S./U.K.
ASTOVL program in the 1980s, then supported by the Nunn-Quayle Research
and Development Initiative, the program has involved international partners.
Although DoD regulations require acquisition managers to pursue international
cooperation for most programs, the F-35 SDD program became a cooperative
development arguably like no other.
For several decades, the operational model for long-term combat and peace-
keeping operations has revolved around tri-service, coalition participation for
both operational and financial burden sharing. However, the historical reality of
this concept had been difficult and was limited by technology, a lack of legacy plat-
form interoperability, and political differences among/between the services and
allied air forces. In the early 1990s, the U.S. services faced severe budget pressures,
and there was an emerging view that significant potential gains could come from
having a common platform. The elimination of interoperability barriers and
reduction of duplicative training and maintenance infrastructure would reduce
both procurement and operational budget requirements. At a lesser level, the
same dynamics existed in most of the participating allied nations.

A. BACKGROUND
In the early 1990s, each of the U.S. services was entering development programs
for the replacement of their respective frontline tactical fighter aircraft. The
USMC was furthest along and was developing prototypes for its ASTOVL repla-
cement for the venerable AV-8B Harrier. The Harrier was a British design that
had been improved and manufactured in the United States for the USMC. The
United Kingdom and United States were co-developing the ASTOVL concept
and were joint signers of the formal ORD that defined the next-generation
requirements. The Italian Navy, a strong U.S. ally in coalition operations, also
operated the Harrier and was expected to join the replacement program.
The USAF was in the early stages of developing operational requirements for a
multirole fighter to replace the multirole fighter inventory, which consisted of the
F-16, A-10, and potentially the F-117. The F-16 had been widely deployed as an
international fighter in 21 countries as the primary Foreign Military Sales (FMS)
54 A. E. SHERIDAN AND R. BURNES

offering by the U.S. government to allied air forces. Key allies requiring modern-
ization of their air forces were prime candidates for joining the JSF program.
The USN had been through several attempts to replace the F-14/A-6/A-7
inventory, including a derivative of the USAF Advanced Tactical Fighter (which
became the F-22), A-12, and AX/AFX competitions. All had been aborted in
favor of a less risky approach to upgrade and modernize the F/A-18 platform.
The F/A-18 had achieved limited foreign air force acceptance with six inter-
national air forces, including principal allies Canada and Australia. Those allies
were also expected to have an interest in joining the JSF program.

B. THE INTERNATIONAL COALITION

The United Kingdom joined the JAST program as a founding member in 1995
under the premise that British industry would suitably share in the industrial
benefits accruing from participating early in a development program that would
lead to substantial production and sustainment industrial opportunities. Italy,
the Netherlands, Denmark, Norway, Canada, and Turkey joined the program at
various times as observers during the CDP program.
At the time of the SDD contract award, the United States and United
Kingdom were committed participants, and the JORD was cosigned by both
U.S. and U.K. senior government officials. The United Kingdom became the
only Tier 1 partner, with an investment in the SDD phase of about $2 billion.
Over the following year, seven additional allies joined the program (Fig. 21).
Italy and the Netherlands joined the SDD phase as Tier 2 partners, each with
an investment of more than $1 billion, and Norway, Denmark, Canada, Australia,
and Turkey joined as Tier 3 partners at about $150 million each. Based on their
financial level of participation, those allies were allowed to provide representatives
in leadership positions in the JSFPO. Tier 1 and Tier 2 partners would also have
the opportunity to participate in Operational Test and Evaluation (OT&E) of the

Fig. 21 Timeline for international cooperative partners joining the F-35 program.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 55

F-35. Each of these government-to-government bilateral agreements also had

unique elements, referred to as side agreements, that identified unique national
requirements of the participating nation. All partner countries would be inte-
grated into the baseline production program seamlessly and benefit from the
economies of scale from larger procurement quantities.
All participating nations would be allowed to participate industrially in the
SDD phase of the program and later phases on a best-value, competitive basis.
This was a significant departure from the traditional offset-based program in
which procuring nations would receive economic benefits to offset their procure-
ment costs. Lockheed Martin and its teammates conducted a comprehensive
survey of industries in partner nations for unique development capabilities that
could enhance the supplier base already selected during the CDP program.
In early 2006, the U.S. government secured the commitment of the nine
partner nations to continue in the JSF program via the Production, Sustainment,
and Follow-on Development (PSFD) Memorandum of Understanding (MoU)
signed by the parties between November 2006 and February 2007. This agreement
eliminated the level-based designations in the SDD agreements so that all nine
participating nations held equal status. The JSF Executive Steering Board
(JESB), with representation from all participants, assumed overall governance
of the program.
The program management complexity of this program structure differed from
any previous program in DoD’s history in terms of the degree of challenges
involved to design, develop, test, and produce the most advanced-technology
fighter in history, integrate the operational requirements of three separate U.S.
services with unique operational environments, include eight international part-
ners in the entire development and production process on a best-value basis,
ensure that all U.S. technology and export control requirements are adhered to,
and maintain cost and schedule commitments made in a competitive proposal
that did not include these complexities when the contract was initiated.
From the perspective of the JSFPO, the challenge was to convince the partici-
pating nations that the governance process was fair and that their unique require-
ments would be met within the boundaries of the funded program. For unique
requirements (e.g., a drag chute for Norway for landing on icy runways), separate
funding would be required. Participating nations would have representation in the
JSFPO in line with their level of investment. Participating nation acquisition
leaders would be full participants in the governance bodies directing program
decisions. Participating nations would procure their F-35s as an integrated
element of the U.S. annual buys, eliminating traditional international procure-
ment processes.
From the perspective of the Lockheed Martin program management team, the
challenge was to identify international industrial partnerships that could provide
best value to the F-35 supply chain while meeting individual national economic
and political requirements for the financial justification to continue with the
program. All economic benefit to the procuring nation was required to be
56 A. E. SHERIDAN AND R. BURNES

direct work on the F-35, as opposed to historic offset-based programs that were
allowed to use indirect or nonassociated trade. All participants wanted high-value
work on a program that was a relatively small platform. To meet this challenge,
significant partnering requirements were flowed down to all teammates and
major suppliers. Figure 22 illustrates the breakdown of airframe component
coproduction from U.S. teammates and International Industrial Participation
(IIP) suppliers.
From the perspective of the participating nations, the requirement for IIP was
imbedded in the PSFD MoU as a fundamental principle of the F-35 program, with
IIP country targets tied to the quantities procured. The agreement addressed the
best-value principles to be used and mandated that the contractors provide oppor-
tunities to partner nation industries. All partner nations insisted that Lockheed
Martin sign industrial participation Letters of Intent (LoIs) with each nation’s
ministries of defense and economic affairs before they would sign the govern-
ment-to-government agreements.
The LoIs identified four categories of IIP opportunities in the production
program:
1. Continuation as a result of being selected in the SDD phase
2. Competitive across all nations
3. Competitive strategic source-directed procurement to a nation’s industry but
subject to best-value, competitive pricing
4. Country-unique for specific capabilities that only one country required

Fig. 22 International production/coproduction of major airframe components.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 57

Fig. 23 International suppliers.

The LoIs also required Lockheed Martin to report semiannually on the IIP per-
formance of the team, which consisted of Lockheed Martin as the prime contrac-
tor, the prime teammates (Northrop Grumman and BAE Systems), and major
system and subsystem manufacturers. Figure 23 shows the international supply
chain created by the F-35 IIP program. Most of these suppliers have greatly
increased their capabilities, facilities, and equipment, as the F-35 is recapitalizing
not only the partner nations’ fighter fleets, but also their defense industries.

C. BUILDING THE F-35 INTERNATIONAL SUPPLY CHAIN: A FEW EXAMPLES

The Tier 1 partner (United Kingdom) and Tier 2 partners (Italy and the Nether-
lands) were given the opportunity to procure test jets and participate in OT&E.
The United Kingdom procured three STOVL aircraft and the Netherlands
procured two CTOL aircraft. Italy decided to forgo the investment in OT&E
jets and instead invested in a Final Assembly and Checkout (FACO) facility in
Cameri, Italy, to produce Italian and Dutch operational jets. The facility was
designed to be converted into a maintenance, repair, and overhaul facility for
European F-35s following the production program (Fig. 24).

1. FOLLOW-THE-SUN ENGINEERING
The United Kingdom, Italy, the Netherlands, and Australia were significant par-
ticipants in the F-35 digital thread design toolset. Other partner countries could
participate on a more restricted basis. The establishment of this virtual design
toolset was a complex and highly controlled infrastructure, but it allowed
58 A. E. SHERIDAN AND R. BURNES

Fig. 24 Italy’s FACO (left) and wing assembly building (right).

continuous design across multiple time zones, leveraging the time dimension of
global participation. Of note was GKN, an Australian enclave of stress engineering
experts who were significant contributors to the design activities of the program.

2. WORLD-CLASS COMPOSITE MANUFACTURING FACILITIES

F-35 manufacturing tolerance control and significant expansion in the use of
advanced composite structures demanded a next-generation manufacturing capa-
bility. New composite manufacturing facilities were established across several of
the participating nations. Turkey, the Netherlands, Denmark, Norway, Canada,
and Australia all invested in building new world-class composite production
facilities.

D. FOREIGN MILITARY SALES

Although neither part of the cooperative development program nor members of
the JESB, other nations have placed orders under the FMS program. These began
to play a significant role with the first production commitments by Israel, Japan,
and the Republic of Korea. Israeli and Japanese industries are already participat-
ing. Israel is providing wing panels and electronic components. Japan has already
rolled out the first aircraft from its own FACO. In the future, other countries may
also be allowed to participate through the traditional FMS processes.

E. IMPACT AND CHALLENGE

Keeping the Why F-35? value proposition in alignment across changing environ-
mental and political factors, and sustaining that value proposition across key
stakeholders, continue to be major strategic challenges that depend on sustaining
a clear economic benefit to participants. In many ways it is a government-to-
government commitment but is generally regarded as an industry responsibility
to sustain it, from an industrial or economic perspective.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 59

Combining this requirement with the intense day-to-day execution challenges

of a major, next-generation-technology engineering development, testing, and
production global program is a unique and unprecedented effort in its complexity
and scale.

VI. TRANSITION TO PRODUCTION

A. BACKGROUND
The JSF production system is absolutely central to achieving the program goals of
affordability and a high production rate. Much like the F-35 air system is recapi-
talizing and transforming a large portion of the west’s air-power capability, the
F-35 production system is recapitalizing and transforming the aerospace industry.
From early in the CDP program, as Northrop Grumman and BAE systems joined
the Lockheed Martin team and numerous countries, in addition to the United
States and United Kingdom, became engaged in the program, it was clear that
the production system would be global. Within the program, Global Production
System was the name given to the manufacturing organization. The envisioned
one-a-day rate would require substantial investment in plants and equipment.
Lean manufacturing was embraced as a core strategy, along with automation
and the digital thread. Producibility was a major emphasis in the air vehicle
design, although the specifics of both the air vehicle and the production system
evolved significantly over the CDP and SDD programs.
In the early concept stages of the program, there was a strong emphasis in the
air vehicle design on rapid final assembly, minimizing part count and tooling,
commonality, and the use of a precision-manufacturing capability to achieve
the geometry controls needed for 5th-Generation capabilities. As described earlier,
many air vehicle design changes had a significant impact on the production
system, but the overall objectives stayed constant, prompting the continuing evol-
ution and refinement of the production system as the aircraft configuration
stabilized.
In the CDP program, preliminary development and risk reduction of the pro-
duction system paralleled the air vehicle design. In SDD the system was actually
built, because one of the objectives of the program was to build even the first test
aircraft on production hard tooling. In fact, there was a strong financial incentive
built into the original SDD award-fee structure based on the actual cost of building
SDD and early LRIP aircraft relative to the desired cost-improvement curve
extended over the entire production program. Reference [10] describes the F-35
production system and ongoing advances.

B. EVOLVING QUANTITY PROFILE

A high degree of development-production concurrency and a steep production
ramp to high rate were built into the program from the outset of SDD, as specified
60 A. E. SHERIDAN AND R. BURNES

Fig. 25 Evolution of production quantity profile.

in the EMD CFI (Fig. 25). Within two years of the last variant’s (CV’s) first flight,
the production rate was originally to be 54 per year, including all three variants.
The production rate was planned to be 168 per year in the sixth and final year of
the LRIP phase. Based on these steep production ramp rates, the prime teammates
and the suppliers invested aggressively in plant capability, and the JSFPO funded a
significant amount of special tooling and test equipment.
As described earlier, a number of challenges with the engineering design, pro-
duction, and testing of the early test aircraft significantly prolonged the SDD
program, and each program replan attracted additional critical scrutiny. Several
factors acted to progressively delay and stretch out planned production-rate
increases in each planning year. First, the governments became wary of buying
significant numbers of aircraft before they had been more completely tested, for
fear that test discoveries would require engineering changes, which would in
turn require high-cost modifications to already fielded aircraft. The U.S. Govern-
ment Accountability Office was strongly critical of these potential concurrency
costs in its annual reports to Congress [20], as was the DoD Director of OT&E.
Second, funds committed to SDD replans were then not available in the acqui-
sition budget. Further limitation to procurement rates were related to the world-
wide financial crisis that strained many participating governments’ budgets.

C. PRODUCTION FACILITIES
The F-35 program has required very large facility investments across the globe in
new facilities, equipment, and tooling. New facilities have been constructed in
Australia, Japan, European countries, Canada, and the United States.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 61

Fig. 26 Lockheed Martin F-35 production facility in Fort Worth, Texas.

Lockheed Martin production is centered at the Lockheed Martin Aeronautics

Company headquarters in Fort Worth (Fig. 26). Throughout SDD and the first 10
LRIP lots, F-35 wing and forward fuselage assembly, mate, and FACO operations
have progressively displaced continuing F-16 production. In 2017 Lockheed
Martin produced the final Fort Worth –produced F-16 and announced that
future F-16 production would be done at its Greenville, South Carolina, facility,
essentially devoting the entire Fort Worth production facility to the F-35 and
completely revamping the arrangement and equipment of the nearly mile-long
assembly building. Major new construction at the site included component and
aircraft final finishes facilities, an acceptance test facility, and a hover pit, as
well as a complete refurbishment of the flight line run stations. Major wing carry-
through assemblies are produced at the Lockheed Martin Marietta plant formerly
used for F-22 production. Wing- and tail-surface edges with special treatments
and embedded sensors, as well as radomes, are produced at the Palmdale plant.
Subassembly work is also done at the Johnstown, Pennsylvania, and Pinellas,
Florida, locations.
Northrop Grumman assembles center fuselages and weapons bay doors at its
Palmdale facility, formerly used for B-2 production. This facility features an inte-
grated assembly line (Fig. 27) that uses automated guided vehicles to progress

Fig. 27 Northrop Grumman F-35 integrated assembly line in Palmdale, California.

62 A. E. SHERIDAN AND R. BURNES

Fig. 28 BAE Systems F-35 production facility in Samlesbury, United Kingdom.

assemblies through the flow line. Production of the complex main-inlet duct is
accomplished at the advanced fiber placement facility in El Segundo.
BAE Systems has developed its Samlesbury site with numerous all-new
facilities dedicated to the F-35 program (Fig. 28). The site produces aft fuselage
and horizontal- and vertical-tail box assemblies for all variants. The CV outboard
wing box is produced for BAE Systems by an IIP supplier in Canada. At Samles-
bury, an all-new assembly hall was constructed in three phases and contains
three overhead-rail flow lines for aft fuselage, horizontal tails, and vertical tails,
with a shared complex of precision milling machines. Each line accommodates
all variants. Assembly of the specialized STOVL aft-nozzle-bay doors is done
in a new facility adjacent to the hot-forming facility where a superplastic-
forming/diffusion-bonding process produces the door detail. Composites are
produced in a preexisting building that was completely gutted and reequipped.
An all-new highly automated hard-metal machining facility was constructed
on the site, as well as a new office building to house management, engineering,
and business operations.

D. AFFORDABILITY: URF IMPROVEMENT

Once the air system design was frozen in the SDD phase and production began,
the affordability program again evolved. Rather than air system requirements
tradeoffs of conceptual design or design-to-cost targets of detailed design, afford-
ability became driven by manufacturing performance, business performance,
continuous improvement, and business cases for change. These factors are in
play at the prime teammates’ sites, but—importantly—within the supply base as
well. Approximately 70% of F-35 URF cost is accounted for by teammates and
suppliers. As successive LRIP lots progressed, strong cost-performance incentives
were created in the form of price-focused contract negotiations and fixed-price-
incentive-fee contract types, both at the prime and subcontract levels. In fact,
the prime contract for LRIP-4 was structured as a fixed-price-incentive-fee
type, beginning with the 32nd production aircraft. Negotiations for this lot
would have been conducted before any production aircraft were delivered.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 63

Fig. 29 F-35A URF trend and BFA projection overlaid with overall production
quantity profile.

Figure 29 illustrates the steady URF cost reduction achieved by the program,
superimposed on the production quantity profile.
Manufacturing performance is improving through the use of producibility
improvements and proactive management of traditional metrics, such as defect
rates, labor hours per unit, material availability, rework, and traveled work.
Business performance involves managing internal cost structure (rates, overheads,
etc.), as well as supplier engagement. Continuous improvement using lean princi-
pals has been applied to manufacturing, engineering, and business operations.
Another element of the affordability strategy has been investment in cost-
saving design improvements. These are termed investments because significant
costs are incurred up front to develop and implement an engineering change,
but the cost savings accrue little by little in recurring production over the long
term, eventually reaching a break-even point, where they begin to produce
returns. The large planned procurement quantities for the F-35 often make
business cases for such changes compelling, with calculated overall return multi-
pliers ranging from 20:1 to 50:1 over the length of the planned production
program. During the SDD program, such affordability initiatives were
implemented as an explicit part of the contract scope. Once that scope was com-
pleted, however, it became less certain how the costs of such investments and
benefits of the returns would be balanced between the government and industry
parties, and what would be the source of investment funds. The savings for a
given change may stretch over a long period of time and be blended with the
effects of other changes, so it can be impractical to validate actual cost savings
due to that specific change. With annual procurement lots, often the timespan
and quantities in a single production contract are not sufficient to implement a
64 A. E. SHERIDAN AND R. BURNES

change and reach a break-even point. This dampened interest in continuing to

invest in affordability initiatives by industry and the government, despite the
obvious long-term benefits.
In July 2014, however, DoD, Lockheed Martin, BAE Systems, and Northrop
Grumman announced an agreement, known as the Blueprint for Affordability
(BFA), intended to drive down F-35 production costs. The agreement provided
for industry investments, with a path to recover those investments as savings to
the government are accrued. The stated objective of the program was to reduce
the purchase price of a 5th-Generation F-35 to the equivalent of a 4th-Generation
fighter by the end of the decade, with a specific target of an F-35A URF of between
$80 and $85 million (then-year dollars) (Fig. 29). Together, Lockheed Martin,
BAE Systems, and Northrop Grumman funded $170 million from 2014 to
2017, with a projected savings of $4 billion over the life of the program. The
program was renewed in 2017, BFA-II, with initial government investment
funds added to those from industry.

E. THE GROWING FLEET

The first production F-35A aircraft was accepted by the USAF in May 2011. In
July that same year, the first aircraft was delivered to Eglin AFB to begin pilot
and maintainer training. A year later, in July 2012, the first F-35B STOVL pro-
duction aircraft were delivered to both the USMC and the United Kingdom,
and that November the first F-35B was delivered to the first USMC operational
base at Marine Corps Air Station (MCAS) Yuma. By December 2013, 100
F-35s had been produced.

Fig. 30 F-35 global basing projection for 2020.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 65

As of January 2018, nearly 300 F-35s had been delivered or were in flight status
preparing for delivery in Fort Worth. Aircraft were operating at eight operational
or training bases in the United States and at five international bases. The largest
contingent was at Luke AFB, Arizona, which housed 118 aircraft for joint inter-
national pilot and maintainer training, nearly four years after the first F-35A
arrived there.
Aircraft deliveries and base standups are climbing rapidly at the time of this
writing. Figure 30 illustrates firm plans through 2020 to increase U.S. bases to
10 and international bases to 6, with double the aircraft fleet, totaling more
than 500 aircraft.

VII. OPERATIONS: ROAD TO INITIAL OPERATIONAL CAPABILITIES

In defense acquisition, the IOC is a point in time during the production and
deployment phase when a system is determined to meet the minimum operational
(threshold and objective) capabilities for the service’s stated need. The operational
capability includes support, training, logistics, and system interoperability within
the DoD operational environment. IOC is a good gauging point to see whether
any refinements are needed before proceeding to full operational capability.
The F-35 is unique in that multiple DoD services and international partners
and FMS participants had a need to declare IOC. One set of criteria could not
satisfy the needs of each customer, so each service defined its unique requirements
for IOC [21].

A. USMC F-35B IOC

In January 2012 the USMC took delivery of its first fleet aircraft at Eglin AFB for
the Marine Fighter Attack Training Squadron 501 (VMFAT-501), the Warlords.
Flight operations began later that year with the initial cadre of USMC pilots.
VMFAT-501 was designated as an F-35 Fleet Replacement Squadron (FRS) in
2010. In July 2014 the Warlords relocated to MCAS Beaufort.
Marine Fighter Attack Squadron 121 (VMFA-121), the Green Knights, were
redesignated as an F-35B squadron in November 2012 in Yuma and became
the first operational F-35 squadron. The Green Knights accepted their first
three F-35Bs and began working toward satisfying the requirements for IOC.
USMC headquarters issued a letter in 2013 identifying criteria to declare
F-35B IOC. The requirements stated the USMC required 10 to 16 aircraft and
that U.S. Marines be trained, manned, and equipped to conduct Close Air
Support (CAS), offensive and defensive counter-air, air interdiction, assault
support escort, and armed reconnaissance in concert with Marine Air-Ground
Task Force resources and capabilities. These requirements enabled the USMC
to declare IOC with an interim capability standard known as Block 2B. Similarly,
requirements were defined for the interim-capability deployable Autonomic
Logistics Information System (ALIS). The initial aircraft selected for the IOC
66 A. E. SHERIDAN AND R. BURNES

squadron were produced in LRIP-4 and had a number of retrofit modifications

required due to engineering changes implemented prior to LRIP-4, including
the invasive OBIGGS change described earlier. Estimates for completing the
modifications fell well into 2016, a year later than needed by the USMC. There-
fore, a series of reassignments of aircraft was made in April 2014 in order to
use later LRIP-4 and LRIP-5 aircraft, which reduced the aircraft modification
workload.
In order to accomplish the extensive aircraft modifications, the program
expanded modification bays at the Fleet Readiness Center – East in Cherry Point,
North Carolina, from two to six and added a contractor field team depot capability
for three aircraft at MCAS Yuma. Two intermediate airworthiness releases
were added prior to the Block 2B fleet release to enable aircraft modifications
and support training with an expanded flight envelope without weapons. The
squadron went through operational testing aboard the L-class amphibious
assault ship U.S.S. Wasp (LHD 1) in May 2015 (Fig. 31). Additionally, they com-
pleted an operational readiness inspection to ensure that they were capable of
performing the five IOC operational mission scenarios.
The USMC declared IOC on 31 July 2015. A comment from General Joseph
Dunford, the Commandant of the Marine Corps, provides a summary of the
events leading up to IOC from the USMC’s perspective.

I am pleased to announce that VMFA-121 has achieved initial operational

capability in the F-35B, as defined by requirements outlined in the June
2014 Joint Report to Congressional Defense Committees. VMFA-121 has

Fig. 31 First F-35B OT&E session aboard U.S.S. Wasp.

F-35 PROGRAM HISTORY: FROM JAST TO IOC 67

ten aircraft in the Block 2B configuration with the requisite performance

envelope and weapons clearances, to include the training, sustainment capa-
bilities, and infrastructure to deploy to an austere site or a ship. It is capable of
conducting close air support, offensive and defensive counter air, air interdic-
tion, assault support escort and armed reconnaissance as part of a Marine Air
Ground Task Force, or in support of the Joint Force. Prior to declaring IOC,
we have conducted flight operations for seven weeks at sea aboard an L-Class
carrier, participated in multiple large force exercises, and executed a recent
operational evaluation which included multiple live ordnance sorties. The
F-35B’s ability to conduct operations from expeditionary airstrips or sea-
based carriers provides our Nation with its first 5th generation strike
fighter, which will transform the way we fight and win. The success of
VMFA-121 is a reflection of the hard work and effort by the Marines in
the squadron, those involved in the program over many years, and the
support we have received from across the Department of the Navy, the
Joint Program Office, our industry partners, and the Under Secretary of
Defense. Achieving IOC has truly been a team effort. [22]

Since declaring IOC, USMC F-35B squadrons have been on an increasing tempo
of deployments over the past few years, including large-force exercises like Red
Flag and Cope Thunder. They have conducted shipboard and expeditionary oper-
ations and, in 2018, they executed their first operational ship deployments aboard
the U.S.S. Wasp and U.S.S. Essex, both in the Pacific. VMFA-121 relocated to
Iwakuni, Japan, in January 2017, where it is now permanently based with 16 air-
craft. The USMC currently has three operational squadrons: VMFA-121,
VMFA-122, and VMFA-211, in addition to the training squadron VMFAT-501.
Formal F-35B OT&E will be conducted by VMX1 with a detachment of six
F-35Bs stationed at Edwards AFB.

B. USAF F-35A IOC

Air Combat Command (ACC) produced the USAF IOC criteria in June 2013
and updated them in January 2015. USAF defined a required schedule for IOC
declaration by 1 August 2016 (objective date) and no later than 31 December
2016 (threshold date). USAF requirements defined a set of mission capabilities
that required the interim capability standard, Block 3i, a later standard than
that required by the USMC for IOC. The mission set included basic CAS,
interdiction, and limited Suppression/Destruction of Enemy Air Defense
(SEAD/DEAD) in a contested environment. The requirements called for the
SDD program-of-record weapons. Inherent in these mission definitions was the
capability to fly throughout the Block 2B flight envelope, day and night, in or
out of adverse weather, and to carry and employ AIM-120 missiles, 2000-lb
joint direct-attack munitions, and 500-lb laser-guided bombs. A total of 12 to
24 deployable aircraft were required with associated support equipment, spares,
verified technical manuals and training programs, and trained pilots and
maintainers.
68 A. E. SHERIDAN AND R. BURNES

The 34th Fighter Squadron at Hill AFB, Utah, began receiving F-35s in the
third quarter of 2015. The squadron’s aircraft, to meet the minimum quantity
of 12, were delivered between the third quarter of 2015 and second quarter of
2016. Seven of these aircraft were LRIP-7 jets and the remainder were LRIP-8
jets. In the months leading up to IOC declaration, several challenges were
addressed by the F-35 enterprise, including on-schedule completion of the aircraft
modifications required to support the Block 3i configuration and development,
and the testing of a Block 3i software release with software stability characteristics
deemed acceptable for warfighting needs, along with the release of the next version
of ALIS, 2.0.2.
A key challenge to achieving IOC was overcoming instabilities experienced
with the Block 3i MS software that required resets of the software. The MS soft-
ware team worked through Block 3i issues and ended up releasing 11 software
versions to integrate Block 3i and address the stability concerns near the end
of development. The final software release for IOC, however, had excellent
stability characteristics.
The 12 aircraft identified for IOC declaration required a series of updates to
bring the configuration into alignment with the IOC declaration configuration.
The engineering design release, procurement of modification kits, and execution
of the on-aircraft modifications spanned 18 months, including the OBIGGS
upgrade. In late June 2016, the final modifications on 12 aircraft were completed
at Ogden Air Logistics Complex, Utah [23] (Fig. 32), and the aircraft ferried back
to the 34th Fighter Squadron at Hill AFB. Outstanding workmanship and focus
from this team provided all 12 jets with just four weeks of flight time available
prior to the USAF IOC objective date.
USAF conducted several operational demonstrations prior to IOC declaration.
The 34th Fighter Squadron at Hill AFB deployed to Mountain Home AFB, Idaho,

Fig. 32 Final F-35A for IOC fleet delivered at Ogden Air Logistics Complex, Utah.
F-35 PROGRAM HISTORY: FROM JAST TO IOC 69

with the newest hardware and software and demonstrated the initial combat capa-
bilities of the F-35A with 88 sorties scheduled and flown. The 422nd Test & Evalu-
ation Squadron also conducted an operation test IOC readiness assessment. This
assessment included the execution of CAS, interdiction, and SEAD/DEAD mis-
sions. The 422nd delivered a report to the Commander of ACC to guide his
decision in declaring IOC.
After notifying Congress, Commander of ACC Gen. Herbert “Hawk” Carlisle
signed off on the IOC declaration on 2 August 2016. Carlisle said, “The F-35A will
be the most dominant aircraft in our inventory, because it can go where our legacy
aircraft cannot and provide the capabilities our commanders need on the modern
battlefield” [24]. Air Force Chief of Staff Gen. David Goldfein stated, “The F-35A
brings an unprecedented combination of lethality, survivability and adaptability to
joint and combined operations, and is ready to deploy and strike well-defended
targets anywhere on Earth” [24].

C. USN F-35C IOC

The USN took delivery of its first fleet F-35C in the summer of 2013 at Eglin AFB.
CF-6 was delivered to the Grim Reapers of Strike Fighter Squadron 101
(VFA-101), which was reestablished as the first F-35C FRS in 2012, seven years
after being disestablished as an F-14 FRS. The new VFA-101 trained the initial
cadre of Navy and USMC instructors and operational test pilots flying the
F-35C. In January 2017 the USN reestablished VFA-125, the Rough Raiders,
a former F/A-18 FRS at NAS Lemoore, California, to become the USN’s west
coast FRS and take on the role of expanding its training capacity to support the
transition of the first deployable USN F-35C squadron, VFA-147. VFA-147 was
in the early stages of transition in early 2018.
USN IOC was first planned for 2015 but was delayed due a combination of
issues, most notably delays in software development and the AHS redesign that
delayed shipboard testing. The revised USN IOC window defines an objective
of August 2018 and a threshold of February 2019. Requirements for F-35C IOC
were delineated in a 2011 Commander, Naval Air Forces letter that stipulated
IOC declaration when the following conditions were met:
1. Ten fully Block 3F capable aircraft are available to the first fully manned,
trained, and equipped operational squadron. (Block 3F is the final capability
standard developed by SDD, to which all variants will be upgraded.)
2. The required ship infrastructure, including ship alterations, is in place to
support F-35C carrier-based operations.
3. The required shore infrastructure, including tools, spares, technical repair and
flight series data, and support equipment, is available to support sustained
training operations.
4. Initial OT&E is complete and declared operationally effective and suitable.
70 A. E. SHERIDAN AND R. BURNES

Although the USN initially required the completion of OT&E as one of the key
elements of IOC declaration in early 2018, the demonstration of adequate war-
fighting capability ahead of official OT&E completion was determined to be
an acceptable alternate criterion. The objective of the criterion is to ensure that
the aircraft delivers expected capabilities prior to first deployment in a carrier
strike group.
Assessment of the Gen III HMD and associated green glow in the demanding
night carrier landing environment also caused the USN to tie IOC to green glow
resolution to the extent it would allow relatively inexperienced pilots to safely
conduct night carrier landings in all conditions experienced at sea. The solution
to green glow has been demonstrated in ground testing as described previously,
and flight trials are planned for later in 2018.

VIII. UPCOMING PLANS: FUTURE DEVELOPMENTS

The F-35 Follow-On Modernization program provides capabilities that ensure
that the F-35 maintains a dominant combat edge in future years, adheres to
future civil aviation mandates, and continuously implements sustainment and
supportability improvements. The program has implemented a requirements gov-
ernance process that identifies the upgrades required to ensure relevance through-
out the air system. This approach provides all F-35 operators with an opportunity
to be involved in the F-35 roadmap and supports maintaining commonality
and interoperability across all F-35 users. A continuous capability development
and delivery framework is used to provide timely, affordable incremental war-
fighting capability improvements to maintain joint air dominance against evol-
ving threats to the United States and its allies.

IX. SUMMARY AND CONCLUSIONS

The JSF program, beginning with JAST, emerged from a unique set of cir-
cumstances in the early 1990s that combined widespread needs to recapitalize
western governments’ fighter/attack fleets with similarly widespread acute
budget constraints. This created an opportunity for a common platform that
could satisfy many or most of the requirements, if feasible. The early phases of
the program demonstrated that the SDLF-based propulsion system uniquely
enabled a common family of configurations to provide the very demanding
STOVL capabilities needed by the USMC and U.K. forces without compromising
the configurations or capabilities of CTOL and CV variants. The Lockheed
Martin F-35 family built upon the unique propulsion system by applying 5th-
Generation low observables and MS capabilities to provide a weapons system
with transformational lethality and survivability. Commonality among the var-
iants, including their support systems, provides unprecedented interoperability
F-35 PROGRAM HISTORY: FROM JAST TO IOC 71

among U.S. and allied services, together with economies of scale that make the
system affordable.
Besides its daunting and disparate technical requirements, the program faced
highly ambitious development and production schedule requirements. Initial
plans for the SDD program relied on significant benefits from acquisition
reform strategies, as well as technical tools and processes that were unproven at
this scale, and a great deal of concurrency to achieve the schedule. As a result,
the development and early production efforts encountered repeated challenges
that slowed—but did not stop—progress. Schedule and cost estimates proved to
be inadequate, forcing three replans of the SDD program, each adding cost,
stretching the schedule, and delaying production. Throughout the development
phase, however, the promise of the system capabilities and the need for the
system remained strong. Although many specific design, manufacturing, and
management features, practices, and approaches were changed over the course
of development, the fundamentals of the propulsion concept and air vehicle con-
figurations remained constant. The missions and requirements also remained
almost wholly intact, with only minor adjustments in areas with little operational
impact.
Massive investments in state-of-the-art production facilities across the globe
enabled nearly 300 aircraft to be produced by early 2018 for all three U.S. services
and eight international customers. More than 170 aircraft per year are expected to
be produced in the mid-2020s. The USMC and USAF declared F-35 IOCs in 2015
and 2016, respectively, and have active operating squadrons in the United States
and overseas. The aircraft have been very successful in live training exercises,
according to service officials.
Although it has been a longer process than planned, with many problems
encountered and solved, the F-35 family is on track to fulfill the original vision
for a lethal, survivable, supportable, and affordable 5th-Generation weapon
system that is interoperable among services and allies.

APPENDIX: ACRONYM LIST

AAI Auxiliary Air Inlet
AFB Air Force Base
AHS Arresting Hook System
AIC Affordability Improvement Curve
ALIS Autonomic Logistic Information System
AOA Angle of Attack
APC Approach-Power Compensation
ASC Aeronautical Systems Command
ASCDR Air System Critical Design Review
ASPDR Air System Preliminary Design Review
ASRR Air System Requirements Review
72 A. E. SHERIDAN AND R. BURNES

ASTOVL Advanced Short Takeoff/Vertical Landing

ATP Authority to Proceed
BfA Blueprint for Affordability
BRAT Blue Ribbon Action Team
BTP Build-to Package
BUW Bottom-Up Weight
CAIG Cost Analysis Improvement Group
CAIV Cost as the Independent Variable
CALF Common Affordable Lightweight Fighter
CAPE Cost Analysis & Program Evaluation
CAS Close Air Support
CDA Concept Demonstrator Aircraft
CDDR Concept Definition and Design Research
CDP Concept Demonstration Phase
CDR Critical Design Review
CFD Computational Fluid Dynamics
CFI Call for Improvements
COPT Cost and Operational Performance Trades
CTOL Conventional Takeoff and Landing
CV Carrier Variant
DAB Defense Acquisition Board
DAE Defense Acquisition Executive
DEAD Destruction of Enemy Air Defense
DoD U.S. Department of Defense
EHAS Electro-Hydrostatic Actuation System
EMD Engineering and Manufacturing Development
ES/G Engine Starter/Generator
FACO Final Assembly and Checkout
FCLP Field Carrier Landing Practice
FETT First Engine to Test
FF Full Funding or First Flight
FMS Foreign Military Sales
FRP Full-Rate Production
FRS Fleet Replacement Squadron
FSETT First STOVL Engine to Test
HMD Helmet-Mounted Display
IBR Integrated Baseline Review
IFR Initial Flight Release
IIP International Industrial Participation
IMRT Independent Manufacturing Review Team
IOC Initial Operational Capability
IPR Interim Progress Review
IPT Integrated Product Team
J/IST JSF Integrated Subsystems Technology
F-35 PROGRAM HISTORY: FROM JAST TO IOC 73

JAST Joint Advanced Strike Technology

JCS JSF Contract Specification
JESB JSF Executive Steering Board
JET Joint Estimate Team
JIRD Joint Interim Requirements Document
JORD Joint Operational Requirements Document
JSF Joint Strike Fighter
JSFPO Joint Strike Fighter Program Office
KPP Key Performance Parameter
KSDI Key System Developments and Integration
LL Long Lead
LoI Letter of Intent
LRIP Low-Rate Initial Production
MoD Ministry of Defence
MoU Memorandum of Understanding
MS Mission Systems
NAS Naval Air Station
NAVAIR Naval Air Systems Command
NTE Not-to-Exceed
OA Operational Assessment
OAG Operational Advisory Group
OBIGGS Onboard Inert-Gas Generation System
OSD Office of the Secretary of Defense
OTB Over-Target Baseline
PAUC Program Acquisition Unit Cost
PBS Performance-Based Specifications
PDR Preliminary Design Review
PEO Program Executive Officer
PSFD Production, Sustainment, and Follow-on Development
PWSC Preferred Weapons System Concept
RAF Royal Air Force
RCS Radar Cross-Section
RF Radio Frequency
RN Royal Navy
SAE Senior Acquisition Executive
SAR Selected Acquisition Report
SBA Simulation-Based Acquisition
SDD System Development and Demonstration
SDLF Shaft-Driven Lift Fan
SEAD Suppression of Enemy Air Defense
STO Short Takeoff
STOVL Short Takeoff/Vertical Landing
SWAT STOVL Weight Attack Team
SWG Senior Warfighters Group
74 A. E. SHERIDAN AND R. BURNES

TBR Technical Baseline Review

TSPR Total System Performance Responsibility
TVEN Telescoping Vectoring Exhaust Nozzle
URF Unit Recurring Flyaway
USAF U.S. Air Force
USMC U.S. Marine Corps
USN U.S. Navy
VAVBN Variable-Area Vane-Box Nozzle
VFA USN Strike Fighter Squadron
VL Vertical Landing
VMFA Marine Fighter Attack Squadron
VMFAT Marine Fighter/Attack Training Squadron
VPA Carrier Powered-Approach Speed
VS Vehicle Systems
3BSD Three-Bearing Swivel Duct/Nozzle

ACKNOWLEDGMENTS
Many people contributed to the completion of this paper. Particularly notable
were Charles T. (Tom) Burbage for the International Participation section, J.D.
McFarlan for the Transition to IOC section, Jeff Catt for the Alternate Engine
Program section, Mark Middlebrook for the Program Replans section, and
Greg Walker for the STOVL Probation section. Others providing significant
data included Bruce Bullick, Wade Cross, Bob Ellis, Carl Fink, David Ford, Jim
Gigliotti, Dean Hayes, Don Kinard, Steve Kopp, Jeff McConnell, Philip Mosley,
David Rapp, Kevin Renshaw, Drew Robbins, Ken Seeling, Kevin Smith, and Art
Tomassetti. Special thanks go to Suzie Pate for cheerfully uncovering many
years’ worth of data in the form of briefing slides.

REFERENCES

[1] Bevilaqua, P., “Genesis of the F-35 Joint Strike Fighter,” Journal of Aircraft, Vol. 46,
No. 6, Nov./Dec. 2009, pp. 1825– 1836. doi: 10.2514/1.42903
[2] Maddock, I. A., and Hirschberg, M. J., “The Quest for Stable Jet Borne Vertical
Lift: ASTOVL to F-35 STOVL,” AIAA Paper 2011-6999, Sept. 2011.
doi: 10.2514/6.2011-6999
[3] Wiegand, C., Bullick, B. A., Catt, J. A., Hamstra, J. W., Walker, G. P., and Wurth,
S. P., “F-35 Air Vehicle Technology Overview,” AIAA Aviation Forum, June 2018
(to be published).
[4] Wurth, S. P., Smith, M. S., and Celiberti, L. A., “F-35 Propulsion System Integration,
Development & Verification,” AIAA Aviation Forum, June 2018 (to be published).
[5] Counts, M. A., Kiger, B. A., Hoffschwelle, J. E., Houtman, A. M., and Henderson, G.,
“F-35 Air Vehicle Configuration Development,” AIAA Aviation Forum, June 2018
(to be published).
F-35 PROGRAM HISTORY: FROM JAST TO IOC 75

[6] Eshleman, J., “Large Scale Testing of the Lockheed Martin JSF Configuration,”
Proceedings of the International Powered Lift Conference, SAE P-306, Society of
Automotive Engineers, Warrendale, PA, 1996, pp. 319– 340.
[7] Nicholas, W. U., Naville, G. L., Hoffschwelle, J. E., Huffman, J. K., and Covell, P. F.,
“An Evaluation of the Relative Merits of Wing-Canard, Wing-Tail, and Tailless
Arrangements for Advanced Fighter Applications,” 14th Congress of the
International Council of the Aeronautical Sciences, ICAS-84-2.7.13, Toulouse,
France, 10 –14 Sept. 1984, pp. 771A– 771L.
[8] Hayward, D., Duff, A., and Wagner, C., “F-35 Weapons Design Integration,”
AIAA Aviation Forum, June 2018 (to be published).
[9] Aldridge, E. C., “Transcript of Briefing on the Joint Strike Fighter Contract
Announcement—Oct. 26, 2001— 4:30 p.m. EDT,” U.S. Department of Defense
[online database], http://archive.defense.gov/Transcripts/Transcript.
aspx?TranscriptID¼2186 [retrieved 4 March 2018].
[10] Kinard, D. A., “F-35 Digital Thread and Advanced Manufacturing,” AIAA Aviation
Forum, June 2018 (to be published).
[11] Ellis, R., Gross, P., Yates, J., Casement, J., Chichester, R., and Nesmith, K.,
“F-35 Structural Design, Development, and Verification,” AIAA Aviation Forum,
June 2018 (to be published).
[12] Schwartz, M., and O’Conner, C. V., “The Nunn-McCurdy Act: Background,
Analysis, and Issues for Congress,” Congressional Research Service, CRS 7-5700,
12 May 2016.
[13] Gates, R. M., “Transcript of DOD News Briefing with Secretary Gates and Adm.
Mullen from the Pentagon—January 06, 2011,” U.S. Department of Defense
[online database], http://archive.defense.gov/transcripts/transcript.
aspx?transcriptid¼4747 [retrieved 25 April 2018].
[14] Marshall, T. C., Jr., “Panetta Lifts F-35 Fighter Variant Probation,” DoD News/
American Forces Press Service [online database], http://archive.defense.gov/news/
newsarticle.aspx?id¼66879 [retrieved 25 April 2018].
[15] Hudson, M., Glass, M., Hamilton, T., Somers, C., and Caldwell, R., “F-35 SDD Flight
Testing at Edwards Air Force Base and Naval Air Station Patuxent River,”
AIAA Aviation Forum, June 2018 (to be published).
[16] Wilson, T., “F-35 Carrier Suitability Testing,” AIAA Aviation Forum, June 2018
(to be published).
[17] Lemons, G., Carrington, K., Frey, T., and Ledyard, J., “F-35 Mission Systems
Design, Development, & Verification,” AIAA Aviation Forum, June 2018
(to be published).
[18] Robbins, D., Bobalik, J., De Stena, D., Plag, K., Rail, K., and Wall, K., “F-35
Subsystems Design, Development, and Verification,” AIAA Aviation Forum, June
2018 (to be published).
[19] Rodriguez, V. J., Flynn, B., Thompson, M. G., and Brelage, S., “F-35 Climatic
Chamber Testing,” AIAA Aviation Forum, June 2018 (to be published).
[20] “Joint Strike Fighter: Additional Costs and Delays Risk Not Meeting Warfighter
Requirements on Time,” U.S. Government Accountability Office, GAO-10-382,
March 2010.
[21] “F-35 Initial Operational Capability,” Report to Congressional Defense Committees,
U.S. Navy, U.S. Marine Corps, U.S. Air Force, June 2013.
76 A. E. SHERIDAN AND R. BURNES

[22] Story, C., “U.S. Marines Corps Declares the F-35B Operational,” Headquarters
Marine Corps [online database], http://www.marines.mil/News/News-Display/
Article/611657/us-marines%20-corps-declares-the-f-35b-operational/ [retrieved
15 May 2018].
[23] Lloyd, A. R., “Ogden Complex Delivers F-35s for Units to Reach Combat Readiness,”
U.S. Air Force News/Ogden Air Logistics Complex [online database], http://www.af.
mil/News/Article-Display/Article/837054/ogden-complex-delivers-f-35s-
for-units-to-reach-combat-readiness/ [retrieved 5 May 2018].
[24] “Air Force Declares the F-35A ‘Combat Ready’,” U.S. Air Force News/Air Combat
Command Public Affairs [online database], http://www.hill.af.mil/News/
Article-Display/Article/883234/air-force-declares-the-f-35a-combat-ready/
[retrieved 15 May 2018].
 CHAPTER 2

F-35 Air Vehicle Configuration

Development
Mark A. Counts , Brian A. Kiger†, John E. Hoffschwelle‡ and
Adam M. Houtman§
Lockheed Martin Aeronautics Company, Fort Worth, TX
Greg Henderson}
Lockheed Martin Aeronautics Company (retired), Beulah, WY

Configuration development of the multinational three-variant F-35

Lightning II fighter program is traced from its earliest beginnings
in the Common Affordability Lightweight Fighter (CALF) effort
through to the F-35 System Development and Demonstration program.
Technologies that were identified and matured during the Joint
Advanced Strike Technology (JAST) program are explored as well as
the parallel development of the X-35 demonstrator aircraft and the Pre-
ferred Weapon System Concept aircraft. Development and maturation
activities along the path to a production configuration, including many
of the key design and integration challenges faced by the design team,
are examined, and discussions are included on the STOVL Weight
Attack Team as well as the unprecedented weight management effort.

I. INTRODUCTION
The U.S. multiservice strike fighter development program that evolved into the
multinational F-35 program held many unique challenges for the air vehicle con-
figuration design team. The program was developed around four pillars: affordabil-
ity, lethality, survivability, and supportability. Effectively balancing the unique
requirements of each service and operator while optimizing performance in each
of the pillar areas proved to be a challenge unprecedented in modern fighter devel-
opment history. Examining the evolution of the F-35 air vehicle configuration

 Senior Manager, F-35 Engineering Project Office.

†
Program Management Senior Staff, Advanced Development Programs.
‡
Principal Systems Engineer, Lockheed Martin Aeronautics.
§
Aeronautical Engineer, Advanced Development Programs.
}
Director (retired), F-35 Mass Properties.

77
78 M. A. COUNTS ET AL.

Fig. 1 The three F-35 variants: F-35C (left), F-35B (center), and F-35A (right).

provides an understanding of the integration challenges faced while developing a

multiservice platform composed of complex, highly integrated systems.
In order to best understand the F-35 development, we will follow a chrono-
logical timeline of the configuration development and expand key events along
the timeline. We will also explore the initial development of the weapon
system, the Concept Demonstration Phase (CDP) that included demonstration
aircraft, and System Development and Demonstration (SDD) that led to pro-
duction configurations of the three variants: F-35A Conventional Takeoff and
Landing (CTOL), F-35B Short Takeoff and Vertical Landing (STOVL), and
F-35C Carrier Variant (CV). The three F-35 variants are shown in Fig. 1.

II. INITIAL DEVELOPMENT: TECHNOLOGIES FOR TACTICAL AIRCRAFT

The Joint Advanced Strike Technology (JAST) program was created in 1993 as a
result of a U.S. Department of Defense (DoD) Bottom-Up Review. The major
tactical aviation results of the review were to continue the ongoing F-22 and
F/A-18E/F programs, cancel the Multi-Role Fighter (MRF) and the A/F-X pro-
grams, curtail F-16 and F/A-18C/D procurement, and initiate the JAST program.
The JAST program’s purpose was to define and develop aircraft, weapons, and
sensor technology that would support the future development of tactical aircraft.
The final goal was to develop a common family of aircraft to replace aging U.S.
and U.K. inventory aircraft.
The JAST program planned to develop shared engines, avionics, munitions,
and other features that could be incorporated into variants of a stealthy airframe
suitable for the individual needs of each service. The goal was to achieve high com-
monality across all variants, regardless of structural configuration. This approach
was predicted to greatly reduce the overall cost of the program.
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 79

When the Defense Advanced Research Projects Agency (DARPA) STOVL

programs were integrated into the JAST program, they brought with them
unique STOVL technology that would ultimately lead to the world’s first oper-
ational 5th-Generation supersonic STOVL aircraft (Fig. 2).
Boeing, Lockheed Martin, McDonnell Douglas, and Northrop Grumman were
each awarded 15-month Concept Definition and Design Research (CDDR) con-
tracts in December 1994. Shortly after the CDDR contracts were awarded, North-
rop Grumman teamed with McDonnell Douglas/British Aerospace (BAE). Each
of the contractors refined their designs and performed a number of risk reduction
activities such as wind tunnel tests, powered-model STOVL tests, and engineering
analyses.
The JAST technologies were available to the Weapon System Contractors
(WSCs). These included programs such as the Air Force Research Laboratory’s
Advanced Compact Inlet Systems (ACIS); Subsystem Integrated Technology
(SUIT); Joint Strike Fighter (JSF)/Integrated Subsystems Technology (J/IST),
which looked to save both weight and cost by taking a more integrated approach
to aircraft subsystems; Advanced Lightweight Aircraft Fuselage Structure
(ALAFS), which looked at concepts to reduce structural weight and cost; and Con-
tracted Research and Development (CRAD) studies.
The “family of aircraft approach” (Fig. 3) was an important element of the
Lockheed Martin commonality story. Commonality is a key enabler to affordability
when implemented in an effective way. Previous development efforts on multiser-
vice aircraft have had only limited success as a result of design inefficiencies and

Fig. 2 JAST technologies led to the F-35B, the world’s first 5th-Generation supersonic
STOVL aircraft.
80 M. A. COUNTS ET AL.

Fig. 3 The F-35 family of aircraft.

scar weight. The challenge was to develop a commonality approach that reduced
cost in key areas without adding complexity and weight to each variant. This
approach separated every part of the aircraft into three categories: common,
cousin, and unique parts.
Common parts were essentially identical parts used for each variant whereas
unique parts were unique to a single variant. Cousin parts were identified as
parts that were not exactly the same but could be designed to take advantage of
similarities in order to maximize cost savings. The cousin part concept can be
explained with a simple example. Different variants with common Outer Mold
Line (OML) might have different loads that, if optimized, would result in unique
composite lay-ups and unique skins. If these skins were made common, the
variant with the lower loads would be saddled with unnecessary scar weight. In
the cousin approach, the OML tooling could be common whereas the composite
skin lay-up would be unique for each of the cousin parts. A significant savings to
tooling costs could be realized without adding scar weight to either aircraft. A
similar approach was used for internal bulkheads and structure. Component com-
monality across all three variants reduced requirements for unique spares as well as
the logistics footprint. The cousin approach also reduced manufacturing and
assembly line footprint and tooling.
Aircraft subsystems have traditionally been designed with a federated approach
that consists of a number of independently designed subsystems. Effectively inte-
grating key subsystems can significantly improve aircraft affordability and
provide warfighting benefits through increased performance and a reduction in
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 81

the number of dedicated standalone components. The J/IST demonstration

program matured integration technologies for aircraft subsystems to enable tran-
sition to the Engineering and Manufacturing Development (EMD, now called
SDD) program [1]. The JSF-sponsored Vehicle Integration Technology Planning
Studies projected life-cycle cost savings that could be achieved with these technol-
ogies vs 1995 state-of-the-art federated subsystems. Integration technologies
included the Thermal/Energy Management Module and its integration with the
engine (Integrated Power Package [IPP] and fan duct heat exchangers), 270
VDC power management and distribution, electric-powered hydraulic flight actua-
tion, and associated controls. These technologies were identified by Lockheed
Martin, Boeing, and McDonnell Douglas/Northrop Grumman/British Aerospace
as providing substantial cost savings and warfighting benefits to JSF weapon system
concepts. The technologies encompassed in the J/IST program allowed 13 major
subsystems to be replaced with 5. Hardware and software components were inte-
grated into major subsystems ground and flight demonstrations in the 1997 –
2000 timeframe.
The goals of the ALAFS program were to reduce the cost of the aircraft build
by 30% and reduce the weight by 20%. These goals were predicted to translate into
6–8% life-cycle cost savings for the JSF program. Technologies and focus areas
encompassed materials, structural design concepts, and manufacturing processes
for improved fabrication and assembly. This program identified and developed
concepts and methodologies to allow much greater integration of advanced com-
posite structures. Unitized composite design concepts were explored to enhance
structural integrity at reduced weight while employing lower cost production con-
cepts and creating volumetric efficiencies. Full-scale test articles were fabricated to
allow direct comparison with traditional structural designs.
The propulsion system used many unique design elements that can be com-
bined into two categories: those associated with the STOVL lift system and
those associated with conventional propulsion systems. For the conventional
system, Lockheed Martin developed the diverterless supersonic inlet to maintain
low signature and reduce weight and cost. With no moving parts and no diver-
ter, this results in a much simpler and better structural integration of the inlet
system. Lockheed Martin also incorporated the Low Observable Axisymmetric
Nozzle (LOAN) that provided a lightweight, low observable nozzle for the
engine [1].
For the STOVL lift system, Lockheed Martin developed a concept that uses a
vertically oriented Shaft-Driven Lift Fan (SDLF). A two-stage low-pressure
turbine on the engine provides the horsepower necessary to power the
Rolls-Royce – designed LiftFanw. The LiftFan generates a column of cool air
that provides approximately 20,000 lb of lift using variable inlet guide vanes to
modulate the airflow. Along with the LiftFan, an equivalent amount of thrust
from the downward vectored rear exhaust provides balanced lift to the aircraft.
The LiftFan utilizes a clutch that engages the shaft drive system for STOVL oper-
ations. Because the LiftFan extracts power from the engine, exhaust temperatures
82 M. A. COUNTS ET AL.

Fig. 4 LSPM testing confirmed the capability of the STOVL propulsion system.

are typically lower than traditional direct lift STOVL systems. The Pratt &
Whitney (P&W) F135 engine leverages the existing F119 core technology with
a larger fan and integrated fan-duct heat exchanger to reduce the number and
size of secondary inlets, exhausts, and heat exchangers. This lift system was
proven using a Large Scale Powered Model (LSPM), shown in Fig. 4, to validate
both the design and supporting analysis [2].
The F-35 aircraft incorporated cutting-edge technologies in the sensor suite to
provide unprecedented pilot awareness. The Distributed Aperture System (DAS)
provides the pilot with a unique spherical view around the aircraft for enhanced
situational awareness, missile warning, and day/night pilot vision. The Multifunc-
tion Advanced Data Link (MADL) provides low probability of intercept ship-
to-ship communications that includes both voice and data sharing. The internally
mounted Electro-Optical Targeting System (EOTS) provides extended range
detection and precision targeting against ground targets, plus long-range detection
of air-to-air threats. Additional information is provided to the pilot through an
integrated Communications, Navigation, and Identification (CNI) suite, an
Active Electronically Scanned Array (AESA) radar, and an integrated Electronic
Warfare (EW) suite [3]. Sensor data are fused together by the core processor to
provide a seamless integrated picture of the battlespace, giving the pilot unsur-
passed situational awareness, positive target identification, and precision strike
capability under all weather conditions [4]. The weapon system also incorporates
a Helmet-Mounted Display System (HMDS), which replaces the traditional
Head-Up Display (HUD), to deliver information directly to the pilot’s helmet.
The JAST program provided the risk reduction needed to mature these new
technologies to the level necessary to be considered for incorporation into the
JSF design [5]. The combination of these new technologies coupled with outstand-
ing aerodynamic performance gives the F-35 unique capabilities and a command-
ing edge over any other fighter in the battlespace.
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 83

III. CONCEPT DEMONSTRATION PHASE

In March 1996, a Request for Proposal (RFP) was released for the Concept Dem-
onstration Phase (CDP) of the JSF program. Design teams from Lockheed Martin,
Boeing, and McDonnell Douglas (with Northrop Grumman and BAE) entered
the competition and put forward uniquely different designs in their proposal
offerings.
Lockheed Martin’s conventional four-tail design was said to resemble a single-
engine version of the F-22 Raptor (Fig. 5). The STOVL version featured a shaft-
driven LiftFan behind the cockpit plus a three-bearing vectored engine exhaust
nozzle aft and two roll control ducts extending from each side of the engine.
The nozzles for the roll control ducts were located on the lower surface of
the wings.
The Lockheed Martin F-35 configuration development lineage, as shown in
Fig. 6, can be traced back to Configuration 100. This configuration was developed
in response to the government’s Common Affordable Lightweight Fighter (CALF)
initiative in the early 1990s that was intended to develop a highly common aircraft
to replace the U.S. Marine Corps (USMC) AV-8B and F/A-18 aircraft, as well as
provide a new fighter platform for the U.S. Air Force (USAF). This was a single-
engine configuration with a delta wing, canards, and twin vertical tails that was
developed with the intent of being a starting point on the path to the LSPM
demonstrator. A series of trade studies matured and refined the design even
though the basic planform remained largely unchanged.
This effort resulted in Configuration 140, which became the baseline for the
LSPM. Configuration 150 was established as the baseline to support wind
tunnel testing. The 150 series of configurations evolved into Configuration 160,
which relocated the vertical tails and changed the control surfaces, but still
retained the basic delta wing/canard planform.
When the CALF program merged with the JAST program, the configuration
took on the added requirement of providing a carrier-based platform for the U.S.

Fig. 5 Lockheed Martin’s X-35 Concept Demonstrator Aircraft.

84 M. A. COUNTS ET AL.

Fig. 6 F-35 configuration evolution.

Navy (USN). This programmatic merge was really a turning point in the configur-
ation design evolution. The CALF configurations had struggled to stay below the
specified weight limit of 24,000 lb, so Configuration 180 adopted several signifi-
cant changes in an effort to reduce weight and improve performance. The caret
inlet was dropped in favor of the diverterless supersonic inlet, and the three-
bearing swivel nozzle replaced the 2D nozzle in the STOVL variant. For weight
savings, the canards were dropped, the twin vertical tails were replaced with a
single tail, and the delta wing planform was modified. These changes were
made for both CTOL and STOVL variants, which received the designations A
(CTOL) and B (STOVL). For the USN variant (C), it was quickly determined
that the baseline delta wing planform would not be able to provide the low-speed
handling qualities needed for a carrier-based aircraft. It was therefore decided that
Configuration 180C would adopt a conventional wing/tail planform.
With weight-related challenges on the A and B variants from the CALF effort,
the addition of the C variant, and the desire for commonality, it was not at all clear
which would be the best planform to carry forward. The 190 series of configur-
ations was developed to answer that question. Configuration 190 was a conven-
tional wing/tail planform with a single vertical tail for all three variants.
Configuration 191 was a trivariant delta wing/canard planform also with a
single vertical tail. The configurations were sized to the same requirements, and
analysis was performed to compare critical design characteristics including
mass properties, aero performance, systems integration, basing, and ship suit-
ability. The conventional wing/tail planform emerged from this competition as
the winner and was carried forward into the 200 series of configurations. Con-
figuration 200 kept the traditional boom-mounted horizontal tails, but went
back to the twin trapezoidal-style vertical tails that had been featured up
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 85

through Configuration 160. With the basic planform set, attention was next
focused on trade studies to mature the integration of subsystems and weapons.
The 210 series of configurations evolved the internal weapon arrangement as
well as changed the trapezoid vertical tail to a swept tail design.
The 220 series of configurations made commonality improvements, which
included the adoption of common wing structure as shown in Fig. 7. The wing
and fuselage were joined as a single structural entity from tip to tip on CTOL/
STOVL and wing fold to wing fold on CV. The design team continued to
mature the configuration with updates to systems integration details, and Con-
figuration 220-2 became the baseline for Lockheed Martin’s CDP proposal. The
focus on commonality is apparent in the Configuration 220-2 internal arrange-
ment cutaways for all three variants, which are shown in Fig. 8.
The McDonnell Douglas team offered a relatively conventional-looking air-
craft design (Fig. 9), except that it replaced conventional horizontal and vertical
tail surfaces with a canted control surface similar to the Northrop/McDonnell
Douglas YF-23 offering for the USAF Advanced Tactical Fighter (ATF) compe-
tition. The STOVL version employed a separate lift engine that was installed aft
of the cockpit.
Boeing offered a delta wing design with twin vertical tails and a single scoop
inlet under the nose of the aircraft (Fig. 10). The STOVL version delivered thrust

Fig. 7 Wing commonality approach.

86 M. A. COUNTS ET AL.

Fig. 8 Configuration 220-2 internal arrangement: CTOL (top), STOVL (center), and CV
(bottom).

from the engine to vectored lift nozzles located on the lower surface of the aircraft,
and the nose intake scoop hinged forward to allow increased airflow.
The government’s CDP acquisition strategy was predominately threefold: 1)
maintain the competitive environment prior to engineering and manufacturing
development down-select while providing for two different STOVL approaches
and two different aerodynamic configurations; 2) demonstrate the viability of a
multiservice family of variants with high commonality among CTOL, STOVL,
and CV variants; and 3) provide affordable and low-risk technology transition
to JSF EMD in FY 2001 [6].
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 87

Fig. 9 Model of the

McDonnell Douglas team’s
CDP design.

In fall 1996, it was

announced that Lock-
heed Martin and Boeing
were each awarded
Concept Demonstration
Phase contracts, with
the McDonnell Douglas
team being eliminated
from the competition.

IV. CONCEPT DEMONSTRATION AIRCRAFT

Each CDP contract called for building two Concept Demonstration Aircraft to
demonstrate the three different configurations of the JSF: CTOL, STOVL, and
CV. In addition, each contractor developed in parallel the Preferred Weapon
System Concept (PWSC), which would be the basis for the contractor’s System
Development and Demonstration (SDD) proposal offering.
The contractors were responsible for defining the ground and flight demon-
strations they believed were critical for their Concept Demonstration Aircraft.
Each contractor was expected to demonstrate STOVL hover, STOVL transition,
and low-speed handling qualities of their CV concept. The results of the demon-
stration program along with the PWSC offering would be the basis of the down-
select to one contractor for the SDD phase.

Fig. 10 Boeing’s X-32 Concept Demonstrator Aircraft.

88 M. A. COUNTS ET AL.

Along with Lockheed Martin and Boeing, Pratt & Whitney received a contract
to provide hardware and engineering support for both Weapon System Concept
Demonstration efforts. The primary propulsion systems being designed for the
JSF program were derivatives of the F119-PW-100 engine that powers the F-22
Raptor. The propulsion system concepts for Lockheed Martin and Boeing used
a P&W F119 core (compressor, combustor, and high-pressure turbine). Both pro-
pulsion system concepts utilized new fan and Low-Pressure Turbine (LPT)
designs that were based on F119 designs, materials, and processes. P&W’s
JSF119-PW-611 was developed to power the Lockheed Martin demonstrator air-
craft, and the P&W JSF119-PW-614 was developed to power the Boeing demon-
strator aircraft. Fabrication of the P&W engine designs began in March 1997, and
engine testing started in June 1998.
Lockheed Martin made the decision to design and build the CDA vehicles at
the company’s Skunk Works facility in Palmdale, California. Starting with Lock-
heed Martin’s current PWSC design, internally called Configuration 230-1, the
Skunk Works team focused on the key demonstrations needed to prove the
concept. Using a commonality approach, the team developed a methodology to
demonstrate all three aircraft with only two demonstration vehicles. The fuselage
for both vehicles was designed to accept the STOVL lift system with the CTOL
and CV versions incorporating a large fuel tank in the location where the lift
fan was installed on the STOVL aircraft. This concept allowed either demonstra-
tor aircraft the flexibility to be converted to demonstrate STOVL capability. Key
demonstrations included short takeoffs, vertical landings, conversions for STOVL,
carrier approaches for CV, and vehicle flight performance for CTOL. With the
basic design approach set, Lockheed Martin launched forward on developing
the demonstrator aircraft (Fig. 11).
Because the X planes were intended as demonstrator aircraft only, there are
many differences between the X-35 and the F-35 production configuration [7].
These differences predominately fall into three major categories: changes to the
STOVL lift system, operational needs of the F-35 that did not need to be demon-
strated, and vehicle improvements that were incorporated after the demonstrator
design and flight test. Some of the most notable differences were in the STOVL
vehicle, which the team continued to mature and refine as the demonstrator
vehicles were being built and tested (Fig. 12 and Fig. 13) [7]. The lift fan inlet
doors of the X-35 were side-hinged bi-fold doors, selected for their light weight
and low-profile design. This design had higher distortion than expected, which
led to changes that were incorporated into the production design. The F-35 lift
fan uses a single aft-hinged inlet door. Although this design is heavier than the
X-35 design, it reduced distortion and improved performance, resulting in a
better overall design. The auxiliary inlet can provide as much as 60% of the air
flow to the engine during low-speed operation. The X-35 auxiliary inlet used a
small two-door design that was hinged on the aircraft’s center line. Although
simple in design, this concept limited air flow at low speeds. The production
design moved the hinge line to the outboard side of the inlet, which yielded
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 89

Fig. 11 Lockheed Martin’s X-35 aircraft during assembly.

Fig. 12 Plan views of X-35 STOVL variant (left) and PWSC STOVL variant (right).
90 M. A. COUNTS ET AL.

Fig. 13 Bottom views of X-35 STOVL variant (left) and PWSC STOVL variant (right).

improved flow characteristics at low-speed conditions. The X-35 lift fan nozzle
used a segmented extending hood configuration to vector the lift fan exhaust
flow. This design was changed to a Variable Area Vane Box Nozzle (VAVBN)
for the F-35 PWSC design. The VAVBN provided not only weight savings
because of more efficient integration into the airframe, but also improved
thrust characteristics by independently controlling the vanes to regulate and
vector the air flow. The STOVL airplane has a roll nozzle outboard of the main
landing gear under each wing to provide roll control during STOVL flight
modes. The roll nozzles on the X-35 used butterfly valves at the engine attachment
to control the roll nozzle flow, and the roll nozzles themselves were completely
uncovered to reduce weight and complexity on the demonstrator. The F-35 pro-
duction design removed the butterfly valves and used the roll nozzles themselves
to control the flow. Aerodynamic improvements in up-and-away flight were made
by adding doors on the lower surface of the wing to cover the roll nozzles.
Many operational capabilities of the F-35 were not required to be demon-
strated in the CDA vehicles. The X-35 had no radar, but instead used this space
to install the vehicle’s flight test boom. Other key Mission Systems components
that were not included in the demonstrator aircraft were the EOTS and the
DAS. The EOTS is a targeting system that can be seen in the chin area of the
lower forebody on the F-35. The DAS includes a set of sensors that, when
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 91

combined with the pilot’s HMDS, provides the pilot with unprecedented all aspect
day/night vision. Lockheed Martin took advantage of the fact that weapon car-
riage was not a demonstration requirement, and used the weapon bay volume
for systems installation and to stow the modified A-6 main landing gear. The
X-35 also incorporated in-flight refuel capability. The design team could have
installed a nonfunctional refuel system to demonstrate flight characteristics
during aerial refueling, but the team decided to incorporate a functioning in-flight
refuel system as an enabler to increase flight test tempo. This proved to be a great
decision because the in-flight refueling capability was used on numerous
occasions to extend flight test time.
To meet stringent timelines and minimize cost, the X-35 demonstrator aircraft
used off-the-shelf components and conventional systems whenever possible. The
X-35 cockpit used off-the-shelf parts from other aircraft for displays and incor-
porated a conventional Head-Up Display (HUD). Conventional hydraulic flight
actuators, a conventional Environmental Control System (ECS), and a conven-
tional Auxiliary Power Unit (APU) were also used. Leveraging the technologies
developed under JAST, the F-35 incorporated a more electric aircraft approach
that utilized an Engine Starter/Generator (ESG) that provides generator and
engine start functionality, as well as Electro-Hydrostatic Actuators (EHAs) that
greatly reduced hydraulic system requirements [8, 9]. A unique Integrated
Power Package (IPP) provides ECS, APU, and emergency power functions all
in one package. An external heat exchanger was incorporated on the lower
center fuselage of the STOVL demonstrator to ensure adequate heat rejection
capability for flight testing in the summer timeframe at Edwards Air Force
Base. The X-35 utilized modified main landing gear from the A-6 Intruder and
nose landing gear from the F-15 Eagle, while the F-35 developed unique
landing gear tailored to save weight and meet packaging requirements. The
X-35 used a modified F119 engine with a new low-pressure fan and turbine com-
bined with a conventional nozzle. The F-35 engine (F135) shares a common core
with the F119 but has a higher bypass ratio and uses a LOAN nozzle design that
was developed during JAST. Fan duct heat exchangers added to the engine bypass
flow path provide a heat rejection path for the IPP [8]. The X-35’s “four-sided”
supersonic diverterless inlet was revised to a “three-sided” design to reduce
weight and drag as well as to improve high angle of attack performance of the
F-35 production configuration. The X-35 CTOL and STOVL versions had a
wing area of 450 ft2 and wing span of 33 ft, whereas the F-35 production versions
of CTOL and STOVL increased the size to 460 ft2 of area and a 35-ft span. The
X-35C wing area was 540 ft2 with a span of 40 ft, whereas the production
F-35C increased the wing area to 668 ft2 and span to 43 ft in order to provide
improved approach speed characteristics for ship-based aircraft.
After meeting the unique design and build challenges of the CDA aircraft,
Lockheed Martin was ready to start flight testing (Fig. 14). The X-35A CTOL
demonstrator took to the skies for the first time on 24 October 2000. This first
flight was quickly followed by subsequent successful flight envelope expansion
92 M. A. COUNTS ET AL.

Fig. 14 The two X-35 aircraft awaiting a fast-paced flight test program.

tests, and the X-35A CTOL flight test effort was completed with all objectives
achieved or exceeded in November 2000. Per the Lockheed Martin plan, the
CTOL vehicle (Aircraft #301) was then modified to the STOVL configuration.
In December 2000, the X-35C vehicle (Aircraft #300) completed first flight. In
parallel to the CV flight test effort, the team continued to work on modifications
to Aircraft #301 to turn it into a STOVL vehicle.
The Lockheed Martin JSF team completed installation of the X-35B’s flight-
ready propulsion system, including the shaft-driven lift fan (Fig. 15) and

Fig. 15 X-35B lift fan installation.

F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 93

engine, in May 2001. Following these modifications, the aircraft was towed to the
hover pit, and BAE test pilot Simon Hargreaves began the STOVL test effort. The
X-35B conducted its first press-up in June 2001, marking the first time that a
shaft-driven lift fan propulsion system had lifted an aircraft into the sky. One
of the most exciting events of the X-35B flight test effort was the successful com-
pletion of what was referred to as Mission X. This particular flight was piloted by
USMC Maj. Art Tomassetti and included a short takeoff, supersonic dash, and
vertical landing. This was the first time in aviation history that a single aircraft
had demonstrated all three capabilities on the same flight. Mission X showcased
the revolutionary capabilities of the X-35 aircraft and the incredible benefits of
the shaft-driven lift fan design concept. On 6 August 2001, Aircraft #301 com-
pleted its 66th and final test flight. The aircraft left the runway at Edwards Air
Force Base and completed a 3.7-hour flight that included six aerial refuelings
and six touch-and-go landings at Palmdale. The X-35B totaled 48.9 hours of
flight time. Upon completion of flight testing, the X-35B eventually found its
way to a permanent display at the National Air and Space Museum (Fig. 16).
The X-35 CDA flight test program successfully demonstrated all three variants
of the Lockheed Martin design and was one of the most successful and efficient
flight test programs ever.

V. PREFERRED WEAPON SYSTEM CONCEPT

The configuration development effort leading up to the CDP proposal resulted
in a solid configuration to carry forward into the Concept Demonstration Phase.

Fig. 16 X-35B and the STOVL propulsion system on display at the National Air and Space
Museum.
94 M. A. COUNTS ET AL.

The basic configuration was in the mold of a classic fighter design. It had a con-
ventional wing/tail arrangement with four tails for high maneuverability, aft
engine placement with a long inlet and short exhaust, and a large internal
weapons bay for lower drag and improved signature. In addition to these key
characteristics, another critical factor in the design of a modern 5th-Generation
fighter aircraft is stealth. Stealth capabilities are difficult to add to an aircraft
after it is designed, providing only limited benefits, and a 4th-Generation aircraft
cannot be turned into a 5th-Generation aircraft. In order to maximize the air-
craft’s effectiveness, stealth must be incorporated into the configuration from
the very earliest conceptual designs, and that is exactly what the Lockheed
Martin design team did. In addition, affordability, supportability, survivability,
and lethality were the four overarching program pillars that influenced design
development activities throughout the PWSC as well as the SDD phase of the
program. Every design decision was carefully assessed to ensure that impacts to
the pillars, either favorable or unfavorable, were understood before incorporation.
The four pillars became ingrained as a mindset that guided the team throughout
the design effort.
The Configuration 220-2 design, which was the basis for the CDP proposal,
was matured with fresh wind tunnel data and systems integration refinement
leading to Configuration 230-1. This new configuration represented the fork in
the road where development of the Concept Demonstrator Aircraft separated
from the development of the Preferred Weapon Systems Concept. The Skunk
Works team took the 230-1 configuration as a baseline and developed the two
X-35 aircraft while a separate team at the Fort Worth facility focused on maturing
the operational design concept. The PWSC team was also bolstered in this time-
frame by the addition of teammates BAE and Northrop Grumman following the
dissolution of the McDonnell Douglas team, which occurred as a result of the
CDP down-select.
The government’s plan was to issue an interim requirements document with
three successive iterations, followed by a draft of the final requirements, and then
ultimately the final requirements document on which the contractor’s SDD pro-
posal design would be based. The government had a formidable task in getting to
the final set of requirements because it not only had to address the needs of the
three U.S. services, but also had to include the needs of the international partners
that had joined the program. Another factor that even further complicated the
government’s task was that in order to maintain the competition, it had to be
mindful of issuing requirement values, particularly Key Performance Parameters
(KPPs), that couldn’t be met by one of the contractors. The government had an
ongoing struggle during this time period to combine all of its customers’ needs
into a single set of requirements. To assist in this process, several series of Cost
and Operational Performance Trades (COPT) were performed. COPT was an
iterative circular process that included the warfighters, simulation-based acqui-
sition, and weapon system – level trade studies and resulted in a carpet plot of
cost as a function of weight, performance, and effectiveness. The resulting data
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 95

were iterated back through the process again with a new set of trades and associ-
ated results. The iterations continued until requirement parameters were ident-
ified. The COPTs led to requirement updates, and each successive release of
requirements yielded changes to range, payload, aerodynamic/maneuver per-
formance, flight envelope, and other critical parameters.
The Lockheed Martin team continued to mature the design while evolving it
to align with each requirement update. Wing area is an excellent example to help
understand how evolving requirements influenced the design. Configuration
220-2 was the baseline when the First Interim Requirements Document (JIRD
I) was released, and the wing area for 220-2 was 450/540 ft2 (the first number
is the area for CTOL and STOVL and the second number is the area for CV).
The Second Interim Requirements Document (JIRD II) was released about a
year later. This led to Configuration 230-2 in which the wing size grew to 500/
600 ft2 in order to improve maneuver performance, as well as improve Short
Takeoff (STO) and CV approach speed (Vpa), both of which were KPPs. The
Third Interim Requirements Document (JIRD III) was released several months
after JIRD II and led to Configuration 230-4. The wing area was reduced to 412
ft2 for CTOL and STOVL while staying at 600 ft2 for CV to maintain Vpa perform-
ance. The smaller wing area reflected an emphasis on lighter aircraft weight, which
included smaller internal weapons. The release of the draft final requirements
document (Draft JMS) brought the focus back to a more capable aircraft, and
Configuration 230-5 was developed to address the Draft JMS with a wing area
of 460/620 ft2. The final requirements document (JMS) was released several
months later and was the basis for the SDD proposal. Configuration 235 was
developed as the SDD proposal configuration and maintained the 230-5 wing
area of 460/620 ft2.
Wind tunnel testing was used throughout the PWSC design period in con-
junction with extensive Computational Fluid Dynamics (CFD) to optimize the
Outer Mold Line (OML) as well as identify performance improvements that pro-
vided additional thrust and control to the STOVL propulsion system. In addition
to refining the external surfaces, this time was also critical for the development of
the aircraft’s internal structure. Major load paths were defined and structural com-
ponents including bulkheads, shear webs, and longerons were modeled. Analysis
was performed, and the component models were iterated to revise load paths
where needed and to support material selection. The Configuration 230-2 struc-
tural arrangements are shown in Fig. 17.
Packaging of the major systems was coordinated with the wind tunnel test
plan to ensure that the configuration was sized large enough to accommodate
all of the necessary systems, payload, and fuel volume, yet small and agile
enough to meet all of the maneuver and mission radius requirements. The internal
weapons arrangement was a critical packaging challenge during this timeframe.
From a fuselage station perspective, the maximum cross-sectional area is influ-
enced by the internal weapons, the engine, and the main landing gear. Reducing
cross-sectional area typically results in better aerodynamics, including lower drag,
96 M. A. COUNTS ET AL.

Fig. 17 Configuration 230-2 structural arrangements: CTOL (top), STOVL (center), and CV
(bottom).

longer range, better maneuver performance, and better acceleration. Integrating

the large set of required internal weapons [10] while trying to minimize cross-
sectional area proved to be particularly challenging. The weapons were arranged
in a staggered manner to avoid the alignment of maximum wing span or tail span
on adjacent weapons. The packaging breakthrough came by toeing the nose of the
weapons inboard while in the staggered arrangement. The resulting arrangement
allowed hinge lines and weapon bay doors that met fall line requirements, freed up
critical volume for door drive mechanisms, and yielded the smallest possible
cross-sectional area.
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 97

The CTOL internal gun influenced several PWSC design iterations. The initial
internal gun was an advanced single barrel gun. After a series of trade studies, the
decision was made to delete the gun in favor of weight and cost. Primarily as a
result of warfighter feedback, the gun was readopted as part of the baseline con-
figuration, and a trade study was initiated to look at the single barrel gun vs a
Gatling-type gun. The Gatling gun emerged as the study winner, primarily on
cost, and the GAU-12 five-barrel 25-mm gun was selected [10]. Significant air
vehicle level impacts resulted from selection of the GAU-12. The cross-section
was larger than the single barrel gun, so a larger external OML bump was
needed. This was an air-to-ground gun, so a negative installation bias was pre-
ferred that pulled the breech up higher resulting in an even larger bump with
steep slopes. The team finally compromised on a zero-bias installation angle as
a balance between the drag of a larger bump vs projectile trajectory. It should
be noted that a subsequent trade study performed during SDD, and as a part of
the STOVL Weight Attack Team (SWAT), deleted one barrel for weight
savings. The resulting four-barrel gun was designated GAU-22.
Integrating key technologies demonstrated in JAST, as well as new and unique
mission systems, added to the PWSC design challenges. The more electric concept
of a 270-V DC aircraft has many advantages; however, each DC motor requires a
controller to be installed in close proximity to avoid excessive wire weight, and
that can present packaging challenges when numerous motors are spread
throughout the aircraft. EHAs also offer many advantages including a reduction
in overall hydraulic system weight and a reduction in vulnerable area, but they
are typically larger than similar load-sized hydraulic actuators, and therefore
require more installation space that often results in larger OML bumps and associ-
ated drag impacts. Another new technology that was incorporated was nonpyro-
technic weapon ejection racks [10]. These racks were developed for both internal
and external weapon carriage and offered improved supportability and combat
turnaround times.
The SDD proposal was prepared and submitted based on the Configuration
235-1 family of aircraft and supported by test data from the X-35 demonstrator
aircraft. In October 2001, it was announced that Lockheed Martin’s proposal
had been selected as the SDD winner, and the company was awarded the contract
to develop the aircraft that would be designated F-35.

VI. SYSTEM DEVELOPMENT AND DEMONSTRATION

The design that was matured during the PWSC effort retained its classic fighter
characteristics and proved to be an excellent configuration to carry forward
into SDD. The design, as shown in Fig. 18, had a conventional wing/tail arrange-
ment, bifurcated long inlet duct, aft engine placement with a fighter optimized
cycle, internal weapons bay, and mid-wing location that took advantage of
legacy-derived shaping.
98 M. A. COUNTS ET AL.

Fig. 18 F-35 design approach.

Immediately following the SDD contract award, Configuration 240-1 was

established as the baseline configuration. This configuration primarily incorpor-
ated maturation activities that had occurred in the timeframe between proposal
submittal and contract award. Changes included standing up the vertical tails
by four degrees for improved signature performance, as well as hydraulic
system sizing changes and revisions to the engine gearbox.
Freezing the OML surfaces (lines freeze) is a critical point in any aircraft devel-
opment program. For the F-35, this milestone occurred with Configuration
240-1.1. The lines freeze configuration was the baseline for fabrication of wind
tunnel models as well as the basis for design-to loads. There were also two key
configuration changes that occurred with 240-1.1. The first was the adoption of
the common weapon bay for STOVL. Optimization of the STOVL variant
during PWSC resulted in a unique weapon bay sized for 1000-lb (1K) class
weapons as opposed to the CTOL/CV weapon bay, which was sized for
2000-lb (2K) class weapons. Adoption of the 2K bay for the 240-1.1 STOVL con-
figuration provided increased capability for the warfighter and improved com-
monality across the variants, but the cost to STOVL was a weight increase of
approximately 240 lb. The second change was an increase of 7 in. to the fuselage
length for all variants: 5 in. in the forward fuselage and 2 in. in the tail boom. The
increased length was needed to provide space for equipment packaging in the
forward fuselage and actuator installation in the tail boom.
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 99

The next step in the F-35 design evolution was Configuration 240-2, which
was the Preliminary Design Review (PDR) configuration. Several key changes
were incorporated into 240-2 including the deletion of dedicated speed-brakes
underneath the engine on the lower fuselage. A three-month intensive co-located
air vehicle integration effort culminated in the deletion of dedicated speed-brakes
and the addition of access panels on the inboard wall of the weapons bay, both of
which improved maintenance access to the engine. The decision reduced the
weight and complexity associated with dedicated speed-brakes while continuing
to provide speed-brake functionality through the use of control surfaces. A critical
feature of the design, which was added to the upper right side of the center fuse-
lage, was a fuel/air heat exchanger to provide the aircraft with greater heat rejec-
tion capability. The design team was also deeply engaged with finding adequate
routing paths for hydraulics, fuel, air, and wiring harnesses, with the STOVL
variant being particularly challenging in the areas adjacent to the lift fan. The
final key change for Configuration 240-2 was the adoption of a single piece lift
fan inlet door as opposed to the previous pair of bi-fold doors. The single door
provided improved airflow characteristics to the lift fan, and two open settings
were adopted to provide 1) max airflow when needed and 2) lower drag when
the max airflow was not required. Following PDR, the baseline was updated to
Configuration 240-2.1, which was defined to support the start of Build-To-
Packages (BTPs) for CTOL, and subsequently Configuration 240-2.2 which was
defined to support the start of STOVL BTPs.
In late 2003, it started to become apparent that the gap between parametric
weight estimates and bottom-up weight estimates was not closing. As a result,
the SWAT activity, which will be discussed in the next section, was initiated.
While SWAT was being planned, Configuration 240-3 was defined and a major
program review was held in early 2004. A key decision coming out of that
review was to continue with the design and build of the first CTOL aircraft (desig-
nated AA-1), but to not continue with detailed design activities on the STOVL or
CV variants until a more weight-efficient configuration was defined. Although not
weight optimized like the post-SWAT configurations, it was determined that
AA-1 could still be a valuable flight test asset to the program while the post-SWAT
configurations were being designed and fabricated. However, the fate of the
program as a whole, and particularly the STOVL variant, was in the hands of
the SWAT team. Configuration 240-4 was defined at the end of SWAT and incor-
porated all of the design changes, refinements, and improvements that were ident-
ified by the SWAT team. The exceptional work of the SWAT team put the
program on a solid trajectory toward the Critical Design Review (CDR).
The SWAT effort highlighted several persistent driving requirements that
influenced a large number of design decisions throughout SDD. Chief among
these were mission radius, signature, Vertical Landing Bring Back (VLBB),
STO, Vpa, internal and external stores, mission systems functionality (including
sensor integration), transonic acceleration, 1% hot day environment, unlimited
angle of attack, internal gun for CTOL, 9 g for CTOL, and powered approach
100 M. A. COUNTS ET AL.

handling qualities for CV. Two other key areas also drove many design decisions.
The first area was providing the maintainer with the best possible ground access to
the aircraft. The design team took advantage of in-flight opening doors wherever
possible to avoid the time associated with removing fasteners on fixed panels.
Examples of areas where this philosophy was used are the weapon bays, the
nose wheel well, and both main wheel wells. A unique feature of the forebody
is the EOTS hinge that rotates the system down to provide maintainer access to
EOTS and other forebody components. Another unique feature is the forward
hinged canopy that enables seat removal without having to remove the canopy.
The second area that drove many design decisions was basing and ship suitability.
Each variant had unique basing requirements. All three had to be land based, with
CTOL having unique shelter requirements. The STOVL aircraft had basing
requirements on several types of smaller deck ships, and CV aircraft had basing
requirements on large deck Nimitz-class aircraft carriers. Ship-based aircraft
also had many unique requirements that impacted the design including compat-
ibility with the hangar decks and elevators on each type of ship, providing an ade-
quate field of view to the pilot for landing approaches to the ship as well as being
able to see personnel and ship features while on the flight deck, self-contained
ladder access, and compatibility with catapult and arresting gear systems, as
well as many other requirements.
BTP activity for the STOVL variant was defined by Configuration 240-4.1 and
started shortly after SWAT in fall 2004. A few months later, CTOL BTP activity
kicked off and was defined by Configuration 240-4.2. CDRs for both the STOVL
and CTOL variants were held concurrently in early 2006. Configuration 240-4.3
was a STOVL variant – only configuration that was defined to support STOVL
CDR. Similarly, Configuration 240-4.4 was a CTOL variant – only configuration
that was defined to support CTOL CDR. The internal arrangement of Configur-
ation 240-4.3 is shown in Fig. 19. Both CDR events were successful with the
decision to proceed toward completion of detailed design and fabrication of the
SWAT-optimized configurations. Following quickly on the heels of the STOVL
and CTOL CDR configurations was Configuration 240-4.5 which defined BTP
activity for the CV variant.
While the BTP design effort on the SWAT optimized configurations was in
full swing, fabrication, assembly, and checkout of AA-1 progressed leading up
to a successful first flight in December 2006 with Lockheed Martin Chief Test
Pilot Jon Beesley at the controls. Flight test activity with AA-1 continued in
Fort Worth, gaining critical insight into numerous systems and features including
handling qualities, fuel dump, electrical system, and flight controls.
Configuration 240-4.7 was defined to support the Critical Design Review for
CV. Although primarily intended as a CV-focused configuration, 240-4.7 included
updates to the baseline for the other two variants as well. The Configuration 240-4.7
internal arrangement cutaways for all three variants are shown in Fig. 20. CV CDR
was held in June 2007. The event was successful and the decision was made to
continue detailed design and fabrication of the SWAT-optimized CV variant.
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 101

Fig. 19 F-35B internal arrangement (Configuration 240-4.3).

Fig. 20 F-35 variants: CTOL (top), STOVL (center), and CV (bottom).

102 M. A. COUNTS ET AL.

Throughout SDD, as with any aircraft development program, design and

integration challenges were encountered ranging from relatively simple issues
that were solved quickly to highly complex issues that required extensive
efforts and coordination across multiple teams, organizations, sites, and custo-
mers over an extended period of time. The F-35 fuel system is highly
complex, with each variant having many unique features including unique
feed tanks. Managing Center of Gravity (CG) through the fuel burn sequence
across the variants with the full range of possible weapon loadings proved to
be quite challenging, particularly for the STOVL variant, which is highly sensi-
tive to CG location while in jetborne flight. Flight test data identified concerns
with the in-flight stiffness of the open auxiliary inlet doors, necessitating a rede-
sign to improve stiffness of this key component of the STOVL propulsion
system. Early testing of the CV tailhook identified concerns regarding its
ability to capture the arresting wire under all required conditions. A significant
redesign effort was undertaken on the tailhook, and the results were shown to be
highly effective during sea trials [11]. Nacelle ventilation was a function that saw
several iterations occur from the start of SDD through Configuration 240-4 as
data from multiple wind tunnel tests were combined with analysis and the matu-
ration of the surrounding areas of the aircraft. A nacelle vent fan was added for
low speed and ground idle conditions, and improved flow conditions drove relo-
cation of the nacelle vent inlets from the upper fuselage to the lower side of the
aircraft near the wing/fuselage intersection. Abrupt Wing Stall (AWS) was
identified as a potential concern early in the SDD program based on experience
with legacy aircraft. The team conducted numerous high-speed and low-speed
wind tunnel tests to characterize critical areas of the flight envelope, test flight
control settings, and test potential configuration changes. After an intensive
effort, it was determined that control law changes could be used to adequately
address potential AWS conditions [9]. Another challenge that had concerns
rooted in legacy experience was the impact of IPP exhaust on the ground
environment. Numerous tests and analyses were performed to show that the
ground temperatures that resulted from running the IPP according to the pre-
scribed procedures were compatible with ground materials used for runways,
taxiways, hold short areas, hangars, and maintenance areas. One of the most
challenging issues turned out to be fuel dump. For low observable aircraft
that have fuel dump capability, dumping tends to be used in case of emergency
as opposed to routine daily use. However, the operation of STOVL and CV air-
craft from ships requires the pilot to dump fuel often in order to control landing
weight (usually on every flight) without requiring postflight maintenance of the
fuel dump system. The combination of day in and day out routine usage coupled
with low observable aircraft requirements drives the need for a robust fuel dump
design. The team conducted several wind tunnel tests followed by extensive
flight testing across a wide flight envelope on all three variants to identify a
design that effectively balanced all of the requirements to meet the needs of
the warfighter.
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 103

VII. STOVL WEIGHT ATTACK TEAM

Most aircraft programs experience weight growth in the early stages of design and
development, and the F-35 program was no exception. In anticipation of this
reality, a weight management plan was instituted following contract award that
conducted periodic bottom-up weight estimates. This period is always a challen-
ging time in any program’s lifecycle as weight prediction methodologies transition
from a parametric weight estimate basis to a calculated weight estimate based on
preliminary design drawings as airframe layout data become more detailed.
Parametric estimates typically make certain assumptions regarding weight
changes, both growth and savings, as the design database matures, so each
bottom-up weight estimate was compared to the parametric estimate in order
to provide a clearer picture of the estimated weight of each aircraft variant. Fol-
lowing the third bottom-up estimate, it was starting to become apparent that
some of the predicted improvements were not being fully realized, particularly
in the areas of fasteners, shims, harnesses, tubing, and system installations.
There were many contributing factors that led to weight issues in the original
F-35 design. As it turns out, the parametric weight prediction analysis tools were
overly optimistic when compared to the bottom-up estimates. The parametric
tools predicted weight assuming a highly optimized configuration with conven-
tional design, fabrication, and assembly techniques; however, the configuration
at this point in time was not optimized to the level that the parametric data
assumed. This resulted in a gap between the estimated values and the achieved
to date values that couldn’t close. Primary load paths are a good example of
this gap. The prediction methodology assumed relatively simple and straight opti-
mized load paths, when in fact the load paths were much more complex and inef-
ficient. Much of the complexity and inefficiency were driven by previous decisions
that had been made on fabrication and assembly techniques, as well as an overall
philosophy to maintain commonality between variants whenever possible. For
example, the wing carry through was not highly optimized from a loads/weight
perspective because some of the bulkheads lacked enough depth to function
efficiently. The wing box also had some inefficient components, particularly
where the trailing edge spar attached to the fuselage.
The major mate joints added to the weight and structural inefficiencies. Too
much emphasis had been given early in the program to manufacturing efficiency
and takt time at the expense of weight and load path simplicity. The forward
and aft mate joints were termed “quick mate” joints in the original assembly
concept. These joints were held together by a small number of large diameter
fasteners. The joints were fast to assemble, but this concept drove additional
weight into the surrounding webs and flanges, which was not adequately pre-
dicted by the parametric weight analysis tools. The wing-to-fuselage mate
joint was on a waterline and incorporated the pi preform technique that essen-
tially aligned a tab in a slot to be bonded with only a minimal number of fas-
teners. The wing mate turned out to be not only complex and time consuming
104 M. A. COUNTS ET AL.

to assemble, but also heavier than originally predicted. The original upper wing
skin design was a one-piece tip-to-tip concept that turned out to be thick and
complex to manufacture, as well as providing only limited access to systems
installed underneath.
Weight was also adversely impacted in this time period by configuration fea-
tures and details associated with maturing systems. Issues associated with trying to
maintain a common weapon bay across all three variants drove weight into the
design, particularly for the STOVL variant. Propulsion system contractor Pratt
& Whitney was maturing the propulsion design in parallel to the Lockheed
Martin air system design maturation. A parallel development effort like this is a
highly collaborative and iterative environment, and there were a few unexpected
weight/CG issues related to the propulsion system that also had to be accounted
for. Other integration issues that occurred during this period that impacted weight
included thermal management, nacelle ventilation, battery technology, flight
control actuators, canopy, and electronic rack components, as well as several
other smaller and miscellaneous issues.
Finally, the parametric methodology didn’t adequately account for weight in
the area of routing. The actual routing details (tubes, lines, and harnesses) were
discretely added instead of being parametrically predicted; however, certain
routing details and accommodations were included in the parametric data,
which assumed relatively simple and optimized routing paths. In fact, the
actual routing was a significant challenge due to more complex configuration fea-
tures, particularly around the highly integrated lift fan bay on the STOVL variant
and the tightly wrapped engine nacelle on all three variants. In order to get routing
around the nacelle, a cross-ship routing tunnel was incorporated, which added
weight as well as displaced fuel volume.
The design team recognized that the parametric tools assumed an optimized
design, and as a result attempted to include factors where F-35 unique features
varied from traditional design techniques. However, by the fourth bottom-up esti-
mate it was determined that a fundamentally different approach to weight needed
to be implemented.
SWAT was formed in early 2004 to address aircraft weight issues that were
identified in the bottom-up estimate. Weight was a concern for all three variants,
but it was particularly critical for the STOVL variant. The simple and overriding
fact of STOVL aircraft design is that thrust has to be greater than weight or the air-
craft does not work. The primary SWAT effort was focused on the STOVL configur-
ation, but the CTOL and CV configurations also benefitted from weight
improvement as a result of commonality in both systems and structure. The
SWAT activity was a focused six-month intensive effort intended to reduce
weight, improve performance, and close on an optimized KPP-compliant STOVL
design. A dedicated, co-located team was identified, and program leadership gave
the team the time necessary to effectively plan and execute the effort. The initial
focus was on the STOVL and common design, and a broad range of trade space
was explored including airframe and systems weight optimization, air vehicle
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 105

performance, propulsion installed performance, and capability tradeoffs. Trade

study results flowed to the other variants when appropriate, and every effort was
made to preserve commonality where possible. The SWAT team was tasked with
the fundamental responsibility of instilling confidence in a viable STOVL design
while developing a solid plan to recover performance on all three variants.
The SWAT attack plan had six key areas of focus:
1. Reduce zero fuel weight. Based on the most recent installed propulsion
performance, the bottom-up weight estimate suggested weight reduction
and/or thrust performance improvements on the order of 2800 lb were
needed at the air vehicle level to recover VLBB performance. Weight optim-
ization was needed for airframe components (such as bulkheads, shear webs,
longerons, brackets, skins, and access panels), vehicle systems components
(such as landing gear, flight controls, the thermal management system, and
fuel), and mission systems components (such as avionics racks, cards, and
fire control and stores).
2. No stone unturned. Consolidate the Weight Savings/Threat database, renew
and strengthen the Weight Incentive Program (WIP), and review all previous
trade decisions. No weight savings idea was too small to be considered.
3. Reduce required fuel weight. Improve drag and aero performance so that less
fuel is required to perform the mission. Less required fuel onboard at the start
of a mission results in a lighter air vehicle at takeoff, which leads to shorter flat
deck Short Takeoff (STO) and ski jump STO distances, both of which are
KPPs.
4. Increase performance. Optimize STO techniques, optimize Vertical Landing
(VL) control allowances, and improve installed thrust.
5. Challenge requirements. Scrub internal requirements, ground rules, and
assumptions. Prioritize and quantify Joint Contract Specification (JCS)
impacts. Maintain KPPs and quantify impacts associated with ground rules
and assumptions.
6. Quash weight increases. Mitigate weight threats and manage growth to zero.
Do not allow weight increases.
Support, participation, and buy-in from the customer were crucial for the
SWAT effort to succeed. The customer community understood the critical
nature of the SWAT effort, and they were great partners throughout the
process. They gave key insight into the relative importance of air system capabili-
ties and provided a prioritized list of operational attributes that was used to guide
the trade study process. Cost as an Independent Variable (CAIV) is a commonly
used design tool that helps determine the relative value of individual items vs their
cost. Plotting the CAIV data typically shows which items offer the largest benefit
for the least amount of impact. The SWAT team adopted the CAIV concept for
weight by creating a Weight as an Independent Variable (WAIV) curve, which
clearly identified which trade study items offered the largest weight savings
106 M. A. COUNTS ET AL.

Fig. 21 WAIV curve.

potential with the least effectiveness impact. A conceptual example of a WAIV

curve is shown in Fig. 21. The best WAIV items were identified, and teams
were assigned to quickly address these high payoff trade studies.
The SWAT team had to move fast in order to build confidence, meet critical
path schedules while still maintaining the program’s four pillars (affordability,
supportability, survivability, and lethality), and ensure that a rigorous change
management process was followed. Potential trade studies were triaged and prior-
itized with full cross-team coordination that included the customer, and compre-
hensive study tracking was coordinated weekly. Another area that was key to the
SWAT team’s success was the implementation of a streamlined decision process.
A two-tier decision board was stood up that consisted of the SWAT Technical
Coordination Meeting (TCM) where technical details of trade studies were coor-
dinated, and a multiboard that combined several decision-making boards into a
single board that had overall decision responsibility for engineering changes, con-
tractual matters, and business matters. The multiboard concept allowed the
SWAT team to incorporate changes much faster than would have been possible
by using the existing established program decision board structure.

A. AIRFRAME
Airframe-related trades were the most numerous of all SWAT trade studies.
More than 500 studies were completed accounting for approximately 1900 lb of
weight savings. These studies ranged in scope from fastener changes to major
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 107

load path changes with values ranging from more than 100 lb all the way down to
a fraction of a pound. Forward fuselage improvements included skin, fastener,
landing gear attachment, and STOVL F-1 fuel tank changes. Center fuselage
changes included optimization of upper skins and fuel tank covers, fuel floors,
lift fan door drive mechanism, keel beams, and weapon bay door hinge fittings.
Improvements to the wing included changes to the upper wing skins, substruc-
ture, bulkheads, nacelle materials, fasteners, and engine rail. The aft fuselage
improvements included longerons, vertical tail attachments, and the F-5 fuel
tank. The empennage changes optimized vertical tails and horizontal tails, and
eliminated fuel vent space in the vertical tail.

B. VEHICLE SYSTEMS
Vehicle systems –related trade studies numbered in excess of 80 and accounted
for approximately 490 lb of weight savings. These studies included technology
updates, material revisions, design optimizations, and margin analysis. Flight
control actuation improvements included changes to the leading edge flap
drive, actuator reservoirs and valves, as well as Electromagnetic Interference
(EMI) filters. Optimization of the electrical power system included changes to
the starter/generator, 28-V battery, power panels, inverter, and electrical distri-
bution unit. Power and Thermal Management System (PTMS) changes were
dominated by closed loop system architecture and component optimization.
Landing gear savings included optimization of the STOVL nose gear and
STOVL main gear as well as material changes. Hydraulic and utility actuation
improvements included changes to the lift fan inlet door actuation, ground main-
tenance pump, hydraulic accumulator, weapon bay door drive, and materials.
Other vehicle systems changes included modifications to fuel system tubing,
probes, valves, and pumps, as well as optimization of wiring.

C. MISSION SYSTEMS
There were more than 40 mission systems – related trade studies, and they
accounted for approximately 100 lb of weight savings. These studies included
architecture changes, installation changes, technology updates, material changes,
and design maturation. Improvements to sensors, processors, and common com-
ponents included changes to racks and antennas, and repackaging of countermea-
sure installations. Pilot systems changes included optimization of the canopy,
canopy actuator, and ejection seat. Fire control and store improvements included
changes to the ejector racks and adding STOVL unique adapters for the ejector
racks.

D. AIR VEHICLE
Improvements made at the air vehicle level often spanned multiple teams and
stakeholders, and many were among the most beneficial of all SWAT weight
108 M. A. COUNTS ET AL.

Fig. 22 Quick mate (left) and integrated mate (right).

changes. Mate joints are perhaps the most discussed SWAT change and are fun-
damental to how the aircraft is assembled. As previously discussed, the pre-SWAT
design used quick mate joints at the forward and aft joints. These joints were
changed to a more conventional integrated design, as shown in Fig. 22. The inte-
grated design improved the load path across the joint and offered the opportunity
to save weight in the adjacent webs, flanges, and stiffeners.
The pre-SWAT waterline mate of the center fuselage and wing is shown
in Fig. 23. This joint turned out to be more difficult and time consuming to com-
plete than originally thought. As a result, the center fuselage was split with
the portion aft of the engine face being combined with the wing. The new mate
joint utilized the conventional integrated joint philosophy similar to the
forward and aft joints. The resultant post-SWAT air vehicle mate concept is
shown in Fig. 24.
STOVL control surface sizing was another area that saw significant change.
The horizontal tail planform was changed while maintaining area. The flaperon
area was slightly reduced, and the vertical tail area was reduced while the
rudder area slightly increased. These changes resulted in weight savings of
approximately 200 lb split about equally between airframe and vehicle systems
flight control actuation.

Fig. 23 Waterline mate (left) vs conventional mate (right) joint.

F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 109

Fig. 24 Revised mate joints.

Numerous configuration changes were made to reduce drag, increase internal

fuel volume, reduce routing, and improve equipment installations. Examples of
these items include relocation of the nacelle vent inlet, OML changes to the fuse-
lage and to the wing-to-body fairing, nose landing gear door changes that enabled
a reduction in vertical tail size, reorienting the IPP exhaust, relocation and inte-
gration of countermeasures, engine feed tank, upper wing access panels, and
many other smaller changes that are too numerous to mention here.

E. PROPULSION
Increased thrust provides performance improvements in jetborne and semi-
jetborne flight that are just as effective as weight reduction. Propulsion-related
improvements were made in several places including modifications to the
STOVL exhaust that lowered weight while reducing aft body drag. The auxiliary
inlet door actuation system was repackaged and the air surface geometry was
slightly modified to smooth the airflow, which improved pressure recovery. The
main engine inlet throat area was increased by 5%, which also smoothed the
airflow while providing increased mass flow and improved inlet/engine compat-
ibility. The final major propulsion improvement incorporated roll post modu-
lation. The pre-SWAT short takeoff design kept the roll posts flowing
throughout the takeoff roll. The modification to add modulation closed the roll
posts during the early portion of the takeoff, increasing axial thrust through the
main nozzle and therefore allowing the aircraft to build takeoff speed earlier in
the roll. Later in the takeoff roll, the roll posts opened to provide roll thrust
prior to the aircraft leaving the deck. The combination of these propulsion
improvements resulted in an increase to hover thrust of approximately 700 lb, a
110 M. A. COUNTS ET AL.

reduction in flat deck STO distance of approximately 100 ft, and an improvement
to mission radius of approximately 26 miles as a result of reduced drag and
improved fuel burn at cruise.

F. GROUND RULES AND CAPABILITIES

Three ground rules were revised in order to provide KPP performance improve-
ments in VLBB. The Hover Weight Ratio (HWR) at the wave-off point in the
pattern was increased by 1% from the more conservative baseline value to
better align with legacy experience and NASA studies. The vertical landing recov-
ery pattern distance was slightly reduced along with a reduction in the on-deck
fuel reserve. The combination of these ground rule changes resulted in an effective
weight improvement of 745 lb. One additional ground rule was revised to improve
KPP performance in STO. This changed a portion of the ingress and egress pro-
files to take better advantage of optimum Mach and altitude, which resulted in less
fuel required to perform the mission and therefore a lighter weight aircraft at the
start of the mission for short takeoff.
Prior to SWAT, the weapon bay size was common across all three variants.
The CTOL and CV variants had requirements that specified 2000-lb-class
weapon carriage whereas the STOVL variant had a 1000-lb-class carriage require-
ment. If the STOVL variant was to remain viable, it was clear that the beneficial
exceedance of 2000-lb-class carriage could no longer be maintained for the
STOVL aircraft. A smaller weapon bay was incorporated that retained the capa-
bility to carry two 1000-lb-class stores with two missiles. The resized STOVL
weapon bay still met the weapon carriage requirements but allowed for more
weight-efficient load paths and more efficient subsystem installations. External
store capability on Stations 2 and 10 was reduced to 1500 lb, which resulted in
improved flutter characteristics and enabled the horizontal tail to be resized. In
addition, external fuel tanks and dual external store air-to-ground loadings
were limited to subsonic carriage, which enabled reductions in flight control
actuators. Unused MIL-STD-1760 wiring at the air-to-air stations was eliminated,
and certain external coating features were modified. The final capability change
was a reduction in limit speed for STOVL. The combination of these capability
trades resulted in weight savings of 540 lb, a portion of which was included in
the airframe and vehicle systems totals previously discussed.

G. SWAT RESULTS
In just six months, in excess of 600 design changes were incorporated capturing
approximately 2600 lb of weight savings, 700 lb of thrust improvements, and
745 lb of improvements associated with ground rule revisions. With these docu-
mented results, it’s clear that the SWAT team returned the technical design back
to a solid foundation and was highly successful in restoring confidence in the via-
bility of the STOVL variant as well as providing significant improvements to the
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 111

CTOL and CV variants. In addition, an important point to recognize is the inter-

dependency and commonality of the variants. At program award, CTOL was the
lead variant followed by STOVL and then CV. Upon exiting SWAT, the lead
variant was switched to STOVL followed by CTOL and then CV. This decision
was fundamental to the success of the entire program. By putting the STOVL
variant first, some of the most difficult program challenges had to be addressed
immediately, which in some cases led to the discovery of other unknown issues
that had simply been lying in wait. Making this firing order change resulted in
a better STOVL aircraft design earlier in the design sequence than previously
planned, and it also resulted in better CTOL and CV aircraft designs due to com-
monality and similarity benefits.

VIII. WEIGHT MANAGEMENT

The F-35 program incorporated a highly successful weight management program,
and the results of that program are unprecedented in modern fighter design
history. Government and industry experts predicted a specified weight growth
curve based on previous experience. Following SWAT, the Lockheed Martin
team adopted a growth curve that was significantly below what was recommended
by the government and industry experts. The team consistently performed at or
below the adopted curve until the final weight was verified, which covered a
period of approximately 12 years.
Concurrent with and following SWAT, Lockheed Martin, in coordination
with its DoD teammates, implemented numerous changes to effect weight
reduction on the F-35 aircraft. Figure 25 shows the weight-empty status of the
F-35 STOVL variant from the period prior to SWAT to the Initial Operational
Capability (IOC) timeframe. Early on, bottom-up weight assessments were con-
ducted to obtain the best possible weight-empty estimate of the aircraft, but it
was determined to be too heavy by approximately 2800 lb in order to meet the
vertical landing requirement. One team of independent reviewers anticipated
that at best, only two thirds of the excess weight could be redesigned out of the
aircraft. The weight goal was achieved, however, primarily through the efforts
of SWAT. An often-overlooked accomplishment is also inherent in the data.
Weight increases that were encountered as a result of maturing the design were
offset with SWAT items, so the gross weight savings associated with SWAT was
actually much greater than the 2800 lb net savings that was needed to meet the
vertical landing requirement.
Perhaps even more challenging than weight reduction was keeping the weight
off as the program progressed. Some experts thought it would be impossible to
take off so much weight and prevent the typical future weight growth. History
has shown that all aircraft gain weight in the layout and detail design phases of
a program, and it is not unusual to see weight growth curves on the order of
8% or more. Further, it is typical for the weight to grow throughout the lifecycle
112 M. A. COUNTS ET AL.

Fig. 25 STOVL plan to perform weight.

of the design. Given that the F-35 program had a STOVL variant incorporating a
vertical lift design, the weight-empty growth curve had to be much more aggres-
sive. Lockheed Martin decided to adopt a weight growth curve of 3%, which was
unprecedented in military aviation for a major development program like the
F-35. Using historical data from numerous military aircraft, the U.S. Navy
Mass Properties Division’s firm opinion was that a weight growth curve of less
than 6% was unlikely to be achieved. No other “clean sheet” military fighter air-
craft design (discounting aircraft mods) had ever come close to this challenge. As
shown, the F-35 program not only achieved the 3% weight growth curve, but also
did so even after incorporating a 200-lb increase later in SDD that was the result of
recategorizing a piece of mission systems equipment from missionized weight to
empty weight.
The unprecedented STOVL results shown in Fig. 25 are superimposed onto
Fig. 26 to illustrate comparative results for several military jet fighter aircraft. In
addition, the CTOL variant has held the weight-empty curve flat for over 70
months. The CV variant has also demonstrated similar results.
Another area of concern common to many programs is the credibility of the
mass properties database. In the early stages of the program, there was skepticism
regarding the accuracy of the database. Compounding this perception was the fact
that tailored software for large program weight databases is generally not com-
mercially available and must be developed in-house. Quality data are fundamental
to an effective weight management system, and Lockheed Martin has expended
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 113

a significant effort in development of the database for the F-35 program. The
robustness of the mass properties database allowed the F-35 program to use the
database as a repository for trade study results. Armed with these data, manage-
ment decisions on design tradeoffs were made routinely and efficiently. This
was further enhanced by a weighted analysis that balanced unit flyaway cost,
implementation time, and other factors.
The analysis methods for estimating the weight of individual parts were
initially validated by actually weighing the parts. After weighing approximately
35,000 lb of parts, the validation practice was terminated. By using high fidelity
digital solid models, the results of the estimation methods were so accurate that
it became clear further weighing of parts didn’t warrant the expense. Furthermore,
at completion of the first production aircraft, the discrepancy between actual
weight of the aircraft and the database derived estimate was only 0.16% across
the approximate 40,000 records per aircraft. The Lockheed Martin database is
now considered among the best robust mass properties databases in existence.
Moving on to a few of the methods used to achieve these results, one of the
first steps beyond SWAT was the total system approach strategy defined for
weight management and control. The strategy called for active engagement by
senior management, and in order to emphasize that engagement, a review of
the weight management and control strategy was kicked off by the company pre-
sident outlining the importance of weight reduction to the program. Senior man-
agement engagement continues to this day.

Fig. 26 STOVL plan to perform weight vs the historical average.

114 M. A. COUNTS ET AL.

A critical activity that led to weight reduction was the Weight Incentive
Program (WIP). Over its four-year existence, 12,140 ideas were submitted by
employees of Lockheed Martin, Northrop Grumman, BAE Systems, and other
companies under contract. A total of 1148 ideas were reviewed for incorporation,
and 855 ideas approved. Weight reductions for individual ideas ranged from just a
few ounces to more than 70 lb. Employees were monetarily rewarded for their
ideas that were incorporated. Suppliers were similarly incentivized through
contractual actions.
Partnering with the customer was critical. Joint evaluation of design changes
often involved compromises in other areas. Impacts were evaluated and approved
or disapproved by joint government/contractor weekly meetings. Paul Kachurak,
NAVAIR Mass Properties Engineering Branch Head, later said of the joint effort:
Reaction to the weight growth problem was swift and decisive—nearly the
entire government and industry team focused on weight reduction resulting
in a more than ten percent net decrease in weight empty in about ten
months, and a program that was back on track to success. Government and
industry leadership emphasis on fixing the problem set the team on the
right path. They instilled confidence and focused efforts on success in spite
of the negativity that assailed the program.
In considering the above methods for weight reduction, it is important to recog-
nize the interdependency of each activity. Lockheed Martin Aeronautics Senior
Manager of Mass Properties Vearl Durrington noted that “It was the exceptionally
strong willingness to execute, combined with the unprecedented data available
that allowed the program to excel in managing weight.” Indeed, the cultural
change that occurred as a result of the emphasis on weight continues to this day.

IX. TEST AND EVALUATION

An essential element of successfully delivering 5th-Generation capabilities to each
customer, both domestic and international, was the maturation of the air vehicle
design and the verification of capabilities during wind tunnel, ground test, and
flight test activities. Figure 27 highlights the maturation of the configuration
design for all three variants aligned with the associated wind tunnel testing
from contract award in October 2001 through the last Critical Design Review
(CDR) in June 2007.
The Lockheed Martin team successfully executed major wind tunnel test pro-
grams for both the CDP and SDD phases of the program. CDP wind tunnel
testing was used to develop the demonstrator configurations of all three variants
as well as support X-35 flight testing. CDP wind tunnel testing totaled 19,472
hours. Maturation of all three variants during the SDD program required an
additional 49,984 hours spanning 17 wind tunnels in the United States, Italy,
Germany, Netherlands, and the United Kingdom (Fig. 28). Details of the wind
tunnel User Occupancy Hours (UOH) for SDD are provided in Table 1 [12].
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 115

Fig. 27 Testing from ATP to CDR.

The team also used water tunnel testing, as shown in Fig. 29, to supplement wind
tunnel activities in both CDP and SDD. The combined wind tunnel test total of
69,456 hours coupled with the extensive use of CFD and flight test data has
resulted in a highly accurate and correlated aerodynamic database that is unsur-
passed by any other fighter program.
Examples of wind tunnel test findings that led to configuration refinements
include the fuselage chine, aft-hinged lift fan inlet door, lift fan inlet bellmouth
geometry, revised wing camber and twist for CV, and fuselage contour changes
to increase fuel volume while optimizing drag for all variants. Additional

Fig. 28 F-35 wind tunnel testing.

116 M. A. COUNTS ET AL.

TABLE 1 SDD WIND TUNNEL TEST SUMMARY

User Occupancy Hour Summary (Actuals thru Mar. 30, 2010)

Actual Hours per Test Discipline

Year Aero Stability & External Store Propulsion Flutter Total

Analysis Control Environment Separation Aero
2002 3975 1435 868 1175 2269 192 9914
2003 3464 3168 677 578 2060 – 9947
2004 1991 2582 810 373 2892 – 8648
2005 4054 3319 556 732 2935 – 11,596
2006 2677 2315 468 342 1253 – 7055
2007 780 – 308 445 250 – 1783
2008 – – – 264 – – 264
2009 – – – – – – –
2010 – – 777 – – – 777
Actual 16,941 12,819 4464 3909 11,659 192 49,984
thru
Mar.
30,
2010
Plan 16,941 12,819 4464 3909 11,659 192 49,984
Complete 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%

changes identified through wind tunnel testing include wing area changes to the
CV variant for improved approach speed, contour refinements for drag improve-
ments, horizontal and vertical tail planform changes, and the split nose landing
gear door, which was incorporated during SWAT for all three variants, enabling
a reduction in vertical tail size.
CDP full scale structural testing subjected the X-35 aircraft to 100% Design
Limit Load (DLL) prior to flight through a total of 12 tests. Maturation of the
three variants to support SDD requirements totaled 547 strength tests and
included 115% DLL, 150% DLL, and store loads greater than 150% DLL [13].
Durability testing of the CTOL and STOVL variants has completed and CV is
scheduled to complete in 2018. The summary of structural testing is provided
in Table 2.
CDP flight testing began with the X-35A in October 2000 followed by the X-35C
starting in December 2000 and concluded with the X-35B between June and August
2001. The demonstrator flight test program totaled 139 flights with variant totals
shown in Table 3. SDD flight testing encompassed a significantly larger scope in
order to provide the data necessary to verify program performance requirements
and to support flight certification for the three U.S. services, international partners,
and Foreign Military Sales (FMS) customers. The SDD flight test program has
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 117

Fig. 29 F-35 water tunnel testing.

totaled 9213 flights through 9 February 2018 utilizing 18 flight sciences and mission
systems flight test aircraft. Details of the SDD flight test program are provided in
Table 3. Testing primarily occurred at the Edwards Air Force Base, California,
and Naval Air Station Patuxent River, Maryland, flight test centers and also
included ship deployments for both the STOVL and CV variants [14].

X. SUMMARY
The Lockheed Martin F-35 design team successfully met the challenge to develop
an excellent performing multinational three-variant configuration composed of
complex, highly integrated systems and solidly anchored by the four pillars of
affordability, lethality, survivability, and supportability. The team effectively
balanced the unique requirements of each service and customer while optimizing

TABLE 2 SDD STRUCTURAL TESTING SUMMARY

Number of strength tests performed on full-scale static test articles:

F-35A, AG-1 174
F-35B, BG-1 214
F-35C, CG-1 159
Total Tests 547
Simulated flight hours performed on full-scale durability test articles:
F-35A, AJ-1 24,000
F-35B, BH-1 16,000
F-35C, CJ-1 18,761 (as of 2 March 2018)
118 M. A. COUNTS ET AL.

TABLE 3 FLIGHT TEST SUMMARIES OF CDP AND SDD

CDP Flight Test Summary SDD Flight Test thru 9 Feb 2018
Variant Flights Variant Flights
X-35A 27 F-35A 3538
X-35B 39 F-35B 3579
X-35C 73 F-35C 2096
Total 139 Total 9213

performance to produce a balanced design that took advantage of commonality

where possible and uniqueness as needed to meet the needs of the individual cus-
tomer in order to deliver a game-changing 5th-Generation platform. The con-
figuration methodically evolved from JAST to CDP and then on to SDD while
taking advantage of an impressive set of test activities and an unprecedented
weight control program to deliver an outstanding family of three variants that
includes the first 5th-Generation supersonic STOVL aircraft and the first
5th-Generation carrier-based aircraft.

ACKNOWLEDGMENTS
The authors would like to thank Kevin Renshaw for his insight into the early
stages of the program, Joshua Harris for the use of his X-35 photograph taken
at the National Air and Space Museum, and Doug Moore for creating the pictorial
lineage of the configuration development. The authors would also like to thank
Dan Sturdevant and Jonathon Curtis for their assistance in the development of
this paper and Thomas Mellies for coordinating the Subject Matter Expert
content review by the F-35 Joint Program Office.

REFERENCES
[1] Wiegand, C., et al., “F-35 Air Vehicle Technology Overview,” AIAA Aviation
Forum, June 2018 (to be published).
[2] Wurth, S., Smith, M., and Celiberti, L., “F-35 Propulsion System Integration,
Development & Verification,” AIAA Aviation Forum, June 2018 (to be published).
[3] Lemons, G., Carrington, K., Frey, T., and Ledyard, J., “F-35 Mission Systems Design,
Development & Verification,” AIAA Aviation Forum, June 2018 (to be published).
[4] Frey, T., Aguilar, J., Engebretson, K., Faulk, D., and Lenning, L., “F-35 Information
Fusion,” AIAA Aviation Forum, June 2018 (to be published).
[5] Sheridan, A., and Burnes, R., “F-35 Program History—From JAST to IOC,”
AIAA Aviation Forum, June 2018 (to be published).
F-35 AIR VEHICLE CONFIGURATION DEVELOPMENT 119

[6] Steidle, C. E., “The Joint Strike Fighter Program,” Johns Hopkins APL Technical
Digest, Vol. 18, No. 1, 1997.
[7] Hehs, E., “X to F: F-35 Lightning II and Its Predecessors,” Code One, Second Quarter,
2008.
[8] Robbins, D., et al., “F-35 Subsystems Design, Development, and Verification,”
AIAA Aviation Forum, June 2018 (to be published).
[9] Harris, J., and Stanford, J., “F-35 Flight Control Law Design, Development and
Verification,” AIAA Aviation Forum, June 2018 (to be published).
[10] Hayward, D., Duff, A., and Wagner, C., “F-35 Weapons Design Integration,”
AIAA Aviation Forum, June 2018 (to be published).
[11] Wilson, T., “F-35 Carrier Suitability Testing,” AIAA Aviation Forum, June 2018
(to be published).
[12] Parsons, D., Eckstein, A., and Azevedo, J., “F-35 Aerodynamic Performance
Verification,” AIAA Aviation Forum, June 2018 (to be published).
[13] Ellis, R., Gross, P., Yates, J., Casement, J., Chichester, R., and Nesmith, K., “F-35
Structural Design, Development, and Verification,” AIAA Aviation Forum, June
2018 (to be published).
[14] Hudson, M., Glass, M., Hamilton, T., Somers, C., and Caldwell, R., “F-35 SDD
Flight Testing at Edwards Air Force Base and Naval Air Station Patuxent River,”
AIAA Aviation Forum, June 2018 (to be published).
 CHAPTER 3

F-35 Air Vehicle Technology Overview

Chris Wiegand , Bruce A. Bullick†, Jeffrey A. Catt‡, Jeffrey W.
Hamstra§, Greg P. Walker} and Steve Wurth
Lockheed Martin Aeronautics Company, Fort Worth, TX

The Lockheed Martin F-35 Lightning II incorporates many significant

technological enhancements derived from predecessor development
programs. The X-35 concept demonstrator program incorporated
some that were deemed critical to establish the technical credibility
and readiness to enter the System Development and Demonstration
(SDD) program. Key among them were the elements of the F-35B
short takeoff and vertical landing propulsion system using the revolu-
tionary shaft-driven LiftFanw system. However, due to X-35 schedule
constraints and technical risks, the incorporation of some technologies
was deferred to the SDD program. This paper provides insight into
several of the key air vehicle and propulsion systems technologies
selected for incorporation into the F-35. It describes the transition
from several highly successful technology development projects to
their incorporation into the production aircraft.

I. INTRODUCTION
The F-35 Lightning II is a true 5th-Generation, trivariant, multiservice air system.
It provides outstanding fighter class aerodynamic performance, supersonic speed,
all-aspect stealth with weapons, and highly integrated and networked avionics.
The F-35 aircraft features many technological enhancements in air vehicle and
propulsion subsystems derived from predecessor programs. These include the
Subsystems Integration Technology (SUIT) studies [1– 7], Joint Advanced
Strike Technology (JAST) program, Air Force Research Laboratory’s (AFRL’s)

 Systems Engineering Senior Staff Specialist, F-35 Power and Thermal Management System.
†
Director, F-35 Air Vehicle.
‡
Systems Engineering Senior Manager, F-35 Vehicle Sciences and Systems.
§
Lockheed Martin Senior Fellow, F-35 Air Vehicle.
}
Systems Engineering Senior Manager, F-35 Propulsion and Flight Controls.
 Lockheed Martin Fellow, F-35 Propulsion and Flight Controls.

121
122 C. WIEGAND ET AL.

Advanced Compact Inlet Systems (ACIS) program, Vehicle Integration Technol-

ogy Planning Studies (VITPS) studies [8, 9], More-Electric Aircraft (MEA) studies
[10], and Joint Strike Fighter (JSF)/Integrated Subsystems Technology (J/IST)
demonstration program [11 – 20]. Additionally, numerous Independent Research
and Development (IRAD) and Contract Research and Development (CRAD) pro-
jects were completed that formed a part of the F-35 design basis [21 – 36]. Many of
these technological enhancements were not incorporated into the X-35 demon-
strator due to schedule constraints to complete the flyable demonstrator aircraft.
They were also postponed due to the results of a technical risk assessment that
made them undesirable candidates for the X-35. Instead, the full development
and integration of these technologies were deferred for incorporation during
the Engineering and Manufacturing Development (EMD) phase (later known
as the System Development and Demonstration [SDD] program). This resulted
in the final F-35 design configuration.
Enormous efforts from these less well-known predecessor projects produced
many of the significant technical achievements that provided necessary technical
maturity and risk reduction. This allowed Lockheed Martin to proceed with them
confidently in the EMD proposal. The production F-35 incorporates a highly inte-
grated air vehicle subsystems architecture that reduces overall aircraft size and
takeoff gross weight. It does so by replacing the federated, individual subsystems
used in other legacy aircraft. Low Observable (LO) technologies are incorporated
into the engine inlet and exhaust nozzle, and the F-35B Short Takeoff and Vertical
Landing (STOVL) propulsion system, with its revolutionary integrated flight pro-
pulsion controls, provides unprecedented capabilities. The system represents a
revolutionary step-increase in vertical lift, compared to predecessor aircraft. Its
fault-tolerant controls are seamlessly integrated with the aircraft control laws,
minimizing pilot workload across the entire flight envelope from hover to super-
sonic flight [36].
The key air vehicle and propulsion systems technologies selected for incorpor-
ation are depicted in Fig. 1. This paper focuses on the evolutionary path followed
to develop these technologies. The final F-35 aircraft subsystems [35], propulsion
[36], and mission systems [37], as well as the SDD program, are discussed in
greater detail in supporting publications. Each of the items featured in Fig. 1 rep-
resents a significant aircraft capability enhancement that added to the overall
success of the F-35 configuration. Successes in the associated development pro-
grams for these led to their incorporation into the F-35 design baseline entering
the SDD program.
The various development projects that evolved into the systems configurations
used in the F-35 spanned the 1990s during the period preceding the winner of the
JSF competition. Figure 2 provides key development milestones leading to the
incorporation of selected technologies into the F-35 program.
The J/IST integrated subsystems development occurred in parallel with the
Concept Development Program (CDP). Interestingly, in it the various JSF compe-
titors cooperated in a collaborative environment, sharing all results and data. This
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 123

Fig. 1 Advanced technologies selected for F-35 air vehicle and propulsion
systems incorporation.

Fig. 2 JSF program air vehicle and propulsion systems technology development roadmap.
124 C. WIEGAND ET AL.

allowed the risk reduction activities associated with the integrated vehicle systems
to be pursued without the need to encumber the Concept Demonstrator Aircraft
(CDA) schedule, and enabled the final results and lessons learned to be incorpor-
ated into the F-35 at the outset of the SDD program.
During the same period, numerous IRAD and CRAD studies evaluating
potential propulsion innovations continued. As with the J/IST results, several
of these were incorporated into the F-35 after the SDD contract award. Significant
technical risks associated with the Diverterless Supersonic Inlet (DSI) and LO axi-
symmetric nozzle, and STOVL propulsion system configurations were retired in
parallel with the CDA work, culminating in flight demonstrations showing the
maturity and efficacy of the concepts. As an example, dual-redundancy features
of the STOVL exhaust nozzle were developed in parallel with the CDA
program and incorporated during SDD.

II. INTEGRATED AIR VEHICLE SUBSYSTEMS

A. EARLY DESIGN STUDIES
Aircraft subsystems have traditionally been designed using a federated approach
consisting of a number of independently designed subsystems. JSF-sponsored
studies showed the potential for significant benefits from integrating these func-
tions. This applied to three subsystems: flight controls and actuation systems, elec-
trical and auxiliary power systems, and the Environmental Control System (ECS).
Results from the studies showed that the effective integration of these three key
subsystems could significantly improve aircraft performance and dramatically
reduce the amount of equipment required. In doing so, it would provide improved
affordability and warfighting benefits essential to F-35 program goals.
In conventional systems, electrical, hydraulic, pneumatic, and mechanical
power are generated and distributed from the engines and auxiliary power
system. The government-funded SUIT [1 –7], MEA, and VITPS [8– 9] studies
showed the potential benefits of eliminating or shrinking the conventional centra-
lized hydraulic system. They also showed the potential benefits of reducing engine
bleed air extraction [1– 7]. AFRL sponsored SUIT in the late 1980s and early
1990s to look into what could be gained from more efficient integrations of
selected air vehicle subsystems [39]. The original SUIT concept was to replace
single-function subsystem equipment with multifunctional hardware, potentially
reducing volume and weight while increasing overall efficiency. Later, the objec-
tive of using the engine’s fan air stream as a sink for waste heat from the subsys-
tems was added. Concurrently, AFRL’s propulsion laboratory was independently
pursuing MEA technology, including electrically powered flight control actuation
and robust electric power generation and distribution system concepts. As a result,
MEA system and component technologies were undergoing development and
testing through several separate and independent initiatives.
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 125

Between 1994 and 1995 the JAST program identified key technology building
blocks to support the development of an advanced strike capability. The idea was
to screen candidate technologies for their applicability based on their respective
payoffs with regard to the four JSF program pillars: affordability, lethality, survi-
vability, and supportability. At that time, three Weapons Systems Contractors
(WSCs) were actively competing to win JSF: Lockheed Martin, Boeing, and the
McDonnell Douglas (later Boeing St. Louis)/Northrop Grumman/BAE Systems
team. The candidate JSF configuration was expected to be a single-seat, single-
engine strike aircraft, largely due to affordability considerations. Originally, the
JSF platform focused on Air Force and Navy customers; however, during JAST
the government concluded that the Advanced STOVL (ASTOVL) concept devel-
opment should be rolled into JAST/JSF. Thus, the STOVL jet was added to the JSF
design space [39]. JAST initiated the VITPS studies [8, 9]. All three WSCs con-
cluded that SUIT and MEA technologies could be combined synergistically in a
strike aircraft. Accordingly, they recommended that the SUIT/MEA combination
be pursued under JAST. Each WSC advocated pursuing integrated subsystem
technology, and from that advocacy the J/IST demonstration program was
conceived.

B. JSF INTEGRATED SYSTEMS TECHNOLOGY DEMONSTRATION PROGRAM

The goals of the J/IST demonstration program were to set the stage for the JSF SDD
program and sharpen the focus for the target JSF platform configuration. Ulti-
mately, that configuration was shaped by the winning proposal for what would
become the largest acquisition program in Department of Defense history. The
J/IST program was to be executed by all three WSCs competing for the JSF.
Each WSC was supporting its own JSF proposal team while simultaneously
being responsible to the other JSF WSCs for the technical results of the J/IST work.
The J/IST government team ensured that each participating WSC team was
contractually accountable to and executing on behalf of the other WSCs. Each par-
ticipant was treated by the government as customers for the technologies being
demonstrated. This arrangement provided a level playing field using WSC
input to make all key decisions within the scope and resources of the program.
This fostered trust, a sense of fair play, and government responsiveness within
the otherwise highly competitive JSF program environment. In turn, that colla-
borative environment enabled the results of each element of the studies to be
capitalized by each competing contractor. Direct involvement of the three
WSCs was fortuitous, because it compelled expanded involvement and
cooperation beyond what otherwise might have been the case [39]. Each WSC
benefitted from the combined J/IST demonstration results, because these pro-
vided information to be used in each WSC’s preferred weapons system configur-
ation aircraft proposals.
The focus of the J/IST demonstration program was to reduce the risks associ-
ated with subsystem integration technologies. This applied specifically to MEA
126 C. WIEGAND ET AL.

technologies, including Switched-Reluctance (SR) Starter/Generators (S/Gs),

Electro-Hydrostatic Actuation System (EHAS) integration, and Thermal/Energy
Management Module (T/EMM) integration through a series of maturation
efforts. The work included developing and flight testing prototype versions of
the SR S/G and Electro-Hydrostatic Actuation (EHA) flight control actuators, suc-
cessfully reducing the risks in these technologies from high to low for the SDD
program. The work was divided into three main focus areas: one focusing on the
electrical power and flight control actuation system (led by Lockheed Martin),
one focusing on the development of T/EMM-related technologies (led by Boeing
St. Louis), and one dedicated to independent benefit assessments (led by Boeing
at its Seattle facility).
Figure 3 compares a conventional federated subsystems architecture to the
F-35 integrated subsystems architectures. The integrated architectures reduced
system parts count, which led to smaller, lighter, and lower-cost aircraft. Based
on those conclusions and the configuration design developed for the SDD
program, these systems were incorporated into the F-35 baseline.
The J/IST program was instrumental in reducing the risk of integrating these
technologies prior to entering the JSF SDD program in 2001. The J/IST con-
clusions indicated an overall reduction in Life-Cycle Costs (LCCs) projected to
be 2– 3%, compared to the LCCs of a similar legacy configuration [19]. The
benefits of these integrated technologies are highlighted in Fig. 4.

1. J/IST ELECTRICAL POWER SYSTEM AND ELECTRO-HYDROSTATIC ACTUATION SYSTEM

The Lockheed Martin –led team achieved the following key accomplishments in
the J/IST program:
a. Development and flight test of an SR, dual-channel S/G system to provide a
fault-tolerant, redundant 270-VDC electrical power source for a single-
engine fighter aircraft

Fig. 3 F-35 integrated vehicle systems.

F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 127

Fig. 4 Benefits of integrated subsystems technologies.

b. Development of a dual-redundant, flight-qualified electric actuation system

c. Development of a flight-qualified 270-VDC battery system to provide unin-
terruptible electrical power to the flight-critical 270-VDC Electrical Power
System (EPS)
d. Development of a flight-qualified emergency power system to provide a sec-
ondary power source
e. Completion of specific technology demonstration tests on the S/G to verify
operation under certain fault conditions
f. Modification, integration, and flight test of the above technologies in a single-
engine Advanced Flight Technology Integration (AFTI)/F-16 aircraft (Fig. 5)
g. Validation of MEA technologies with an AFTI/F-16 demonstration for the
F-35 subsystems suite

J/IST Electrical Power System

Immediately following J/IST contract award, Hamilton Sundstrand (now a UTC
Aerospace Systems [UTAS] company) was selected to develop and qualify the SR
generator system. UTAS was involved in the initial Air Force SR technology devel-
opment and proposed using the same 250-kW, 270-VDC generator design.
However, UTAS rated it for 80 kW and packaged it for the F-16 application.
The primary focus was on using an SR generator system, but the fault tolerance
and redundancy of the 270-VDC power were emphasized as well. This included
128 C. WIEGAND ET AL.

Fig. 5 AFTI/F-16 J/IST aircraft MEA technologies and transition to F-35.

an emergency generator and a 270-VDC battery. The AFTI/F-16’s EPS was modi-
fied significantly to support the J/IST program’s MEA technologies. These con-
sisted of one 270-VDC EHAS for the five primary flight control surfaces, two
270-VDC fuel pumps, one 270-to-28-VDC converter, and one 270-VDC-to-
115-VAC inverter.
AFTI/F-16’s pre-J/IST electrical system was a combination of an F-16 Block
15 production system and a digital flight control power system (production Block
40), receiving only 115-VAC and 28-VDC power. Therefore, to support the legacy
F-16 equipment and MEA systems, the electrical system was modified to provide
270-VDC, 115-VAC, and 28-VDC power. During the initial program’s design
phase, it was determined that the MEA systems would consume the most
power. Accordingly, the primary power type would be 270-VDC power. Also,
because the EHAS is a flight-critical system, the 270-VDC system would be
designed to be fault tolerant and provide limited uninterruptible power. For
EPS integration, the primary challenges were:

. SR S/G flight certification

. Fault-tolerant power generation and distribution
. 270-VDC power system stability with multiple variable power loads
. 270-VDC fill-in battery operation
. Electromagnetic Interference (EMI) and Electromagnetic Compatibility (EMC)
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 129

The following combination of components provided the baseline design for the
J/IST Engine S/G (ES/G) system. The Inverter/Converter/Controller (ICC)
for the ES/G system was taken from the UTAS LV100 SR S/G system. The
power electronics converter and the SR generator were taken from the UTAS/
General Electric Integrated High-Performance Turbine Engine Technology
(IHPTET) research program. The ES/G system was required in two tests: the
AFTI/F-16 flight test and a ground test. For the flight test it was used to demon-
strate generation capability in an MEA application. For the ground test, it demon-
strated motor, start, and generation capabilities. The ground test application
included a demonstration of starting a Pratt & Whitney F119 engine and transi-
tioning to generate mode. The ES/G system was designed to accommodate all
applications; however, the ICC package design was driven by the AFTI/F-16
installation. Both the flight and ground demonstrations used nearly identical
hardware, with minor changes made to adapt to their respective operating
environments. The resultant J/IST MEA architecture is shown in Fig. 6.
The ground demonstrations exercised all the capabilities of the ES/G system,
specifically its engine motoring and starting, power generation, and fault toler-
ance. The T/EMM motor/generator system provided 270-VDC power to the
ICCs for engine motoring and starting. The ES/G system was transitioned
from start to generate to provide 270-VDC power to two primary motor buses.
Modifications were made to the ES/G and ICCs from the AFTI/F-16 configur-
ation to enable bidirectional power flow and increased speed range operation.

Fig. 6 AFTI F-16 J/IST electrical power and actuation flight demonstration architecture.
130 C. WIEGAND ET AL.

J/IST Electrically Powered Flight Control System

Parker Aerospace was selected to provide the EHAS for the F-16. The company
had previously developed prototype EHAs for various ground and flight test pro-
grams; among them were the flight tests of an aileron actuator on the Lockheed
Martin C-130 high-technology testbed. Previous electric actuation integration
programs focused on single-surface operation or non-flight-critical control sur-
faces. Two examples are the F/A-18 Electrically Powered Actuation Development
(EPAD) and F-15 Fly by Light Advanced Systems Hardware (FLASH) programs.
By contrast, J/IST focused on integrating all primary flight control surfaces,
with no hydromechanical backup system. This bold approach was essential to
convincingly prove that the MEA concept supported the JSF technology transition
criteria.
Lockheed Martin worked closely with Parker Aerospace to define the EHAS
system-level requirements, integration test requirements, and software configur-
ation management. The development and testing of the EHAS was the most chal-
lenging of the contracted tasks. The selection of the F-16 as the flight vehicle
provided known air vehicle and system-level requirements, which made the devel-
opment task easier. However, multiple hardware and software development chal-
lenges were encountered that provided valuable lessons learned for the
SDD program.
The Lockheed Martin team conducted a detailed design of two EHASs: one to
replace the F-16 flaperon and horizontal tail Integrated Servo-Actuators (ISAs),
and one to replace the F-16 rudder ISA (Fig. 7). The designs incorporated a
common power electronics package for all five actuation systems. The Lockheed

Fig. 7 Key AFTI F-16 modifications.

F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 131

Martin team also provided an analog interface with the existing F-16 Digital Flight
Control Computer (DFLCC). To do so, it developed and integrated a separate
interface box that required no major flight control software or hardware
changes to the DFLCC. This approach resulted in four major component
designs: an F-16 flaperon/horizontal tail dual-tandem EHA, an F-16 rudder dual-
tandem EHA, a common power drive electronics package, and an EHA interface
controller electronics unit.
EHAS integration testing of all five actuators with electronics was successfully
completed at Parker Aerospace’s facility. The testing verified that the hardware
and software met all design requirements. Further, it provided detailed visibility
into the system’s operation and added confidence in the EHAS software. It also
reduced risk in the integration of the system into the AFTI/F-16 aircraft. The
few technical issues discovered were resolved in the laboratory environment,
avoiding the risks of discovery on aircraft. Throughout the on-aircraft flight
control integration and test, minimal EHAS changes were required.
The power and actuation flight demonstration provided key technical proof to
mature MEA technologies for the JSF EMD phase. This demonstration tested the
external S/G, the 270-VDC power distribution system, and the EHAs in a realistic
aircraft environment. It provided valuable integration and installation data for
thermal environment, EMI/EMC, and supportability, among other areas. The
AFTI/F-16 modifications were successfully demonstrated in the aircraft, with
the key highlights including:

. Nine flights accomplished totaling 13 flight hours, performing realistic JSF-like

mission profiles
. Flight envelope including Mach 1.3 (600-kt engine limitation), up to 30,000 ft
altitude (ferry)
. Test pilots reporting no observable difference in handling qualities, compared
to a baseline F-16

For its achievements, the AFTI/F-16 flight demonstration won Flight Inter-
national magazine’s 2000 Aerospace Industry Award for Engineering, Mainten-
ance and Modification, presented at the Paris Air Show in 2001 [40].

2. J/IST THERMAL/ENERGY MANAGEMENT MODULE

Boeing St. Louis (formerly McDonnell Douglas) led an industry team responsible
for the subsystem architectures and ground demonstrations of power and cooling
integration, and electric actuation and power system integration [11]. Major par-
ticipants in the team were Honeywell Aerospace (formerly AlliedSignal Aerospace
Inc.), BAE Systems (formerly Astronics Corporation), Moog Inc., Northrop
Grumman, and Pratt & Whitney. The team achieved the following key accom-
plishments in the J/IST program:
132 C. WIEGAND ET AL.

1. Development of an integrated T/EMM system:

a. T/EMM turbomachine
b. High-temperature air/fuel heat exchanger
c. Air/liquid heat exchanger
d. Engine fan duct air-to-air heat exchanger
e. Dual-mode recuperator heat exchanger
f. T/EMM integral S/G
g. T/EMM controller and vehicle management computer interface
2. Development of the T/EMM system controls and modes of operation
3. Demonstration of integrating the power and cooling subsystems into a
stand-alone integration environment
4. Demonstration of the integration of the T/EMM integrated subsystems with
the engine

The team selected a generic, JSF-like aircraft for assessing and defining the J/IST
requirements. These included requirements for power, cooling, and actuation
systems of the aircraft for ground, flight, and emergency operating conditions.
The subsystem architecture that was developed resulted in the consensus architec-
ture configuration depicted in Fig. 8. This architecture used the engine fan duct as
a heat sink and integrated the functions traditionally performed by the ECS,
Auxiliary Power Unit (APU), and Emergency Power Unit (EPU). The system was
designed to support requirements for electrically powered flight control actuation
and electric start of the main engine using T/EMM power to drive the ES/G.

Fig. 8 J/IST T/EMM architecture.

F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 133

Thermal/Energy Management Module Requirements

The T/EMM system architecture was intended to perform all functions normally
accomplished by a traditional aircraft ECS, auxiliary power system, and emer-
gency power subsystems. The T/EMM was required to provide additional capa-
bilities expected to be required in the F-35, including cooling high-powered
electronics. This requirement drove the requirement in J/IST for a backup
cooling capability to accommodate inflight failures and provide active cooling
for Navy hangar deck maintenance. The operating requirements created a need
to develop multiple modes of operation and reconfigure the system between
these modes as required to perform the differing functions. Consequently, the fol-
lowing operating modes were developed and demonstrated in the J/IST effort:
. Mode 1.0: Electrically powered hangar deck cooling
. Mode 2.0: Stand-alone T/EMM start
. Mode 3.0: Cooling and electrical power for ground maintenance
. Mode 5.0: Electrical power for main engine start
. Mode 6.0: Engine bleed air-driven cooling and triplex electrical power (normal
flight)
. Mode 7.0: Engine bleed air-driven cooling and emergency electrical power
(failure of the ES/G)
. Mode 8.0: Emergency electrical power (failure of the ES/G)
. Mode 9.0: Stand-alone emergency electrical power—stored air (engine failure
at high altitude)
. Mode 10.0: Stand-alone emergency cooling and electrical power—ambient air
(engine failure at low altitude)
. Mode 11.0: Shutdown
. Mode 12.0: Emergency cooling—fuel heat sink (not demonstrated in J/IST)

Thermal/Energy Management Module Turbomachine and Component Development

The design and development of a full-scale T/EMM turbomachine was accom-
plished by Honeywell Aerospace in Phoenix, Arizona. The turbomachine con-
figuration incorporated a power turbine, compressor, SR Integral S/G (IS/G)
rotor, and cooling turbine on a single spool that operated at speeds up to
60,950 rpm (Fig. 9). The compressor and power turbine were mounted on a
single shaft, and the IS/G and cooling turbine were mounted on a separate
single shaft. The two shafts were connected by a floating quill shaft, resulting in
a design that provided sufficient margin from high-speed shaft bending modes.
This split allowed each section to be developed and tested independently.
The power turbine used a variable geometry stator with movable vanes that
allowed for performance optimization at the various operating conditions. The
134 C. WIEGAND ET AL.

turbomachine included a canister-type combustor developed specifically for the

T/EMM that featured three distinct modes (tri-mode combustor) and was sub-
sequently patented by Honeywell.
Additional components making up the T/EMM configuration requiring sig-
nificant development were the SR I/SG configuration, T/EMM controller, and
various other heat exchangers, valves, and related components.
In parallel with the Honeywell T/EMM development projects, Pratt &
Whitney integrated the Honeywell-designed fan duct heat exchangers into a
modified F119 engine. This was another key success supporting the claim that
the T/EMM concept was viable and supported the JSF technology transition
criteria.

T/EMM System Demonstrations

Several significant demonstrations were accomplished in support of the J/IST
team effort. Figure 10 summarizes the major demonstrations, building up from
the component level to successively higher levels of integration. One of the
three most significant demonstrations was the integration of the power and
cooling subsystems into a stand-alone integration environment. Another of the
three was the electrical integration of T/EMM and ES/G, high-power EHA inte-
gration testing, and EMI testing. The third was the demonstration of the inte-
gration of the T/EMM integrated subsystems with the engine.
The stand-alone integration demonstration began in the summer of 1999 at
Honeywell in Torrance, California. The demonstration, which simulated a
mission profile, included ground maintenance operation with 25 kW of electrical

Fig. 9 T/EMM turbomachine cross-section.

F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 135

Fig. 10 Key T/EMM-related J/IST demonstrations.

power and 12 kW of cooling. It also included normal-engine-bleed air-driven

operation during taxi, takeoff, acceleration, climb, cruise, loiter, descent, and
landing. In addition, it showed emergency operations that simulated the loss of
the engine and ES/G failure. The results demonstrated the capability to
perform all the required modes and successfully reconfigure between modes as
required. Post-test comparisons of test data with the dynamic model confirmed
the model’s ability to predict results. The comparisons also reinforced the value
of the virtual modeling effort prior to commencing hardware operations.
The electrical power management integration of T/EMM and ES/G was
conducted at Northrop Grumman’s facility in El Segundo, California. The
136 C. WIEGAND ET AL.

testing demonstrated that the IS/G-integrated EPS could provide the needed
capability to power the ES/G using IS/G power to start the F119 engine. The
testing also showed that the system could provide the necessary emergency
power requirements, and that its generators could handle the dynamic loads
imposed by the high-power EHAs.
The final engine integration demonstration was performed in 2000 at Pratt &
Whitney’s facility in West Palm Beach, Florida. It brought Honeywell’s T/EMM
system, Northrop Grumman’s EPS, and Hamilton Sundstrand’s (UTAS’s) ES/G
together with a modified Pratt & Whitney F119 engine. With this combination,
it validated the integrated subsystems concept. The Hamilton Sundstrand
(UTAS) ES/G was connected to the engine using a speed-increaser gearbox.
The Honeywell T/EMM system was connected to the engine with low-pressure
drop plumbing. The demonstration included testing using a simulated mission
profile incorporating operations in all ground, flight, and emergency modes.
Mission segments included taxi, climb, cruise, loiter, descent, dash, and combat.
The demonstration validated the integrated subsystems concept by operating
the system successfully in all required conditions. It showed that the T/EMM
could be successfully driven by engine bleed air, and it demonstrated the success-
ful integration of engine and airframe systems. It also demonstrated electrical
engine starting and motoring and showed that the required emergency power
could be provided within 50 milliseconds without using a battery. The
maximum starting torque of 131 lb/ft using electric power furnished from the
T/EMM was also shown.

C. TRANSITION TO THE F-35 PROGRAM

J/IST was one of the largest and most successful subsystems technology demon-
stration programs ever accomplished. Every major element of the J/IST architec-
ture was transitioned to the F-35’s design [39]. The program was composed of a
series of cascading development and demonstration projects of increasing com-
plexity. Each demonstration built upon a prior one to gradually reduce com-
ponent, subsystem, and air vehicle – level risks. These efforts showed the various
technology elements in the most appropriate environment, considering the
desired risk reduction, available resources, and perceived potential. The two
final J/IST demonstrations were the most significant. One was a comprehensive
ground demonstration of the engine and relevant (associated) subsystems.
The other was a flight test of an F-16 with all primary flight control surfaces
powered electrically.
The F-35 employs the key elements derived from J/IST, including the EHAS
and EPS architectures. It also incorporated numerous elements from the T/EMM
architecture and equipment. The Pratt & Whitney F135 engine uses integral fan
duct heat exchangers and additional bleed manifolds to support the Power and
Thermal Management System (PTMS) configuration. The F-35 EPS features the
Hamilton Sundstrand (UTAS) ES/G and General Electric 270-VDC and
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 137

28-VDC lithium-ion batteries. The Flight Control System (FCS) features Moog/
Parker Aerospace EHAS hardware.
The resultant F-35 PTMS uses many of the control modes and requirements
developed in J/IST. The components used in the system derive directly from the
J/IST configuration. Many lessons learned in J/IST were applied to the design,
development, and test of the turbomachine and its subcomponent systems, the
heat exchanger arrangements, the valve designs, and other system aspects.
The F-35 turbomachine was redesigned to eliminate the variable area nozzle
and replace the SR generator with a permanent magnet generator. The F-35
PTMS has a robust, highly reliable electrical power and cooling system, providing
an electrically driven engine start system and supporting inflight emergencies.
The F-35 aircraft subsystems enable stand-alone ground maintenance power
and cooling for all systems maintenance and checkout operations. The single-
stage power turbine was replaced with a two-stage radial/axial configuration,
and the variable area power turbine nozzle was eliminated. The SR generator
was replaced by a permanent magnet generator configuration. Numerous design
lessons learned were incorporated into the lubrication system, sump sizing, rotor
element design and clearances, system controls, and operation. As a result of
J/IST, Honeywell developed significant improvements for the modeling and
simulation techniques applied when developing the SDD program. The F-35
PTMS heat exchanger configuration used a thermal cycle similar to that in the
J/IST system. However, significant optimization in the packaging was made by
Honeywell. This resulted in a highly integrated compact heat exchanger configur-
ation with multiple cores in a single assembly that proved essential to meeting
the installation design requirements. F-35 design requirements also permitted
the elimination of the stored-energy system and avoided the need to integrate
the T/EMM turbomachine exhaust into the main engine exhaust. The require-
ments also mitigated other complexities discovered during J/IST.

1. F-35 FLIGHT CONTROL ACTUATION SYSTEM

The F-35 FCS uses the EHAS to power its primary and secondary flight control
surfaces. This is a departure from predecessor legacy combat aircraft powered
by hydraulics. The conventional system design was reliable and had a mature
design concept, but it added substantial weight and volume and drove the hydrau-
lics system’s sizing and redundancy. The SUIT and MEA studies showed that the
more-electric architecture promised significant improvements [4]. The F-35 flight
control actuation system provides control to position both of the primary and sec-
ondary flight control surfaces. The FCS architectures are largely common for all
three F-35 variants for the primary control of horizontal tails, flaperons,
rudders, and leading edge flaps. However, the F-35A and F-35C variants also
incorporate horizontal tail centering actuators, and the F-35C variant also incor-
porates conventional hydraulically powered ailerons. The architecture of the F-35
flight control actuation system derived from the MEA studies and J/IST
138 C. WIEGAND ET AL.

demonstration program. Integrating the flight control actuation system with the
power and cooling systems was key to the overall success of the F-35 flight
control development. Additional discussion of this can be found in Ref. [35].

2. F-35 ELECTRICAL POWER SYSTEM

The F-35 EPS provides the generation, distribution, control, and protection of
electrical power for various equipment. Its key power management functions
for the air vehicle include primary flight control power, power to numerous air-
craft systems, main engine starting, emergency power control and distribution,
and power for ground maintenance functions. Key system elements include a
single ES/G, two ICCs, 28-VDC and 270-VDC batteries, and the additional
elements required to provide the uninterruptable power sources required for
safety of flight. The EPS provides 270-VDC, 28-VDC, and 115-VAC power.
The architecture of the F-35 EPS was derived from the MEA studies and J/IST
demonstration program. Integrating the main engine starting function and
flight control power sources matured risk reduction activities during the F-35
development program. One major change made during the SDD program was
the replacement of the original SR machine ES/G with a synchronous generator
to accommodate a broader power generation capability. The F-35 EPS design and
development is further explored in Ref. [35].

3. F-35 POWER AND THERMAL MANAGEMENT SYSTEM

The F-35 PTMS, sometimes referred to as the Integrated Power Package (IPP),
integrates the conventional functions of the ECS, emergency power system,
engine starting system, and auxiliary power systems into a single, highly inte-
grated system. The system features two primary modes of operation: stand-alone
combusted-mode operation and bleed-driven operation when the F135 engine is
running. The primary power and cooling enables stand-alone ground mainten-
ance with no required external power and cooling carts. The PTMS also provides
primary power for on-ground engine starting, followed by seamless reconfigura-
tion into bleed-driven operation to support flight operations. The PTMS likewise
supports inflight emergencies and automatically reconfigures into combusted-
mode operation to support flight control, emergency electrical power, and inflight
engine-start assist power.
The PTMS provides 270-VDC and 28-VDC electrical power, as well as
forced-air cooling for flight-essential systems. It also provides liquid cooling for
aircraft avionics systems and pressurization for the cockpit, fuel system, and
other aircraft systems. During flight, waste heat generated by onboard systems
is rejected overboard via engine-mounted heat exchangers embedded within the
F135 engine fan air duct. This eliminates the weight and volume penalties associ-
ated with conventional ram air heat-sink systems.
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 139

The system incorporates many elements of the architectures developed in

the J/IST demonstration program and MEA studies; however, substantial
additional system design maturation was accomplished during F-35 development.
Several of the concepts studied in J/IST were not adopted because they were not
needed to meet the program requirements. Among these were a stored-energy
backup system and an integrated T/EMM exhaust system integrated with the
main engine exhaust [35]. The resultant configuration provides improvements
in aircraft range due to reduced bleed air consumption and improved thermal
management. The integrated PTMS eliminates the need for the separate aircraft-
mounted accessory drive gearbox, air turbine starter, EPU, APU, and ECSs
employed in legacy aircraft. The fan duct heat exchanger configuration avoids
the increased weight and drag penalties and LO impacts associated with a dedi-
cated ram air circuit.

III. PROPULSION TECHNOLOGIES

A. OVERVIEW
The conventional F-35 propulsion system features critical propulsion/airframe
integration technologies that began with technology development programs.
These technologies were used in large-scale demonstrations, incorporated onto
the X-35 demonstrator, and transitioned to the production F-35 program. Key
propulsion features (Fig. 11) include a DSI and LO axisymmetric nozzle configur-
ation (LO Axi) common among all variants. The F-35B STOVL variant includes a

Fig. 11 F-35 advanced propulsion system technologies.

140 C. WIEGAND ET AL.

three-bearing swivel module (3BSM) that integrates the LO Axi configuration

with thrust vectoring, providing vertical lift. The lift is augmented by a shaft-
driven lift fan system providing additional vertical lift. The DSI design is charac-
terized by a detailed, shaped compression surface and forward swept cowl, with
twin inlet apertures feeding a bifurcated, serpentine duct. The design eliminates
the need for boundary-layer diverter or bleed-system inlet subsystems, reducing
cost and weight. The LO Axi configuration minimizes radar reflections by using
serrated trailing edges, a serrated interface with the airframe, and tight gap
and seam control. It also has specialized high-temperature coatings on internal
and external surfaces to provide an excellent balance of signature, weight, and
performance requirements.
The F-35B STOVL propulsion system utilizes the same main engine turboma-
chinery as the Conventional Takeoff and Landing (CTOL)/Carrier Variant (CV)
configurations with the addition of a shaft-driven lift fan, a roll control system,
and an auxiliary inlet. The 3BSM adds thrust vectoring to the LO Axi nozzle.
This permits engine exhaust to either pass straight through for forward propulsion
in conventional flight or be deflected downward to provide aft vertical lift. The
3BSM can move through 95 deg of motion seamlessly with no change in engine
operation. The nozzle also provides yaw control during hover and transitions to
hover. Additionally, the F-35B STOVL variant incorporates Rolls-Royce’s shaft-
driven LiftFanw system. In lift mode, horsepower is extracted from the F135
engine’s low-pressure turbine via a drive shaft, clutch, and gearbox, and is used
to drive the LiftFan. The exhaust is discharged through a thrust vectoring
nozzle on the underside of the aircraft to provide balanced lift with the 3BSM.
Engine fan air ducted to outboard roll nozzles provides roll control.

B. F-35 DIVERTERLESS SUPERSONIC INLET

The F-35’s primary engine air induction system features a DSI design conceived
and matured through several technology programs and then transitioned to
the JSF program. The DSI is characterized by a 3-D compression surface and
forward-swept cowl (Fig. 12). These features enable high aerodynamic perforce,
boundary layer diversion, and inlet stability without using a boundary layer diver-
ter or bleed system. Eliminating these inlet subsystems reduces cost and weight
compared to prior state-of-the-art designs. Although many different configur-
ations were explored in the technology programs, the F-35’s DSI is embodied
in twin, side-mounted inlets that feed a compact bifurcated duct. A detailed
history of the DSI concept can be found in Refs. [21] and [22].

1. BACKGROUND
Tactical aircraft have always posed a formidable challenge for engine inlet
designers, and the incorporation of modern affordability and survivability
requirements have made the challenge more difficult [23, 24]. The inlet must
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 141

Fig. 12 F-35 engine air induction system features.

provide the engine with high-quality (high pressure, low distortion) airflow over
a wide range of speeds, altitudes, and maneuver conditions. At the same time, it
has to accommodate the full range of engine airflow from idle to maximum after-
burning power. The inlet designer must also consider the constraints imposed by
configuration features, such as nose landing gear, weapon bays, equipment bays
and access panels, and forebody shaping.
In addition to these general considerations, two key aerodynamic require-
ments are at the forefront in the design of any supersonic inlet system. The first
requirement is for flow compression. The inlet system must reduce the airstream’s
speed while increasing its static pressure as airflow approaches the engine. For
combat aircraft, this is usually done with a series of external shockwaves and
internal flow area expansion. As freestream speeds approach Mach 2, elaborate
compression schemes, including movable compression ramps, were historically
used to reduce losses and enable high inlet efficiency.
142 C. WIEGAND ET AL.

The second key issue is Boundary Layer Control (BLC). This is the means by
which the inlet system will account for a layer of low-energy air that forms on
the surface of the fuselage and compression surfaces. This must be managed at
both subsonic and supersonic speeds. The boundary layer can create chaos
when disturbed by shockwaves created during flow compression. Shockwave/
boundary layer interaction can lead to severe airflow distortion at the engine
face, which may subsequently lead to engine stall. Several methods can be used
for BLC. The inlet can be physically isolated from the fuselage by a boundary
layer diverter, a feature found on most of today’s combat aircraft. Another
primary technique is boundary layer bleed. Bleed systems may be fully fixed or
involve mechanical variation, such as movable exit louvers, to optimize perform-
ance. Many of today’s tactical aircraft use a combination of bleed systems, diver-
ters, and compression ramps.
Variable compression and bleed systems can provide the aerodynamic
functionality required for a high-performance inlet. However, such features
also introduce mechanical and structural complexity, weight, and cost into the
system [24].

2. DIVERTERLESS SUPERSONIC INLET CONCEPTUAL DEVELOPMENT

In the early 1990s Lockheed Martin began an IRAD project to develop a Mach 2
class combat aircraft inlet concept. The concept would embody traditional aero-
performance levels and advanced survivability features. Further, it would improve
affordability (reduced cost and weight) compared to state-of-the-art design con-
cepts. To meet these goals, the concept would need to incorporate flow com-
pression and BLC functionality, advanced shaping, high structural efficiency,
and minimal or no moving parts. These studies were conducted primarily with
Computational Fluid Dynamics (CFD) tools augmented with small-scale wind
tunnel testing.
The DSI emerged as the preferred concept early in the IRAD program. It is
distinguished by two main physical features: a fixed, 3D compression surface
(bump) and an edge-aligned, forward-swept cowl. The bump compression
surface derives from the flow field produced by a reference axisymmetric body
in supersonic flow. The reference body (virtual cone) may be a simple cone, a
double or isentropic cone, or any of these bodies at incidence angle to the oncom-
ing stream. In the latter case, the flow field is 3D, not axisymmetric. A set of CFD
particle traces are released along a locus of points representing the intersection of
the shock field and aircraft surface. As the particles travel into the shock field, they
are deflected away from the virtual cone by internal flow field pressure gradients.
A 3D contour is then defined by a surface faired through the particle traces. When
introduced into an identical supersonic flow field, this contour produces a shock
structure identical to that of the virtual cone. The bump surface not only achieves
flow compression, but also creates a span-wise static pressure gradient that assists
with boundary layer diversion (Fig. 13).
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 143

Fig. 13 CFD simulation of DSI supersonic boundary layer diversion.

The second physical feature distinguishing the DSI is a forward-swept cowl.

Forward swept denotes that the cowl structure is cantilevered from the aircraft
and closes against the forebody at its aft-most points. This geometry enables
low-pressure boundary layer air to be spilled out the side of the inlet as mass
flow ratio is decreased. Many CFD studies showed the effects of combining a
bump compression surface and forward-swept cowl. The studies demonstrated
that the combination could provide the aerodynamic functionality of traditional
inlets without a boundary layer diverter or bleed system.
Small-scale testing of isolated inlet aperture models (Fig. 13) was conducted
throughout the conceptual development program to augment CFD studies.
Whereas CFD was best at providing detailed flow physics at a few specific operat-
ing conditions, testing produced key operating data at a broad range of conditions.
These tests were used to evaluate inlet pressure recovery, distortion, and shock-
wave instability (buzz) boundaries for the key bump design parameters.

3. INTEGRATED AIRCRAFT DESIGN AND VALIDATION

Having verified the basic DSI principles, the focus shifted to integrating the inlet
into a realistic aircraft configuration and assessing air vehicle system-level
impacts. These activities were accomplished with synergistic studies on the
JAST and AFRL ACIS programs. JAST studies investigated the best way to inte-
grate the unique DSI features. As with the eventual F-35 configuration, the JAST
aircraft featured twin, side-mounted inlet apertures with a bifurcated duct and
single engine. On ACIS, Lockheed Martin performed three tasks that helped to
build the overall knowledge base. We analyzed the benefits, cost, weight, and
maintainability of the DSI vs a reference system. We also assessed the perform-
ances of different subsonic diffuser concepts integrated with the DSI. In addition,
we performed a large-scale wind tunnel validation of a forebody/inlet/duct con-
figuration, shown in Fig. 14.
144 C. WIEGAND ET AL.

Fig. 14 Large-scale wind tunnel testing validation of CFD design methodology and CFD test
model (inset).

System-level trade studies were performed to quantify the weight, cost,

and benefits of the DSI, compared to more conventional inlets (e.g., F-22 and
F/A-18E/F caret inlet systems). In these studies, a 30% inlet weight reduction
was estimated for the DSI, relative to the reference caret inlet. The largest con-
tributing factor was the elimination of the bleed and bypass systems. Studies per-
formed by other ACIS contractors [25] indicated similar savings for diverterless/
bleedless systems.
The ACIS subsonic diffuser study evaluated the DSI with both straight and
serpentine diffusers and different flow area curves. Results from this task provided
guidelines on how to best shape the diffuser, given the rather unique incoming
flow field from the DSI.
A 0.125-scale inlet/forebody model of the ACIS configuration was tested at
subsonic, transonic, and supersonic conditions in two test entries. A wind
tunnel test was conducted at the NASA Langley Research Center 16-Foot Transo-
nic Tunnel. Another was conducted at the Lockheed Martin High Speed Wind
Tunnel, which is a 4-by-4-ft tri-sonic blow-down wind tunnel. Test data were
obtained over a wide range of angle-of-attack and angle-of-sideslip conditions at
speeds up to Mach 2. Test results indicated that the integrated inlet configuration
satisfied typical goals of high pressure recovery, manageable distortion, and good
supersonic stability. Pressure recovery for the DSI with a serpentine duct was
shown to be comparable to that of the F-16 modular common inlet duct recovery.
This was at static conditions and speeds ranging from Mach 0.6 to 1.2. At super-
sonic speeds, the DSI’s recovery was shown to exceed F-16 recovery.
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 145

The development of the DSI was based on CFD. As such, it was necessary to
verify the accuracy of CFD to predict the complex inlet flow fields associated with
integrated forebody, aperture, and duct geometries. In general, excellent agree-
ment between the CFD and test data was noted (Fig. 14).

4. DIVERTERLESS SUPERSONIC INLET FLIGHT DEMONSTRATION PROGRAM

The next step in the development of the DSI was to conduct a flight demonstration
on a relevant combat aircraft platform. A Block 30 F-16 powered by an
F110-GE-129 engine was selected for this demonstration due to its modular
inlet construction and consequently low cost of modification. One objective of
this effort was to demonstrate engine/inlet compatibility throughout the
combat aircraft envelope. Another was to demonstrate stable inlet operation up
to Mach 1.8. Although integration in the F-16 chin location dictated an inlet
design different from the twin, side-mounted ACIS/JAST design, the aerody-
namic design principles were identical.
Prior to the flight test, an inlet wind tunnel model was fabricated and tested at
one of the Arnold Engineering Development Complex propulsion wind tunnels in
November 1995. Wind tunnel test results verified CFD-predicted inlet perform-
ance and demonstrated engine/inlet compatibility. Inlet lines were frozen based
solely on CFD results, and flight hardware fabrication had begun prior to any
wind tunnel test verification of the CFD-based design. The flight test program
consisted of 12 flights flown in nine days in December 1996. The flight tests
covered the entire F-16 flight envelope and achieved a maximum speed of
Mach 2 (Fig. 15). With this flight demonstration, the viability of the DSI as a
Mach 2 class, diverterless, bleedless, highly survivable inlet concept was proven.

Fig. 15 DSI technology was matured through rigorous F-16 flight demonstration.
146 C. WIEGAND ET AL.

5. TRANSITION TO THE JSF PROGRAM

Risk had been significantly reduced through IRAD efforts, ongoing work and the
flight demonstrations on ACIS, and extensive program trade studies. As such, the
DSI system was selected as the baseline inlet for Lockheed Martin’s JAST program
in early 1995, replacing the previous F-22-like inlet. Configuration refinement
continued through the lines freeze on the X-35 concept demonstrator aircraft.
Because of the sophisticated yet simple DSI design, it was possible to support
this milestone with only a single additional up-and-away inlet wind tunnel entry.

C. LOW OBSERVABLE AXISYMMETRIC NOZZLE

Before the development of the F-35 Low Observable Axisymmetric Nozzle
(LOAN), signature demands typically drove nozzles to fixed, structurally inte-
grated affairs (e.g., F-117). They had to have very high aspect ratio designs (e.g.,
F-117) or highly capable but heavy two-dimensional systems (e.g., F-22), as illus-
trated in Fig. 16. Departing from what was then the state of the art, industry and
CRAD efforts developed multiple nozzle configurations to create a LOAN for the
F-35. The F135 engine with a LOAN balanced the requirements of LO and effi-
cient aeromechanical performance. This resulted in a lightweight configuration
with reduced radar cross-section.

1. BACKGROUND
The F-16’s and F-15’s exhaust systems are examples of what was used in classic
4th-Generation tactical supersonic aircraft. These systems are axisymmetric for
low weight and structural efficiency, with variable geometry to maintain stable
and efficient engine operation. Specifically, afterburning requires a large increase
in the minimum nozzle flow area (throat area) to retain engine stability. This is
particularly necessary when fuel is being dumped into the exhaust and the resul-
tant flow density decreases. These convergent/divergent nozzles are composed of
overlapping flaps and seals. The exit area and throat area are mechanically linked
and scheduled with a power setting for efficient flow expansion. Externally, over-
lapping flaps provide a fairing between the air vehicle aft body and nozzle exit for

Fig. 16 A spectrum of fighter aircraft nozzle designs.

F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 147

reduced drag. However, the nozzles were not generally considered capable of
meeting LO requirements.
The subsonic F-117 stealth fighter and B-2 bomber exemplify how exhaust
system designs can be dominated by LO features. This effect results in nontradi-
tional exhaust systems that are driven by air vehicle shaping. The F-117 incorpor-
ated an airframe-mounted, structurally integrated, fixed exhaust system with
planform-aligned edges. The exhaust system transitions from axisymmetric to a
high-aspect two-dimensional design and forgoes the variability needed to accom-
modate afterburning. Although these highly integrated designs can reduce drag,
they are less aerodynamically and structurally efficient than axisymmetric designs.
IRAD and CRAD investments were critical to positioning the industry to
develop world-class solutions targeted at major program innovations. Well
before the JSF X-planes were contracted, multiple vertical lift and exhaust
system configurations were developed and matured. These were then available
on call to meet the evolving needs of both new and existing aircraft designs.
Some of the nozzle technologies explored were conformal fully fixed aperture
nozzles, fixed aperture nozzles, and LOANs.
Complexly shaped, fixed-geometry nozzles were matured for reduced drag
and signature. These designs allowed flexibility to implement full line-of-sight
obscuration. With few moving parts, they allowed for a great diversity of LO
and thermally tolerant materials.
Innovations in convergent section mechanical manipulation and shaping were
explored as well. These were intended to reduce leakage, weight, and structural
integration penalties while enabling the nozzle exit to remain motionless.
Throat skewing and inducing shocks in the divergent section were assessed for
thrust vectoring potential. Variable cycle engines were evaluated to reduce the
need for large-scale nozzle throat area control, even during augmentation.
During this period the notion that axisymmetric nozzles were not amenable to
LO (or even thrust vectoring) was being challenged. A new generation of LO
nozzles emerged that was characterized by shaping features to minimize radar
reflections. Namely, they used a serrated trailing edge, serrated interface with the
airframe, and interleaved external seals to complement the external flaps. They
retained tight control of gaps and seals and had specialized high-temperature
coatings on internal and external surfaces. In the years leading up to the X-35,
various LOAN configurations were developed, ground tested, and flight tested
on F-16s. Testing was done in concert with both Pratt & Whitney F100 and
General Electric F110 engines. These versions incorporated a nod to an ejector
feature that introduced nacelle bay airflow near the nozzle throat to cool divergent
flaps. This improved divergent seal durability and reduced the infrared signature.

2. LOW OBSERVABLE AXISYMMETRIC NOZZLE RAPID PROTOTYPE EVALUATION

In the early 1990s, engineers at Lockheed Martin used CFD and an astatic thrust
measurement facility to perform rapid prototype evaluation and 3D printing. This
148 C. WIEGAND ET AL.

Fig. 17 Lockheed Martin rapid nozzle prototyping capability in Fort Worth, Texas.

allowed for a rapid assessment of nozzle aerodynamic performance (Fig. 17). The
facility had a flow-through six-component balance housed within an altitude
(pressure reduction) chamber. Varying ambient backpressure allowed continuous,
low-load testing with very high nozzle pressure ratios with minimal variation in
mass flow and minimal model loads. This also allowed the balance to operate
within an optimum band of its calibrated mass flow and force measurement
range. It had excellent overall accuracy and repeatability and reduced the variation
in the Reynolds number. Multiple airflows could be independently controlled and
metered via a bank of critical-flow venturis.

3. AIRCRAFT AERODYNAMIC INTEGRATION

Intentional aft body and nozzle integration is critical to balance drag and
weight, which are driven by aft body and nozzle boattail angles. Finding the
optimum length and external shape is important to net propulsion perform-
ance. Additionally, the aft body is typically not axisymmetric, and integrating
an axisymmetric exhaust may result in large aft body boattail angles and base
regions. This would introduce drag, effectively stealing from the net thrust of
the propulsion system. This is evident in most twin-engine configurations in
which a base or dead region is normally found between two closely spaced
nozzles. In certain circumstances, the low local pressure endemic to these
base regions can serve as an exit for secondary flow systems. For example,
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 149

base regions at the root of the F-16 horizontal and vertical tail surfaces are used
to promote nacelle ventilation.

4. GROUND AND FLIGHT TESTING

The LOAN configuration developed by Pratt & Whitney was developed under a
precursor to the JSF contract to evaluate advanced, affordable technologies appli-
cable to the F-35. Ground testing was completed in 1996 (Fig. 18). A Lockheed
Martin/Pratt & Whitney team modified an Air Force F-16 and F100-PW-200
engine with a bailed LOAN from the F-35 Joint Program Office. This was used
for the ground test in a two-day rapid prototype operation. During tests from
idle to maximum afterburner, infrared images, nozzle temperatures, and nacelle
inlet pressures and airflow velocities were measured. The Pratt & Whitney
LOAN configuration significantly reduced radar cross-section and infrared signa-
ture emissions from the engine, as well as maintenance costs. The result was a
low-cost nozzle system that reduced the chance of radar and infrared detection
and applied to both retrofit and new-production aircraft.
The solution developed by General Electric was designated the LO Axi Nozzle,
demonstrated by Lockheed Martin on an Air National Guard F-16C (Fig. 18).
Dramatic temperature reductions provided by the LO Axi Nozzle were expected
to greatly improve F-16 exhaust system durability. F-16 flight certification was
conducted at Edwards Air Force Base in the summer of 2001. That system was
offered as an F-16 upgrade option, reinforcing the vibrancy of the F-16 platform.
These F-16 LO nozzle configurations included an ejector feature to address
thermal degradation of nozzle divergent section seals and flaps. That remains a

Fig. 18 General Electric LO Axi Nozzle (left) and Pratt & Whitney LOAN (right) ground and
flight testing on the F-16.
150 C. WIEGAND ET AL.

Fig. 19 Radio frequency test fixture with the F-35 LO nozzle at the Pratt & Whitney
facility in West Palm Beach.

top driver of maintenance on modern jet engines. The ejector provided effec-
tive film cooling to reduce nozzle temperatures and improve component life
using engine nacelle bay bypass air. These techniques were expected to double
or quadruple nozzle divergent flap life, resulting in significant maintenance
cost savings.

5. RADIO FREQUENCY TESTING

For the F-35, the ability to see before being seen provided a fulcrum upon which to
balance sensor/radar capability and LO technologies; strengths in one area pro-
vided flexibility in the other. The strength of emerging technologies in the radar
and sensor suites opened up options for using the LO nozzle. This was comparable
to the manner in which the development of high off-boresight missiles reduced
the appetite for thrust vectoring and nose pointing.
Radar testing was employed for the JSF program to evaluate the unique shape
and special coatings evident in the Pratt & Whitney LOAN. The configuration
achieved stealth through a combination of technologies, including geometric
shaping, advanced cooling, and special coatings on internal and external surfaces.
One major radio frequency model (Fig. 19) developed under IRAD was suitable
for integrating inlets, nozzles, apertures, edges, and other subsystems. It was
designed for installation on the 30,000-lb rotator at the Meridian Antenna Test
Range in Meridian, Texas.

6. TRANSITION TO THE JSF PROGRAM

For the F-35, the Pratt & Whitney F135 engine and LOAN balanced the require-
ments of LO and efficient aeromechanical performance. It offered a lightweight
(especially for the F-35B STOVL variant), low-cost configuration. The F-35A
and F-35C variants use the same nozzle configuration. A shorter version was
readily configured for compatibility with the F-35B STOVL 3BSM to meet
ground clearance needs while landing vertically. Because the engine exhaust
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 151

system is the primary contributor to aft sector infrared signature, engine and
nozzle design needed to incorporate effective methods to reduce infrared emis-
sions. This was accomplished using reduced radar cross-section-compatible tech-
niques, including hiding, shaping, and temperature control. The F-35 exhaust
system employs a cooled turbine face blocker, effectively eliminating the tempta-
tion to employ more impacting techniques like a serpentine exhaust duct. The
F135 exhaust system does use a cooled nozzle to significantly reduce the aft
sector infrared signature. With these techniques, the cooled blocker and nozzle
tail-on infrared signature is significantly less than the signature of an uncooled
exhaust system.

D. F-35B STOVL LIFT SYSTEM

1. BACKGROUND
For more than 50 years, fighter aircraft designers have vigorously pursued the
speed and range of a conventional jet while achieving the basing flexibility of
Vertical Takeoff and Landing (VTOL). Numerous STOVL concepts have been
developed over the decades, all with compromises that limited the effectiveness
of the aircraft. The F-35B STOVL lift system successfully achieved a breakthrough
that redefines the relationship between conventional thrust and vertical propulsive
lift. Moreover, it achieves that with major increases in performance, efficiency,
and safety. This elegant integration results in a relatively simple engine-driven
LiftFan. It has an enabling engine powerful enough to achieve a lift-to-thrust
ratio of approximately 1.5-to-1 (Fig. 20)—a significant increase over direct lift
designs. The shaft-driven LiftFan provides high levels of thrust augmentation

Fig. 20 Revolutionary step increase in vertical lift.

152 C. WIEGAND ET AL.

with a cool, low-pressure footprint, sufficient control power, and efficient packa-
ging in the airframe design. The main engine is primarily optimized for conven-
tional flight, so the propulsion system performance is not compromised for its
vertical lift capability. The LiftFan augments vertical flight similarly to the way
an afterburner augments high-speed performance [31]. The LiftFan provides an
additional ingenious benefit: the (relatively cool) thrust exhaust protects the
main engine inlet and forward portions of the aircraft from hot gas re-ingestion
or damage.
In addition to achieving powerful lift thrust, a STOVL aircraft must achieve
sufficient control power in each axis to successfully transition through the wing-
borne, semi-jetborne, and jetborne flight phases. The F-35B STOVL lift system
accomplishes this through several key components (Fig. 21):

1. LiftFan clutch and driveshaft: To selectively transfer power from the main
engine to the LiftFan
2. Variable Area Vane Box Nozzle (VAVBN): To control the LiftFan exit area
and fore-aft thrust vectoring
3. Roll post nozzles: To redirect main engine fan air through under-wing nozzles
for roll control
4. 3BSM: To vector the main engine nozzle fore – aft and laterally for yaw control

Fig. 21 Comparison of F-35 conventional and F-35B STOVL lift systems.

F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 153

2. LIFTFAN DEVELOPMENT
The F-35 LiftFan system is the overarching characteristic of the F-35B STOVL
variant. It underwent years of technology development and maturation by Lock-
heed Martin and F-35 propulsion contractors Pratt & Whitney and Rolls-Royce.
Initial work began in the late 1980s with STOVL JSF studies sponsored by the
Defense Advanced Research Projects Agency. Lockheed Martin, General
Dynamics, Boeing St. Louis (then McDonnell Douglas), and Boeing all developed
concepts with different technologies for generating vertical lift [34]. These studies
led to the ASTOVL competition that Lockheed Martin won with the shaft-driven
LiftFan propulsive concept. This effort eventually evolved into the JSF concept
demonstration phase resulting in the X-35B flight demonstration.
Rolls-Royce’s LiftFan is a novel, counterrotating concept with a bladed disk
(blisk), two sets of stationary vanes, and a set of Variable Inlet Guide Vanes
(VIGVs). The VIGVs provide the thrust variation from maximum to idle necess-
ary for the VTOL application. The gearbox distributes 29,000 hp to the LiftFan
rotor stages. The load capacity and envelope characteristics were key to providing
an industry-first 30-to-1 horsepower-to-weight ratio. The previous norm (in
earlier aircraft) was a ratio of no more than 15-to-1, which was then doubled.
The gearbox is integral to the LiftFan unit and employs counterrotating output
shafts to simplify geometry and reduce gear and bearing loads. VIGVs on the first
fan stage provide thrust modulation. Lubrication for the LiftFan bearings and
gearbox is provided by the LiftFan lubrication system, which is independent
from the main engine lubrication system. The Rolls-Royce LiftFan is designed
to operate throughout the entire speed range of the main engine.
One of the key challenges in transitioning the concept development to pro-
duction was in the LiftFan’s aeromechanical rotor modes. These caused operating
restrictions (time at certain LiftFan speeds) on the X-35B. The spatial pressure dis-
tortions in the inlet flow field excited resonance modes in the LiftFan turboma-
chinery, becoming a high-cycle fatigue or aeromechanics concern. This was
addressed in the F-35B by redesigning the upper LiftFan door configuration to
reduce flow angularity and distortion. It was also addressed by redesigning the
LiftFan rotor (hollow blades, blisk) that attenuated the modal responses.

3. LIFTFAN CLUTCH DEVELOPMENT

Shaft/clutching functionality was achieved with both hardware and software
functionality. Pioneering shaft, clutch, and gearbox designs permitted the devel-
opment of a lightweight, high-speed (8000 rpm) drive train. A unique closed-loop
clutching system provides precise control resulting in smooth, reliable power
transmission to the LiftFan. This innovative clutch design, leveraging aircraft
brake technology, produced a dry clutch plate arrangement. This achieved the
required rapid engagement performance time, while providing durability that
exceeded program requirements. The clutch is mounted to the LiftFan, with the
154 C. WIEGAND ET AL.

input directly coupled through the main drive shaft and couplings to the main
engine low-pressure rotor shaft. The clutch consists of a pack of dry disk plates.
When driven together by aircraft-powered hydraulic actuation, the pack
couples the main engine low-pressure rotor via the drive shaft to the LiftFan.
Lubrication for the clutch bearings is provided by the LiftFan’s lubrication
system. The driveshaft couplings can flex to take up misalignment between the
main engine and the clutch.
The LiftFan clutch allows the engagement and disengagement of the LiftFan
from the main engine. It achieves this through two devices, each providing a
torque path from input to output. During engagement, speed synchronization
and acceleration of the fan rotors at low power is achieved by applying pressure
to a pack of five carbon-carbon plates, operating dry. Subsequent engagement
of a locking spline is required for high power transmission. Engaging the
splined lock requires synchronizing the clutch input and output shaft speeds
within a few rpm. An indexing mechanism insures against a failure to engage
due to mating splines contacting end to end. During disengagement, the clutch
plate pack unloads the splines to enable them to be retracted.
One of the key challenges experienced during the X-35 development was
obtaining smooth clutch engagement with minimal transition time. Early clutch
control design encountered a chatter phenomenon as the clutch plates came in
contact. Through innovative closed-loop control modes, a combination of clutch
clamping force and longitudinal position feedback solved the chatter problem,
permitting smooth and precise engagements. Continued maturation during
the F-35 program intended to complete the conversion in the minimum time
(operational flexibility) and obtain a full-life clutch (minimized maintenance
interval). The F-35 clutch can complete an engagement cycle within nine
seconds from command to engage. With improved clutch plate material, the
system will accommodate more than 1500 engagements.

4. VARIABLE AREA VANE BOX NOZZLE DEVELOPMENT

Prior to the F-35 development phase, the X-35 LiftFan nozzle vectoring was
accomplished via a three-hooded telescoping nozzle. Although very precise in
directing the LiftFan thrust vector, it was heavy, required a lot of volume, and
was difficult to integrate into the aircraft signature. This prompted the pursuit
of a more compact design through a series of parallel vanes that could be
hidden behind lower fuselage doors. The development of the F-35 VAVBN capi-
talized on five years of prior efforts by Rolls-Royce on vane box nozzles. These had
been developed for earlier lift engine concepts that were considered in the JAST
program [35]. It was tested in a 27%-scale demonstration (Fig. 22).
Many nozzle design variables were studied including duct geometry, and
the number, spacing, and profiles of the movable vanes. Additional design par-
ameters that influenced nozzle integration with the LiftFan included the
gearbox profile, the location of six support struts, and the size and location of
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 155

Fig. 22 Twenty-seven-percent-scale
F-35B STOVL VAVBN test, with (inset)
VAVBN close-up.

the vane actuator mechanism.

Studies were also conducted to
evaluate the tradeoffs of per-
formance, vane actuation, and
airframe integration. From these
it became apparent that the
vane box configuration with six
highly cambered vanes with a
low thickness-to-chord ratio was
most promising. In addition to
supporting flow path pressure, vane aerodynamics, and actuation loads, the
nozzle box is designed to contribute to the airframe structural stiffness. The
nozzle vane box is airframe-mounted, with the vane box sidewalls serving as
aircraft structure keel members.
The VAVBN (Fig. 23) provides directional control of the LiftFan thrust
vector and an additional effector to the VIGVs for turndown. Turndown refers
to the commanded position of the lift fan inlet guide vanes used to control
lift fan thrust. Three VAVBN
vanes are driven by dual-tandem,
linear, hydraulic actuators. Drive
is transferred to the other three
vanes through bar linkages.
With this system, nozzle thrust
may be directed in an arc of
41.75– 104 deg (fore/aft aircraft
coordinate system), at a rate of
40 deg/s. Independent control of
the three VAVBN actuators pro-
vides the capability to vary the
nozzle throat area independent
of the vector angle. VIGV and
VAVBN area variation are both
used to control LiftFan perform-
ance, manage the LiftFan stall
margin, and minimize thrust-
thrust split coupling effects.

Fig. 23 F-35 VAVBN as seen

from below.
156 C. WIEGAND ET AL.

Thrust split is defined as the ratio of main engine thrust over lift fan thrust, typically
used to represent the propulsion system pitching moment applied to the aircraft.

5. ROLL POST NOZZLE DEVELOPMENT

The F-35B STOVL lift system uses roll nozzles in each wing to provide roll control
in powered-lift operation. The roll nozzle controls thrust by varying the nozzle
area using two hinged flaps. Unlike the reaction control systems on legacy Vertical
and/or Short Takeoff and Landing (V/STOL) aircraft, the F-35 roll posts produce
about 10% of the vertical thrust through redirected engine fan air. The port and
starboard side roll post systems are part-number common and interchangeable,
providing improved maintenance flexibility. Actuation for the nozzle flaps is pro-
vided by a twin-motor, hydraulic, rotary actuator. External, hydraulically actuated
aircraft doors on the underside of the wing are opened in powered-lift operation
to provide an exit aperture for roll post thrust.
A key challenge in transitioning to the F-35B was providing adequate roll
control authority and rate for store asymmetries and fuel imbalance. The roll
post nozzles were positioned as far outboard as the internal wing structure
would allow to maximize the moment arm. Main engine fan air to the roll
posts was increased to the extent possible while maintaining adequate flow to
cool the exhaust liner. Architectural changes in full-authority digital engine
control were made to minimize the time delay from roll moment command to
roll nozzle actuator response.

6. 3BSM DEVELOPMENT
The original design for the primary nozzle on the ASTOVL was a two-
dimensional single expansion ramp nozzle. In this design, one nozzle flap is
longer than the other. The nozzle vectors the primary thrust by deflecting the
upper flap through at least 90 deg. To control the nozzle exit area in hover, the
lower flap was designed as a sliding panel that would retract as needed to
adjust the backpressure on the engine. This was a critical control needed to
make the shaft-driven LiftFan turbine work.
As Lockheed Martin began construction and tests of the nozzle, the shortcom-
ings of the design became more apparent. The abilities to turn the flow through
90 deg under high loads and control the nozzle exit area would have resulted in
a very heavy design. This resulted in the pursuit of a lighter design that traced
its roots to an early 1970s nozzle design from the proposed Convair Model
200 V/STOL fighter aircraft concept. A three-bearing swivel nozzle was devel-
oped by Pratt & Whitney and became part of the Convair Model 200 design
that never continued into development. Following joint studies by Pratt &
Whitney and Lockheed Martin, the 3BSM concept was integrated into the
X-35B design and shown to be lighter. It also provided a very efficient means
for turning the aft thrust post with minimal losses [36].
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 157

The F-35B 3BSM consists of a STOVL LOAN and a three-bearing swivel

mechanism. The mechanism can deflect the exhaust flow through 95 deg in the
pitch axis and +12.25 deg in the yaw axis as a function of pitch angle. The
3BSM can vector up to 23,900 lb of thrust at the maximum rearward thrust
split. The 3BSM forward (No. 1) bearing is powered by twin fueldraulic actuator
motors through a gearbox and drive train. The middle (No. 2) bearing is likewise
powered by a twin fueldraulic actuator motor and gearbox/drive train system. A
transfer gearbox links the middle and aft (No. 3) bearings with an efficient,
compact, epicycle gear train. The twin actuator motors on the No. 1 and Nos. 2
and 3 bearings, respectively, are designed with a fail-degraded capability (full
torque, half rate). This is one of the key differences between this design and
that of the X-35B. In the X-35B, the Nos. 2 and 3 bearings were braked following
a first failure, with no ability to continue vectoring the aft thrust post. This did not
satisfy operational requirements requiring an ability to perform a shipboard ver-
tical landing following a first failure. The dual redundancy on the fueldraulic
motors enabled that fault tolerance on the F-35B.

7. TECHNOLOGY DEMONSTRATION PROGRAM AND TRANSITION TO THE JSF PROGRAM

The X-35B STOVL lift system completed more than 1200 hours of ground testing,
culminating in the successful concept flight demonstration in August 2001. The
aircraft accomplished this impressive performance under demanding hot, high-
desert conditions and substantiated the robust performance capabilities of the
shaft-driven LiftFan concept. Particularly impressive were the precise aircraft
dynamics enabled by the responsive and accurate control of the STOVL lift
system. Thirty-nine flights were conducted on the X-35B, including 22 hovers,
17 vertical takeoffs, 18 short takeoffs, 57 STOVL mode transitions, 27 vertical land-
ings, 116 conversions (95 ground, 21 inflight), 63 clutch engagements, and 21.5
flight hours. This performance far surpassed the vertical operation goals and
demonstrated sufficient maturity to proceed to production aircraft development.
The transition to a production F-35B principally centered on evaluating the
STOVL lift system design changes and demonstrating a full-life propulsion
system. More insights into the transition and full system development for the
production F-35 system can be found in Ref. [38].

IV. CONCLUSION
The F-35 combines numerous technologies that have significantly advanced the
state of the art in combat aircraft. This is particularly pronounced in the areas
of integrated air vehicle subsystems and propulsion systems. The resultant aircraft
provides exceptional performance with unparalleled capabilities, enabled by the
air vehicle and propulsion systems.
The integrated air vehicle subsystems architecture selected for incorporation
was based on a continuum of progressively refined development projects. Each
158 C. WIEGAND ET AL.

of these further refined the concepts and validated the approach. The SUIT and
MEA studies from the early 1990s gave the JSF contractor teams confidence in
the concepts. The J/IST studies then provided the final proof of the viability of
the designs. They also validated the conclusion that the overall air vehicle
takeoff gross weight and cost could be reduced by 2 –3%. The T/EMM system
development project in J/IST contributed to the development of the turboma-
chine, fan duct heat exchangers, and other key elements used in the current
F-35 PTMS. Without these elements, the chosen configuration might have been
deemed too risky to pursue in the SDD program. Likewise, without the great suc-
cesses of these development programs, many elements of the F-35’s integrated
systems, EHAS, and EPS would likely have been substituted with more conven-
tional federated configurations. In such an instance, the benefits of the integrated
systems might never have been realized. Instead, the resultant systems incorpor-
ated into the F-35 have been proven to provide excellent technical performance
and reliability. They also provide a backbone for future systems growth through
the expected long life of the F-35 program.
The F-35 propulsion systems incorporating the numerous technology
upgrades have driven an unprecedented capability in performance. This has
enabled the aircraft’s unique performance capabilities, particularly in the F-35B
STOVL variant. The final F-35 configuration incorporated a DSI, LO axisym-
metric engine thrust nozzle, and unique STOVL propulsion system integrating
the LiftFan and three-bearing swivel nozzle. These systems enabled the develop-
ment of the F-35 variants, each providing exceptional performance and serving as
the basis for long-term growth and capability improvements.

ACKNOWLEDGMENTS
The authors thank Jeff Hamstra, Rick Mange, and J.D. McFarlan for their leader-
ship in pursuing the entire family of related papers. The authors give special
thanks also to graphic artists Daniel Buck and Mark Lawhon, and to Joey Sikorski
for help in pulling together a lot of details that went into this report.

REFERENCES

[1] Barhydt, J., Carter, H., and Weber, D., “Subsystem Integration Technology (SUIT)
Phase I,” WL-TR-93-3031, July 1993.
[2] Blanding, D., Schlundt, D., and Aldana, J., “Subsystem Integration Technology
(SUIT),” WL-TR-92-3105, December 1992.
[3] Hill, B., Stem, P., Ho, Y., and Martin, D., “Subsystem Integration Technology
(SUIT),” Executive Summary, Vol. 1, WL-TR-92-3115, October 1992.
[4] Hill, B., Stem, P., Ho, Y., and Martin, D., “Subsystem Integration Technology
(SUIT),” Phase I Report, Vol. 2, WL-TR-92-3116, October 1992.
[5] Hill, B., Stem, P., Ho, Y., and Martin, D., “Subsystem Integration Technology
(SUIT),” Phase II Report, Vol. 3, WL-TR-92-3117, October 1992.
F-35 AIR VEHICLE TECHNOLOGY OVERVIEW 159

[6] Hill, B., Stem, P., Ho, Y., and Martin, D., “Subsystem Integration Technology
(SUIT),” Phase III Report, Vol. 4, WL-TR-92-3118, October 1992.
[7] Carter, H., Rupe, K., Mattes, R., and Wiese, D., “Subsystem Integration Technology
(SUIT),” Phase II Report, WL-TR-96-3065, May 1997.
[8] Terrier, D., Hodge, E., Eicke, D., and Roberts, J., “Vehicle Integration Technology
Planning Study (VTIPS),” WL-TR-96-2077, April 1996.
[9] Weber, D., Hill, B., and Van Hom, S., “Vehicle Integration Technology Planning
Study (VTIPS),” WL-TR-97-3036, February 1997.
[10] Eicke, D., “Joint Strike Fighter Integrated Subsystems Technology (J/IST) Power and
Actuation Flight Demonstration Program,” AFRL-PR-WP-2001-2036, January 2001.
[11] Carter, H., “Joint Strike Fighter (JSF)/Integrated Subsystems Technology (J/IST)
Demonstration Program,” Executive Summary, Vol. 1, AFRL-VA-WP-TR-2001-
3028, June 2001.
[12] Rupe, K., and Carter, H., “Joint Strike Fighter (JSF)/Integrated Subsystems
Technology (J/IST) Demonstration Program,” Vol. 2, Program Integration,
AFRL-VA-WP-TR-2001-3029, June 2001.
[13] Heydrich, H., and Roddiger, H., “Joint Strike Fighter (JSF)/Integrated Subsystems
Technology (J/IST) Demonstration Program,” Vol. 3, Thermal/Energy Management
Module (T/EMM) Turbomachine, AFRL-VA-WP-TR-2001-3030, June 2001.
[14] Dastur, N., and Roddiger, H., “Joint Strike Fighter (JSF)/Integrated Subsystems
Technology (J/IST) Demonstration Program,” Vol. 4, Thermal/Energy Management
Module (T/EMM) System Integration, AFRL-VA-WP-TR-2001-3031, June 2001.
[15] Anderson, G., and Waffner, W., “Joint Strike Fighter (JSF)/Integrated Subsystems
Technology (J/IST) Demonstration Program,” Vol. 5, High Power Electric
Actuation, AFRL-VA-WP-TR-2001-3032, June 2001.
[16] Torchia, D., and Rupe, K., “Joint Strike Fighter (JSF)/Integrated Subsystems
Technology (J/IST) Demonstration Program,” Vol. 6, Power Management
Integration, AFRL-VA-WP-TR-2001-3033, June 2001.
[17] Anderson, J., and Roddiger, H., “Joint Strike Fighter (JSF)/Integrated Subsystems
Technology (J/IST) Demonstration Program,” Vol. 7, Engine Integration, AFRL-
VA-WP-TR-2001-3034, June 2001.
[18] Eicke, D., “Joint Strike Fighter Integrated Subsystems Technology (J/IST) Power and
Actuation Flight Demonstration Program,” Executive Summary, AFRL-PR-WP-
2001-2077, January 2001.
[19] Iya, S., Harlan, R., and Hawkins, L., “LCC and War Fighting Assessment of J/IST
Suite,” AFRL-VA-WP-TR-2000-3016, March 1999.
[20] Hamstra, J. W., McCallum, B. N., McFarlan, J. D., and Morehouse, J. A.,
“Development, Verification and Transition of an Advanced Engine Inlet Concept
for Combat Aircraft Application,” NATO RTO Paper MP-121-P-43, October 2003.
[21] Hehs, E., “Diverterless Supersonic Inlet,” Code One, July 2000, pp. 8 –13.
[22] Scharnhorst, R. S., “An Overview of Military Aircraft Supersonic Inlet
Aerodynamics,” AIAA Paper 2012-0013, January 2012.
[23] Sobester, A., “Tradeoffs in Jet Inlet Design: A Historical Perspective,” Journal of
Aircraft, Vol. 44, No. 3, May 2007, pp. 705 –717.
[24] Gridley, M. C., and Walker, S. H., “Inlet and Nozzle Technology for 21st Century
Aircraft,” ASME Paper 96-GT-244, June 1996.
[25] Gridley, M. C., and Cahil, M. J., “ACIS Air Induction System Trade Study,” AIAA
Paper 1996-2646, July 1996.
160 C. WIEGAND ET AL.

[26] Catt, J. A., Welterlen, T., and Reno, J., “Decreasing F-16 Nozzle Drag Using
Computational Fluid Dynamics,” 29th Joint Propulsion Conference and Exhibit,
1993. doi: 10.2514/6.1993-2572.
[27] Catt, J. A., Welterlen, T., and Reno, J., “Evaluating F-16 Nozzle Drag Using
Computational Fluid Dynamics,” 32nd Aerospace Sciences Meeting and Exhibit,
1994. doi: 10.2514/6.1994-22.
[28] McWaters, M., “F-35 Conventional Mode Jet-Effects Testing Methodology (AIAA
Paper 2015-2404),” 31st AIAA Aerodynamic Measurement Technology and Ground
Testing Conference, 2015. doi: 10.2514/6.2015-2404.
[29] Miller, D., and Catt, J. A., “Conceptual Development of Fixed-Geometry Nozzles
Using Fluidic Injection for Throat Area Control,” 31st Joint Propulsion Conference
and Exhibit, 1995. doi: 10.2514/6.1995-2603.
[30] McFarlan, J. D., McMurry, C. B., and Scaggs, W., “Computational Investigation of
Two-Dimensional Ejector Nozzle Flow Fields,” 26th Joint Propulsion Conference,
1990. doi: 10.2514/6.1990-2148.
[31] Bevilaqua, P. M., “Inventing the F-35 Joint Strike Fighter,” AIAA Paper 2009-1650,
Jan. 2009.
[32] Maddock, I., and Hirschberg, M., “From ASTOVL to JSF: Development of the
STOVL Propulsion Systems,” 11th AIAA ATIO Conference, AIAA Centennial of
Naval Aviation Forum, Aug. 2011.
[33] Simons, T., and Sokhey, J., “Performance Optimization of VAVBN the Variable Area
Vane-Box Nozzle for JSF F-35B LiftFan Propulsion System,” International Powered
Lift Conference, Philadelphia, PA, Oct. 2010.
[34] Renshaw, K., “F-35B Lightning II Three-Bearing Swivel Nozzle,” Code One, Aug. 2014.
https://codeonemagazine.com/article.html?item_id=137
[35] Robbins, A., Bobalik, J., De Stena, D., Plag, K., Rail, K., and Wall, K., “F-35
Subsystems Design, Development, and Verification,” AIAA Aviation Forum,
Atlanta, GA, June 2018 (unpublished).
[36] Wurth, S., Smith, M., and Celberti, L., “F-35 Propulsion System Integration,
Development & Verification,” AIAA Aviation Forum, Atlanta, GA, June 2018
(unpublished).
[37] Lemons, G., Carrington, K., Frey, T., and Ledyard, J., “F-35 Mission Systems Design,
Development & Verification,” AIAA Aviation Forum, Atlanta, GA, June 2018
(unpublished).
[38] Counts, M., Kiger, B., Hoffschwelle, J., Houtman, A., and Henderson, G., “F-35 Air
Vehicle Configuration Development,” AIAA Aviation Forum, Atlanta, GA, June
2018 (unpublished).
[39] Dietrich, R., “Brief History and Perspective on J/IST,” JSF Joint Program Office
(Ref: DM #444383) (approved for public release, distribution unlimited)
(JSF14-304), Distribution A.
[40] “Lockheed Martin Receives Prestigious Aviation Award for Joint Strike Fighter
Technology,” Lockheed Martin Aeronautics Company [online database], https://
news.lockheedmartin.com/2001-06-20-Lockheed-Martin-Receives-Prestigious-
Aviation-Award-for-Joint-Strike-Fighter-Technology [retrieved 11 May 2018].
 CHAPTER 4

F-35 Digital Thread and Advanced

Manufacturing
Don A. Kinard
Lockheed Martin Aeronautics Company, Fort Worth, TX

The F-35 program has delivered more than 270 aircraft to the U.S. Air
Force, Marine Corps, Navy, and partner countries. This was accom-
plished during the Low-Rate Initial Production (LRIP) phase of the
program. LRIP facilitated the completion of structural and flight
testing, and enabled pilot and maintainer training. Arriving at LRIP
for this quantity of F-35 jets surpasses the total production quantities
achieved at the same stage for the F-117 and F-22. Also, initial operating
capability has already been reached for the F-35 by the Air Force, Marine
Corps, and Israeli Air Force. The F-35 program is now advancing toward
full-rate production and reinforcing key strategy tenets. Included among
them are lean manufacturing deployment, the use of low-risk materials
and supportable low observables, and the implementation of digital
thread technologies.

I. INTRODUCTION
In 2001, when the Joint Strike Fighter System Development and Demonstration
(SDD) program commenced, the challenges ahead for manufacturing were daunt-
ing. Not since the days of the McDonnell Douglas F-4 Phantom II in the 1960s
and 1970s had a program attempted to produce three variants of an advanced
fighter aircraft to satisfy requirements for the Air Force, Navy, and Marine
Corps on a single assembly line (Fig. 1). Not since the early F-16 days had
advanced fighter aircraft been scheduled to be produced at the rates envisioned
by the F-35 Joint Program Office (JPO). Affordability was a fundamental chal-
lenge, as was the need to support international participant aircraft configurations,
manage a global supply chain, and initiate Final Assembly and Checkout (FACO)
facilities in three countries. As the newest 5th-Generation fighter, the F-35 intro-
duced and incorporated many evolutionary and revolutionary technologies. These
included sophisticated avionics, supportable low observables, an unprecedented

 SeniorFellow, F-35 Production Operations, AIAA Sr. Member, [email protected]—Approved for public
release 5/17/18, JSF18-402.

161
162 D. A. KINARD

Fig. 1 F-35 multivariant final assembly line in Fort Worth, Texas.

amount of aircraft and support software, and complex vehicle systems develop-
ments, such as the LiftFanw and Electro-Hydrostatic Actuation Systems (EHASs).
To date, Lockheed Martin, our partners, and our suppliers have delivered
more than 270 aircraft, and there are currently approximately 175 additional
new aircraft in work around the world. These aircraft were built during the
Low-Rate Initial Production (LRIP) phase of the program to allow the completion
of testing and pilot and maintainer training. The F-35 LRIP deliveries are already
greater in number than the total production aircraft for either the F-117 or F-22
programs. Further, the Marine Corps, Air Force, and Israeli Air Force have all
declared initial operating capability with the F-35.
The Fort Worth facility’s layout, nearly one mile in length, is shown in Fig. 2.
Wing components and the fuselage are produced in Fort Worth. These Lockheed
Martin components are then mated to the center fuselages from Northrop
Grumman and the aft fuselages from BAE Systems along with the center wing
assemblies built at Lockheed Martin in Marietta, Georgia, before going to final
assembly. This is followed by coatings, radar cross-section testing, and fuel
testing before finally sending them to the flight line for delivery operations. The
complex F-35 supply chain is illustrated in Fig. 3. There are more than 1400
F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 163

Fig. 2 F-35 facility layout in Fort Worth.

suppliers, more than 80 of which are in non-U.S. countries, as well as three FACO
facilities. Orchestrating this supply chain requires 24/7 operations and other
efforts to satisfy the requirement to support and supply parts to the more than
270 aircraft already delivered to the field.
The F-35 aircraft is completing development testing, and changes have
been incorporated into the design. The aircraft is meeting its key performance
points, and trained pilots and maintainers complement the aircraft’s performance
and capabilities well. The F-35 is now starting to transition operations from LRIP
(Fig. 4) to full-rate production. The engineering, manufacturing plan, and supply

Fig. 3 F-35 global supply chain.

164 D. A. KINARD

Fig. 4 F-35 commencing the climb to full-rate production with approximately 1400 suppliers.

network have been tested in low-rate production in preparation for the transition
to full-rate production, but there are many remaining production challenges.
Thousands of employees around the world need to be hired, and the last of the
required tooling and capital needs to be installed. In addition, efforts must con-
tinue to improve quality and reduce costs, and to climb the rate production curve.
Several key strategies have been implemented as part of the overall F-35 fighter
production system. Lean manufacturing principles have been adopted, including
flow-to-takt production lines, point-of-use materials delivery, and emphasizing
standard work. A low-risk approach has been taken to select materials, structures,
and supportable low observables technologies. The development and implemen-
tation of engineering and manufacturing technologies has been enabled by the
digital thread.

II. F-35 LEAN MANUFACTURING

The Lockheed Martin F-35 assembly lines in Fort Worth; Palmdale, California;
and Marietta, Georgia, implemented a flow-to-takt lean manufacturing philos-
ophy [1 –3]. Lockheed Martin wanted to create rhythm and reinforce the sense
of urgency in the factory. To do this, it moved the individual components at
the same pace as the average delivery rate. Takt is the pace of the factory expressed
in manufacturing days (i.e., calendar days not counting weekends or holidays).
For example, five aircraft per month equates to one aircraft every four manufac-
turing days, which is a takt time of four.
On the F-35 program, this sometimes requires large material handling systems
to efficiently transfer components from one station to another at this takt pace. An
example is shown in Fig. 5 for the Fort Worth wing box assembly line. This wing
F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 165

box line is effectively duplicated in Italy at the Leonardo plant in Cameri, and in
Israel at the Israel Aerospace Industries plant. Lean assembly lines are also the
standard at Northrop Grumman and BAE Systems sites, at sites in participating
nations, and at supplier locations. Lockheed Martin is responsible for ensuring
that all components produced for the F-35 fighter production system conform
to U.S. government engineering requirements. Other examples of F-35 lean
manufacturing are:
. Point-of-use materials delivery
. Tasks broken into four-to-eight-hour segments
. A factory-wide radio frequency identification system for tracking part kits
and tooling
. Automated dispensing units for providing perishable and hand tooling
Lockheed Martin anticipated that single-piece flow manufacturing was initially
going to be difficult, and this turned out to be the case. Flow requires a stable man-
ufacturing system, and the initial F-35 flow was disrupted often by multiple
factors. Among these were:
. Normal developmental corrections to the Build-to Packages (BTPs)
. Implementation of changes to correct deficiencies found during structural and
flight testing

Fig. 5 Flow-to-takt assembly line with highly common tooling and material handling.
166 D. A. KINARD

. Natural progression of the learning curve reducing the number of hours per
unit by 75% since production start
. Startup and expansion of the global supply chain
. Variable production rates
Instability in the manufacturing system and the disruptions often result in
out-of-station work being sent downstream in the flow-to-takt areas, increasing
cost and span [4]. Lockheed Martin expects to realize the benefits of flow after
incorporating SDD-related changes, stabilizing the production rate, and fully
maturing the supply chain.
The trivariant production line was less disruptive than expected, and the tri-
variant assembly line exhibited a shared learning curve as a result of high parts
commonality. There was nearly 100% commonality among the mission
systems, 70% among the vehicle systems, and 20% among the airframes. In this
way, each variant contributed to an overall reduction in the learning curve. The
F-35C Carrier Variant (CV) and F-35B Standard Takeoff and Vertical Landing
(STOVL) variants in particular experienced a reduction in hours per unit from
this commonality. They benefitted from the F-35A Conventional Takeoff and
Vertical Landing (CTOL) variant’s higher production rate. This was likely due
to the common materials, processes, and assembly, which led to similar work
being done in each manufacturing position regardless of variant type.
Lockheed Martin determined another arrangement that increased production
efficiency: spreading the F-35B/C models among the F-35A models. The F-35B
and F-35C models each constitutes 15% of the total build and requires more
work to produce, and building them in blocks would disrupt staffing and learning.
Also contributing to production efficiency, Lockheed Martin was able to use
many common tools, such as the large assembly tool holding the wing in Fig. 5.
The tool’s gold and blue stripes indicate that it can be used for both F-35As
and F-35Bs. Building all three variants on a single assembly line with high com-
monality saved an estimated 30% in overall production costs, including those of
capital and tooling. This was an improvement over the traditional approach,
with which each aircraft is built in a different factory, by a different company.

III. F-35 MATERIALS AND STRUCTURES: SUPPORTABLE

LOW OBSERVABLES
As far as material and process selections are concerned, the F-35 program adopted
a low-risk approach using well-characterized materials and processes. In some
respects, this decision was a foregone conclusion because of the lack of material
(metallic and composite) development in the aerospace industry today. The
industry is looking at novel processing technologies, such as additive manufactur-
ing, but the basic composites and metals available in the marketplace have been
F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 167

there for more than 20 years. The cost of qualification and inherent risk aversion
on the part of customers and companies have reduced the emphasis on develop-
ment. Exceptions to the low-risk philosophy for the F-35 were the aluminum for-
gings in Fig. 6 and the formed titanium engine doors from BAE Systems on the
F-35B. Another exception was the incorporation of the latest in low observable
materials and structures technologies. In general, the F-35 program used alumi-
num wherever possible and titanium wherever necessary, such as in hot areas
and areas of concentrated load introduction. Composites were used wherever
weight savings were sufficient to justify the cost.
Another fundamental aspect of the F-35 involves the technologies controlling
the Outer Mold Line (OML) of the aircraft or simply the panel-to-panel and
panel-to-skin mismatches. Mismatches on the surface of a 5th-Generation
fighter (F-22 and F-35) are detrimental to the aerodynamics and low observable
performance. Therefore, the F-35 applied advanced manufacturing technologies
to control part dimensional tolerance tightly and thus control the mismatches.
As an example, controlling surface mismatches requires tightly controlling the
skin thickness on both sides of the panel/skin joints. Composite parts typically
have a cured ply thickness variation insufficient to control F-35 mismatch toler-
ances. Therefore, we must correct for this thickness variation in the cured parts.
One approach was to add sacrificial plies to the Inner Mold Line (IML) of the
composite skins and machine the IML using very tight tolerance Numerically
Controlled (NC) machines. These NC machines feature laser-compensated

Fig. 6 Aluminum forgings from Alcoa for unitization and weight savings.
168 D. A. KINARD

Fig. 7 Laser scanning the

thickness measurement of
composite parts.

calibration and temperature

control to produce the final
skin thickness to the desired
tolerance. Another approach,
patented by Lockheed
Martin, is to use a laser
radar system (Fig. 7) to
measure the composite
thickness after cure. This is
done where the skins are
designed to be at or below
the final required tolerance.
These thickness measure-
ments are automatically
sent to a ply-cutting
machine to produce com-
pensation plies. Compen-
sation plies are located on the skin IML using laser projection systems and then
the part is recured to provide the final thickness-controlled part. Both of these
approaches may seem expensive, but the goal was to produce an aircraft that
was supportable and affordable long term, for Lockheed Martin customers.
Trade studies indicated that it was beneficial to the overall F-35 air system to
reduce maintenance requirements over the life of the aircraft, specifically the
maintenance hours per flight hour. In addition, because OML is controlled so
tightly, Lockheed Martin was able to develop simplified coating applications.
These not only reduced the cost of initial application, but also significantly
reduced the repair costs when a skin incurred cosmetic or maintenance damage
in the field.

IV. THE DIGITAL THREAD

The F-35’s development and early production benefitted significantly from the
phased adoption of a digital thread philosophy. Designers produced 3D solid
models, constructed them to support factory automation, and facilitated their
consumption by downstream manufacturing and sustainment functions.
Recently, technology has allowed the rapid validation of as-designed to as-built
configuration verification through the use of laser scanning and structured light
technologies.
F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 169

The term digital thread was reportedly coined by the Air Force Research Lab-
oratory (AFRL) and Lockheed Martin during the early days of F-35 development
[5]. The overall F-35 digital thread philosophy is depicted in Fig. 8. Lockheed
Martin defines it as the creation, use, and reuse of the 3D models by engineering
and downstream functions, including manufacturing and sustainment. In Phase 1
of the digital thread implementation, engineering produced exact 3D engineering
models and 2D drawings. Partner and supplier models, 3D tool designs, drawings,
specifications, and related analysis data were released into a common product
life-cycle management system for accessibility and configuration integration.
Manufacturing produced 3D models for tools and factory layouts that improved
facilities development and installation. For many of the airframe parts, engineer-
ing was able to produce reduced-dimension drawings that decreased engineering
costs and facilitated supplier NC machining. Fiber placement was used for com-
posites based on the digital thread. Coordinate-measurement machines’ inspec-
tion points were programmed directly into the solid models because the solids
contained the master engineering data. These models also supported the suppor-
table low observable structures processes discussed earlier involving the machin-
ing of the IML/OML and cured laminate compensation.
Three-dimensional models were used for virtual mockups, manufacturing,
and sustainment simulations. Significant successes of the 3D solids led to large
reductions in the quantities of engineering and tooling changes. Solid models
reduced engineering changes, compared to historical numbers, because of the
ability to provide accurate surfaces and improved integration between parts.
Because of solid model engineering and solid model tooling, Lockheed Martin
was able to reduce tool design changes due to tool interferences with released

Fig. 8 The F-35 digital thread.

170 D. A. KINARD

Fig. 9 Producibility analysis and key characteristics.

parts. Interferences between parts were also reduced, compared to nonsolid

model programs, and suppliers were able to produce and validate machined
parts to the released solid model masters. These improvements were especially
important to the F-35 because there were three variants to design. Assembly inter-
face control drawings, typical of legacy programs, were not required because the
solid models facilitated coordination among partners and suppliers.
Virtual manufacturing simulations, encouraged by the digital thread technol-
ogy, turned out to be time consuming and expensive, so only a few areas were even
attempted. This is one area requiring tool development for future programs. Phys-
ical mockups were used for specific bays depending on complexity. Sustainment
engineering was able to use virtual reality helmets to assist in simulating mainten-
ance actions with the completed 3D engineering models.
The digital thread allowed significant increases in producibility and variation
management analysis during the F-35 SDD program (Fig. 9). Geometric dimen-
sioning and tolerancing was introduced on the F-35, and variation management
analysis was performed by manufacturing engineering. This included using
specialized 3D software to perform complex assembly variation studies. Critical
installations were identified, followed by the collection of process capabilities facil-
itating the variation studies and the creation of variation management documents
containing assembly datum schemes. Variation analysis led ultimately to the defi-
nition of engineering tolerances that were flowed into the models, drawings, and
tooling, and to the identification of Key Characteristics (KCs).
A KC [6] is a feature of a material, process, or part (including assemblies)
whose variation within the specified tolerance significantly influences product
F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 171

fit, performance, service life, or manufacturability. The F-35 identified many KCs
early on, but would likely revise selections for future programs. KC selections
should be made with the understanding that their selection will drive costs into
the production system. They will have this effect by requiring the development
and imposition of KC management plans, the formal collection of data by fabri-
cators and assemblers, and reporting and analysis of deliverables. KCs should only
be selected under certain conditions. Plans must first be in place to alter the engin-
eering or manufacturing BTP designs. Similarly, requirements must first be ident-
ified and implemented to better control or take advantage of the variation
reduction. KCs that simply accentuate but do not drive changes to engineering
accept/reject criteria or BTPs are not KCs. For example, KCs were put on the
diameters of holes on previous programs, but there was never a plan to change
the tolerances. As such, these should not have been regarded as KCs.
Phase 2 of the digital thread transformation is about constructing the engin-
eering data to support factory automation. Examples are automated drilling
(Fig. 10) and robotic coatings applications (Fig. 11). Automated drilling is used
by all the F-35 partners, and we drill 20% of the total holes using automation.
This includes 80% of the accessible OML holes. Automated drilling is about
four times faster than manual drilling and its quality is nearly perfect, with
remarkable repeatability.
Lockheed Martin uses automated drilling for the wing boxes, forward
fuselage skins, and upper skin to center wing skins. We also use it for the
center wing at Marietta, Georgia. Northrop Grumman uses a metrology-assisted
robot to drill the narrow inlets on the F-35. BAE Systems drills its empennage
skins and structure separately using its high-precision machining centers. This
is a remarkable feat of high-
precision machining, consid-
ering the bolt-to-hole toler-
ances it requires. There is a
plan in place to implement
even more automated drilling
in the future as part of the con-
tinual effort to drive costs
down and improve quality.
Other automation used
for the F-35 includes fiber
placement technology, which
Lockheed Martin, Northrop
Grumman, and other suppli-
ers use to lay up complex

Fig. 10 Automated drilling is

enabled by the digital thread.
172 D. A. KINARD

Fig. 11 Robotic coatings applications.

inlets, nacelles, and large wing parts. Recently BAE Systems also introduced a
robot to countersink its composite skins (Fig. 12).
Phase 3 provides the
digital thread directly to the
mechanics to create such pro-
ducts as work instruction
graphics. These graphics
were facilitated by the 3D
solid models, which can be
used to create graphics
through visualization soft-
ware tools. Ideally, they
would visually instruct mech-
anics on the floor or maintai-
ners in the field and reduce
the time it takes them to
understand their tasks;
however, for production this
intended benefit on the floor
ended up being very difficult
to maintain. This was
because graphics are static
images that are incapable of

Fig. 12 BAE Systems

robotic countersinking.
F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 173

affordably being updated as engineering or manufacturing changes. For the F-35

program specifically, with concurrent engineering and manufacturing develop-
ment, graphics must frequently be updated to accommodate a steep initial learn-
ing curve and significant engineering, tooling, and planning changes.
An unexpected factor in this process was the impact of flow-to to-takt man-
ufacturing on graphics creation. As the takt time (production rate) changed,
new tool positions were added and hours per unit decreased. As a result, the man-
ufacturing sequencing needed to be adjusted constantly by breaking up planning
cards and redoing graphics. In an effort to circumvent static graphics costs,
graphics can be made available on the factory floor. This is done by granting
the mechanics access to the visualization tools from their work terminals.
Harness installations are a good example of how this can work because harness
routings are especially difficult to understand from a 2D drawing. Early in
SDD, large television monitors were placed in some of the work areas, and the
mechanics used these until experience was gained with the installations.
Another downside of graphics is that the mechanics typically only need them
for a short time. Despite this, however, graphics do help mechanics who are new
to the F-35 program during ramp-up to full-rate production. Recently, movies of
the critical installations were produced that could be accessed as needed through
the electronic work instruction terminals that mechanics use on the floor.
One of the unique ways for mechanics to use the digital thread is through
optical projection technology. As Fig. 13 demonstrates, mechanics can use this
to visually project work instructions directly onto the aircraft. In the figure, fas-
tener locations and part numbers are being projected onto the inlets being built

Fig. 13 Optical projection of work instructions directly onto the work surface.
174 D. A. KINARD

at Lockheed Martin’s Marietta facility. The conventional procedure requires

mechanics to look at drawings and write down part numbers and inlet locations
to do their work. With the digital thread’s procedure, mechanics can instead view
the projected instructions while performing their work.
Continuing work is being done to capture the actual fastener grip during auto-
mated drilling operations. This information will be used to eliminate grip vali-
dation time and support fastener projection. In addition, it will be used to kit,
clean, and promote fasteners and deliver these to the point of use.
An additional example of Phase 3 is the technology for laser ply projection in
the composites shop. This was one of the first Phase 3 digital thread technologies
used in the aerospace industry. Bulkhead marking was initiated on the F-35 where
ink jet markings for bracket locations were printed directly onto large bulkheads.
This saved span and costs and eliminated thousands of tools that had to be
designed and maintained.
The production of tools for the mechanics and sustainment modifications
using additive manufacturing (Fig. 14) is another digital thread success. Lockheed
Martin has produced more than 5000 tools for the floor and field using Fused
Deposition Modeling (FDM) of polymers. FDM provides a fast, low-cost
approach to producing temporary tools (temporary because of the FDM material’s
durability issues) to assist the mechanics. The development of more durable FDM
materials continues in the industry. The objective is to enable producing perma-
nent tooling and replacing more expensive metal, fiberglass, and other typical
hole-drilling tooling.
Phase 4 of the digital thread is the validation of engineering as-designed to the
as-built structure using advanced noncontact metrology techniques, including
laser scanning and structured light (Fig. 15). This technology can identify

Fig. 14 Additive manufacturing of tools.

F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 175

Fig. 15 Noncontact metrology.

deviations from engineering early in the build or fabrication process and rapidly
correct them, reducing cost by stopping defects from traveling downstream. A
truly revolutionary technology, it may eventually replace coordinate measuring
machine inspections and become a requirement for suppliers prior to shipping
parts, tools, and equipment.
Additively manufactured tools configured to mimic various F-35 weapons
normally take about a full shift to be installed and pass clearance checks. Now,
a laser can scan the bays and compare the as-built aircraft to the engineering
models, which takes only a few hours. As another example, when there is a fit
problem with a tube on the floor, the cause is not readily apparent. The tube
could be bad, the bracket location could be incorrect, or there might be a
problem with the structure. To overcome the problem, the tube can be brought
to a scanner and rapidly validated, or equipment can be brought to the aircraft
to have the brackets and structure checked.
Current scanning technologies typically depend on targets being placed on
the aircraft or parts. However, this will eventually be replaced by feature-based
recognition as the digital thread connections to the 3D scanner technology
matures. This scanning technology can also replace the thousands of manual
mismatch, gap, and flushness measurements required today (Fig. 16). Further,
it can inspect detail parts, tooling, and assemblies on a first-article basis today.
Identifying defects during first-article inspections will significantly reduce the
cost and flatten the learning curve for concurrently developed products like
military aircraft. Early identification will also reduce the recurring cost for
measurements. In the future, the use of 3D noncontact metrology for recurring
real-time validations in an assembly, as well as for supplier acceptance, may
become routine.
176 D. A. KINARD

Fig. 16 Applications of noncontact metrology.

Applying the digital thread on the F-35 program has brought significant
benefits, including:

. Development of the BTPs

. Use of automation to reduce touch labor and improve quality
. Integration of the digital thread on the factory floor
. Opportunity to validate the configuration using digital thread technologies
(e.g., laser scanning and structured light)

There has been an explosion of digital technologies in the past five years and a
tremendous amount of continued development in the industry. A future
additional phase of the digital thread is discussed in the following section. It
refers to the ascension of Industry 4.0 [7], the fourth industrial revolution: the
revolution of data.

V. THE FUTURE OF ADVANCED MANUFACTURING

Affordability is and will continue to be a focus for the F-35 program in the years to
come. Lockheed Martin, our suppliers, and the customer community have
invested significantly in affordability through the Blueprint for Affordability
Program (BFA). As BFA Phase 1 concludes, it is showing an impressive return
on this collective investment, as measured by dividing dollars saved ($2 million
per aircraft) by dollars invested. BFA Phase 2 began in 2018 and will provide
funding for additional projects (Fig. 17).
F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 177

The F-35 program aims to achieve an $80-million aircraft cost at full-rate

production. As with most programs like F-35, the supply chain is responsible
for more than 70% of the aircraft’s cost. This is why Lockheed Martin is develop-
ing and transferring technology to simplify the supply chain. Doing so enables its
suppliers to take advantage of manufacturing technology advancements.
One of the features of the F-35 program is its long-term production forecast
(Fig. 4). This gives the F-35 the ability to develop and implement new technologies
for cost savings now and for years to come. Multiyear buys are anticipated in the
near future, which will also help to reduce the cost of the aircraft.
One particular focus is on how the mechanics consume engineering data on
the factory floor. Three-dimensional digital models and access to the 2D paper
drawings created from them are provided to the mechanics on the floor. Work
instructions have graphics in some instances but can be expensive to maintain
and tend to be limited to complex installations, as previously discussed. One
remedy is to provide engineering data in augmented-reality-style glasses
(Fig. 18). The ability to provide work instructions hands-free with voice-activated
commands has shown promise in improving data accessibility; however, it is com-
plicated by the same problems mentioned earlier for static graphics: obsolescence
and maintenance. In the example in Fig. 18, the mechanic reads the wire number
aloud and the pin location lights up in the pin diagram view in his glasses. (In the
future, the glasses could be reading the wire number directly.) In cases like harness
routing, mechanics need 3D views of the engineering models in order to route the
harnesses. Movies for complex installations are now available to the mechanics at
their work instruction terminals and eventually perhaps in glasses. The efficient
creation, consumption, change, and utilization of the engineering data may
provide affordability savings for current and future programs. Such efforts

Fig. 17 Blueprint for Affordability investments reducing aircraft cost.

178 D. A. KINARD

Fig. 18 Consumption of the engineering data using hands-free, voice-activated glasses.

include the development of digital twin technologies by manufacturing and

sustainment.
The advent of Skypew, Facetimew, and similar applications has opened up
the potential for real-time communication among suppliers, customers, and
engineering from remote locations. Support personnel are typically located
onsite with the customers and are often sent to supplier and customer locations
to investigate and resolve problems. Communication through the use of services
like Skype, using real-time photos and two-way video links, provide opportunities
for cost savings and rapid response times. Figure 19 shows experts located at
remote locations who can interact with maintenance, production, or supplier
personnel onsite to resolve issues faster and at a lower cost. The key issue with
this technology is data security, which is high on the customer’s concern index,

Fig. 19 Remote augmented reality.

F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 179

especially for active bases. This will likely be resolved in the future because of
customers’ pronounced interest in faster, lower-cost support and the private
sector’s focus on data security solutions.
The digital thread, noncontact metrology, vision systems, artificial intelli-
gence, and machine learning technologies will bring new opportunities for auto-
mation and robotic applications. This is especially the case for routine operations,
such as sealant and coatings application, fastener installation, hole drilling, and
inspections. Also, drones are being developed to perform external aircraft inspec-
tions. Additive manufacturing has been extensively used to produce tools as well.
Its uses in fracture-critical structures are still years away due to certification
requirements; however, there are opportunities for additive manufacturing in
support equipment. It can also be used to produce nonstructurally critical parts
that cannot be produced using conventional technologies.
Lockheed Martin has been working with 5ME to implement cryogenic
machining technologies (Fig. 20) for titanium part cost savings. Liquid nitrogen
is applied at the cutting surface and has shown potential for increasing tool life
and machining speeds. This technology is a result of Creare LLC’s work on the
U.S. government Small Business Innovation Research (SBIR) program. It is an
example of Lockheed Martin’s interest in developing technologies to support
cost savings in the supply base. SBIR projects have also developed and
implemented the following:
. BTG Labs surface energy measurement system for nutplate bonding
. Creare fastener fill measurement systems
. Twin Coast Metrology and Delta Sigma Company fastener projection systems

Fig. 20 Cryogenic machining of titanium.

180 D. A. KINARD

The F-35 JPO, AFRL, Naval Air Systems Command, and Office of Naval
Research have collaborated extensively for such efforts. Their cooperation has
been instrumental in securing support for SBIR and other contracted research
and development efforts for technology development.
Another innovation that could impact future F-35 affordability is the advent
of Industry 4.0, which is the fourth industrial revolution: the revolution of data.
Steam and water power drove the first industrial revolution, electrical power
drove the second, and computers drove the third. The fourth is about the strategic
and tactical use of data resident in our systems.
Industry is recognizing that the data in IT systems can provide tactical insight
to drive efficiency in operations. It can also lower support costs for data collection,
analysis, and performance visibility and transparency. The connected enterprise
depicted in Fig. 21 is the key to enterprise efficiency. It improves the integration
of systems data, facilitates automated data collection and dashboard metrics, and
supports descriptive, predictive, and prescriptive analytics. It also connects factory
equipment with the IT systems, driving efficient usage and secure data transfer.
As previously discussed, data security is one of the enabling technologies for
the connected enterprise. This revolution is in its early stages but will rapidly drive
industry to become more efficient in above-the-factory-floor functions. Further, it

Fig. 21 The connected enterprise.

F-35 DIGITAL THREAD AND ADVANCED MANUFACTURING 181

will provide needed insight for continued productivity gains on the factory
floor and in the supply base.
Lockheed Martin has already deployed a phone/computer application for air-
craft on the production floor. It provides information about locations, schedule
performance, part shortage, nonconformances, and other factors for every aircraft
and component in work. This pertains to not only those in the Fort Worth factory,
but also the aircraft and components in Italy and Japan. The application is being
combined with a factory-wide part kit and tool Radio Frequency Identification
System (RFID). Ultimately, it will be able to automate updates of each aircraft
on monitors located at each tool position with status and performance data.

VI. CONCLUSION
The F-35 has benefitted greatly from new technologies and is well-positioned
to invite further technology development. It will do so through affordability
investments and nonaerospace and defense industry commercial developments
in artificial intelligence, augmented reality, machine learning, and other areas,
as well as the rise of Industry 4.0.
With a focus on future technologies, it will also benefit the broader defense
industry and other Lockheed Martin programs by helping them to maintain a
technological edge. Manufacturing technologies have applications in the sustain-
ment of aircraft in the field, including automated measurement technologies,
drone inspections, and data integration. The F-35 program will continue to inno-
vate with the addition of warfighter capabilities and the continued reduction
in cost through advanced manufacturing implementations. It will also provide
significant benefits in sustaining delivered aircraft.

REFERENCES

[1] Ohno, T., Toyota Production System: Beyond Large-Scale Production, Productivity
Press, Portland, OR, 1988.
[2] Womack, J. P., Jones, D. T., and Roos, D., The Machine that Changed the World:
The Story of Lean Production, Harper Perennial, New York, 1991.
[3] Spear, S., and Bowen, H. K., “Decoding the DNA of the Toyota Production System,”
Harvard Business Review, Sept. 1999, pp. 97 – 106.
[4] Cochran, D., Kinard, D., and Bi, Z., “Manufacturing System Design Meets Big Data
Analytics for Continuous Improvement,” Proceedings of the 26th CRIP Design
Conference, Stockholm, Sweden, 2016.
[5] Kraft, E., “Expanding the Digital Threat to Impact Total Ownership Cost,” U.S. Air
Force Presentation, 2013 NIST MBE Summit, National Institute of Standards and
Technology [online database], https://www.nist.gov/sites/default/files/documents/
el/msid/1Kraft_DigitalThread.pdf [retrieved 8 May 2018].
182 D. A. KINARD

[6] Aerospace Standard AS9103, “Variation Management of Key Characteristics,”

SAE International, Warrendale, PA, 2001.
[7] Smit, J., Kreutzer, S., Moeller, Carolin, and Carlberg, M., “Industry 4.0 Analytical
Study,” Policy Department A: Economic and Scientific Policy, Directorate General for
Internal Policies, European Parliament [online database], http://www.europarl.
europa.eu/RegData/etudes/STUD/2016/570007/IPOL_STU(2016)570007_EN.pdf
[retrieved 8 May 2018].
 CHAPTER 5

F-35 Weapons Design Integration

Douglas M. Hayward and Andrew K. Duff†
Lockheed Martin Aeronautics Company, Fort Worth, TX

Charles Wagner‡
F-35 Joint Program Office, Arlington, VA

Tactical aircraft operations in a high-threat, hostile environment require

an internal weapon carriage capability that supports stealth operations
while ensuring aircrew survivability and mission success. Once the
threats have been reduced, a multirole, increased external carriage capa-
bility is then required to prosecute an expanded target set. The F-35
meets both roles with internal carriage of air-to-air and air-to-surface
stores. It has a robust external carriage capability of multirole stores,
an internal 25-mm Gatling gun for the F-35A, and a missionized
25-mm gun pod for the F-35B and F-35C. It also has a stores manage-
ment system designed to control current and future store interfaces.
The F-35 weapon carriage design is based on experience gained from
developing our successful 4th-Generation F-16 and 5th-Generation
F-22 fighter aircraft. The design is highly common across the three
F-35 variants and provides interoperability and operational flexibility
to the theater commanders to engage a wide array of targets. This paper
describes the weapons requirements, design development, evolution,
and trades performed to define and mature the weapon carriage
layout. It also details the associated systems required to load, carry,
release, and interface with F-35 weapons.

I. INTRODUCTION
The design and integration of weapon carriage provisions on a stealth aircraft,
such as the F-35, are a fundamental part of the overall aircraft design. As such,
they must be considered during the earliest stages of configuration development
and layout. The weapons suite and carriage requirements are key drivers for
both the fuselage and wing designs. Internal carriage affects fuselage sizing and

 Director,
F-35 Systems Engineering.
†
Senior Manager, F-35 Armament and Stores Management.
‡
GS-14, F-35 Weapons Integrated Product Team Lead, F-35 Joint Program Office.

183
184 D. M. HAYWARD ET AL.

structural arrangement, and external carriage affects wing load and flutter require-
ments. The weapons also define the interfaces required to employ the weapons,
which drives system capabilities from a targeting, wiring, power, and computing
perspective. The mass properties and resultant weapon inertias are drivers in the
overall flight control system and impact control surface definitions. Even the
landing gear design and placement are affected by ground clearances and
center-of-gravity relationships. In spite of these design constraints, the system
does not contradict the F-35 program’s pillars of affordability, lethality, surviva-
bility, and sustainability.

II. F-35 WEAPONS SUITE AND GENERAL ARRANGEMENT

F-35 weapon requirements were defined by the F-35 Joint Program Office (JPO)
in the Joint Contract Specification (JCS). They were later refined during the
Concept Definition and Design Research (CDDR) and Concept Development
Program (CDP) phases, prior to the System Development and Demonstration
(SDD) production aircraft selection. The initial weapons suite from the SDD con-
tract contained effectively all stores in the U.S. and U.K. inventories, as well as
smart and older, unguided weapons. Early in the SDD program, we refined the
weapons suite to eliminate some of the older weapons that were becoming obso-
lete, such as the AIM-120/B Advanced Medium-Range Air-to-Air Missile
(AMRAAM), naval mines, current generations of the AGM-84 Harpoon anti-ship
missile, and the AGM-88 High-Speed Anti-Radiation Missile (HARM).
Newer weapons, such as the Small Diameter Bomb (SDB) and Guided Bomb
Unit (GBU) GBU-49 Enhanced Paveway II munitions, were added to the require-
ments list. This keeps the F-35 weapons suite as modern, relevant, and flexible as
possible. The suite enables the F-35 to perform multirole missions, providing
flexibility to the operational commander to prosecute a wide variety of targets.
Figure 1 shows the resultant internal and external weapons suite.
In addition to the weapons the F-35 is required to carry, the JCS also included
other specific requirements that drove the weapons carriage arrangement and
weapon station capacities, such as being able to carry external fuel tanks. It also
issued mixed parent and dual-carriage requirements for each service. Other
influential factors were the variant-dependent mission requirements, such as the
F-35A’s need for an internal gun and the F-35B/C’s needs for a missionized gun.
Additional requirements focused on safety and flexibility, such as the requirement
for MIL-STD-1760 interfaces at all stations and the ability to safe the weapons
systems from outside the cockpit. These design requirements resulted in four
internal and seven external weapon stations. Figure 2 shows the general arrange-
ment of the weapon carriage, and Fig. 3 shows the station type and weight capacities.
The commonality of the weapon carriage design across the three aircraft var-
iants was maximized to reduce the differences in the certification process and to
minimize cost. The internal weapon bays are common on the F-35A and F-35C
F-35 WEAPONS DESIGN INTEGRATION 185

Fig. 1 F-35 internal and external weapon carriage requirements.

Fig. 2 General weapon carriage arrangement.

186 D. M. HAYWARD ET AL.

variants. These bays are capable of carrying one internal 2000-lb class store with
one AIM-120C AMRAAM missile in each bay. The F-35B bay is unique and
capable of carrying one 1000-lb class store with one AIM-120C missile in each
bay. Air-to-ground or air-to-air stores are carried on adapters on the roof of
the bays. Each bay design includes a door-mounted AIM-120C station that
rotates with the door to improve separation clearance and loading access. The
external wing stations are located on nearly common buttock line stations
across all three variants. The suspension and release equipment is common
across all variants, with the addition of a unique 14-in. bomb rack tailored to
the 1000-lb class internal bays for the F-35B aircraft. The Stores Management
System (SMS), interfaces, safing features, and system operation are common.
As part of the SDD program, a subset of these weapons and requirements
was required to be fully certified. The weapons in the subset needed to provide
a warfighting capacity in time for each service’s declaration of initial operational
capability. The weapons and specific loadouts we certified during the SDD
program are shown in Fig. 4. The certification effort for these capabilities and
specific loadouts to be fully certified was based on extensive analysis and
testing. The design was developed and validated using a combination of three-
dimensional (3D) design tools, wind tunnel tests, and computational fluid
dynamics analyses. The F-35’s preliminary layout and detail designs were devel-
oped using a 3D digital thread to ensure that all systems were effectively integrated
and complete. Wind tunnel testing formed the backbone of the validation tests
and included 11 test entries with more than 3900 user-occupancy hours and
more than 1700 air-on hours. In these tests, we assessed the weapon bay design
from both separation and acoustic perspectives. We also defined performance

Fig. 3 Weapon station capacity.

F-35 WEAPONS DESIGN INTEGRATION 187

Fig. 4 SDD certified weapon loadouts.

requirements for the Suspension and Release Equipment (S&RE) and validated
our separation models to support flight test continuation criteria. The weapon cer-
tification effort culminated in the successful separation of 110 air-to-ground and
73 air-to-air stores during the flight test program. The effort also led to 46 Weapon
Delivery Assessment (WDA) engagements in which the total end-to-end systems
were exercised to validate the overall air system employment’s accuracy. The
details of the weapon separation and certification efforts are described in Ref. [1].

III. INTERNAL WEAPON BAY DESIGN

The final F-35 configuration was heavily influenced by choices made with respect
to the weapon bay design. From a weapon arrangement standpoint, designers
often discuss packing efficiencies associated with rotary, stacked, flat, and
tandem weapon bays; however, it is not until these arrangements are integrated
into an aircraft concept that the resultant system can be evaluated. As with all
designs, a balance of competing requirements must be struck. From a weapons
standpoint, the final configuration must result in:
. Tight packaging solutions while also optimizing weapon loading
. Open bay acoustic environment
188 D. M. HAYWARD ET AL.

. Employment timelines
. Station failure immunity
. Separation risk
. Loadout flexibility

Embedded in the arrangement decision are lower-level impacts that may signifi-
cantly influence air vehicle weight, handling qualities, and performance. These
include weapon bay door configuration, the need for acoustic suppression
devices, provisions for maintainability, and mission reconfiguration. Finally, the
F-35 must provide these 5th-Generation stealth features while also providing a
solution across three separate variants and for a variety of users. This implies
additional constraints on loading solutions and structural load considerations.
Key to making F-35 weapon arrangement decisions was the list of weapons
required for internal carriage. Many of the tightly packed arrangements often
advertised by weapon suppliers and advanced design advocates depict a limited
number of weapon configurations. This allows more efficient options to be
employed when it comes to rotary or stacked bays. However, once the full list
of weapons is explored, many of these options eliminate themselves. Additionally,
rotary and stacked configurations often imply a deep weapons bay. Historically,
deep weapons bays drive higher acoustic levels. This leads to the need for
acoustic suppression devices to comply with weapon qualification levels, and
adversely impacts the structural integration in a fighter [2]. Early F-35 wind
tunnel testing evaluated the acoustic environment associated with the relatively
shallow bay and intended operational envelope. The testing indicated that the
environment would allow for the incorporation of all required weapons without
needing acoustic suppression devices. Because the use of ramps and spoilers
adds weight, complexity, and bay length, the decision was made to move
forward without dedicated acoustic suppression devices. This option could have
presented some risk if the actual environment proved to be incompatible with
the weapon qualification levels; however, the benefit to the aircraft was sufficiently
significant and long lasting to justify taking the risk. Further, the qualification
levels of most weapons are based on the existing aircraft environment. Accord-
ingly, increasing the weapon qualification (if required) is typically a matter of
retesting to higher levels, and does not necessarily indicate an inherent limitation
of the weapon.
Aside from rotary and stacked weapon bay configurations, tandem weapons
bays were also an option. However, from a weapons standpoint, these bays are
less desirable due to their impacts on loading. They are especially problematic
when loading the Navy’s F-35C Carrier Variant (CV) aircraft with the tail
parked over water. Additionally, tandem bays cause undesirable aircraft center-
of-gravity excursion, given the weight of the weapons to be carried. The final
F-35 configuration is a roof-mounted air-to-ground weapon and a door-mounted
air-to-air missile that offers a good mix of each weapon bay type. It has the
F-35 WEAPONS DESIGN INTEGRATION 189

efficiency of stacked bays without the acoustic and loading issues, and
the flexibility of a purely flat bay. After the overall configuration was selected,
detail decisions needed to be made with respect to design constraints to establish
the final aircraft sizing. One unique feature of the F-35 weapons bay is the
air-to-ground stores are angled nose inboard. This was done to improve the inte-
gration of the bays with the fuselage and was validated through wind tunnel
testing.
One of the most critical and controversial decisions to be made involved
weapon envelope size. This refers to the volume to be allocated and protected
around the weapons to allow for adequate installation, carriage, and separation.
Again, the large mix of weapons to be carried by the F-35 played a major role
in these decisions. Historically, aircraft weapons bays were pristine boxes with
fixed S&RE locations. The bays were not encumbered with hydraulic, pneumatic,
and electrical routing or an abundance of vehicle system components. However,
overall volume in a fighter aircraft is at a premium. As such, insisting that all
routing and systems be located outside the weapons bay significantly increases
the size of the aircraft. Ultimately, the decision was made to allow systems and
weapons to coexist in the weapons bay. This required the qualification of all
systems to weapon bay environmental levels. The resultant weapon bay design
is highly integrated with systems, routing, and weapons to ensure the smallest
package and highest aircraft performance possible, as Fig. 5 illustrates.

Fig. 5 Bottom view looking into the weapons bay.

190 D. M. HAYWARD ET AL.

Sharing the weapon bay volume necessitates tight controls to ensure that the
evolving systems and routing do not intrude on space needed for weapons. To
minimize the weapons package volume, we performed a tradeoff between location
commonality and bay length. If a singular location was allocated in the bay, this
would provide an easy and efficient installation from a structural standpoint. This
is because all loads would be carried by the same structure in the same way, no
matter which weapon was installed. However, weapons come in all shapes and
sizes, and hook locations are not always located in the middle of the weapon. Con-
sequently, having a common hook location would result in a relatively large
weapon volume.
By contrast, we could locate all weapons in the minimum volume required to
allow the resultant hook and bomb rack location to be a fallout. In this way, the
bay volume and overall size of the aircraft could be minimized. Unfortunately, this
would create a need for a variety of Alternate Mission Equipment (AME) adapters
and hook locations within the bay. However, even with the weight and logistic
penalties for this arrangement, the benefit to the overall aircraft performance
was undeniable. Therefore, the final weapon configuration embodied multiple
bomb rack locations. To accommodate this solution, we identified several adapters
and developed a system to quickly latch them into the bay. We designed the Ord-
nance Quick Latch System (OQLS) to permit quick bay reconfiguration by latch-
ing and snubbing adapters in a manner similar to that used by bomb racks to latch
and snub stores. Figure 6 shows the OQLS layout and a typical bomb rack adapter
when clipped into the OQLS.
With the location of the stores established, the final volume required for car-
riage and employment could be established. First, however, we needed to decide
how much volume was needed around each store as a stay-out zone. Key factors to
consider were the volumes required for maintenance, weapon motion (deflection)
during carriage, and uncertainty associated with weapon separation. Historically,
military specifications, such as MIL-STD-1289 and MIL-I-8671, would be used
to establish clearance requirements. However, these specifications are often

Fig. 6 OQLS arrangement and typical bomb rack adapter installation.

F-35 WEAPONS DESIGN INTEGRATION 191

conservative in some areas and vague in others, resulting in larger bay volumes.
In general, the aircraft performance reductions that would be driven by strict
adherence to military standards were deemed unacceptable. The F-35 JCS did
not require the use of historical specifications for weapon spacing. However, early
versions of the JCS did require the maintenance of a 2-in. stay-out volume around
all weapons. Some exceptions were allowed between the hook points where
relative motion would be at a minimum. As the program progressed, there
were multiple identified and proven instances in which this blanket requirement
would significantly penalize the air vehicle if strictly followed. Therefore, the spe-
cifications were ultimately changed to preclude store-to-store or store-to-aircraft
contact, with no specific clearance requirement identified. This approach required
the development of complete, specific design guidance to establish and protect the
weapon stay-out zones.
Once again, a balance of risk and benefit was required in establishing the
weapon stay-out zones. Too much volume allocated to the weapons would
result in an overly large and heavy aircraft, decreasing performance as a result.
Too little volume would result in high-risk weapon carriage and employment,
with the potential for aircraft operational envelope limitations or configuration
restrictions. The general approach chosen for the F-35 was to create a volume
that reflected tolerances, freeplay, and a deflection of the store throughout the
operational envelope. In addition, we added a small margin to account for analysis
fidelity. This margin was based on differences discovered between analysis
accuracy and flight test verification, as proven on the F-22 program.
In the end, decisions made to optimize the overall air vehicle made the weapon
volume allocation more difficult. This is because weapon attachment involves
multiple interfaces (S&RE, adapters, OQLS) with associated flexibility and free-
play. Deflections associated with these interfaces are significant when fractions
of an inch are important to the overall integration of the weapons.
Ultimately, the weapon bay volume will be challenged again and again until it
has been reduced to its bare minimum. Therefore, we performed a detailed mod-
eling of each weapon, to include loads associated with its specific carriage and
employment envelope. The deflection of a store depends on where and how it
is attached, so the envelopes that result from the deflection analysis are larger
when farther away from the attach points. When viewed from above, the final
weapon envelopes appear narrow in the middle, near the bomb racks, and
larger toward the nose and tail of the stores. As a result, these envelopes are
referred to as bow ties. We rolled together and consolidated the bow ties from
each weapon to create the overall weapon volume. This clearance approach,
shown in Fig. 7, is a unique design feature of the F-35.
When looking aft at the volumes, the ejection sweep was largely influenced
by the design of the S&RE, which constrains the store during the ejection
stroke. The stay-out volumes also reflected a narrow band of deflection at
the top that flared to a 10-deg sweep at the bottom. However, the width
assumptions made for the weapons bay required that the lateral motion of
192 D. M. HAYWARD ET AL.

Fig. 7 Weapon stay-out volume bow tie arrangement.

the stores never exceed 0.75 in. total at End of Stroke (EOS). This was particu-
larly critical between the suspension hooks in order to accommodate tolerance,
freeplay, and deflection. This resulted in a significant design stiffness driver for
the S&RE.
After establishing the volume, we needed to address other issues due to the
highly integrated relationship between the required weapon stay-out volume
and the volume used by systems and structure. Namely, we needed to determine
how to ensure adherence to the volumes during the design, manufacture, and field
service phases. Each phase of the volume life cycle required individual treatment.
The design phase required establishing configuration control processes, to include
design release requirements and volume conflict dispute resolution boards. For
the manufacture phase, the central question was whether standard manufacturing
tools and processes could ensure adherence to the volumes. This was especially
important in areas where there is some allowed variation in harness routing
and support sizing. Stay-out volume maintenance during in-service use also
needed to be accommodated to prevent clearance conflicts due to normal aircraft
operations, modifications, and maintenance. To ensure that adequate clearance
was retained, we created a physical tool that included all required weapons,
AME, deflections, and fall clearances.
The tool we created for the field service phase is now used on every F-35
that leaves the factory to ensure that the volume is conflict free. Figure 8 shows
one of the fit-check tools used to inspect the bay and verify the bay volume.
One key feature of the tool is its inclusion of removable sections, which enables
access to visualize and measure specific interfaces. Currently, we are investigating
the efficiency of using a laser mapping of each bay as an alternative to the
physical tool.
Another concern is that maintenance activities in the field may impact the
final configuration of systems near the stay-out zones. This is especially difficult
for maintainers in the field to recognize because not all weapons are installed
during maintenance, making it difficult to understand the potential impact. To
F-35 WEAPONS DESIGN INTEGRATION 193

address this, aerospace equipment instructions have been created to identify

critical areas for inspection as the aircraft is maintained.
The ability to load weapons was a key consideration when establishing the
configuration for 5th-Generation internal carriage aircraft, especially for the
limited volumes of internal weapon bays. Loading considerations for the F-35
were complicated by the fact that many different services and countries would
be using the aircraft with a variety of support equipment. One of the toughest
loading requirements was associated with shipboard operations. In this situation,
push-up loading equipment (e.g., the legacy Air Force MJ-1 munitions jammer) is
not suitable due to space requirements. It is also ill-suited due to the need to better
control the store being loaded during rough weather conditions. Historically, the
solution for shipboard loading has been the Single Hoist Ordnance Loading
System (SHOLS). However, the use of SHOLS within the limited confines of
the F-35 weapons bay is difficult due to volume and access requirements.
Access to both sides of the weapon is required to permit the installation and
disconnection of the cable system. Originally, this was to be accommodated on
the F-35 through access panels located in the main wheel wells; however, as the
design progressed and more weapon locations were defined, this arrangement
was no longer feasible.
As a result, the Ordnance Hoist System (OHS) was developed, originally
conceived as a permanently installed onboard hoist system. The OHS would
raise the adapter, bomb rack, and weapon into the bay using straps located at
the front and aft ends of the adapter. The hoist spools were to be driven by
small, handheld battery-powered motors by two maintainers. Coupled with the
bomb racks in adapters, the OHS would allow the weapon to be fully mounted

Fig. 8 Weapon bay clearance fit-check tool.

194 D. M. HAYWARD ET AL.

Fig. 9 Weapon bay maintainer access study.

on the rack and lanyards rigged outside the weapons bay. Once completed, the
whole assembly would be raised into the bay and quickly latched into the
OQLS. The system worked so well in demonstrations for the F-35C and F-35B air-
craft that the OHS was incorporated onto the F-35A as well. This was because the
MJ-1 loading of the stores left little room for maneuvering error. While exploring
various AME, S&RE, and weapon loading equipment, the F-35 team used compu-
ter simulations and mockups, along with virtual reality, extensively. Figure 9
depicts one of the early studies showing that maintainers had room to conduct
required operations with an onboard hoist system.
As the design of this system progressed, additional requirements for ship-
board operation evolved. The concept of using handheld drill motors for
power was lost, and a requirement for a stand-alone 270-V battery was added.
The battery was sized to load multiple aircraft and included a sophisticated
controller that automated many aspects of the loading process. This increase
in complexity, along with changes made to receive a Conformité Européene
(CE) marking, ultimately led to the elimination of the OHS from F-35A aircraft.
Following this, a manual drive replacement was developed for the powered
components of the OHS. Although the OHS has evolved, it remains the only
approved method for loading weapons shipboard. With the OHS removed
F-35 WEAPONS DESIGN INTEGRATION 195

from the F-35A, a new Length-Extender Table Adapter (LETA) was developed to
install on the MJ-1 load table and improve clearances.

IV. EXTERNAL WEAPON CARRIAGE DESIGN

The emphasis of the F-35 is on internal weapon carriage, but there will always be
ferry and lower threat missions that benefit from external carriage. As a result, all
F-35 variants include seven external hardpoints: three on each wing and one on
the fuselage centerline. As noted previously, stores specified for external carriage
include a variety of weight classes and types. Station type and weight capacities
(shown in Fig. 3) reflect the differences in station capabilities of the F-35B aircraft
for structural weight considerations. Fuel tank carriage is accommodated at the
inboard stations, and each of the four inboard stations can accommodate
both parent and multiple-store carriage. The outboard stations are for air-to-air
weapons and were designed to improve stealth performance over traditional
external pylons and launchers, while still balancing the cost and serviceability
of traditional AME. Figure 10 depicts an F-35A heavy external mission loadout
with external air-to-air missiles on the dedicated missile stations. It also shows
2000-lb air-to-ground parent carriage stores at the multirole air-to-ground or
air-to-air stations.
From a configuration standpoint, the locations of the pylons were established
in a manner similar to that for legacy fighter aircraft; fowever, the fact that the
design needed to accommodate three variants meant that additional factors

Fig. 10 Heavy weapon mission loadout.

196 D. M. HAYWARD ET AL.

needed to be considered. For example, all external designs need to consider

locations to accommodate wing flaps, landing gear doors, and ground lines for
all the landing gear and pitch angle variations. Owing to this, the addition of
the seaborne F-35B/C variants added some challenges. Specifically for F-35C,
the deck of the carrier has multiple obstructions that can become an issue.
These obstructions must be considered when evaluating high roll angles and
large stores. Furthermore, the deck obstructions, cables, and launch shuttle can
create additional clearance challenges for some of the designs. Examples are the
gun pod on a centerline station and outboard pylon clearances. Additionally,
the plume impingement from the F-35B roll post is another design driver that
has to be considered. These are columns of hot gas ejected from the engine at
the wing-to-fuselage intersection used to control roll attitude when in vertical
flight. It is important to minimize store heating in the basic positioning of the
stores and pylon design. Additionally, the carriage pitch angle of the stores is
usually established to eliminate drag in cruise flight conditions. Therefore, the
stores are nose down to accommodate Angle of Attack (AoA) and can be nose
inboard to accommodate wing wash. The mix of variants, missions, and required
stores made all these compromises a challenge.
One of the tools frequently used in F-35 development from a human factors
standpoint was virtual reality. Many of our questions concerning maintainer
access, vision, and safety with conducting operations were addressed using tra-
ditional mockups. For the remainder, we supplemented those assessments with
virtual assessments. Figure 11 depicts a maintainer in a mixed (i.e., having
some of the stores and physical constraints actually connected with virtual hard-
ware) virtual environment. The F-35 weapons team used this tool extensively early
in the concept evaluation phase to validate the acceptability of initial design.

Fig. 11 Maintainer virtual environment clearance study.

F-35 WEAPONS DESIGN INTEGRATION 197

There is a high degree of commonality among the external AME. However,

there were differences required due to the fact that the F-35C wing and flutter
requirements were significantly different from those of the F-35A/B wing. This
meant that pylons used on the F-35C aircraft needed to carry the stores farther
forward, with respect to the wing hardpoints, than the F-35A/B pylons. Despite
this difference, however, there remained a high degree of commonality; for
example, the same pylons are used at inboard and outboard stations. Rather
than changing the pylon structural core to accommodate the significant differ-
ences in wing twist, we provide an adapter. Other commonality-driven features
include the fact that the sway braces can be reversed to allow the installation of
the pylon on either the left or right wing. In addition, such components as the
pivot fittings, fairings, and the entire Pneumatic Power Source (PPS) are also
common. Commonality in the design has significant advantages for provisioning
and spares, as well as cost reductions that come from manufacturing economies of
scale. Affordability has always been one of the foundations of the F-35 program.
The air-to-air pylons at Stations 1 and 11 presented a unique challenge,
and commonality was once again a key consideration. Ideally, from a structural
standpoint, the pylon would accommodate a launcher directly below the hard-
points. However, adjacent store station clearance and structural provisions on
the three variant wings meant that the missile could not be located directly
below the hardpoints and needed to be offset. Rather than creating a unique
canted pylon, the pylon was designed with two launcher attach points at its
base. This allowed for launcher offset while also allowing the same pylon to be
used on the left or right wing by simply switching the location of the launcher.
Additionally, the same pylon could be used on all three variants’ wings—
despite significant angular differences—by simply changing a set of bushings
used during the attachment.
The rail launcher used at the base of the pylon was originally designated as a
non-low observable LAU-148 and LAU-149. The only difference between the two
was that only the LAU-149 contained a gas generation system. This was needed to
provide continuous cryogenic cooling to the AIM-132 Advanced Short Range
Air-to-Air Missile (ASRAAM) seeker head. By contrast, the LAU-148 contains
ballast. Either launcher can be used with an AIM-9X. The fact that the
LAU-148 used ballast was done to create similarity between the two launchers
such that extensive flutter and loads testing would not need to be repeated. As
the F-35 program matured, the desire for improved Low Observability (LO)
of the air-to-air mission evolved. This new requirement resulted in the design
of the LAU-151 and LAU-152 rail launchers and an associated upgrade to the
air-to-air pylon (now designated as a SUU-96). Figure 12 depicts these designs.
To limit the financial impact of potential regression flight testing, a significant
effort was expended to design the new AME to be very similar (structurally) to
the initial design. The cost savings of this approach over a completely new quali-
fication and certification program were significant.
198 D. M. HAYWARD ET AL.

Fig. 12 SUU-96 pylon and launcher installation.

V. SUSPENSION AND RELEASE EQUIPMENT DESIGN

F-35 S&RE is a family of components that carries and employs air-to-ground and
air-to-air weapons. Safe release and separation is achieved through two eject
pistons powered by 5000-psi nonpyrotechnic pneumatics. The primary
air-to-ground bomb rack is the BRU-68/A, which can carry stores from 500- to
2000-lb class using either 14- or 30-in. conventional suspension hooks. It is
used internally on the F-35A/C variants, and on all three variants in the
air-to-ground pylon. The BRU-67/A is similar to the BRU-68/A but is only
used in the F-35B internal weapons bay, and it can carry 500- to 1000-lb class
stores using conventional 14-in. hooks. The BRU-67/A is a product of the
Short Takeoff and Vertical Landing (STOVL) Weight Action Team (SWAT)
program, and it reduced air vehicle weight by 20 lb. The LAU-147/A is specifically
designed to eject-launch the AIM-120 missile from any of the internal weapons
stations on all three F-35 variants. Since the LAU-147/A is uniquely designed
to eject only the AIM-120 missile, the sway brace is customized to the diameter
of the missile. It incorporates an anti-rotation feature in the sway braces to
decrease missile roll tendencies discovered during early wind tunnel testing.
S&RE also has unique store control requirements to accommodate the tight
internal carriage volume allocations. Historically, two methods have been
employed in S&RE to ensure that store clearances are maintained during the ejec-
tion event. The first is to push the store hard enough to achieve high-EOS velocity,
and the second is to restrain the store’s lateral motion during the ejection event.
F-35 WEAPONS DESIGN INTEGRATION 199

The first concept is limited by the structural loads that are imparted to the aircraft
and store, as well as the physical constraints of the stored energy package. The
second concept relies on the stiffness of the S&RE and backup structure to
provide lateral constraints. Reducing deflections by increasing system stiffness
increases weight.
A hybrid approach combining both the restraint and velocity concepts was
chosen for the F-35 for the following reasons. The variety of weapon locations
in the internal weapon bays requires movement of S&RE. This requirement to
translate the S&RE necessitates locating the pneumatic power storage bottle in
the S&RE housing. This constrained the maximum volume of gas and resultant
energy available for ejection. Whereas actual EOS performance is a function of
many variables, two of the primary drivers are ejection stroke length and stored
energy content. Packaging issues tend to drive the stroke length. As a result, orig-
inal versions of the bomb racks contained multistage pistons to achieve the desired
stroke length while fitting in the rack housing. To constrain the stores, the sway
braces were designed to move with the store and provide lateral restraint
through inertia only. The alternative was to constrain the store throughout the
stroke and release the hooks at EOS. Historically, racks with true lateral constraint
have been complex, heavy linkages that open the store attach hooks at EOS. The
F-35 hybrid approach eliminated much of the weight concern from these linkages
by using simple pistons with moving sway braces; however, it still required signifi-
cant tailoring of the design to accommodate the required stiffness. We built an
analytical model of the weapons bay structure and S&RE, and we performed
complex dynamic simulations. With these simulations, we calculated the lateral
deflection of stores during the ejection event throughout the operational envelope.
Unfortunately, it was shown that the BRU-67/68 could not achieve the required
clearance with multistage pistons. This is the reason for which the ejector towers
of the BRU-67 and BRU-68 were extended beyond the rack housing to eliminate
the multistage pistons. The LAU-147 retained the original two-stage piston
because the door-mounted location required minimal launcher height. Moreover,
the available volume could accommodate the AMRAAM ejection path. This
hybrid design provides the mix of performance and constraint needed to live
within the very strict lateral confines of the F-35 bay.
S&RE was designed to accommodate nonpyrotechnic ejection while conform-
ing to the conventional bomb rack load and form requirements of MIL-A-8591
and MIL-STD-2088. Because the air-to-ground stores vary in diameters
(ranging from 8 to 23 in.), the bomb rack units incorporate a semi-automatic
pivoting sway brace. This sway brace allows the maintainer to quickly adjust
the rack to the many F-35 stores. This feature simplifies store installation and
removal procedures in the limited space of a tight weapons bay. All three S&RE
units incorporate the same reversible in-flight lock that provides mechanical
blocking and electrical interrupts to prevent inadvertent store ejection. These
enhanced safing features provide at least two levels of protection whenever the
maintainer is working around the rack.
200 D. M. HAYWARD ET AL.

TABLE 1 F-35 SUSPENSION AND RELEASE EQUIPMENT

BRU-68 BRU-67 LAU-147

Envelope 36.0  10.9  4.0 32.0  10.9  4.0 36.9  6.9  4.0
(L  H  W) (in.)
Specification 87.5 67.5 63.9
Weight (lb)
Store Class (lb) 500/1000/2000 500/1000 350
EOS Velocity (fps) 20/15/11 20/15 25
Piston Type/ Single/6.9 Single/6.9 Dual/7.5
Length (in.)

The level of Prognostic Health Management (PHM) incorporated into the

S&RE is consistent with maintainability goals for the F-35. Ejector technology
continues to advance, and other options to resolve 5th-Generation fighter inte-
gration issues may appear within the next few years. However, F-35 S&RE
(Table 1) represents a significant departure and advancement from 4th-Gener-
ation pyrotechnic systems.
Nonpyrotechnic ejection was a key requirement for F-35 S&RE. Legacy ejector
systems incorporated in 4th-Generation (and earlier) aircraft relied on pyrotech-
nic cartridges to provide ejection power. By changing to a nonpyrotechnic power
source, many cost and maintenance issues can be eliminated; for example, pyro-
technic cartridges require frequent rack disassembly to remove debris and
replace damaged parts from the burning combustibles. Furthermore, inconsistent
burn profiles caused by slow- or fast-burning cartridges generate performance
anomalies that can be eliminated through the use of stored gas systems. In
addition, shipping and storing pyrotechnics creates safety issues and an expensive
logistics footprint. These issues were evaluated in a variety of feasibility studies
performed in the early 1990s. The Life-Cycle Cost (LCC) advantages of pneuma-
tic racks for the F-35 program were estimated at between $2.3 and $2.6 billion in
savings for internal carriage [3]. If external pylons were also considered,
an additional $1.67 billion in savings was estimated. As a result, the F-35 air
vehicle was required to incorporate a nonpyrotechnic power source into the S&RE.
There are many ways to generate and maintain the high gas pressures needed
for store ejection. The solutions generally comprise three types of systems:
1. Stored gas systems that must be charged on the ground
F-35 WEAPONS DESIGN INTEGRATION 201

2. High-pressure gas generation systems that provide onboard pumping

3. Hybrid systems that use hydraulic intensified gas systems, both closed and
open loop

Aside from the requirement to provide a high-pressure volume of gas, the

thermal variations present in a high-performance aircraft will drive pressure vari-
ations that need to be accommodated as well. Given the temperature differences
between sea level and 40,000 ft, the system loses as much as one-third of the
pressure when flying at altitude. Therefore, the system must be either regulated
or supplemented to accommodate the reduction in pressure.
The time required for the system to recharge or be serviced in between mis-
sions is also a design driver. For the F-35, the time allotment for an integrated
combat turn did not accommodate servicing each weapon station’s pneumatic
power module. Therefore, we decided to eliminate the need for ground charging
of all racks and launchers, both internally and on the wing stations. As a result,
the F-35 has an onboard PPS that generates the required 5000 psi source pressure
for the racks and launchers. The components of this system are depicted in
Table 2.
Although the pumps and controllers needed to accommodate this require-
ment add weight and complexity to the system, the task and support equipment
benefits to the maintainer outweigh the costs. To minimize the system sizing, the
PPS was designed as a maintenance system. In it, the pump sizing and power
requirements assume that S&RE reliability only requires the PPS to pump four
empty racks at once, four times in the life of the aircraft. Most pressurizing oper-
ations aim to maintain pressure due to low temperatures, or to repressurize the
partially discharged S&RE following a store release. The PPS concept of oper-
ations is to recharge all racks following a store release to full operating pressure
during the return flight. This must be done before the aircraft is chocked to
begin the integrated combat turn. This operational concept requires that the

TABLE 2 F-35 PNEUMATIC POWER SOURCE EQUIPMENT

Motor Filter/ Pylon Electronic

Compressor Manifold Filter Control Unit

Envelope 12.4  4.3  4.4 13.1  4.1  5.3 5.6  3.1  4.7 8.5  3.9  4.4
(L  H  W)
(in.)
Weight (Ib) 11.0 7.0 3.0 2.7
202 D. M. HAYWARD ET AL.

racks be normally charged even when in storage. This concept was a major driver
of S&RE requirements because the racks had to be qualified for handling, storage,
and shipping while fully charged to 5000 psi.
The PPS was sized primarily by the needs of the internal weapon bay. A single
PPS consisting of a motor/pump assembly, electronic control unit, and filter/
manifold assembly provides high-pressure pure air to all four internal weapon
stations. To maintain performance at any altitude, the PPS receives inlet air
from the aircraft’s Power and Thermal Management System (PTMS). The
PTMS typically provides air at around 18 psi, no matter what the ambient air
pressure. The pump charges a filter/manifold assembly that subsequently distri-
butes the air to each weapon station through a series of valves, as directed by air-
craft commands. Each station is isolated through specific valving and sequential
distribution rather than plumbing to a common bottle. This isolation eliminates
the potential to lose charging capability for all stations due to a failure (leak) at
one. For added reliability, each rack contains a check valve to ensure that it is
not possible to back-flow out to the PPS once engaged.
Compression of ambient air results in significant amounts of water expulsion
because the relative humidity will exceed 100% at high pressure. As a result, the
system contains devices to clean, dry, and expel excess moisture from the com-
pressed air. Removing this moisture prevents contamination of the racks that
may prevent operation from icing. For commonality, the motor/pump assembly
and electronic controller used for the weapons bay is the same motor/pump
and electronic control unit used in the air-to-ground pylon. Because the PPS
supports a single rack in the air-to-ground pylon, the filter assembly does not
incorporate a manifold or isolation valves, allowing air to flow directly into the
rack reservoir. Unlike the weapons bay with PPS inlet air from the PTMS, the
pylon PPS inlet air is from ambient. Accordingly, pressure varies as a function
of altitude. Although the pylon PPS performance is reduced due to altitude
effects, recharge times are equivalent to charge times for all the main bay racks
because it only services one rack. Therefore, the pylon and all internal racks are
fully recharged by the time the aircraft returns to the chocks to begin an integrated
combat turn.
Mission reconfiguration is a key requirement for weapons in the bay, and
there is a wide variety of S&RE locations within the bay. To accommodate
these, multiple flexible hoses are required to connect the S&RE to the PPS after
the adapter and bomb rack have been loaded. The hoses are equipped with
Quick Disconnect (QD) fittings and incorporate a safety device that prevents
de-mating the QD until the hose has been depressurized. As testing of the F-35
progressed, it was apparent that even the smallest of air leaks can have a profound
impact on overall system performance. As a result, the system design has evolved
over time. Improvements incorporated range from design changes in the hose
QD fittings to process changes in PPS manufacturing methods. To improve the
maintainer experience, enhanced courseware is being deployed explaining
system operation, troubleshooting, and normal operational expectations.
F-35 WEAPONS DESIGN INTEGRATION 203

VI. GUN SYSTEM DESIGN

The gun system installation was defined by the requirements of each aircraft
variant. The F-35A Conventional Takeoff and Landing (CTOL) aircraft required
an internal gun system. The F-35B STOVL and F-35C CV aircraft required a mis-
sionized gun system that could be installed when necessary to meet specific
mission requirements. The primary requirement of the gun system is that of a
Close Air Support (CAS) mission, but it must also be able to accomplish an
air-to-air combat role. The program specification includes an open fire range of
9000 ft, a specified strafe dive angle and speed, and a requirement for three effec-
tive passes per ammunition loadout. Due to the dive angle and speed requirement,
a limited time is available to complete each effective pass.
During the competition phase of the F-35 program, a trade study was com-
pleted to determine the best gun system to meet the aircraft specification require-
ments. The trade study reviewed five possible gun solutions that included three
Gatling guns: the M61A2 20-mm, the GAU-12 25-mm, and the advanced
cased-telescoped round CT-525 25-mm. The possible solutions also included
two single-barrel guns: the BK-27 27-mm and the Aden 25-mm. Each system
was compared with respect to its ability to be integrated onto the aircraft. This
was based on weight, volume, interface loads, round capacity, and power require-
ments. It also factored in the ability to be integrated internally into the Air Force
aircraft and into a missionized pod for the Marine Corps and Navy.
The systems were evaluated with respect to performance capabilities that
included system firing rates, round accuracy, and round effectiveness against
the target set stipulated in the aircraft system specification. Development and
Unit Recurring Flyaway (URF) costs were also compared for each of the systems.
After reviewing the data, we determined that the Mauser BK-27 27-mm system
was the best system for integration for both the internal gun system and missio-
nized gun system. This determination was based on the following factors:

. It is the lightest-weight solution.

. It requires no hydraulic power from the aircraft.
. It has the lowest predicted development and URF costs.
. It requires only 152 rounds.
. It has high round accuracy and effectiveness.

The cased, telescoped CT-525 had the potential advantages of increased lethality
and reduced weight; however, the design was not mature enough for the F-35 and
would have introduced unacceptable risk in terms of development and cost
if selected.
During the early detailed design phase of the production aircraft, a concern
surfaced with respect to the LCC and lethality of the BK-27 and its ammunition.
As a result, we conducted another trade study to ensure the selection of the
204 D. M. HAYWARD ET AL.

best gun system solution. The new trade study included the BK-27, GAU-12,
and GD-425 (a four-barrel variant of the GAU-12). The trade study evaluated
the three gun systems based on the same criteria used in the initial trade
study that had led to the BK-27 being chosen; however, this study focused on
lethality and LCC as well, along with the availability of 25-mm and 27-mm
rounds.
The results of the trade study showed that the GAU-12, shown in Fig. 13, was
the best solution. It had lower development costs, lower URF costs, significantly
lower LCC, and improved lethality. Further, the 25-mm rounds were available
to the services. The GAU-12 selected was to be hydraulically driven, fire 4000
shots per minute, and use the 25-mm rounds already in the United States’ inven-
tory. The GAU-12 is a reverse-clearing gun system. It reverses direction after each
burst and brings unfired rounds from the last burst back through the gun and into
position to be fired on the next burst. This reverse-clearing procedure ensures that
the highest number of rounds in the ammunition handling system will be fired
during normal-burst firing.
GAU-12 development design and testing was in progress when the F-35
program entered into the SWAT weight-reduction program. To support SWAT
program goals, the GD-425 four-barrel Gatling gun was reviewed to determine
the weight savings that could be achieved over the five-barrel GAU-12 system.
It was determined that a weight savings of 35 lb could be achieved by utilizing
the GD-425 system for the F-35A. Although this change did not affect the
empty weight of the F-35B variant, it was determined to be a good trade for the

Fig. 13 GAU-12 gun.

F-35 WEAPONS DESIGN INTEGRATION 205

Fig. 14 F-35A internal GAU-22/A and ammunition handling system installation.

overall F-35A configuration. The weight reduction was accomplished by removing

one barrel and one bolt assembly and its supporting guides. We also reduced the
diameter required for the new gun housing and rotor. The GD-425 (officially
referred to as GAU-22/A) is a hydraulically driven gun with a firing rate of
3000 shots per minute. The ammunition feed system is a hydraulically driven
single-bay, linear-linkless system capable of supplying a 181-round total system
capacity. It enables three operational employment opportunities of 60 rounds each.
Figure 14 shows the F-35A internal GAU-22/A and linear-linkless Ammuni-
tion Handling System (AHS) installation. The AHS was originally designed to
accommodate three types of rounds: PGU-20/U armor-piercing incendiary,
PGU-23/U target practice, or PGU-25/U high-explosive incendiary. Later,
more rounds were certified, including PGU-32 semi-armor-piercing, high-
explosive incendiary; PGU-47 armor-piercing explosive; and PGU-48 fragmented
armor piercing. The gun has a rotary action mechanism into which the
percussion-primed 25-mm ammunition is fed from the AHS by the linkless
carrier into the breech bolt assemblies. The rounds are chambered and fired,
and spent cases are extracted and ejected back into the AHS as the gun rotates.
The gun is reverse rotated to clear unspent rounds after burst firing. It includes
provisions for automatically arming and safing the gun on each firing burst.
Gun system boresighting is accomplished with software rather than hardware
adjustments. Tooling is used for boresighting to inspect and record the exact
aiming and alignment of the airframe installation interface points during aircraft
manufacture. Any airframe aim point corrections are recorded and stored in the
aircraft software. The gun assembly aim point and dynamic impact point of
the rounds are measured and recorded during acceptance test procedures by
the gun system manufacturer. This aim point and dynamic impact point are sup-
plied with each gun for installation into the aircraft software. The gun and attach
point correction factors are then coupled with canopy and other correction maps
through an algorithm to establish the correct aiming queues for the pilot.
206 D. M. HAYWARD ET AL.

Fig. 15 Missionized gun system installation.

The Marine Corps and Navy require the gun system to serve as AME. Under
this requirement, the gun system is only carried on an aircraft during missions
requiring its use. For use on the F-35B/C it has been designed into a Missionized
Gun System (MGS) pod. As Fig. 15 shows, the gun pod is installed on the aircraft
centerline, between the weapon bay doors. This location is limited in space
because both the inboard weapons bay doors must be allowed to fully open
when the MGS is installed on the aircraft. The lower surface of the Navy’s
MGS is also limited by the need to clear the launch shuttle used during carrier
launch of aircraft. The limited volume of the pod required the supplier to
develop a helical feed system to store the ammunition. The helical system is
double helical, surrounds the gun assembly barrels, and allows for the carriage
of 220 rounds of ammunition. The MGS is installed on the centerline store
station on the bottom of the aircraft, with two guide pins and a four-bolt interface.
A requirement was levied early in the program to make as many components
as possible common between the internal gun system and the MGS. This require-
ment resulted in common gun assemblies and gun system control units, and a
common hydraulic drive motor. The MGS has the same reverse-clearing oper-
ation of the internal gun system and uses the same technique for boresighting.
It uses a gun assembly boresight number supplied by the manufacturer and a cor-
rection number based on the MGS installation hardpoint alignment. There
remained differences in the aircraft surfaces, door contours, and airframe struc-
tures, however. As such, minor changes to the aft fairing, door bumpers, and
access panel were required for the F-35B MGS and F-35C MGS.

VII. STORES MANAGEMENT SYSTEM DESIGN

The F-35 SMS benefits from the evolution of a standardized weapon electrical
interface defined by MIL-STD-1760. This standard enables a shift in the operating
premise of the weapon electrical interface from an energy control mechanism to a
data message control mechanism. Although many of the weapon detail functions
F-35 WEAPONS DESIGN INTEGRATION 207

remain energy-driven mechanisms, the actual control circuits have become small
and inexpensive enough to implement within the weapon itself. This enables the
electrical interface between the weapon and the aircraft to become functionally
generic and flexible.
Safety interlock features are preserved via dedicated function circuits within
MIL-STD-1760 such that hazard probabilities are maintained at acceptable
levels. The aircraft/weapon interface is primarily a data message mechanism, so
most weapon delivery functions are controlled by software programs in both
the aircraft and the weapon. This has established a basis for further standardizing
that data message interface such that a weapon plug-and-play mechanism can be
defined. This is now known as the Universal Armament Interface (UAI). The F-35
will be hosting the UAI in incremental aircraft software releases as the develop-
ment program concludes and follow-on block upgrades are fielded.
Architecturally, the F-35 SMS is composed of interface units installed near the
weapon point of use and designed as slaved equipment items. The items’ serial
data interfaces to the aircraft fire control software are located in the avionic com-
puters, which constitute the Integrated Core Processor (ICP). The ICP supplies all
SMS processing resources and connectivity with other mission systems and
vehicle management systems functions, as well as appropriate redundancy for
these functions. Figure 16 shows the architectural arrangement of the SMS
components.
The SMS Fuselage Remote Interface Unit (FRIU) and Missile Remote Inter-
face Unit (MRIU) are hardware components that form an input/output structure.

Fig. 16 Stores management system architecture.

208 D. M. HAYWARD ET AL.

Through this structure, the Fire Control and Stores (FC&S) software controls the
weapons complement. The FRIUs provide hardware interfaces to control the
S&RE ejector racks and pneumatic equipment, as well as the MIL-STD-1760
circuits (except for power). The FRIUs are located in the main landing gear
wheel wells as a central location relative to the weapon stations, minimizing
the lengths of wire harnesses. The MRIUs provide hardware interfaces for
peculiar air-to-air missile signals and are installed within the LAU-151/A rail
launchers.
Weapon MIL-STD-1553 data message interface hardware is located on the
general-purpose input/output modules within the ICP. This location affords
minimum communication latency between the FC&S software threads and
smart weapons. The traditional approach is to employ a single MIL-STD-1553
bus controller for the entire aircraft weapons suite, but the F-35 SMS implements
eight controllers instead. It employs a separate controller for each air-to-surface
weapon station (six channels), and another controller for the internal and external
dedicated air-to-air weapon stations (one controller channel for four stations).
Adding to this, it employs a final controller for the internal gun and centerline
station, totaling eight controllers. This enables significantly faster bandwidth for
MIL-STD-1760 weapons.
Operational benefits that arise from the multiple-controller approach include
the rapid initialization and targeting of weapons, flexibility to support multiple
carriage subracks (e.g., the BRU-57), and an enhanced ability to simultaneously
employ multiple weapons. The approach also provides a functional isolation
benefit. Through it, the FC&S domain software that manages a store at any
station can execute independently of FC&S domain software that manages
stores at other aircraft weapon stations. This feature enables superior consistency
of weapon event timelines, as well as ease of integration.
For the electrical circuit, the F-35A/C implement a full Class 1 interface, as
defined in MIL-STD-1760, at all weapon stations. The SWAT weight reduction
exercise implemented early in the development program resulted in the high-
bandwidth signals HB1, HB2, HB3, and HB4 wiring being omitted from the
F-35B at the air-to-air missile weapon stations. It also led to omitting the 115
AC power and low-bandwidth wiring from the centerline weapon station. This
achieved a 10-lb weight savings at the aircraft level. All corresponding electronics
were preserved, and the aircraft design provided for restoring the omitted wiring.
The retained F-35B MIL-STD-1760 Class 1 circuits at these stations enable all
required weapons and a UAI capability.
Single-point safing, nonpyrotechnic S&RE, and the elimination of ground
support equipment were other F-35 requirements that significantly influenced
the SMS architecture. Single-point safing requires the entire armament system
to be capable of being safed from a single location on the aircraft. This drove
the SMS design to be implemented such that an absence of electrical energy
would safe the weapon articles (bomb racks, missile launchers, and gun). This
design aspect affected equipment articles as well, compared to those of legacy
F-35 WEAPONS DESIGN INTEGRATION 209

equipment. Specifically, the new design prohibited the use of REMOVE BEFORE
FLIGHT pins, levers, or mechanisms at the weapon stations. Therefore, the F-35
equipment was designed to require specific activation energy to solenoid mechan-
isms in the bomb racks, missile launchers, and gun from the SMS. This was
needed to advance from a safe state to an employment state.
To satisfy the nonpyrotechnic requirement for S&RE, the bomb racks in the
F-35 are operated via pneumatic solenoid mechanisms. The SMS electrical inter-
faces for those solenoids are tailored specifically for solenoid operation. They
incorporate tickle-test features that enable periodic confirmations of end-to-end
circuit connectivity.
Another primary requirement for the F-35 air system was the elimination of
most ground support equipment. This presented a challenge to the SMS design
because most weapon interface circuits are open-ended wires at the weapon
stations, with aircraft electronics attached to the internal end. Further, there is a
companion F-35 SMS requirement that the aircraft be capable of validating the
integrity of the armament system without support equipment. During discussions
with the F-35 JPO, we observed that the open-ended aircraft wires required by
MIL-STD-1760 presented a validation impossibility without some form of aircraft
support equipment. Our solution for the F-35 was to incorporate a voltmeter
capability into selected circuits of the MIL-STD-1760 signal set and wrap
adaptor support equipment. These wrap adaptors are small, low-complexity
items with internal passive components that wrap the voltages from circuits in
the MIL-STD-1760 signal set back to the circuit with the voltmeter capability.
This design approach confirms the integrity in the aircraft wiring and, to a
lesser degree, the functionality in the MIL-STD-1760 signal electronics. Over
the life of an aircraft, a primary root cause of downtime is wiring issues, such
as open circuits and shorts to structure. The F-35 weapons wrap test is designed
to detect and isolate these problems.

VIII. FUTURE WEAPONS GROWTH

The F-35 weapons systems will be operational for many years and have been
designed to evolve to counter ever-changing threats. The internal bays have
been designed to maintain the maximum volume available for future weapons
growth, and can accommodate the integration of new stores. Maximum bay
volume is important because the F-35 will be the primary internal weapon carriage
attack aircraft in any conflict. As such, future weapons will be largely designed to
be compatible with the F-35 weapon bays. Only 500-lb-class air-to-ground
weapons were certified externally during SDD. However, the F-35 and its requisite
AME are structurally designed to accommodate up to 2000-lb-class munitions
and external fuel tanks. This includes dual carriage of 1000-lb munitions and
single and dual carriage of external air-to-air weapons to augment this flexibility.
Figure 17 illustrates a potential maximum air-to-air loadout.
210 D. M. HAYWARD ET AL.

MIL-STD-1760 compatibility at all stations provides the electrical and logical

interface for the current and future stores that implement it. For legacy gravity
weapons and their carriage systems, the design accommodates a pylon remote
interface unit that converts applicable MIL-STD-1760 signals to analog discretes
recognized by legacy carriage systems. External air-to-air carriage is accomplished
by integrating external missile adapters onto the existing F-35-unique air-to-
ground pylons and rail launchers. The external missile adapters provide for
single or dual missile carriage of existing or future missiles.
To further simplify future weapon incorporation, the F-35 is incorporating a
UAI: a functional and logical interface definition between platforms and stores. It
is intended for decoupling the integration of new stores (weapons, carriage
systems, and sensors) on platforms (aircraft, ships, ground vehicles) from plat-
form Operational Flight Program (OFP) update cycles. The UAI standard is
being certified in SDD through laboratory integration with the UAI Certification
Tool. The SDB II will be the first weapon to be integrated using a UAI in the
post-SDD Follow-on Modernization (FoM) program. Figure 18 illustrates other
FoM weapons being considered for future integration.
The UAI is platform-agnostic and independent of program specifics; however,
F-35 personnel hold key positions in several UAI teams: Systems Engineering,
Platform Store, Mission Planning, and Launch Acceptability Region. The teams’
F-35 insight into the interfaces and processes they developed will enable efficient
and effective integration as new weapons are brought into the program. To further
enhance future growth capabilities, we assessed current and future developmental
weapons that could enhance weapon system effectiveness. Additional volume
reserves were identified, and we are modifying the weapons bays to accommodate
the new weapon volume, where required.

Fig. 17 Potential maximum air-to-air weapon loadout.

F-35 WEAPONS DESIGN INTEGRATION 211

Fig. 18 Candidate follow-on weapons.

IX. CONCLUSION
F-35 weapon carriage systems have been designed and validated through a suc-
cessful SDD certification program. They provide tactical commanders with the
flexibility to prosecute a wide array of targets in either a stealth mode with
internal weapons or using an enhanced carriage configuration with external
weapons when the threat level allows it. We designed the internal weapon
bays for maximum flexibility for weapons and aircraft systems integration to
limit aircraft performance impacts across the variants. In addition, we incorpor-
ated loading access into the design with the inclusion of an onboard hoist
system. Supporting this, we mounted air-to-air missiles on the weapon bay
door mechanism, thereby opening up maintainer access to load the weapons.
We also included robust external weapons provisions to enable the carriage
of a wide variety of weapons on seven hardpoints. Internal and missionized
gun systems, tailored to service needs, provide lethal target engagements for
close combat with inherent air-to-air capability. Nonpyrotechnic bomb racks
and missile ejectors provide consistent separation performance over a wide
range of temperatures, providing significant LCC savings over legacy pyrotech-
nic systems. The adaptable SMS we designed provides the conditioning and
control of all required weapons. Further, its flexibility accommodates the devel-
opment of a UAI to easily integrate future weapons. As F-35 weapons
212 D. M. HAYWARD ET AL.

development and integration continue, growth provisions will enable future

smart weapon certification to respond to evolving threats and necessary
weapon/target pairings.

ACKNOWLEDGMENTS
Many members of the Weapons Integration, Graphics, and Multi-Media teams
provided input for this paper to capture the history, design evolution, and
lessons learned. Douglas Hayward and Andrew Duff would like to thank the Lock-
heed Martin team of Jubal Arey, Pete Paciorek, Bill Whitener, Keith Spalding,
Rick Wild, Greg Cochran, Keith Schulz, Daniel Buck, Dennis Soultaire, David
Gilland, and Russel Baldonado for their support in documenting the evolution
of F-35 weapons.

REFERENCES
[1] Hetreed, C., Carroll, M., Collard, J., and Snyder, R., “F-35 Weapons Separation Test
and Verification,” 2018 AIAA Aviation Forum, Lockheed Martin, Fort Worth, Texas,
2018 (submitted for publication).
[2] Dix, R. E., and Bauer, R. C., “Theoretical Study of Cavity Acoustics,” Arnold
Engineering Development Complex, AEDC-TR-99-4, 2000.
[3] Rhodes, L. R., “Weapon Carriage Technology (WCT), Life Cycle Cost Study Phase II
for Wright Laboratory/MNAV,” F-35 Joint Program Office, Oct. 1996.
 CHAPTER 6

F-35 System Development and

Demonstration Flight Testing at
Edwards Air Force Base and Naval
Air Station Patuxent River
Mary L. Hudson and Michael L. Glass†
Lockheed Martin Aeronautics Company, Edwards Air Force Base, CA

Lt Col Tucker “Cinco” Hamilton‡

United States Air Force, Edwards Air Force Base, CA

C. Eric Somers§ and Robert C. Caldwell}

Lockheed Martin Aeronautics Company, Naval Air Station Patuxent River, MD

Since the deliveries of the first F-35s to Naval Air Station Patuxent River
(PAX) in late 2009 and Edwards Air Force Base (EDW) in mid-2010,
more than 9000 flights, 16,000 flight hours, and 65,000 test points have
been completed. This huge feat utilized 18 flight test aircraft to
demonstrate the flight sciences and mission systems capability of all
three F-35 variants. An Integrated Test Force team composed of
government, military, and contractors orchestrated test ranges, tankers,
chase aircraft, flight clearances, run cards, parts, support equipment,
control rooms, and aircraft configuration. Extensive flight test
instrumentation collected the required data to verify modeling and
simulation and support the certification of F-35 capabilities. This paper
describes the many aspects of flight test successes, challenges, and
lessons learned at EDW and PAX to position future F-35 flight testing
for continued success and improved efficiency.

I. INTRODUCTION
The F-35 is the first-ever stealth dual-role fighter designed and built to conduct all
fighter mission sets, both air-to-air and air-to-ground: Offensive Counter-Air
 Lockheed Martin Sr. Manager, F-35 Test Operations.
†
Lockheed Martin Director, F-35 Test Operations.
‡
Commander, 461 FLTS, AIAA Sr. Member.
§
Lockheed Martin Sr. Manager, F-35 Test Operations.
}
Lockheed Martin Director, F-35 Test Operations, AIAA Sr. Member.

213
214 M. L. HUDSON ET AL.

(OCA), Defensive Counter-Air (DCA), Close Air Support (CAS), air interdiction,
maritime interdiction, and Suppression of Enemy Air Defenses (SEAD). To
conduct these missions, the F-35 needed to be able to penetrate enemy air defenses
virtually undetected. Additionally, the F-35 actually comprises three distinct air-
craft: the F-35A Conventional Takeoff and Landing (CTOL) variant; the F-35B
Short Takeoff and Vertical Landing (STOVL) variant, which can land on aircraft
carriers previously occupied by AV-8B Harriers; and the F-35C Carrier Variant
(CV) for use on large aircraft carrier ships currently occupied by F/A-18
Hornet aircraft. In order for the F-35 to conduct these mission sets, the airframe
and mission systems imbedded in the airframe needed to be developed, integrated,
and proven. Each variant required rigorous testing to ensure that it could safely fly,
maneuver, and land with all types of stores. This scope of flight test is defined as
flight sciences. In addition, the aircraft needed to prove it could conduct each of
these missions against the most advanced threats and in concert with other
friendly battlespace platforms. This scope of flight test is defined as mission
systems testing. The first of each flight test variant is shown in Fig. 1.
To accomplish delivering this 5th-Generation capability, a flight test fleet was
established coast to coast, as shown in Fig. 2. At Edwards Air Force Base (EDW)
and Naval Air Station Patuxent River (PAX), the Integrated Test Force (ITF)
teams tested all three F-35 variants with one goal: to deliver unrivaled capabilities
to the warfighter. Flight testing at EDW and PAX provided data to multiple dis-
ciplines for both model verification and demonstration [1– 11]. From flight
sciences to mission systems testing, the critical work of the F-35 test teams
cleared the way for capabilities to be delivered to the training and operational
F-35 fleets. Flight testing obtained the data required to verify the F-35s’ key per-
formance capabilities and validated extensive modeling and simulation work. This
supporting evidence was required in the packages delivered with the F-35 aircraft
to define and authorize warfighter capability. EDW and PAX ITF teams success-
fully executed the ground and flight tests required to deliver U.S. Marine Corps

Fig. 1 Firsts of each flight test variant: AF-1, BF-1, and CF-1.
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 215

Fig. 2 F-35 Integrated Test Force coast to coast.

Initial Operational Capability (IOC) with Block 2B software in 2015 and U.S. Air
Force IOC with Block 3i software in 2016. They also completed flights supporting
the end of the System Development and Demonstration (SDD) phase in 2018 and
delivered U.S. Navy IOC in 2019 with Block 3F software. Three IOCs in this
period of time are unprecedented for any weapons system.
Eighteen F-35 aircraft at EDW and PAX have tested flight sciences and mis-
sions systems capabilities since 2009. Over a period of extensive replan, with rede-
fined engineering requirements and revised flight test metrics, nine aircraft at each
test site were determined as the right-sized flight test fleet to complete flight test for
the SDD phase of the F-35 program. In addition to the number of flight test aircraft,
the team to maintain the aircraft and execute the flights had to be the right size and
have the right composition. The F-35 ITF team included military, government, and
contractor personnel working as one team to plan and execute flights. Every
employee on the ITF team had a unique function that utilized his or her specific
skillsets. The ITF conducted flight test missions every day, and the criticality of
the work, coupled with pushing the limits, required performance at the highest
level from the entire team. This realization, along with a seemingly overly con-
strained task, created a very dynamic and challenging work environment. Every
day, the team had to be flexible and adaptable to deal with changes in range avail-
ability, cancellation of tankers or test assets, weather, engineering discoveries, and a
host of other constraints that often made it seem a miracle to execute the mission.
The F-35 Joint Program Office (JPO) was also an integral part of delivering
F-35 capability and interacted daily with each test site. The F-35 JPO had an
216 M. L. HUDSON ET AL.

impact on the day-to-day operations and overall processes that dictated flight test.
Through this unique test structure, the team discovered some key lessons:
. The F-35 JPO needed to identify one conduit to prioritize tasks to the test sites.
. The F-35 JPO needed to understand and respect the autonomy of the test sites
and not dictate changes too close to execution.
. Clear lines of responsibility between the F-35 JPO and test sites needed to be
defined and a process put into place that checked that balance.

Given 18 flight test aircraft stationed at two test sites, nearly 1000 employees at each
test site, and all the possible flight test constraints, the F-35 flight test team accom-
plished more than 9000 flights, 16,000 flight test hours, and 65,000 test points.
Although flight test metrics are not the end goal, they lead to the ultimate result
of demonstrated and delivered warfighter capability. The following assessment
from Exercise Northern Edge/Distant Frontier 2017 best exemplifies the flight
test effort and resultant capability that each ITF team member contributed to F-35.

During NE [Northern Edge Alaska], the F-35 repeatedly demonstrated the

capability to use its high fidelity sensors and Link-16 capabilities to signifi-
cantly increase the situational awareness of 4th Generation assets. This
support was a major force multiplier and enabled 4th Generation assets to
be more lethal and survivable. In many cases, the F-35B was the last remain-
ing aircraft in the fight (often even after it’s simulated A/A or A/S [air-to-air
or air-to-surface] load out was expended). F-35s were able to remain on
station to support other assets non-kinetically and still perform in the afore-
mentioned “quarterback” role. . . . Regardless of ordnance load out, the F-35 is
a key node in an A2/AD [any access/area denial or contested] environment
and its mere presence on station makes the entire package more lethal and
survivable. [12]

Over the eight years of flight tests at EDW and PAX, the following key flight test
capabilities were demonstrated and summed toward the delivered capability to the
warfighter:
. High Angle-of-Attack (AoA) missions
. Aerial refueling qualification with KC-10; KC-130; KC-135; F/A-18; and U.K.,
Australian, and Italian tankers
. High sink rate and arrested landings
. Flutter/loads/buffet flights
. Numerous vehicle systems and mission systems software release
regression flights
. Ground and airborne gun fire on all three variants
. Weapons environment, store separation, and Weapons Delivery Accuracy
(WDA)
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 217

Fig. 3 F-35 testing at EDW.

. Shore-based ship integration with jet blast deflector, catapult, arrestment gear,
and ski jump
. F-35B deployments onboard USS Wasp (LHD 1) and USS America (LHA 6)
. F-35C deployments onboard USS Nimitz (CVN 68), USS Dwight
D. Eisenhower (CVN 69), and USS George Washington (CVN 73)
. Climatic laboratory testing in the McKinley Climatic Laboratory at Eglin Air
Force Base (AFB), Florida
These and other major test events are shown in a timeline depiction at EDW and
PAX in Figs. 3 and 4, respectively. The EDW and PAX test sites have unique
capabilities allowing unique testing and accomplishments. The data obtained at
each site were required together to certify warfighter capability on each F-35
variant. This was truly an integrated test force with not only cross-teams, but
also cross-country data gathering and analyses from California to Maryland
and Texas and in between. The expanse of aircraft, flights, data, and people led
to the F-35 ITF’s selection as a finalist for the 2017 Collier Trophy.

II. EDWARDS AIR FORCE BASE

EDW has an incredible history of flight test firsts, and the F-35 is proud to join this
history of engineering innovation demonstration. EDW is a unique flight test
location in the Mojave Desert with excellent year-round weather and dry lake
beds providing a natural extension to the base’s runways (Fig. 5).
Nine F-35 flight test aircraft resided at EDW to support SDD flight testing.
Three CTOL aircraft were dedicated to flight sciences testing: AF-1 to obtain
218 M. L. HUDSON ET AL.

Fig. 4 F-35 testing at PAX.

store separation data, AF-2 to obtain loads data by expanding the envelope, and
AF-4 to understand the complex fuel system. Six aircraft, using all three variants,
comprised the tails supporting mission systems flights: AF-3, AF-6, AF-7, BF-17,
BF-18, and CF-8. Each tail had some claim to a flight test first; for example, AF-1

Fig. 5 Aerial view of F-35 AF-1 over EDW.

F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 219

Fig. 6 EDW ITF team with the nine flight test aircraft.

was the first F-35 aircraft to arrive at EDW to begin SDD flight testing on 19 May
2010. AF-2, aka Work Horse, holds the record for flights and flight hours at 689
and 1509, respectively (as of December 2017), accomplishing loads, flying qual-
ities, buffet, dry and wet runway directional control and antiskid braking, air
data tower fly-bys, propulsion, software regressions, and gunfire. AF-3 was the
first to conduct countermeasures testing. The mission systems aircraft were the
first of their variants to employ weapons using their complete avionics suites.
In addition to firsts, these nine F-35 aircraft accomplished more than 4500
flights, 10,000 flight test hours, and 30,500 flight test points.
The people were the key to orchestrating and executing the incredible number
of flights, supporting the nine F-35 flight test aircraft, and delivering
5th-Generation warfighting capability. The nine flight test aircraft and ITF
team at EDW are shown in Fig. 6. At the peak of F-35 SDD flight testing at
EDW, there were nearly 1000 men and women in the ITF made up of government
civilians, military service members, and contractors (Lockheed Martin, Northrop
Grumman, BAE Systems, Raytheon, Boeing, and Pratt & Whitney). The F-35 test
pilots came from the military, U.S. government contractors, and Lockheed
Martin. The ITF organization evolved over time to improve team communication
and test efficiency. The ITF was organized into Flight Sciences Engineering,
Mission Systems Engineering, Maintenance, Sustainment/Logistics, Test Oper-
ations, Instrumentation/Data Processing, and Test Pilots. The successes, chal-
lenges, and lessons learned of each team are summarized in the following sections.

A. FLIGHT SCIENCES ENGINEERING

Flight sciences engineers test the aircraft’s critical systems, analyze data, and man
the control rooms advising test pilots when they reach the edge of the envelope or
providing situational awareness of system functionality. Engineering works in
conjunction with the Test Operations team to develop the Joint Test Plans
(JTPs) that define the test objectives, test plan, and required flight test points.
There are JTPs and engineers specializing in the aircraft fuel system, loads and
220 M. L. HUDSON ET AL.

Fig. 7 Loads testing with different weapon loadout configurations.

dynamics, flying qualities, landing gear, hydraulics, electrical power system, pro-
pulsion, mass properties, store separation, and low observables.
Flight sciences envelope expansion testing commenced in summer 2010.
During this phase, the aircraft performed maneuvers under a variety of configur-
ations to experience aerodynamic forces to test the limits of the design envelope.
Testing factored in environmental conditions, such as temperature, pressure, alti-
tude, and airflow. The aircraft were subjected to high-speed/high-g-load con-
ditions and flutter and buffet to characterize structural responses, as shown in
Fig. 7.
To validate the aircraft’s flight control laws and handling qualities, high-AoA
testing was performed, as shown in Fig. 8. The tests characterized the aircraft’s
performance beyond the operational envelope and thus demonstrated the
ability to recover from uncontrolled flight.
The aircraft’s vehicle subsystems were also tested to their limits in both ground
and flight scenarios. Some of the critical tests included inflight aerial refueling
(Fig. 9), inflight engine restart, high sink rate landing, crosswind landing, high brake
energy, wet runway performance, arresting hook, and ground gunfire (Fig. 10).
A key lesson learned from these teams was test planning; for example, several
of the most difficult-to-acquire test points remaining at the closure of SDD were
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 221

Fig. 8 High angle of attack: aircraft flies at various uncontrolled flight conditions.

Fig. 9 Aerial refueling tests from different tankers, clockwise from upper-left: nighttime
refueling using a KC-10 tanker, F-35 refueling from an Italian KC-767, and refueling from an
Australian KC-30.
222 M. L. HUDSON ET AL.

Fig. 10 Ground testing on the F-35, clockwise from upper-left: arresting hook, wet runway
performance, and ground gunfire test.

those for flying qualities crosswind testing. This testing can be especially difficult
because of the limited number of days on which the atmospheric conditions meet
the test objectives. A handful of test points could take months to complete because
of uncooperative weather, namely wind speed and direction. While waiting for the
weather, it became apparent that overly constrained test plans and tight airfield
restrictions had driven this predicament. The test cards could have been prepared
and carried to every sortie to catch the rare opportunity for these tests. The test
plan also specified tight tolerances on weights and Center of Gravity (CG) that
significantly decreased test efficiency. In the first attempts to capture test
points, the weight band was burned through so quickly that aerial refueling was
often needed. Upon further data review, the weight and CG conditions were
deemed irrelevant for this testing. These changes to the JTP were doable but
incurred schedule impact due to the lengthy change process and approvals.
Updating the test plans was a common occurrence because the testing require-
ments changed as the system matured and became more understood.
There were discoveries on a near-daily basis: exceedances, hardware issues and
software issues that led to new constraints limiting testing, analysis that reprior-
itized the schedule, or new understandings of how the systems operate that were
reason to pause and reevaluate planned points. A lesson learned was to review the
entire test point requirements to most efficiently missionize test points and to
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 223

write a test plan that allows as much flexibility and agility as possible early in
the program.

B. MISSION SYSTEMS ENGINEERING

Mission systems discipline engineers verify and validate the mission systems capa-
bilities, such as communication, navigation, and interrogation; datalinks and
interoperability; displays and core processing; weapons integration and accuracy;
electro-optics and infrared; offboard mission planning; radar; fusion; electronic
warfare and countermeasures; and signature and survivability. They document
anomalies and work with the software engineers and teams on root-cause analysis
and resolution. The mission systems engineer also works directly with the main-
tenance and control engineers to troubleshoot and resolve aircraft anomalies.
Here are a few lessons learned from the Mission Systems engineers.

1. SOFTWARE RELEASES
Inflight discoveries, related to verifying millions of lines of code enabling F-35
mission systems capability, triggered numerous software releases. A flexible,
agile flight clearance process was required to support a fly-fix-fly concept for
revised software. The Quick Reaction Capability (QRC) flight clearance process
was developed to address this issue and allow flight clearance turnaround in a
couple days vs several weeks for a full release. If the software revision did not
touch any airworthiness software, a board reviewed and approved a QRC flight
clearance that enabled the fly-fix-fly concept and greatly improved flight test
efficiency while maintaining safety.

2. WEAPONS SURGE
A new test concept was used for complex weapons testing called surge due to the
numerous assets and personnel required for a short period of time. The F-35 flight
testing showed that for complex testing, such as WDA flights, maximum test effi-
ciency was achieved with a focused surge approach including exclusive use of
ranges, dedication of tankers, secure communication and datalinks for control
rooms, chase and photography, targets, off-site staffing, and a seven-day team
workweek. During Block 3F, there were two such surge events that directly sup-
ported the Air Force declaring IOC and the closeout of SDD. During the first
surge event in August 2016, 25 missions, composed of 12 WDAs and 13
weapon separation tests, were executed during a month-long period. This
required laser focus by the entire test team to ensure that the aircraft were avail-
able every day and to maintain priority for use of the range and support assets.
Historically, the execution of such events would happen once a month given
the extensive coordination required. Prior to the surge, the highest number of
weapons events accomplished in a single month was three in November 2014
during the Block 2B software testing. During the unprecedented period in
224 M. L. HUDSON ET AL.

August 2016, a total of 30 weapons were dropped or fired in 31 days, including the
Joint Direct Attack Munition (JDAM), Advanced Medium-Range Air-to-Air
Missile (AMRAAM), Small Diameter Bomb (SDB), AIM-9X Sidewinder air-to-air
heat-seeking missile, and GPS/laser-guided munition.
The second surge event occurred in August 2017. Unlike the prior surge event
in 2016, there were only a handful of missions required to complete the F-35
developmental test for Block 3F. These final missions demonstrated the
complex air-to-air capabilities of the F-35 with the AMRAAM, showing the full
capability of the F-35. The AMRAAM missiles used during these test missions
were equipped with live motors and guidance systems, but the warheads were
exchanged for telemetry units. The open-air testing was conducted over the
water at Naval Air Station Point Mugu’s range in Ventura County, California
(Fig. 11). The WDA testing was the graduation exercise before the aircraft was
delivered to the operational test organizations to prove combat readiness.

3. MISSION EFFECTIVENESS
The final and most complicated demonstration of the F-35 5th-Generation capa-
bility was the mission effectiveness, multiship missions (Fig. 12). These missions
required maximum coordination for successful execution. As many as 15 airborne
assets participated in a single test mission including F-35s, F-16s, F/A-18s, F-15s,
KC-10 tankers, KC-135 tankers, and airborne command and control platforms.
The various mission types executed during Block 3F included CAS, SEAD,
Destruction of Enemy Air Defenses (DEAD), OCA, and DCA. The mission effec-
tiveness missions highlighted the interoperability of the F-35 and its ability to
command the technical situation. This ability was further demonstrated in
several Large-Force Exercises (LFEs) by providing command and control situa-
tional awareness of the battlefield using the various arrays of onboard sensors.

Fig. 11 Weapons delivery accuracy flights.

F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 225

Fig. 12 Four-ship mission effectiveness flight.

C. AIRCRAFT CONFIGURATION AND MAINTENANCE

Maintainers and Flight Test Control Engineers (FTCEs) repair, inspect, configure,
and modify the aircraft to ensure safe and correct functioning during every flight.
Aircraft maintenance was performed using the tail-team concept composed of an
aircraft supervisor, FTCE, quality assurance inspector, field and service mech-
anics, and avionic technicians. The FTCE ensured that the aircraft was properly
configured to perform a flight test mission by providing a liaison between engin-
eering and maintenance. They used many tools, including the Autonomic Logis-
tics Information System (ALIS) and Product Data Management (PDM) and
worked with maintainers to ensure that parts that had been replaced on the air-
craft were within the proper aircraft effectivity. The team reviewed the air vehicle’s
status and developed the near-term maintenance requirements during a daily
maintenance scheduling/coordination meeting at the beginning of each shift.
This meeting set the tone for the day and provided critical aircraft status infor-
mation to ITF leadership. Below are some details of lessons learned regarding air-
craft configuration, maintenance procedures, inspections, and ALIS.

1. AIRCRAFT CONFIGURATION
The required aircraft configuration for each flight was defined by the Mission Prep
Sheet provided by the Flight Test Engineer (FTE) and specified the software ver-
sions, weapons loadout, instrumentation configurations/settings, and mass prop-
erties Form F defining the alternate mission equipment. The FTCE was the focal
226 M. L. HUDSON ET AL.

point to coordinate with all involved teams to ensure that the aircraft was config-
ured per the Mission Prep Sheet requirements prior to releasing the aircraft
for flight.
Perhaps the most valuable and accurate tool to ensure proper aircraft con-
figuration and tail effectivity of parts was PDM. For example, if a -0007 part
was issued to an aircraft, it was the FTCE’s responsibility to ensure that the
part was applicable to that aircraft. After checking in PDM, if the -0007 part
number was not listed, the FTCE reverted to the previous dash number (-0006)
and checked whether a service data note was present stating that a -0007 could
be used for all replacements of a -0006. Then, the -0007 was authorized to be
installed on the aircraft. The Configuration Management team ran a Release
Authorization Notice report before every flight to verify that the aircraft was in
an authorized configuration.

2. MAINTENANCE PROCEDURES AND INSPECTIONS

The ability to properly maintain a new aircraft from Day One is important and
requires correct technical data to exist early on. Joint Technical Data (JTD)
define maintenance procedures and must be verified and validated before use.
Dedicated aircraft were intended to provide early verification of JTD but were
not available early in the program. This required continued use of Aeronautical
Engineering Instructions (AEIs), which are the unverified source data for JTD,
to accomplish tasks on the aircraft in a timely manner. This delayed the use of
some JTD to operational test and fielded units, requiring them to use field
service representatives with instructions in the form of an action request that
could delay an answer.
Another challenge early on was well-documented and communicated inspec-
tion requirements that are defined in Engineering Inspection Requirements (EIRs)
for SDD aircraft and in Production Aircraft Inspection Requirements (PAIRs) for
Low-Rate Initial Production (LRIP) aircraft. Both EIRs and PAIRs are designed to
provide inspection criteria and timelines for parts that may have been identified
with unknown life limits, manufacturing, or flight failure/damage issues seen in
testing or fleet operations. In addition, Time Compliance Technical Directives
(TCTDs) specify inspections or parts replacements/modifications. TCTDs are
released in JTD dataset releases and managed and installed by ALIS administra-
tors. The FTCE then performs the defined action based on aircraft scheduling/
availability and time compliance definition. It is critical that an accurate and
reliable tool exist to track all the inspection requirements.

3. AUTONOMIC LOGISTICS INFORMATION SYSTEM

The F-35 is the first tactical aviation system to have sustainment tools engineered
in concert with the aircraft for efficiency and cost effectiveness. Compared to
previous aircraft, a higher fidelity of information about the F-35 fleet is tracked
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 227

Fig. 13 ALIS: integrated application suite supporting multiple air system functions.

within ALIS to reduce operations and maintenance costs and increase aircraft
availability. ALIS consists of the system, application, and network infrastructures
required to provide global integrated and autonomic support, as illustrated in
Fig. 13. ALIS integrates a broad range of capabilities, including operations, main-
tenance, prognostics, supply chain, customer support services, training, and tech-
nical data. It gives F-35 operators the ability to plan ahead and to maintain and
sustain its systems over the life of the air vehicle. A single, secure information
environment provides users with up-to-date information on any of these areas
using web-enabled applications on a distributed network.
Just like the airframe, ALIS needed time to mature in order to reach full capa-
bility during SDD. The ITF used ALIS as a tool to manage aircraft maintenance
while concurrently evaluating it and identifying issues and improvements. This
dual task made it challenging to use a development tool for reliable aircraft main-
tenance. The FTCE or support FTCE used ALIS to check whether a part had been
replaced per a PAIR, and whether a TCTD was valid for the assigned aircraft.
Because of downtime during ALIS development, backup techniques were devel-
oped, including a paper system with which all the data were entered into ALIS
when it was back up and running. A lesson learned with developing ALIS was
to have users involved with developers from the beginning. Technicians could
tell the developers what they needed to effectively accomplish their job, leading
to an improved initial concept.
The Sustainment team at the flight test site used a specific module of ALIS to
manage the location’s inventory in support of aircraft maintenance. This team
228 M. L. HUDSON ET AL.

received, stored, issued, and moved retrograded F-35 assets in direct support of
flight operations, back shop, and flight line maintenance. ALIS periodically received
software upgrades to fix anomalies and add capabilities, so additional tools and pro-
cesses were developed to keep parts flowing and aircraft flying. To adjust spares
levels based on requirements, assets from the EDW ITF needed to be reallocated.
To manage these limited assets, EDW developed a tracking tool for top degraders
and a series of metrics to provide a snapshot of wellness in the following areas: top
degraders (batteries, wheels, cartridge/propellant-activated devices), target stock
levels, inventory effectiveness, historical usage comparison, support equipment
analysis, tool room analysis, and a warehouse effectiveness summary. The use
and feedback of ALIS performance in a flight test setting was critical to its develop-
ment, resulting in a superior fielded product.

D. SUSTAINMENT/LOGISTICS
Parts and spares were an ongoing critical requirement and issue to support the
pace of SDD flight testing. The Sustainment/Logistics team managed the Material
Master Catalog, spares/repairable parts, sustainment data, transportation inven-
tory, property, and shelf life. They were responsible for initial provisioning as
well as demand forecasting. This team assisted with issues related to purchase
orders, consignments, asset transfers, inventory audit and management, electronic
equipment logbooks, and action requests; ensured that material was processed
promptly to avoid potential work stoppages; clarified original equipment manu-
facturer/supplier data to clear frustrated cargo; and served as the liaison
between the F-35 Maintenance team and the Sustainment Modification Kit Man-
agement team in Fort Worth, Texas, for supply-related issues. ALIS was also an
integral part of the sustainment supply and logistics concept for supporting
F-35 operations as described previously.

1. WHEEL AND TIRE MITIGATION PLAN

For flight test, tire tread limits are stringent to support test execution, causing
Wheel and Tire (W&T) assemblies to be changed more often than in a standard
squadron. The F-35 initial concept was to accomplish the buildup at the supplier
facilities, which took weeks to complete a refurbish turnaround. To meet the
demand, the Sustainment team developed a plan to train and certify the local
tire shop to have the ability to refurbish F-35 W&T assemblies. With this plan
in place, the refurbish turnaround was reduced from weeks to three to five days.

2. TOOL CONTROL
The Sustainment/Logistics team was also responsible for tool control. There are
strict processes for Foreign Object Damage (FOD) and tool control, but a process
and comprehensive training were required for visiting mod teams and vendors.
Five documents were combined into one all-encompassing standard operating
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 229

instruction for visiting mod teams and vendors to review and acknowledge prior
to being released to work on the aircraft. Additionally, all tooling was processed
through the tool room and went through a process of accountability that included
quality team buy-in prior to assets being released to the flight line or hangar floor.
The new single-document direction and comprehensive training effectively
emphasized the importance of FOD and tool control and highlighted unique
aspects of working at the test site. FOD and tool control are critical concerns of
flight test safety and can cause not only safety issues, but also program delays if
not given due attention.

E. TEST OPERATIONS
Test Operations consisted of a group of FTEs and schedulers who converted data
requirements into test missions and executed these missions to obtain the
required data. Test procedures and run cards were developed for the pilots to
execute based on requirements in the JTPs. FTEs were key players in developing
the JTPs and were skilled in recognizing what-ifs and requirements. They were a
vital link between the pilots and engineering and were the gatekeepers from test
planning through actual execution to ensure safe, smart, and efficient testing.
To execute a flight test mission, FTEs must be aware of limitations defined in
multiple sources: test safety package, flight clearance, system release memo, and
aircraft operating limitations. They must coordinate physical assets, including:
test aircraft configuration, mission control room, instrumentation, telemetry,
radio frequencies, airspace, test ranges, tanker, chase/target aircraft, weapons,
ground vehicle targets, radar reflectors, terrain types, air temperature, wind
speeds, day/night, and moon condition. To orchestrate these many moving
parts, some key factors were identified to safely, successfully, and efficiently
execute daily F-35 flight tests.

1. TEAMWORK AND TRAINING

Development work and daily flight test execution are challenging and difficult by
nature and are best accomplished by a dedicated group of people who work
together, respect each other, have a solid work ethic, and care about the work
and each other. Teamwork has been shown in the form of knowledge, skill, and
trust. The Test Operations team gained knowledge through detailed training; suc-
cessful training on a program of this size required a lead instruction coordinator to
oversee the effort. Per the EDW Lockheed Martin Chief Test Pilot David “Doc”
Nelson: “The person training FTEs, test conductors, and test directors should
be a leading world class expert in the role they are instructing. There is no
amount of study or theory that can replace the knowledge and lessons learned
by successful and especially unsuccessful experience” [13].
FTEs were trained to be active leaders in premission preparation including
card writing, maneuver development, mission planning, and execution. Training
230 M. L. HUDSON ET AL.

was accomplished on not only the aircraft systems, but also the basics of flight
test operations as well as the unique nature of testing the F-35. The control
room training program worked well in establishing a capable group of Test
Conductors (TCs), Test Directors (TDs), and discipline engineers. The program
was successful in taking individuals with no control room experience and pre-
paring them for control room duty (Fig. 14) through emergency procedure
simulations and mentorship and guidance from instructors. Their skills were
proven through real test missions to the point that trust was gained. The trust
was then knowing that, no matter the challenge, the team would work seam-
lessly together to accomplish goals that could not be achieved alone. Some
lessons learned were to thoroughly train the team, empower them, listen to them,
and trust them.
Despite being part of an excellent team, personnel attrition rates occurred with
one cause being the stress and workload. There were averages of six sorties a day of
testing of all types and disciplines. Maintenance personnel often worked week-
ends, and overtime was common. The departure of trained, independent
workers resulted in a loss in the organization’s knowledge base, which required
veteran personnel to focus on training new personnel rather than test planning
and execution. Future programs should determine what workload is sustainable
for the number of personnel over the course of the entire program and may con-
sider offering incentives (e.g., financial, promotions) early in the program to help
create and sustain a large group of veteran personnel.

Fig. 14 EDW control room.

F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 231

2. ROLES AND RESPONSIBILITIES

With the large number of F-35 personnel with multiple affiliations (military, gov-
ernment, and contractors), it was especially important to define roles and respon-
sibilities. Lacking this definition can lead to duplicated work, missed tasks, and
miscommunication among groups, which leads to increased workload and per-
sonnel conflict. The Test Operations and the discipline engineering groups experi-
enced some conflicts based on their differently perceived roles and responsibilities.
With high pressure to succeed, it was tempting for personnel to develop an atti-
tude of doing whatever it takes to get results, which disrupts the organization’s
leadership and schedule stability. Future programs of similar size and of a joint
organizational nature need to define and establish the roles and responsibilities
of their groups early on to avoid conflict and the impact on testing.

3. COMMUNICATION
Good communication was an essential key to success, both internally within the ITF
and externally at other test sites, Fort Worth, JPO, vendors, and foreign partners.
This is especially critical during large developmental test programs for which any
printed schedule is soon inaccurate due to the dynamic nature of the work. Two
lessons learned were to make the effort to communicate more and to communicate
effectively. Communication success resulted from the core communication
elements of speaking and listening to each other and using clear, concise, and
direct language. This proved invaluable, especially in communications with
ranges and asset providers to understand how their organizations worked. Lines
of communication must also be established among FTEs, relevant engineering
groups, and leadership. As discoveries are made, solutions must be evaluated and
decided upon. A designated integrator/decision maker who has the big picture
and required authority will be key to moving forward with minimal test delay.

4. TEST REQUIREMENTS
Each F-35 variant had unique test requirements that presented flight test chal-
lenges. This was most evident in the flight sciences engineering disciplines due
to the differing wing planforms between CTOL/STOVL and CV, and the different
weapon bay sizes between CTOL/CV and STOVL. Also, each variant had unique
fuel system and structural designs. All CTOL flight sciences testing was performed
at EDW, and all STOVL and CV flight sciences testing was done at PAX.
The majority of mission systems testing was variant agnostic. F-35A, F-35B,
and F-35C aircraft participated in multiship scenarios together, seamlessly execut-
ing tactics for which the aircraft were designed. At no point during SDD did the
aircraft operating limits match among the CTOL, STOVL, and CV aircraft
because flight sciences testing was concurrent with, not ahead of, mission
systems testing. Extensive maneuver development at EDW and Fort Worth simu-
lators was essential to execute edge-of-the-envelope testing. Even within a single
232 M. L. HUDSON ET AL.

variant, an individual tail’s limits differed from one another due to hardware and
software modifications. There also existed customer-specific requirements that
constrained particular testing to certain variants, depending on which combi-
nation of variant and weapon the customer was purchasing. Although the F-35
concept of commonality was beneficial to manufacturing affordability, it chal-
lenged engineers with complex and intertwined test requirements.

5. PLANNING, SCHEDULING, AND EXECUTION

The optimal forecasting period for scheduling both mission systems and flight
sciences testing was determined to be between four and six weeks out from the
current date. This allowed the team to adequately plan the major test missions
(i.e., those with multiple test aircraft, ranges, and assets), as well as logistically
easier missions. This time also allowed outside resources, including ranges, air,
ground, and sea assets, to be coordinated, scheduled, and configured as required.
The EDW test wing locks in the test schedule approximately two weeks out,
whereas many external ranges lock in their schedule at different times. By starting
the initial coordination and scheduling efforts further out, changes and refine-
ments can be made to the schedule as needed by the pace of the program,
thereby having a solid plan at the time of scheduling.
New, creative ways to schedule, plan, and execute missions were developed to
support the fast-paced F-35 flight test schedule. Single-ship range blocks at Point
Mugu’s sea range, the White Sands Missile Range, Naval Air Weapons Station
China Lake and Echo ranges, and EDW Precision Impact Range Area evolved
into concurrent operations sharing one piece of airspace as the program
matured. F-16 chase and target aircraft, oftentimes in short supply, were shared
among separate missions, shifting those missions’ particulars to accommodate.
Hot pits, the ability to refuel with the engine running and aircrew remaining in
the aircraft, were stood up when tankers fell out. A summary of flight test schedul-
ing and execution lessons learned is as follows:
. Brief the plan and fly the brief.
. Schedule and plan for both success and failure.
. Be ready to keep moving if things are going well, and be ready with backup
plans if they are not.
. Take tactical pauses when the test team needs them, because pilot safety could
depend on taking just one more day to study data.
. Spend time in the simulator with the engineers, aircraft, and the people who
maintain it because the more you research and evaluate what-if’s in advance,
the better you can prepare for the unknowns.
. Understand everything that you can about the systems and the data needs of
the customer.
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 233

. Ask questions, then ask more questions.

. Identify a drop-dead time to make changes and, once past that time, do not
allow additional changes even if the mission has to cancel. (The team needs
stability; if they are always on the hook for making changes, safety can be
compromised.)

F. INSTRUMENTATION/DATA PROCESSING
Robust instrumentation and timely data processing are central to flight test.
Flights are executed to obtain data to support engineering disciplines to under-
stand and verify system design. The extensive number of flights and test points
drove a huge number of test data requests. In 2016 more than 41,000 test data
requests were processed, with a total end-SDD of more than 279,000. New data
analysis tools were also required to evaluate 5th-Generation capability, such as
the Fusion Analysis Tool—Measures of Performance (FATMOPS) processing
capability, which was brought online at EDW in 2016. The following are
lessons learned from instrumentation, with some specific examples.

1. INSTRUMENTATION CHALLENGES AND LESSONS LEARNED

Instrumentation on the F-35 provided a unique challenge for the flight test commu-
nity. Not much has changed in the analog world of evaluating the strain on the
landing gear and the robustness of the flaps. The analog Orange Wire solution is
used time and time again for aircraft structural data with high success. The F-35
has an intricate digital computer network. For the first time in aviation history,
the avionics system under test had multiple styles of high-speed data buses, and
the cost of an instrumentation system to bulk-capture all the aircraft traffic, as
was traditionally done on legacy aircraft, far exceeded the feasibility of an affordable
solution.
The biggest challenge of the digital instrumentation on the F-35 was to meet
the data demands of the flight test community. There were extensive essential data
to be captured to validate the functions under test and additional data to evaluate
any unexpected anomalies. The instrumentation system recorded about six times
the amount of data that it was initially designed to record. This was partly due to
technology-refresh upgrades made over time with better processing power, RAM,
and throughput. The Data Acquisition, Recording and Telemetry (DART) instru-
mentation pod was in the left weapons bay of all mission systems aircraft, as
shown in Fig. 15. Different DART recording modes were selectable, which
allowed the aircraft to have multiple data maps in which some data were recorded
for some testing modes and not in others, with the intention of making more
bandwidth available. The data maps required intense mission planning across
all aircraft disciplines. This proved a successful solution, but the analysts are
always in need of more data.
234 M. L. HUDSON ET AL.

Fig. 15 Data Acquisition, Recording and Telemetry (DART) instrumentation pod.

The aircraft systems themselves have also been modified over time to provide
less bandwidth in specific cases. For example, high-rate diagnostic data sent to the
recorders may have previously published all zeros when no diagnostic information
was available. At design time, this made sense because the data requester would
know that the system was still alive and working based on the fact that data
was received, albeit zeros. These have since been modified to integrate the instru-
mentation system and to only publish data as required, freeing up immense band-
width for other messages to be recorded. The designers embraced the mechanism
of making the instrumentation system more efficient and stretched it to making
the aircraft network more efficient.
For the future, these lessons learned will probably lead to an even more inte-
grated solution of instrumentation digital buses. The pitfalls are now known and
can be harnessed on future upgrades and for future avionics platforms. The flexi-
bility and adaptability of the instrumentation system directly contributed to the
fast-paced test plan of the F-35, and any future platform will need to keep expand-
ability as a top priority.

2. MULTIPLE AIRCRAFT TELEMETRY SYNCHRONIZATION

Although the six original mission systems aircraft’s telemetry systems were crafted
as tail-specific constructs, the closure of Block 2B testing spawned the invention of
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 235

the Multiple Aircraft Telemetry Synchronization (MATS) scheme. All six air-
craft’s instrumentation systems were rebaselined and made compatible with one
another. This allowed for testing multiple aircraft simultaneously from one
control room and swapping telemetry from one aircraft to another when
needed. The turn time for pilots stepping from a ground-aborted aircraft to a
crew-ready cold spare was reduced, as well as the turn time for the control
room, which no longer needed to shut down and repower with the cold spare’s
unique telemetry parameters. The ITF’s telemetry footprint was reduced by pro-
gramming wingmen with the same telemetry frequencies. The benefits of the
MATS effort continued to be realized in Block 3F. Resource limitations, such as
control rooms, personnel, and frequencies, that would have otherwise prevented
the successful completion of 18 or more multiship mission effectiveness narratives
were now easily overcome.

G. TEST PILOTS’ PERSPECTIVE

Test pilots push the envelope and fly the F-35 through every situation possible,
doing it first to clear the way for the fleet pilots. The test pilots are convinced
that as a result of their work, more fleet pilots will return successfully from
combat missions. The first F-35 pilot to reach 1000 flight hours was David
“Doc” Nelson at EDW on 6 January 2017 in AF-3. The following are some of
his comments on F-35 flight testing at EDW [13].
A personalized objective of the F-35 test pilots has been to maintain solidarity
with the operational pilot end-user. High risk testing has been conducted by test
pilots at each edge of the F-35 envelope in a controlled and instrumented flight
test environment so that the operational pilot will never have to experience some-
thing in tactical flying that has not been explored. Discipline, pilot, and engineer-
ing skill along with team-wide trust have been critical to the success of the testing.
Every test pilot who has flown the F-35 has tactically relevant experience
allowing that pilot to suggest improvements or call for needed changes to make
this stealthy attack aircraft successful in combat and able to return its pilot
safely home. Success has come from using appropriately aggressive approaches
to flight testing that preserve safety while making efficient use of available
resources and time. High-fidelity simulation and a robust safety process have
been instrumental in balancing and enabling an aggressive approach. Examples
include:
. Reaching the maximum AoA limit on only the fourth mission of high-AoA
testing. This has taken weeks on previous programs.
. Intentionally putting the aircraft out of control in different ways more than 150
times to confirm its ability to recover to controlled flight.
. Turning off, then restarting the single engine multiple times, including near
Mach 1 at 40,000 ft to verify its restart capability. (An innovative Self-Contained
236 M. L. HUDSON ET AL.

Underwater Breathing Apparatus [SCUBA] system was invented for air-start

flight test to maintain cockpit pressure with the engine off.)
. Releasing dozens of weapons, including missiles, bombs, and bullets, and
including partner country and multiservice weapons, to verify safe release
and lethality.
. Repeatedly achieving maximum Mach and maximum g to expand the F-35 by
more than double the first flight envelope.

Each of these achievements gives confidence that the fighter pilots who will fly the
F-35 into combat have the greatest probability of success and survival throughout
the entire flight envelope of the aircraft. Test pilots have stressed several innova-
tive features that have their first significant application in the F-35. Examples
include:
. The Helmet-Mounted Display (HMD): The helmet is fully integrated with air-
craft sensors and allows well off-boresight use of target designator boxes for
air-to-air designation and weapon aiming, as well as seamless visual desig-
nation for ground targets. It also includes an integrated night vision camera.
The Gen III HMD is shown in Fig. 16. The main difference between the
Gen III and Gen II is the larger aperture for the improved night vision and
the sensors that talk to the fixed-camera assembly mounted on the glare shield.
. The electro-hydrostatic actuators used in the F-35 flight control system are the
first to be used in a widely produced aircraft.
. Electronic flight series data are used exclusively on the F-35.
. The multifunctional dis-
tributed aperture system
serves as a missile launch
detector and provides a
second means of night
vision to the pilot in a
separate spectrum.

Flight test of the F-35 has

been a high-tempo, aggres-
sive, long-term task for the
test pilots (Fig. 17). As a
result of the sustained effort
over time and the resultant

Fig. 16 Gen III

helmet-mounted display.
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 237

Fig. 17 F-35 flight test pilots at EDW.

increase in flight envelope and sensor performance, confidence in the airframe

and systems increased; the services, partner, and participating countries were
encouraged; the press became friendlier; and the unit cost of the aircraft continues
to decrease.

III. NAVAL AIR STATION PATUXENT RIVER

Continuing a tradition begun in 1943 with the creation of the Naval Air Test
Center (NATC) at PAX (Fig. 18), SDD flight test of the Navy’s (F-35C) and the
U.S. Marine Corps’ (F-35B) first 5th-Generation aircraft commenced on 15
November 2009 with the arrival of BF-1 from Fort Worth.
STOVL envelope expansion testing began soon after, kicking off an ambitious
flight test program. The PAX ITF conducted flight sciences and weapons envelope
expansion testing of the STOVL and CV variants in a similar fashion as the team
at EDW, with the additional challenge of demonstrating unique basing and ship
suitability requirements. PAX was the natural location for this testing due to the
unique facilities it offered: a shore-based TC-7 catapult, Mk-7 land-based arrest-
ment system, and centerfield Vertical Takeoff and Landing (VTOL) complex. The
VTOL complex includes a 1900-ft AM-2 expeditionary airfield, a hover pit, two
hover pads, a concrete outline of an LHD with an adjoining 38-ft hover position
indicator, and a shore-based ski jump. Additional shore-based catapult and arrest-
ment facilities were used at nearby Naval Air Warfare Center Aircraft Division’s
(NAWCAD’s) site in Lakehurst, New Jersey. Over the next eight years, nine
238 M. L. HUDSON ET AL.

Fig. 18 Aerial view of the F-35 BF-1 over Naval Air Station Patuxent River.

permanently assigned aircraft supported up to six test missions per day to achieve
multiple program milestones. Figure 19 shows two of the nine flight test aircraft
with the PAX ITF team.
The STOVL and CV basing and ship suitability demonstrations (Fig. 20) were
accomplished through a progression of shore-based testing and culminated with

Fig. 19 PAX ITF team with flight test aircraft.

F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 239

Fig. 20 Ship integration testing of the F-35B and F-35C.

each variant conducting a series of three sea-based developmental testing sea

trials of increasing complexity. This required the team to deploy to conduct
testing at various off-site locations with specialized facilities. The PAX ITF sup-
ported major detachments, including three F-35C deployments to CVN carriers,
three F-35B deployments to LHA/LHD carriers, and numerous detachments to
Lakehurst and EDW. The team also conducted climatic chamber testing in the
McKinley lab at Eglin AFB, Florida.

A. FLIGHT SCIENCES ENGINEERING

The PAX ITF was responsible for flight sciences envelope expansion testing for
the F-35B and F-35C variants, with the addition of highly specialized STOVL
and CV suitability testing. This led to four envelope expansion test programs
that ran in parallel, each with its own lines of wind-dependent testing and temp-
erature considerations for STOVL. The PAX ITF was also responsible for all
weapons environment and safe separation testing for the F-35B and F-35C var-
iants, with more than 50 total separations performed by the end of the SDD
program. Test planning had to be adaptable to engineering discoveries, including
various loads envelope expansion pauses, weapons integration issues, UHF
antenna redesign, F-35B Auxiliary Air Inlet (AAI) door redesign, and F-35C
tail hook and outer wing redesign. With an integrated team effort, all these tech-
nical challenges were overcome and successfully demonstrated.
In addition to conventional Stability and Control (S&C), flutter and loads
expansion, and handling qualities testing, F-35 testing at PAX included STOVL
mode. STOVL mode testing was a major effort that required significant planning
and integration among propulsion, flying qualities, performance, and external
environment disciplines. STOVL testing included both S&C and handling qual-
ities testing to clear a STOVL mode envelope separate from the conventional
flight envelope. A build-down approach was used to achieve the first hover and
vertical landing. First, the aircraft achieved first conversion from conventional
flight into STOVL mode and subsequently conducted Integrated Test Blocks
240 M. L. HUDSON ET AL.

(ITBs) to evaluate the S&C at flight conditions decreasing in speed. Once sufficient
ITBs were conducted at a certain speed, slow landing and short takeoff testing
began at those speeds. Then, once jetborne ITBs were conducted clearing the
hover, the aircraft could complete the first vertical landing. The envelope was
then cleared for additional stores configurations. The shore-based envelope was
then used as a basis for conducting STOVL expansion on LHA/LHD class car-
riers. Vertical takeoff and ski-jump short takeoffs were conducted later in the
program.
A significant amount of S&C data were required for door-open testing. During
conversion into STOVL mode, the STOVL doors open before the propulsion
system fully converts to STOVL mode. The doors have a significant impact on air-
craft stability during that transition. Collecting S&C data to accurately character-
ize the aerodynamics of the door-open state had to be done carefully. Frequent
data reviews were required to expand the envelope in this transition condition
and provide a complete aerodynamic database with door effects.
Takeoff and landing testing required more testing than a conventional
program. Conventional aircraft typically take off and land at relatively fixed
speeds. The STOVL propulsion system allows the aircraft to land at a wide
range of speeds, but this affects the takeoff and landing dynamics. Additional
testing was required to verify the flying qualities and aircraft loads at these
conditions.

B. TEST OPERATIONS
The PAX ITF needed to collect data to characterize the operation and perform-
ance of the F-35 in the shipboard environment. This environment consisted of
a combination of ship electromagnetic effects, ship motion (pitch, roll, yaw,
surge, heave, and sway), coupled with ship burble and winds of varying magni-
tudes and directions. These test conditions were not easily manufactured shore-
based due to the complex environmental factors of launching and recovering
fixed-wing aircraft at sea. The data collected by the PAX team during these sea
trials have been used to generate fleet launch and recovery bulletins, flight
manual information, and ship/aircraft Naval Air Training and Operating Pro-
cedures Standardization (NATOPS) in use by operational F-35 squadrons
deployed aboard U.S. Navy ships around the world today.
Conducting a developmental testing sea trial was no trivial matter due to the
data requirements, safety risk, associated mitigations, required efficiency, and
limited time available onboard a ship. The six sea trials conducted by the PAX
ITF were supported by an average of 200 ITF personnel who were deployed
onboard ship for sea trials from three- to five-week periods of time.
Because the nature of the developmental testing sea trials was so complex,
integration with the ship’s company was critical to ensure that the test operations
at sea were safe and operated as efficiently as possible. These efforts kicked off
with an Initial Planning Conference (IPC) held a year prior to the ship going
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 241

underway with the F-35. Additional planning conferences followed the IPCs, as
required. A portion of the ship’s company would travel to PAX for an aircraft fam-
iliarization that consisted of training briefings conducted by the PAX test team.
This trip would include an observation of live F-35 test events, such as short take-
offs and vertical landings for the F-35B and catapults and arrested landings for the
F-35C. Hands-on operations and demonstrations were conducted on such items
as chain-down procedures while noting consideration of such items as instrumen-
tation on landing gear. Upon arrival at the ship, the PAX ITF would complete
flight deck training, a ship familiarization, and become familiar with normal
ship operations/drills in order to integrate with ship’s company.

1. DETACHMENTS PLANNED AND ACCOMPLISHED

PAX testing included frequent and extensive detachments to other locations
(Table 1) due to unique test requirements. These detachments required personnel
to commit large amounts of time to plan and train. The ITF developed the skill to
plan these detachments efficiently through repetition, with FTEs taking lead roles
in this coordination. Ship detachments were very high paced, with two aircraft
flying two flight periods a day, seven days a week, all done in a very complex
and unforgiving environment. The detachments shown in Fig. 21 were all accom-
plished while continuing test operations at PAX.

2. STOVL SHIP SUITABILITY TESTING

U.S. Marine Corps requirements for the F-35B variant led to unique shore-based
ship suitability and compatibility testing in the following areas: external environ-
ment (vibro-acoustics), short takeoff/short landing/vertical takeoff testing in
multiple weapon configurations, and ski jump testing. Additional ship-based
testing was performed, including the three major sea trials that proved initial
capabilities and then expanded the wind and weapon envelopes for fleet delivery.
Smaller ship testing efforts occurred for Joint Precision Approach and Landing
System (JPALS) testing and certification and helmet night compatibility.
In January 2010, the BF-1 conducted the first airborne conversion from fully
wingborne flight to semi-jetborne flight. This milestone was followed by the
STOVL initial transitions test program, which featured the STOVL aircraft
demonstrating its ability to fly at progressively slower speeds throughout the semi-
jetborne flight regime. As the aircraft transitioned to slower airspeeds, it also
demonstrated its ability to perform short landings and short takeoffs. On 17
March 2010, the BF-1 performed its first hover, which was followed by the pro-
gram’s first vertical landing the very next day. Since that day, more than 1400 ver-
tical landings have been performed by PAX flight test aircraft.
One of the reasons to flight test is to identify issues early so that they can be
fixed prior to entering full-rate production. In 2010 the PAX team identified a
vibration issue with the AAI door during high-speed semi-jetborne testing. This
242 M. L. HUDSON ET AL.

flight test discovery, in part, led to the STOVL variant being placed on probation
by the U.S. Secretary of Defense. A redesign of the AAI door quickly ensued.
While the AAI door was being redesigned in 2011, the PAX team continued to
expand the STOVL mode flight envelope in areas not impacted by the door
vibration issue. In fact, the team was able to expand sufficient semi-jetborne
and jetborne flight envelope that year to allow the Initial Developmental Test
(DT-I) sea trial to commence onboard the USS Wasp (LHD 1) in October 2011
with test aircraft BF-2 and BF-4. During STOVL DT-I, the team demonstrated
daytime STOVL operations in the clean configuration, as well as aircraft spotting,
deck handling, MIL and MAX power short takeoff launches, and recovery vertical
landings from stern to bow approaches. Logistics operations were demonstrated

TABLE 1 PAX DETACHMENTS

Year Detachments
2011 1. Lakehurst, NJ: F-35C initial F-35C catapults and traps
2. STOVL DT-I (USS Wasp, East Coast)

2012 1. Lakehurst, NJ: F-35C traps

2. EDW: F-35B air-starts

2013 1. STOVL DT-II (USS Wasp, East Coast)

2014 1. Lakehurst, NJ: F-35C catapults and traps

2. EDW: STOVL wet runway and crosswinds
3. CV DT-I (USS Nimitz, West Coast)

2015 1. EDW: STOVL crosswinds and CV wet runway and crosswinds

2. Eglin AFB: McKinley Climatic Laboratory
3. CV DT-II (USS Dwight D. Eisenhower, East Coast)

2016 1. Lakehurst, NJ: F-35C catapults and traps

2. NASA Wallops: Ops temporarily moved due to PAX runway construction
3. CV DT-III (USS George Washington, East Coast)
4. STOVL DT-III (USS America, West Coast)

2017 1. Lakehurst, NJ: F-35C catapults and traps

2018 1. MCAS Cherry Point/MCALF Bogue Field, NC: Mode 4 sloped pad testing
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 243

Fig. 21 PAX detachments.

and external environment data were gathered to characterize the operation of the
STOVL aircraft in the carrier-based environment.
After completing a successful initial sea trial, a redesigned AAI door system
was installed on the BF-1 late that same year. The redesign was successfully ver-
ified through regression flight testing, which resulted in the STOVL probation
being rescinded by the Secretary of Defense. Throughout 2012 and the first half
of 2013, the team expanded the flight envelope into night operations and began
carrying internal stores.
In August 2013, the team returned to the USS Wasp (LHD 1) with test aircraft
BF-1 and BF-5 for the second Developmental Test (DT-II) sea trial. During this
detachment, the team of more than 200 PAX personnel successfully demonstrated
the ability to conduct 94 shipboard sorties in day and night operations while flying
in a clean-wing configuration with internal stores.
The final SDD STOVL sea trial took place onboard the USS America (LHA 6)
in the fall of 2016 with test aircraft BF-1 and BF-5 in the Pacific Ocean. The USS
America was the first Navy vessel designed and built to accommodate the F-35.
This final sea trial demonstrated day and night operations in the presence of
higher sea states and deck motion. The test team demonstrated the ability to
launch and recover the aircraft with both symmetric and asymmetric external
store configurations. This sea trial was also significant in that we no longer
enjoyed the luxury of an exclusive flight deck; our developmental test operations
were conducted in conjunction with those of VMX-1, an operational test squa-
dron performing carrier qualifications.

3. CV SHIP SUITABILITY TESTING

Navy requirements for the F-35C led to unique shore-based ship suitability and
compatibility testing in the external environment (vibro-acoustics), Jet Blast
Deflector (JBD) and flight deck compatibility, approach handling quality survey
in multiple weapon configurations, roll-in and fly-in arrested landing structural
244 M. L. HUDSON ET AL.

Fig. 22 F-35C shore-based catapult spotting.

survey in multiple weapon configurations, and a catapult structural survey in mul-

tiple weapon configurations (catapult steam ingestion testing).
In parallel to the STOVL ship suitability and envelope expansion testing
program, the F-35C was conducting its own flight test program and facing its
own unique challenges. The first CV F-35, CF-1, arrived at PAX in November
2010. Early the next year, it began catapult spotting evaluations using the TC-7
site at PAX, as shown in Fig. 22.
In June 2011, the CF-2 traveled to NAWCAD Lakehurst to complete JBD
testing. This testing was important to verify that the installed engine function
was not degraded by hot gas ingestion when the aircraft was placed in front of
or behind the JBD (Fig. 23). This test included JBD modifications to change
water cooling designed for the exhaust impingement of smaller twin-engine air-
craft to accommodate the larger single-engine F-35. This was followed by a
month-long period of testing at Lakehurst that encompassed a structural survey

Fig. 23 F-35C jet blast deflector integration testing.

F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 245

using the C13-2, steam ingestion testing, and roll-in arrestments using the
Mk-7 and E-28 arrestment systems.
The following year, the CF-3 conducted the first fly-in arrestments with the
Mk-7 and E-28 arrestment systems at Lakehurst. The PAX team then discovered
that the F-35 Arrestment Hook System (AHS) and shoe geometry were
inadequate to reliably catch an arrestment wire. These test results triggered modi-
fications and updated designs of both the AHS and the hook shoe. In January
2014, the CF-3 resumed roll-ins and arrested landings with a redesigned AHS
at Lakehurst. These test results were sufficient to allow the team to progress to
a Mk-7 structural survey in the clean configuration with internal stores.
In November 2014 the initial CV developmental test sea trials commenced
onboard the USS Nimitz (CVN 68) with CF-3 and CF-5 test aircraft. This was
a west coast sea trial, which further complicated the logistics of the east coast
PAX detachment. During this sea trial, for which we enjoyed exclusive use of
the flight deck, the ITF conducted 124 catapults/traps and 222 touch-and-go
landings with no bolters (i.e., failure to catch an arresting cable when landing).
In doing so, the ITF successfully demonstrated day operations, deck handling
and logistics, launch and recovery handling, minimum end-speed testing, MIL
power launches, and some initial night operations. The aircraft was launched
from both the bow and waist catapults. The team did identify an adverse roll
when the aircraft was launched from the outer waist catapult (Cat. 4) that
would have to be resolved, but overall it proved to be a very successful initial
sea trial.
The team resumed sea trials onboard the USS Dwight D. Eisenhower (CVN 69)
in October 2015 (Fig. 24), but the team no longer enjoyed the luxury of having
exclusive use of the flight deck. This at-sea detachment focused on operations
with increased aircraft weight and forward CG. During DT-II, 66 catapult
launches were performed from all four catapults, along with 66 arrestments and
40 touch-and-go landings. Operations demonstrated during DT-II included an
ability to operate in day and night conditions, internal weapon carriage, MIL
and MAX power launches, and powered aircraft operation in the hangar bay.
Additionally, external environment data were collected and the team performed
a risk reduction of the JPALS.
In 2016 the test team began the shore-based work-up with external stores.
Between February and May, the Mk-7 structural survey and arrestments were
evaluated with a variety of external store loadouts. In August, the third and
final CV developmental test sea trial was performed onboard the USS George
Washington (CVN 73), shown in Fig. 25. The team shared the deck with
VFA-101, an operational squadron performing carrier qualifications. During
this final sea trial, 121 catapults and traps were performed, along with 67
touch-and-go landings. Operations were conducted with full internal bays and
external stores.
These highly successful sea trials were a result of the PAX ITF’s thorough
shore-based envelope expansion test planning and execution. During the last
246 M. L. HUDSON ET AL.

Fig. 24 PAX ITF personnel and test aircraft onboard the USS Dwight D. Eisenhower (CVN 69).

Fig. 25 CF-3 catapult launch from the USS George Washington (CVN 73).
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 247

eight years of flight testing at PAX, the shore-based testing was very demanding
and uncovered several issues prior to taking the F-35B and F-35C to operate on
U.S. Navy ships. These issues, their fixes, and the additional learning that took
place during these sea trials have made the F-35B and F-35C more capable
platforms.

C. INSTRUMENTATION/DATA PROCESSING
Test instrumentation data are the blood flow of the flight test heart. The primary
purpose of mission execution at PAX is to provide the truest and most accurate
data required for air vehicle system verification and to demonstrate that the
F-35B and F-35C variants operate efficiently, effectively, and safely. Keeping the
data system operational and productive had its challenges, but over the past
several years, this team overcame several obstacles.
Integrating commercial, off-the-shelf systems with unique data bus protocols
that matured over the course of the SDD program resulted in several flight test
data system upgrades to increase throughput and recording capabilities. PAX
worked out a restrictive approach to data throughput by asking the integrated
product teams to narrow their scope of measurements and sample rates to a
mission-critical and safety data list.
With a limited amount of noninstrumented weapons available at PAX, instru-
mented weapons were flown continually for dummy stores. This led to multiple
incidents in which strain gauges, transducers, and wiring were compromised.
When instrumented weapons were required for safety of test/safety of flight,
delays in mission readiness became an issue, often after weapons had been
loaded onto the aircraft. Resolution of this issue required a cross-team effort
and a written process for inspecting, repairing, and tagging instrumented
weapons after every flight.
Unique airframe instrumentation, paired with a limited number of instrumen-
ted landing gear, LiftFanw, and engines, posed unique constraints. The required
instrumentation configurations, particularly for carrier integration work, were
not always available due to competing needs. This issue resulted in several gear
swaps between/among aircraft during the SDD flight test program. Similarly, a
limited number of instrumented LiftFans and engines posed test configuration
constraints that had to be effectively managed, particularly when this instrumen-
tation became damaged or needed repair off wing. It was not uncommon during
SDD to have a variety of lines of test on different airframes competing for instru-
mented propulsion system hardware, sometimes necessitating the swap of hard-
ware between/among aircraft.
Before sea trials commenced, some modifications were required to aid the
team in gathering the required data. Mobile control rooms, such as the Mobile
Instrumentation and Telemetry System (MITS-12) and a deployable debrief facil-
ity, were installed in the hangar bay to allow engineers to monitor aircraft flight
test activities in real time. High-speed deck edge cameras were installed to
248 M. L. HUDSON ET AL.

gather launch data. Cameras were also installed about the stern of the ship to
document where the aircraft touched down during recovery. A 30-ft anemometer
was installed on the bow of the ship. This, along with a ship motion package,
helped to characterize the dynamic ship environment that the F-35s would
call home.

D. TEST PILOTS’ PERSPECTIVE

PAX flight operations included a cadre of test pilots representing Lockheed
Martin, BAE Systems, and the U.S. Navy, U.S. Marine Corps, and U.K. Ministry
of Defence (Fig. 26). Beginning with the arrival of the first aircraft in 2009, 35
different pilots flew test missions at PAX, completing more than 4400 flights total-
ing 6600 hours and 35,000 test points. All pilots were qualified to fly all variants of
the F-35, and most of them carried additional qualifications in the F/A-18. With
no available off-site F-35 training and no experienced F-35 pilots coming into the
ITF, all training was developed and conducted within the organization. The team’s
goal to fly 10 hours per pilot per month in the F-35 was never achieved; rather, the
average was close to 7 hours per pilot per month. However, this rate was accep-
table due to the mixture of pilot experience, extensive use of simulator-based prep-
aration, and a significantly extensive suite of automated maneuvers built into the
control laws.
Manning considerations provided a challenge throughout SDD. The histori-
cally useful pilot-to-test-aircraft ratio of 1.5-to-1 was used as the basis for
manning decisions. With nine assigned aircraft, manning levels generally kept

Fig. 26 F-35 flight test pilots at PAX.

F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 249

to 13 pilots flying at any time (4 –5 contractors and 8– 9 government pilots).

Unique test requirements, specifically STOVL and CV envelope expansion, mul-
tiple long deployments, and the availability of flight time and requirements for
chase support, led to occasional deviations.
There is a cultural and historical belief that safety chase is required for devel-
opmental flight test. As a result, early in the program, a safety chase aircraft was
required for nearly every flight. It soon became clear that, in many cases, the pres-
ence of another aircraft in the airspace and in close proximity to the test aircraft
added risk to the test event. One example involved chasing very slow-speed
STOVL missions with F/A-18s. The missions were flown at low altitude in
busy airspace at speeds that were incompatible with the F/A-18, which left the
Hornet pilot in a compromised position. The ITF methodically went through
the process of reevaluating each test plan for residual risk and applied appropriate
mitigations for each one, which was rarely a chase asset. The most effective miti-
gators included mission preparation requirements, maneuver standardization,
and control room monitoring of critical parameters and air vehicle health. The
result was a reduction in the complexity of ITF testing, lower overall risk associ-
ated with each mission, and a significant cost savings.
PAX ITF flight operations were successful in meeting all requirements with
the following major accomplishments:
. Successfully contributing in the proof of design for the first-ever use of the
dynamic inversion model for the control laws.
. Extending the STOVL concept to deliver STOVL capability, including verify-
ing the STOVL performance models and developing Level 1 handling qualities
throughout the STOVL envelope in nominal conditions.
. Recognizing CV hook design and pitch pivot point problems early enough that
they had timely resolution to meet SDD completion requirements.

The immense amount of work performed in a short time by the PAX ITF is extra-
ordinary and unprecedented. This work has already enabled the Marine Corps
warfighter to deploy aboard U.S. Navy L-class vessels today and will soon do like-
wise for the U.S. Navy onboard CVNs. These efforts also significantly contributed
to the Marine Corps declaring F-35B IOC in 2015, and they will contribute to the
Navy declaring F-35C IOC in 2019.

IV. CONCLUSION
The SDD phase of the F-35 program was in work for 16 years. During that time,
eight years of flight sciences and mission systems flight testing with 18 aircraft at
two test sites yielded and delivered an unprecedented 5th-Generation warfighter
capability, summarized in Fig. 27. It also provided invaluable experience and
many lessons learned: test plans must be flexible and agile early on; a surge
concept with focused and concentrated resources accomplished complex testing
250 M. L. HUDSON ET AL.

Fig. 27 The F-35 Block 3F delivered 5th-Generation warfighter capability.

in a short period of time; the optimized flight test scheduling forecast is four to
six weeks out; instrumentation must be flexible, adaptable, and expandable; train-
ing by an expert trainer is key to successful test conductors; and teamwork, with
clearly defined roles and responsibilities, and communication are essential. With
these lessons learned, future F-35 flight testing can now expand on the historical
accomplishments achieved during the SDD phase of the F-35 at EDW and PAX,
and continue to deliver unrivaled F-35 capability.

ACKNOWLEDGMENTS
This paper is dedicated to the EDW and PAX Integrated Test Force (ITF) team
who executed an unprecedented, historic flight test program. Special thanks to
those who provided inputs to capture the history, flight test accomplishments,
and lessons learned. Many thanks to the EDW team: David “Doc” Nelson,
Dave Holcomb, Teo Avram, Brittany Galloway, Don Lewis, Ron Teague, Mike
Shreeve, Tyler Sanders, Brian “Grumpy” O’Melia, Torrey Given, Bernie Taylor,
Petar Simich, Jim Bishop, Ryan Mankin, Stan Bogumil, Rita Jones, Peter Liu,
Brandon Dees, Jason McCoy, George Hicks, Lane Vaught, Caitlyn Fahey, Jen
Schleifer, Ed Sabalburo, Wade Cross, Susan Bishop, and others. Many thanks to
the PAX ITF team: Darrell Carney, Jerry Courville, Leigh DePiazza, Eric
Faidley, Dan Levin, Trey Mangan, Larry Martin, Dane Wiedmann, Andy
Wolfe, Mike Jackson, and others. Also, thanks to J.D. McFarlan for his leadership;
F-35 SYSTEM DEVELOPMENT AND DEMONSTRATION FLIGHT TESTING 251

graphic artists Daniel Buck, Justin Simmons, and Laine Miller; and Chad Bellay
for the awesome flight test video.

REFERENCES
[1] Ellis, R., et al., “F-35 Structural Design, Development, and Verification,”
AIAA Aviation Forum, June 2018 (to be published).
[2] Harris, J., and Stanford, J., “F-35 Flight Control Law Design, Development, and
Verification,” AIAA Aviation Forum, June 2018 (to be published).
[3] Wurth, S., Smith, M., and Celiberti, L., “F-35 Propulsion System Integration,
Development, and Verification,” AIAA Aviation Forum, June 2018
(to be published).
[4] Robbins, D., et al., “F-35 Subsystems Design, Development, and Verification,”
AIAA Aviation Forum, June 2018 (to be published).
[5] Lemons, G., Carrington, K., Frey, T., and Ledyard, J., “F-35 Mission Systems Design,
Development, and Verification,” AIAA Aviation Forum, June 2018
(to be published).
[6] Frey, T. L., Aguilar, J. C., Engebretson, K. R., Faulk, D. K., and Lenning, L. G.,
“F-35 Information Fusion,” AIAA Aviation Forum, June 2018 (to be published).
[7] Wilson, T., “F-35 Carrier Suitability Testing,” AIAA Aviation Forum, June 2018
(to be published).
[8] Parsons, D. G., Eckstein, A. G., and Azevedo, J. J., “F-35 Aerodynamic Performance
Verification,” AIAA Aviation Forum, June 2018 (to be published).
[9] Canin, D. G., McConnell, J. K., and James, P. W., “F-35 Weapons Separation Test
and Verification,” AIAA Aviation Forum, June 2018 (to be published).
[10] Hetreed, C. F., Carroll, M. D., Collard, J. E., and Snyder, R. C., “F-35 Weapons
Separation Test and Verification,” AIAA Aviation Forum, June 2018
(to be published).
[11] Parsons, D. G., et al., “F-35 STOVL Performance Requirements Verification,”
AIAA Aviation Forum, June 2018 (to be published).
[12] Commanding Officer, Marine Fighter Attack Squadron 121, “MAG-12 After Action
Report for Exercises Northern Edge/Distant Frontier 2017,” Marine Fighter
Attack Squadron 121, U.S. Marine Corps Official Report, FPO AP 96310-7335,
July 2017, p. 8.
[13] Nelson, D., “F35 SDD Lessons Learned,” Lockheed Martin Aeronautics Company,
F-35 System Development and Demonstration Presentation, Palmdale, CA,
August 2017.
 CHAPTER 7

F-35 Structural Design, Development,

and Verification
Robert M. Ellis , Philip C. Gross†, Joseph B. Yates‡ and
John R. Casement§
Lockheed Martin Aeronautics Company, Fort Worth, TX
R.H.L. (Tad) Chichester}
Lockheed Martin Aeronautics Company (retired), Fort Worth, TX

Kathryn Nesmith
F-35 Joint Program Office, Arlington, VA

This paper will discuss the structural design, development, and verifica-
tion of the F-35 Joint Strike Fighter, guided by the standard practices of
MIL-STD-1530 Aircraft Structural Integrity Program (ASIP) to success-
fully produce a versatile air vehicle platform that meets the varied per-
formance and airworthiness requirements of our worldwide customers.
This approach resulted in the simultaneous development of three air-
frames that are exceptionally robust, and will meet the demanding
requirements of war fighters worldwide that are choosing the F-35.
The “Five Pillars of ASIP,” as defined in MIL-STD-1530, provide the fra-
mework for development of an air vehicle platform from design through
analysis, testing, and force management. ASIP applies throughout the
entire aircraft life cycle from requirements definition through retire-
ment. This paper will follow the progression through each task for the
development of the F-35 used to enhance the development of this
extraordinary product.

I. INTRODUCTION
This paper will discuss the structural design, development, and verification of the
F-35 using the standard practices of MIL-STD-1530 [1] Aircraft Structural

 Senior Manager, F-35 Structures Development Integration.

†
Deputy Director F-35 Structures Development and Chief Structures Engineer.
‡
Aeronautical Engineering Principal, F-35 Service Life Analysis.
§
Aeronautical Engineering Principal, F-35 Structures Development Integration.
}
Retired Senior Manager, F-35 Structures Development.
 Air Vehicle Lead, F-35 Joint Program Office.

253
254 R. M. ELLIS ET AL.

Integrity Program (ASIP) to successfully produce a multiservice, versatile air

vehicle platform that meets the varied performance and airworthiness require-
ments of our worldwide customers, as depicted in Fig. 1 showing unique require-
ments for each of the three variants. The “Five Pillars of ASIP,” as defined in
MIL-STD-1530, provide the framework for development of an air vehicle plat-
form from design through analysis, testing, and force management. ASIP
applies throughout the entire aircraft life cycle from requirements definition
through retirement. This paper will follow the progression through each task as
shown in Table 1, used to enhance the development of the F-35 product.
The Joint Strike Fighter (JSF) program, later to be designated the F-35, is
the largest single defense acquisition program in U.S. government history. The
program was awarded in a new era of acquisition reform. The primary tenant
of this reform was the utilization of Performance-Based Specifications (PBSs),
intended to give the awarded contractor unprecedented freedom in the develop-
ment of the weapon system by merely specifying end-state capabilities, rather
than utilizing legacy practices of employing layered military specifications,
which dictated detailed requirements to the process of developing and verifying
such capabilities. The approach gave the government’s Joint Program Office
(JPO) insight, but not oversight, of the efforts of the contractor, Lockheed Martin.
Although the approach had been successfully employed in the development
of aircraft subsystems and other smaller acquisition programs, the use of
performance-based specification to the development of an air vehicle structure
was unprecedented. This precedent has been, and remains, the cornerstone in
the certification of aircraft structures. To adhere to the intent of the PBS approach,

Fig. 1 F-35 multiservice structures design.

F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 255

TABLE 1 FIVE PILLARS OF ASIP

Task I Task II Design Task III Task IV Task V Force

Design Analyses & Full-Scale Certification Management
Information Development Testing & Force Execution
Testing Management
Development
ASIP Master Material and Joint Static Tests Certification Individual Airplane
Plan Allowables Testing First Flight Analysis Tracking (IAT)
Design Service Loads Analysis Verification Strength Program
Life and Design Service Ground Tests Summary & Loads/
Design Loads Spectra Flight Tests Operating Environment
Usage Design Chemical/ Durability Test Restrictions Spectra Survey
Structural Thermal Damage (SSOR) (L/E SS)
Design Environment Tolerance Force Structural ASIP Manual
Criteria Spectra Tests Maintenance Aircraft Structural
Durability & Stress Analysis Climatic Tests Plan (FSMP) Records
Damage Damage Tolerance Interpretation & Loads/ Force
Tolerance Analysis Evaluation of Environment Management
Control Durability Analysis Test Results Spectra Survey Updates
Program Corrosion Assessment (L/E SS) Recertification
Corrosion Sonic Fatigue Analysis Development
Prevention Vibration Analysis Individual
and Control Aeroelastic & Airplane
Program Aeroservoelastic Tracking (IAT)
(CPCP) Analysis Program
Nodestructive Mass Properties Development
Inspection Analysis
(NDI) Survivability Analysis
Program Design Development
Selection of Tests
Materials, Production NDI
Processes, Capability Assessment
Joining Initial Risk Analysis
Methods,
and
Structural
Concepts

JPO and Lockheed Martin worked together to develop the JSF Joint Contract
Specification (JCS) to define expected structural performance parameters, while
also clearly defining design criteria, verification analyses, and testing, which tra-
ditionally had been employed to achieve such performance.
In this joint government –contractor effort, development of a thoroughly inte-
grated and executed ASIP became the foundation to achieve challenging perform-
ance requirements for airframe safety and service life. ASIP provides an umbrella
framework that ensures coordination of critical engineering and management
processes, specifications, and plans—such as quality assurance, manufacturing
256 R. M. ELLIS ET AL.

engineering, corrosion control, fracture control board, materials and processing,

structures building block development test plan, and so forth—leading to air
vehicle certification. The challenge facing the JSF program was to come to agree-
ment on an ASIP that sufficiently addressed the concerns of the multiple service
and partner country airworthiness authorities. Although each partner country
had its own unique legacy to structural certification, all agreed that the most
comprehensive foundational approach to the problem was the adaptation of
MIL-STD-1530, with its prescription under the Five Pillars of ASIP, to define
the certification path for the F-35.
The Five Pillars of ASIP are a planned series of tasks employed to ensure
aircraft safety and structural integrity requirements are met and maintained
throughout the fleet operational service life. Much of the affordability aspect of
ASIP is accomplished by following a prescribed rigorous approach that attempts
to avoid design shortfalls while building in robustness to avoid costly structural
rework. The ASIP seeks to optimize cost of ownership and aircraft availability
while maintaining aircraft safety. The foundation of ASIP is a set of test and analy-
sis plans defined to meet the program requirements. The evidence that these
requirements are met supports the structures integrity and certification position
discussed herein.

II. TASK I: DESIGN INFORMATION

The Design Information task encompasses those efforts required to apply existing
theoretical, experimental, applied research, and operational experience to specific
criteria for materials selection and structural design. Elements of Task I include
developing the ASIP Master Plan, Structural Design Criteria (SDC), and Dura-
bility and Damage Tolerance (DADT) control plan; design loads, selection of
materials, processes, and joining methods; and defining the design service loads.
The primary Task I objective is to ensure the appropriate criteria and planned
usage are applied to the design of the F-35 so the specific operational requirements
will be met.
Design information and data used to develop the basic F-35 airframe struc-
tural design under the System Development and Demonstration (SDD) phase
of the program is complete. Air system –level design requirements were specified
in the F-35 JCS. The JCS is a performance-based specification that defines vehicle
requirements in terms of required operational parameters like speed, range, or
payload. The specific “how-to” actions to be performed to achieve the required
performance are defined in lower level requirements documents developed by
the contractor and JPO, and, in the case of structures, following the ASIP frame-
work. Most of the structural requirements flow from the contractual requirements
to provide a safe, durable, and damage tolerant air vehicle that will endure for 30
years or 8000 flight hours. Additional requirements are imposed by the need to
meet performance-based requirements that encompass combat radius, payload
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 257

requirements, austere basing requirements, flight performance characteristics,

and so forth.
The F-35 Structural Design Criteria was one of the first and most fundamental
of the structure requirements documents generated in Task I. The joint JPO –
Lockheed Martin team employed the Joint Services Specification Guidance
(JSSG-2006) to provide the basic framework for the SDC, tailoring the guidance
of the vehicle-level performance requirements (speed, weapons, etc.) and antici-
pated peacetime training usage defined in the JSF JCS. The completed F-35
SDC document contains all the aircraft design weights, weapons carriage and
employment, speeds, load factors, aeroelastic stability requirements, landing
sink rates, ship compatibility parameters, and the like required to develop struc-
tural design loads that, when applied and verified, ensure each F-35 variant meets
the specification performance. Of note is the early program policy decision to
include projected empty weight growth-to-Initial Operational Capability values
in the structural design weights, reducing the risk of “loads creep” and associated
life and performance limits that have plagued other aircraft programs.
The F-35 SDC document also decomposes the required 8000 flight hour
service life and design mission utilization specified in JSSG-2006 Appendix A-3
(anticipated mission profiles, mission mix, ship vs land basing, etc.) into a com-
plete definition of the operational parameters required to determine the baseline
load spectrum for each of the variants. The definition includes for each variant:
. Numbers and types of each mission per lifetime
. Mean and 90th percentile maneuver load factor exceedances per 1000 flight
hours, as determined from analysis of measured usage data from applicable
legacy aircraft
. Numbers and types of landings (vertical, shipboard arrested, etc.) as well as the
distribution of landing sink rates
. Numbers and types of takeoffs including catapult launches, ski-jump launches,
and short takeoffs
. Taxi and ground handling spectra for taxi, turning, braking, and runway
roughness

The importance of the baseline usage and resulting spectrum cannot be over-
stated—every ASIP task, from detailed design to full-scale test to force life man-
agement, relies on or is impacted by this information as a foundational
starting point.
Finally, the SDC defines the verification requirements for each of its require-
ments, including whether the verification method will be via inspection, analysis,
demonstration, test, or a combination of these. These verification requirements
are linked to the corresponding JCS requirements.
To support SDD contract closure, a Verification Cross Reference Matrix
(VCRM) was established that details the verification objectives and jointly
258 R. M. ELLIS ET AL.

developed Success Criteria (SC) that verify each JCS requirement. Closure of the
SC are tracked in the Verification Test and Evaluation (VT&E) database. SC clo-
sures support closure of the parent JCS paragraphs, which in turn support SDD
contract closure, as shown in Fig. 2.
The establishment of the operating environment was critical for use in the
design and analysis of the structure. Figure 3 highlights how the static and
service external loads were developed using wind tunnel data and an established
aerodynamic database. The PBS for the F-35 requires that 90% of the delivered
aircraft meet the full service life requirement of 8000 hours and 30 years. This
was accommodated by developing intentionally severe spectra based on legacy
fleet usage severity factors such that a 90th percentile usage spectrum was devel-
oped for design and test. The analysis criteria on which the service life was to be
determined was not specified. At the outset of the F-35 program, a common
service life analysis criteria was proposed for the three variants; however, after dis-
cussions among JPO, U.S. Air Force, and U.S. Navy, no common criteria
were found to be acceptable to all of the services. It was then decided to adopt
service life analysis criteria that reflected the legacy experience base of the
primary U.S. user for each variant. Thus, the F-35A was analyzed using crack
growth– based criteria combined with a 90th percentile mission-based spectrum
for durability analysis and a mean of 50th percentile spectrum for damage toler-
ance analysis—similar to previous Air Force aircraft. The F-35B and F-35C were
designed using a crack initiation criteria coupled with a Critical Point in the Sky

Fig. 2 F-35 structures certification process.

F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 259

Fig. 3 Loads wind tunnel testing and aerodynamic database development.

(CPITS) based severe usage spectrum for durability analysis, and the damage
tolerance analysis was performed using crack growth analysis coupled with the
same severe CPITS spectrum.
The CPITS approach is based on a single critical point in the sky for
each major airframe component, where a point in the sky is a unique Mach –alti-
tude combination. A single reference aircraft weight, with multiple configurations
and centers of gravity as shown in Fig. 4, was used for each such point in the sky. A
damage reference level based on that single point in the sky was established for
each major airframe component. Then, a single spectrum of multiple critical
points in the sky was developed, with the goal that no single component’s
damage was less than 80% of its reference level. Catch-up cycles were added to
bring each of the critical component damage levels as close to the 80% goal
as feasible.
For the F-35B and F-35C, the CPITS approach resulted in a mission mix
with 50% supersonic usage. This is far more supersonic usage than is expected
in actual usage, which makes it especially severe for aft fuselage and empennage
structure. The remaining usage includes 30% of the time spent at sea level,
which is critical for the wing structure. In contrast to CPITS, the mission-based
spectra used for the F-35A variant is based directly upon the Mach numbers,
speeds, altitudes, and mass properties for the mission profiles provided in
the JCS.
For the F-35 program to be successful under the new format of acquisition, it
was imperative a robust ASIP Master Plan be developed. This plan outlined the
260 R. M. ELLIS ET AL.

Fig. 4 CPITS load control points.

criteria for material selection, structural design and analysis, and certification
testing requirements. The plan provided a roadmap to ensure all evidence for
certification and verification was identified, planned, and controlled. The F-35
program was awarded to three companies with Lockheed Martin as the lead.
Northrup Grumman and BAE Systems joined Lockheed Martin as the partners
on the F-35. Lockheed Martin, as the lead, was responsible for overall program
integration. The Fort Worth, Texas, facility held design authority for the
forward and wing structure. The Palmdale, California, facility was responsible
for all control surfaces and edges. The Palmdale teams worked closely with the
BAE Systems teams who held design authority for the Horizontal Tails (HT),
Vertical Tails (VT), and aft fuselage. Northrup Grumman was assigned design
authority for the center fuselage structure.
The tri-company teaming arrangement for the F-35 program required signifi-
cant emphasis on planning and coordination to ensure the establishment of
consistent design and analysis criteria. F-35 – specific design and analysis
manuals and common toolsets were developed using the best practices of the
three companies. Direction to the structures disciplines was provided through
more detailed subordinate documents; for example, the Structural Analysis
Methods and Design Criteria (SAMDC), DADT Guidelines and Control Plan,
and Drawing Requirements Manual were published as unique F-35 –specific
direction. The SAMDC defines the structural analysis methods, policies, and cri-
teria used by the development team to conduct structures analysis. A recurring
coordination meeting was established to ensure all structural analysis disciplines
from all three companies met and discussed resolution of issues with best
practices selected. Frequent reviews were held at each partner site, as well as at
major subcontractor sites, during the layout and design phases to review the
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 261

design and analysis of the F-35 structure to ensure that the policies and criteria
were being applied properly and that the resulting structure was the lightest
weight structure that would meet the requirements. The multisite collaboration
was enabled by the selection of a common design tool suite and carefully con-
trolled interface control drawings. A centralized database, or Product Data
Manager (PDM), was established to allow all partners to work on the latest
designs. This “digital thread” allowed the three partners to develop and share
digital models within the team and, upon release, to send them to suppliers
for fabrication.
The design usage requirements in the SDC discussed previously were
broken down further in the Environmental Description Document (EDD),
which contained detailed usage and environmental data needed to flow down
to any engineering group requiring it, particularly those responsible for procure-
ment and qualification of mechanical and electrical systems for the aircraft.
Lifetime and extreme usage information, such as total time in Mach-altitude
and temperature bins, time at angle-of-attack, and so forth is defined in the
EDD. That data, in turn, became the basis for usage and environments data
for items such as subsystems design usage; vibratory and acoustic environ-
ments; temperature environments; and humidity, moisture, and precipitation
environments.
The F-35 DADT Control Plan identifies and assigns responsibility for all of
the fracture control tasks necessary to ensure compliance with the DADT
design requirements and continued structural integrity. The Fracture Control
Board ensures consistent implementation of the DADT Control Plan require-
ments at all F-35 team sites and provides guidance to the product teams
with respect to the DADT requirements. Board membership includes the dis-
ciplines of stress, DADT, Materials and Processes (M&P), design, manufactur-
ing, quality assurance, and an airframe Integrated Product Team (IPT)
representative for a direct, authoritative tie to the subordinate IPTs. Similarly,
the F-35 Corrosion Prevention Control Plan identifies and assigns responsibil-
ity for the corrosion control tasks. The Corrosion Prevention Advisory Board
oversees the consistent implementation of the design requirements needed
for corrosion prevention and control at all F-35 team sites. The NDI Require-
ments Review Board (NDIRRB) guides the development of NDI capabilities
across the F-35 team and at all subcontractors to ensure that the quality
requirements of the aircraft structure are met. It also ensures that the NDI
capabilities needed to support the aircraft in service are well defined and avail-
able in a timely manner.
Materials, processes, and joining methods are defined, reviewed, and
program approved for airframe structural applications. Building block tests of
successively increasing complexity, as shown in Fig. 5, were conducted to
provide engineering data used to demonstrate compliance with structures
design requirements. A structures building block development plan was defined
after contract award and is documented as part of the overall F-35 Systems
262 R. M. ELLIS ET AL.

Fig. 5 Building block test approach.

Test Plan. Coupon- and element-level tests were conducted as part of the F-35
building block test program to generate material allowable properties, qualify
materials and processes, characterize joining methods, and develop strength,
durability, and damage tolerance properties for design-specific features of the
F-35 aircraft.
Physical, mechanical, and chemical properties and attributes are documented
and controlled by material specifications and verified through required incoming
material receiving and quality assurance processes. All materials used in design
applications are covered by released specifications. Processes used in airframe
construction, from detail part fabrication through final assembly, are controlled
by specification reference on the face of the drawing and by manufacturing
work instructions or process bulletins.

III. TASK II: DESIGN ANALYSIS AND DEVELOPMENT TESTING

Once the framework for design and analysis procedures was established, the
team was ready to move into structural layouts to mature the structural
configuration.
The initial focus on affordability and structure commonality forced the air-
craft design into a configuration that was too heavy and lacked efficient wing
carry though structures load paths. The program went through a weight reduction
effort as described by Counts et al. [2], which resulted in a greatly improved struc-
tures platform; however, this significant change in design created an extremely
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 263

challenging schedule with all three variants concurrently in various phases of the
design process, as represented in the schedule in Fig. 6. The tri-company teaming
arrangement became a very important lever to the success of completing the three
variant designs per schedule.
To enable this monumental task, the structures engineering team needed to
find ways to engage engineering talent from a worldwide pool. Lockheed Martin
and the partners hired engineers from multiple engineering firms around the
world, as depicted in Fig. 7. A common saying on the program was, “The sun
never sets on the F-35 program.” This was enabled by the digital thread mentioned
earlier. All data were continuously updated and available to allow work 24/7. The
framework of ASIP as established in Task I allowed for the inclusion of engineers
from multiple different companies, each with their own methods of design and
analysis, to conform to the standard process used on the F-35 program. The enfor-
cement of the defined criteria and plans ensured consistent design and analyses
processes were used by all participating engineers.
A rigorous and disciplined design maturation process was followed as
depicted in Fig. 8. Elements of those phases are discussed in this section.
The broad scope and complexity of the program presented an organizational
challenge for the F-35 Structures Technology group, which was responsible for
external loads, structural dynamics, and flutter, as well as all vehicle-level Finite
Element Analysis (FEA) and internal loads. The first task was to determine
how to efficiently and effectively transfer external and internal loads data for
three variants to a large number of engineers on multiple teams in various
locations around the world. The form and content of design loads databases
had to be consistent across variants to avoid misinterpretations and ensure

Fig. 6 F-35 structures design timeline.

264 R. M. ELLIS ET AL.

Fig. 7 Worldwide engineering effort.

Fig. 8 Airframe structural maturation.

F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 265

proper use. All types of design loads, regardless of the method used to produce
them, had to be expressed quasi-statically so that differences were transparent
to the design teams and to allow simple translation to ground test loads. Efficient
ways of combining loading actions within the loads databases, such as fuel press-
ures in combination with air loads, were also developed.
Flutter prevention and aeroelastic stability were strongly emphasized early in
the configuration development. Optimization of control surface and empennage
planforms, hinge lines, surface thickness, hardpoint locations, and weapons
pylon geometry was carried out based on flutter and divergence analysis using
a combination of Commercial Off the Shelf (COTS) software and Lockheed
Martin proprietary methods. A dedicated wind tunnel flutter model representing
the F-35C aft fuselage, VT, HT, and rudder was tested in the NASA Langley
Transonic Dynamics Tunnel (Fig. 9) to collect empirical data for correlation
of the flutter analysis models. Data were obtained to help define HT and
VT rotational stiffness requirements; characterize the unsteady aero interaction
among the HT, VT, and rudder; develop transonic dip corrections for use
in the flutter analysis; and allowable HT and rudder free play values. Once the
required control surface rotational stiffnesses were established, flutter team
members worked with design, stress, and flight control hardware engineers
to allocate the required stiffness to each individual part in the most weight-
efficient manner. This process also included the allocation of allowable weights
including reserves for future repairs. Dedicated Finite Element Models (FEMs)
for each variant optimized for dynamic analysis and derived from the master
Air Vehicle (AV) FEMs were used throughout the program. By the Concept
Development Review (CDR) for each variant, all of the designs were shown to
be fully compliant with the aeroelastic stability requirements of the SDC.
Work also began early in the program ensuring that empennage, wing, and
control surface buffet loads were fully accounted for in the initial design for
both strength and life. Buffet caused by forebody vortex interaction with tails,

Fig. 9 F-35C aft fuselage and empennage flutter model in the NASA Langley Transonic
Dynamics Tunnel.
266 R. M. ELLIS ET AL.

as well as wing buffet from unsteady transonic flow, has challenged all recent
legacy fighter designs. This has typically not been discovered until flight testing,
making fixes expensive and disruptive. Early water tunnel testing was performed
to characterize the aerodynamic drivers of tail buffet and point the way for more
thorough wind tunnel tests. Wind tunnel tests were then conducted to measure
unsteady pressures acting on the tail and wing surfaces. Summary of both
water and wind tunnel testing used as the basis of buffet loads development is
shown in Fig. 10. The unsteady pressures, after proper scaling, were used as the
forcing function in a dynamic transient response analysis to calculate resulting
buffet loads. These results were then combined with appropriate steady maneuver
loads and used to generate internal loads for design. As the analysis matured,
aircraft AA-1 was used to conduct a high force-level empennage Ground
Vibration Test (GVT) to measure the extent that VT and HT buffet response pro-
pagated throughout the structure. Results of this test were used to define more
representative structural damping at high force levels for use in design loads
development.
A preliminary full airframe FEM was developed to provide internal load sets
to the structures analysis teams to further mature the major structural load paths.
At this stage in the process the primary focus is on defining a stable, airworthy
platform; however, the service life of the aircraft is also a consideration as dis-
cussed in the Five Pillar process. During this phase, a preliminary service load
spectra is developed to ensure features of the structural arrangement meet all
Service Life requirements.

Fig. 10 Buffet loads development.

F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 267

As the design teams matured the structural arrangement during Phase 1

layouts, part commonality was considered, where possible, to reduce the number
of unique parts and save design span time for all three variants. Material selection
for primary structure was also further developed. During this phase, the systems
design teams received system component definitions from the Vehicle and Mis-
sions Systems teams and identified locations for major systems and reserved
space for routing paths.
The exit criteria for Phase 1 was an airframe architecture with major load
paths established and understood, as well as all major subsystems located and
attachment configuration established.
The external static and service loads continued to be developed as part of
Phase 2. A full aircraft preliminary FEM was constructed using the Phase 1 con-
figuration and updated external loads to provide internal loads for use by the
structural analysis teams to further develop the sizing of the structural parts.
During this phase all subsystems Interface Control Packages (ICPs) were finalized
to ensure adequate space and attachment features were identified. These ICPs also
provide definition of electrical, hydraulic, fuel, and cooling requirements. The
routing paths for these systems were established to ensure adequate clearances
within the structural arrangement. This included wiring harnesses, fuel delivery
tubes, hydraulic lines, and liquid cooling tubes. The effort in Phase 2 produced
a mature, well-defined air vehicle architecture.
During the final phase, all loads information was finalized to incorporate
environmental conditions, such as temperature and vibration. As the parts
became more defined, a fully represented FEM was developed to incorporate all
Phase 2 sizing of structures parts, as depicted in Fig. 11.
Combined with the updated external loads, a full set of internal loads were
published to perform final sizing of all airframe structural parts and assemblies
to facilitate release of the Build-to Packages (BTPs), or drawings. All systems
installation drawings were also fully matured and released. The total number of
drawings released for all three variants was over 55,000. The release of all draw-
ings, as shown in Fig. 12, represented the closure of the structural maturation
process and established the baseline SDD configuration produced and used for
flight testing.

IV. TASK III: FULL-SCALE TESTING

The objective of Task III full-scale airframe ground and flight testing is to provide
data necessary to validate the design strength, durability, vibration, flutter, aeroser-
voelasticity, and other structures analyses to provide a safe operating envelope for
the warfighter and to provide the data needed to manage the service life of the fleet.
A rigorous structural test program was developed to verify contractual
requirements and to produce certification evidence for the strength and durability
of the airframe for each variant. The building block test approach mentioned
268 R. M. ELLIS ET AL.

Fig. 11 Air vehicle finite element models for internal loads.

Fig. 12 Total original BTPs released.

F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 269

previously sets the base leading to the full-scale ground and flight test programs.
The test program was designed to include the following testing:
. Component-level certification tests
. Full-scale static and durability test articles for each variant
W Three full-scale static test articles (fuselage, wing, and vertical tail)
W Three horizontal tail static test articles
W Three full-scale durability test articles (fuselage and wing)
W Three horizontal tail durability test articles
W Three vertical tail durability test articles
W Full-scale nose and main landing gear static tests
W Full-scale nose and main landing gear fatigue tests
. One combined drop test/barricade/live fire test
. Flight test aircraft
W Proof testing and calibration of loads instrumentation: AF-2, BF-3, CF-2
W Flutter, including GVT: AF-1, BF-2, CF-1
W Flight loads, buffet, store ejection response: AF-2, BF-3, CF-2
W STOVL doors and lift system, landing, and ski-jump loads: BF-1
W Shipboard launch and recovery loads: CF-3
. Individual ground tests to aid in design development
These tests are designed to provide the necessary information to address material
characterization development and risk reduction, manufacturing risk reduction,
structural analysis correlation and calibration, material allowables, qualification,
and certification.
The F-35 program tri-variant and tri-company organization and the program
schedule provided unique challenges and opportunities for the full-scale static and
durability testing. Seven dedicated airframes were initially planned as part of
SDD. As the program matured, the dedicated drop test for the carrier variant, per-
formed at Vought Aircraft Industries in Grand Prairie, Texas, was combined
with the static test article, reducing the number of ground test articles to six.
For the first time ever in an aircraft development program, three different variants
were scheduled for testing concurrently. To accomplish this, full-scale testing was
performed at two sites, as shown in Fig. 13. A further cost saving efficiency was to
utilize the same test frame for both static and durability testing. A common test
frame design was used for all static and durability testing with modifications as
needed for unique local testing requirements.

A. STATIC TEST
The F-35 static test program consisted of three full-scale test articles as well as
three independently tested horizontal tails. Each test article was subjected to a
set of critical test conditions to evaluate the strength of the structure at limit
270 R. M. ELLIS ET AL.

Fig. 13 Ground testing at multiple locations.

and ultimate load. In some cases, the measured deflection of the structure was
also a major goal of the test. Each test article was instrumented with strain
gages and load bridges during the build of the test article to collect data with
which to correlate the FEM and to allow live monitoring of the test to ensure
the safety of the article. The sequence of the test conditions was carefully opti-
mized to support the needs of the flight tests and to reduce the cost of setup
changes for the test where possible.
The requirements for flight clearance for the flight test program were also
unique to the F-35. For multiple legacy fighter aircraft platforms, first flight clear-
ance was based on an 80% envelope substantiated structural analysis. The F-35
program only allowed a 40% envelope for flight testing prior to proof or static
test confirmation of the analysis. This resulted from a statistical analysis of the
risk of a safety of flight failure that was based on prior ground test program
results. To stay within the Program Executive authority for flight risks, the envel-
ope was limited to 40% prior to confirmation of the analysis via test. To support
the flight test schedule, proof tests were conducted on the loads-instrumented air-
craft to provide the evidence needed to open the flight test envelope and to provide
calibration data for the flight test instrumentation. Ground vibration and free play
tests were also conducted on the loads and flutter flight test aircraft to assess the
dynamic performance of the air vehicle.
To expand the flight testing envelope and support the flight clearance need
dates, the static testing of all three variants needed to be completed as quickly
as possible. Early decisions regarding the design of the common test frames and
the deployment of real-time monitoring of the test at the partner sites supported
the rapid pace of the tests. It was decided to include sufficient load rams in the test
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 271

frame to be able to move from one test condition to the next with minimum
reconfiguration. This investment in test equipment was offset by reductions in
the duration of the test setup for each condition, thus enabling more efficient
testing. It was possible to progress through many conditions with no physical con-
figuration change to the test frame load introduction hardware. When possible,
multiple critical conditions were combined into “hybrid” conditions and tested
simultaneously. For example, a maximum hammer shock condition, which was
critical for the inlet duct, could be combined with a maximum full airframe con-
dition. Testing in the common test frames and testing concurrently provided the
efficiency required to complete static test milestones well ahead of the flight clear-
ance need dates. The F-35B test was completed first in Fort Worth, and the lessons
learned from that test were used to accelerate the planned test rates for the F-35A
and F-35C tests.
In the case of the F-35A test article, an early portion of the testing was accom-
plished at Fort Worth prior to shipment to BAE Systems in the United Kingdom
to support the needs of the initial flight clearance. The shipment of the test
article to the United Kingdom was a unique experience. During the transport
from Fort Worth to the Port of Houston, there were many UFO sightings
reports as the test article, wrapped to protect it during shipment, made its way
along the highway. Further, the logistics of barging the article to a location
where it could be removed from the barge (Fig. 14) required timing of the tides
of the Humber River.

Fig. 14 Shipment of CTOL static test article to Brough, United Kingdom.

272 R. M. ELLIS ET AL.

Figure 15 shows the pace of the F-35A testing as it accelerated and outpaced
even the ambitious schedule needed to support flight test needs. This was due
in part to the robust airframe, which allowed for a virtually trouble-free test
program, as well as the lessons learned from the F-35B testing accomplished in
Fort Worth.
Although there were some test findings on the article, all three full-scale static
test programs were completed ahead of schedule with no major findings or any
findings that would jeopardize safety of flight. The F-35B static test identified
an interference issue between the inboard weapons bay door when it was under
load that required a redesign for the production articles and a minor trim for
the SDD flight test assets. The other noteworthy finding was the failure of the
auxiliary air inlet door downlock mechanism, which carried less than 150%
DLL prior to failure. This was redesigned for production and SDD flight test
jets and retested independently. The F-35C static test had one noteworthy
finding: The FS 503 frame segment cracked at slightly less than 150% DLL
leading to a repair for the test article and a redesigned part for production and
SDD flight test jets.
The successful early completion of the baseline static test programs allowed
the F-35 program to perform additional expansion testing of the stores hardpoints
to enable future expansion of the F-35’s stores capability. Figure 16 shows the pace
of all variants of the F-35 static test program relative to recent, relevant static test

Fig. 15 AG-1 testing to support flight test need.

F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 273

Fig. 16 F-35 all-variant static test rate compared to legacy aircraft.

programs. The pace of these tests could not have been achieved without a robust,
well-designed airframe.
The three HT static tests were performed in two separate test rigs. This
allowed these tests to be completed well in advance of the needs of the flight
test. The success of the horizontal tail structure at withstanding the limit and ulti-
mate loads during the test demonstrated the robustness of the structure and
afforded the program the ability to stretch the test program and apply loads
that were well in excess of the ultimate loads to support potential future flight
test loads envelope expansion.
The F-35A and F-35C tails were taken beyond the 150% DLL ultimate loads
up to 200% of DLL for multiple conditions. Figure 17 shows the F-35C HT being
loaded to 200% of its design limit load. The HT for the F-35B variant was taken
successfully to 200% of DLL as well and then the test was taken on to failure,
which occurred at greater than 200% of DLL, significantly exceeding the strength
requirements for the structure. This proved to be a very wise investment for the
program because the data were utilized to expand the flight clearance envelope
when later flight test discoveries demanded more capability out of the tails.

B. DURABILITY TEST
The F-35 durability structural test program included three full-scale airframe
test articles, three full-scale horizontal tail component tests, and three full-scale
274 R. M. ELLIS ET AL.

Fig. 17 F-35C horizontal tail maximum down bending test.

vertical tail component tests. It was realized early in the planning that the test
program would need to address load cycling due to maneuver loads and buffet
loads because both were significant contributors to the damage. Coupon-level
testing, as described in the Building Block process, showed that the two components
could be applied in separate, sequential blocks without significantly affecting the
results. The program made the decision to test the full-scale articles with alternating
maneuver and buffet blocks, which represented 1000 flight hours each so that when
the two blocks were performed, the resulting damage represented the damage that
would be accrued by 1000 hours of flight with both loading sources. In the case of
the F-35C, the carrier variant, a third set of load cycling blocks, were applied, which
represented three lifetimes of catapult loads and two lifetimes of arrestment loads.
Each block represented 1000 hours of arrestments and 1500 hours of catapults. This
was done to accelerate the testing program. It was faster to apply the buffet cycles in
a block rather than interspersing them throughout the maneuver spectrum.
Another test efficiency was replacing the VT on the full-scale airframe with a
stiffer “dummy tail” that would not deflect as far under the applied load. This
enabled the loads, in particular buffet loads, to be applied more efficiently to the
vertical tail backup structure. As a further cost savings, a handful of specific features
were tested on the full-scale test article as local tests.
Because of the timing of the durability tests relative to the delivery schedule for
three F-35 variants and the inherent concurrency, the F-35 program had to
address the need to develop both redesigns and retrofit solutions to findings on
the test that were determined to be deficiencies. This concurrency was driven
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 275

by program decisions regarding the production schedule for the F-35, which sup-
ported the affordability of the F-35. To address this concurrency, the program
utilized data regarding the test findings from the recent, relevant durability tests
to predict the quantities and types of findings that the tests would encounter.
This was then used to create predictions of the number and magnitude of
design changes that would need to be incorporated. This plan enabled the
program to accurately estimate the budget needed for these changes and make
plans to support the redesigns and to plan for modification periods to make the
necessary changes to the delivered aircraft. These predictions of findings were
then used as a basis for comparison for the F-35 durability test program.
Another mechanism to address the concurrency was to invoke a rigorous inspec-
tion program and careful monitoring of the test instrumentation to ensure that
test findings were discovered and acted upon at the earliest opportunity.
The F-35 HT durability tests were performed in the two static test fixtures
following the successful completion of the HT static strength tests at the BAE
Systems test lab in Brough, as shown in Fig. 18. These were also tested using inten-
tionally severe test spectra that were applied in separate, repeating 1000-hour
buffet and maneuver blocks. These tests were originally planned to include two
lifetimes (16,000 hours) of testing; however, the Program Office issued a change
to the contract specification to include additional cycling beyond two lifetimes.
All three of the HT test articles successfully completed the three lifetimes of com-
bined maneuver and buffet testing prior to being shipped back to the United States
for teardown inspections.
The F-35 VT durability tests were performed in multiple fixtures at the BAE
Systems Test Lab as well, as shown in Fig. 19. There was one maneuver test rig and
two buffet/vibration test rigs. The maneuver loads were applied quasi-statically

Fig. 18 Horizontal tail test fixtures in Brough, United Kingdom.

276 R. M. ELLIS ET AL.

Fig. 19 Vertical tail test fixtures in Brough, United Kingdom.

using hydraulic rams, and the buffet loads were applied in sound-dampening test
chambers using electromagnetic shakers to excite the desired vibrational modes.
These tests were also performed by alternating the buffet and maneuver loads,
but due to the effort required to switch from one mode to the other, the F-35
program opted to use blocks that were 2000 hours long. Again, the three vertical
tail tests successfully completed three lifetimes of combined maneuver and buffet
testing and were shipped back to the United States for teardown inspections.
The F-35A, AJ-1 durability test encountered roughly half of the findings, as
shown in Fig. 20, that were found on the most recent relevant durability test
program. At the end of two lifetimes, the F-35A durability test had experienced
a little over a third of the findings encountered on recent legacy aircraft. Following
successful completion of the third lifetime of testing, this total had risen to just
under half of the amount of findings experienced on the legacy aircraft durability
test. This small number of findings is a testament to the diligence of the Structures
team, the advancements made in the tools used to analyze the airframe structure,
and the robustness of the crack growth –based service life analysis criteria used in
the development of the CTOL variant. The discovery of these findings prior to
teardown through the planned inspections has allowed the F-35 program to
break in design changes in the early production aircraft to avoid more costly retro-
fits. Additionally, findings on each variant were used for inspections and design
reviews on the other variants to determine if similar structural details were
affected. This allowed for early break-in of changes on those aircraft as well.
The F-35B, BH-1, and F-35C, CJ-1, durability test program discovered findings
consistent with the quantity and rate exhibited by recent legacy programs, also
shown in Fig. 20. The most significant finding during the full-scale durability test
program was on the primary carrythrough bulkhead on the F-35B variant, which
is aluminum rather than titanium as on the other two variants Although this
resulted in a crack through a significant portion of the bulkhead, the wing suc-
cessfully carried limit load several times after the cracking occurred due to the
robust load redistribution capability of the wing structure. Root cause analysis of
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 277

Fig. 20 F-35 durability test findings vs legacy predictions.

278 R. M. ELLIS ET AL.

another crack finding on that part during the second life of testing led to the
discovery that the etch process used to prepare the part for inspection during
production, along with the anodizing process used for corrosion protection
on aluminum, caused very small surface pits that were not adequately captured
by the material properties used for service life analysis. This led to a test program
to characterize the impact of these small surface pits on the aluminum alloys
used in the design of the F-35. This knowledge was then used in a review of all
aluminum parts on all three variants to identify potential life shortfalls and to
break in design changes where required at the earliest point in production that
was possible.
During the tests, as findings are identified and correlated back to test results,
updated designs are incorporated into the earliest production configuration to
reduce the need for modifications to fielded jets. Designs for modifications of
existing aircraft structure are also released to be implemented on delivered jets.
These modifications are then grouped together for the efficiency of installation
and to maximize the availability of the aircraft for the warfighter.
Upon completion of testing, the test articles are carefully inspected and then
completely torn down to inspect all parts and identify the final set of findings that
may need to be addressed. The posttest inspection prior to teardown allows find-
ings to be discovered as early as possible in order to introduce required design
changes at the earliest possible break-in point to reduce the need for modification
after delivery of the aircraft.
As of this writing, disassembly and detail inspection of the F-35B durability
article commenced at the National Institute for Aviation Research (NIAR) in
Wichita, Kansas. The F-35A article is currently being removed from the test
fixture in the United Kingdom and is being prepared for shipment to NIAR in
April 2018 to begin teardown of the article. The F-35C durability test article is
still cycling in its third lifetime.

C. F-35C DROP TEST

A full-scale F-35C drop test program consisting of a series of simulated shipboard
landings was conducted using the static test article CG-1, shown in Fig. 21. The
drop tests simulated aircraft landing conditions—including sink speed, aircraft
attitude (pitch/roll), wing lift force, and wheel spin-up speed—associated with
extreme high energy carrier landings. Full-scale drop testing is typically required
by the U.S. Navy to clear the aircraft for high sink rate landings in flight test. The
test program, conducted in first and second quarter of 2010, consisted of more
than 40 aircraft drops to measure landing gear loads and kinetic energy absorption
characteristics as well as dynamic response of the airframe, stores, and other
critical equipment locations.
High-level objectives of the full-scale drop test were to: 1) verify that the struc-
tural integrity and functional design requirements for shipboard landing impacts
as specified by the F-35 air vehicle structural design criteria were met; and 2) to
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 279

Fig. 21 F-35C drop test.

collect loads and dynamics response data for correlation with analytical predic-
tions. Specific test objectives were:

. Measurement of landing gear loads and energy absorption characteristics,

including the effect of airframe flexibility, under simulated carrier landing
conditions up to the maximum design sink speed.
. Measurement of dynamic response of the airframe, landing gear, stores, and
other large mass items during simulated carrier landings.
. Confirmation of the functionality of the landing gear retract/extend, landing
gear door open/close, and weapons bay door open/close mechanisms after
application of critical carrier landing loads.
. Confirmation of fuel tank sealing after application of critical carrier landing
loads.
. Measurement of internal weapons deflections during simulated carrier
landings.

Results from the CG-1 drop test were excellent. The testing confirmed the validity
of the landing gear and static airframe loads. Pretest landing gear load predictions
showed excellent correlation with measured test values. No changes to the landing
gear metering pin designs were required as a result of the test measurements,
proving the value of the high-fidelity dynamic models developed in the design
phase. A full account of the test and results are found in Ref. [3].
280 R. M. ELLIS ET AL.

D. FLIGHT TESTING
As mentioned in a previous section, the requirements for flight clearance for the
flight test program were also unique to the F-35. To stay within the Program
Executive authority for flight risks, the envelope was limited to 40% prior to con-
firmation of the analysis via test. To support the flight test schedule, the required
static testing and proof tests were conducted on the loads instrumented aircraft
to provide the evidence needed to open the flight test envelope and to provide
calibration data for the flight test instrumentation.
Structural flight testing for all three F-35 variants was conducted during
the SDD program to collect the data required to correlate analytical models
for final certification, as well as to demonstrate specification compliance with
certain requirements that could not be verified analytically, such as control
surface free play effects. Conduct of the flight testing program is described com-
pletely in Ref. [4]. The full scope of structural flight tests were:
. Flight flutter tests to verify requirements of SDC, including damping, and to
collect data for dynamic model correlation
. External flight and ground loads measurements for correlation with models
used for certification
. Aerodynamic buffet and other dynamic loads and response data, also for
model correlation
. Measurement of vibroacoustic environments to verify levels used for design of
structure and qualification of vehicle and mission systems components
Flutter flight tests were performed using AF-1, BF-2, and CF-1 after ground
vibration tests verified aircraft vibration characteristics and provided data for
dynamic model correlation. Dedicated flutter test points were conducted at
increasingly critical airspeeds through an orderly buildup of Mach number at
progressively decreasing altitudes. At each test point, critical structural modes
were excited using a flutter excitation system that perturbed the airframe
through forced oscillation of control surfaces at specific frequencies or at
random. Decay of the structural response was measured to determine damping
at that flight condition. Initial clean wing tests were done in parallel with flying
qualities and propulsion flight tests as part of an integrated envelope expansion
program. Flutter testing with various external stores was also conducted.
Results of the flutter flight testing, as well the final flutter analyses based on cor-
related dynamic models show that all three F-35 variants are flutter free through-
out the required envelope, including margins, and meet all of the damping
requirements of the SDC. No Limit Cycle Oscillations (LCO) were observed
with any of the external store configurations flown. Towards the end of the test
program, flutter flight tests were conducted to significantly expand the horizontal
tail free play limits, increasing the in-service interval between maintenance
actions.
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 281

All maneuver loads flight tests were conducted using three dedicated instru-
mented and calibrated loads jets: AF-2, BF-3, and CF-2. Primary instrumen-
tation comprised groups of strain bridges placed on structural components
and airframe sections corresponding to the locations for which design loads
were generated. The strain bridges were calibrated in a test fixture by applying
distributed loads of known magnitude, measuring strain bridge output, and
developing regression equations relating that output to the applied loads.
These aircraft were also used for buffet and other in-flight dynamic response
tests, particularly ejection of stores and weapons bay vibroacoustics. Flight loads
tests were performed by executing a block of defined maneuvers (pullups, push-
overs, sideslips, rudder kicks, 1-g and elevated-g rolls, etc.) at numerous Mach
numbers and altitudes covering the entire flight envelope. Air loads on each
structural component and section were then extracted and used to correlate
the wind tunnel – derived air loads database. Results were continuously moni-
tored as the flight envelope expanded, compared against the allowable strength
envelopes, and action taken when adverse trends emerged. For example, early
VT loads measurements indicated that limit loads would be exceeded as tests
progressed into the transonic region. To counter this, the flight control laws
were modified to introduce a rudder deflection bias in the critical Mach
number range, relieving the VT air load moment and allowing testing to continue
without having to modify the structure. The strategy of tailoring flight control
laws to alleviate loads was used successfully on all three variants throughout the
loads flight test program. Final loads analyses conducted using test-correlated
air loads databases and the final optimized flight control laws show that loads
on nearly all structural components were within limits, with a few notable excep-
tions described below.
Loads testing unique to the F-35B variant was conducted on BF-1, the test air-
craft dedicated to Short Takeoff and Vertical Landing (STOVL) mode propulsion
and flight controls. These tests were performed to measure loads and environment
on the many in-flight operating doors that make up the STOVL mode propulsion
system: Upper Lift Fan (ULF) door, Lower Lift Fan Doors (LLFDs), Auxiliary Air
Inlet Doors (AAIDs), Roll Control Nozzle (RCN) doors, and three-bearing swivel
module (3BSM) doors. Loads and dynamics testing were typically conducted con-
currently because of the high noise and high vibration environment combining
with steady air loads.
Some of the most challenging flight tests conducted during SDD were to
measure buffet loads. Empennage, control surface, and wing buffet dynamic
response were measured over the full range of flight conditions where buffet
is present. Many of the required flight conditions were at high angles of
attack that were difficult to reach and maintain long enough to gain statistical
confidence in the high-frequency, random responses being measured. The
tests were done with the loads jets using the loads instrumentation augmented
with accelerometers in key locations on the wing, empennage, and control sur-
faces. Buffet data was normally collected during the maneuver loads tests, but
282 R. M. ELLIS ET AL.

many dedicated sustained windup turns and constant AoA descents were also
conducted to gather data at constant flight conditions. Measured response was
used for correlation of the dynamic models and results combined with the
attendant steady loads to yield final results for certification and verification.
The results were generally good with the measured frequencies and modal
response well predicted; however, several cases of higher-than-predicted buffet
loads emerged that required significant mitigation. For all three variants,
rudder hinge moments at certain airspeeds at approximately 20 deg AoA were
underpredicted and required extensive structural analysis and several flight
control law changes to bring them within the limits of the existing structure.
Measured buffet loads on the F-35B AAID were significantly higher than initially
predicted and required structural redesign to mitigate. Redesign consisted of
higher hinge line rotational stiffness to place the rotation mode frequency well
above the excitation frequency as well as increased strength in the door and
attach structure. Finally, significantly higher-than-predicted buffet loads on the
F-35C wing tip AIM-9X hardpoint were measured during transonic windup
turns. After limited progress with other mitigations, a modified wing tip hard-
point design of increased strength was developed and adopted for the final
design. Weapons bay acoustics measured concurrently with the loads, buffet,
and other structural flight tests were as predicted, alleviating early concerns
that a spoiler at the front of the bay might have to be incorporated to control
bay noise.

V. TASK IV: CERTIFICATION AND FORCE MANAGEMENT

DEVELOPMENT
The final phase of the F-35 SDD program includes the verification of the contract
requirements through final certification analysis products, development of fleet
loads tracking (L/ESS) and Individual Aircraft Tracking (IAT) tools, and prep-
aration of the Force Structural Maintenance Plan (FSMP) using the updated
analysis to prepare for execution as defined in Task V.
Each of the F-35 contractual requirements has associated, predefined Success
Criteria (SC) established at the beginning of Task I. Each of these SC is closed once
all supporting reports are signed off by the contractor and F-35 JPO. In many
cases, these reports are updates to Task II design analyses that have been validated
using Task III test results.
The cornerstone of structural certification on the F-35 program is analysis that
has been correlated to test results. There are many effects that are impractical to
simulate throughout the structure during a full-scale ground test, such as system
interface loads, distributed internal pressure loads including fuel head pressure,
and thermally induced internal loads. Distributed aircraft inertia loads must be
simulated by discrete loads applied to the surfaces of parts, which also leads to
local differences between flight and ground test loads.
F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 283

The design analyses performed during Task II were updated during Task IV
by correlation to the results of the full-scale testing that was performed in Task
III. First, the initial design external loads were updated by correlation to flight
test measurements. Internal loads were verified by correlation of finite element
model predictions to strain gauge measurements taken during the full-scale
static tests. Next, the strength analysis was updated to reflect flight test correlated
external loads and the results of the full-scale static tests. The flight test correlated
loads and updated strength analysis were then used to update the Strength and
Summary Operating Restrictions (SSOR) reports for each variant.
Similarly, the durability analysis from Task II will be updated to reflect
flight test correlated external loads and findings during the full-scale durability
tests. Root Cause Analysis (RCA) is performed on each full-scale durability test
finding in order to understand what analysis updates are required. The updated
durability analysis will be summarized in a final DADT Report at the conclusion
of the SDD program.
When the updated Task IV analysis indicated that structural design require-
ments were no longer met, production design changes and retrofit modifications
were undertaken as necessary to restore capability. These changes were typically
minor perturbations of the baseline design and were analyzed using the latest test-
correlated methods. Thus, they were able to be certified without additional testing
of the revised configuration.
As part of this effort, the crack initiation life of aluminum structure in the
STOVL and CV variants was reassessed using updated material properties that
account for the impact of etching and anodizing, as discussed earlier. This corre-
lated approach will be the basis for the certified service life analysis published in
the final DADT report, and for the FSMP. Service life deficiencies resulting from
this analysis update are being addressed as required using production design
changes and structural retrofits.
Completion of the final SSOR and DADT analysis reports are key milestones
that support closure of SC for structures requirements. The closure of all SC sup-
ports closure of the associated JCS paragraphs, which in turn enables SDD
contract closure.
The L/ESS and IAT systems are developed and verified as part of Task IV. IAT
equations are developed for critical control points that are verified or discovered
during the full-scale durability tests. These equations will be used by the PHM
system during Task V to support adjusting maintenance intervals on each aircraft
based on their actual usage severity.
The initial FSMP for each F-35 variant was developed and put in place to
provide the basis for the services to effectively manage the fleets. The FSMP
defines when, where, and what type of maintenance is required, along with esti-
mated costs. The report describes all durability and damage tolerance critical
locations and provides a summary of their analysis inputs.
These L/ESS, IAT, and FSMP reports will be updated on a regular basis as part
of the contractual ASIP process.
284 R. M. ELLIS ET AL.

VI. TASK V: FORCE MANAGEMENT EXECUTION

The final pillar of ASIP is the execution of the Force Management Plan as depicted
in Fig. 22. This is the culmination of all the efforts that have taken place leading
up to this point. Real-life fleet usage data are collected and analyzed to provide
more realistic life estimates for the services. As the pilots become more familiar
with the capabilities of the aircraft, they find where in the flight envelope it per-
forms best and how to best execute the missions that it was designed for. It may
also take on new missions that affect the usage as new strategies and tactics are
employed by the services. These changes may lead to usage that differs from
the original usage spectrum used to project the life of the aircraft. When
enough data are captured, the baseline operational spectrum can be updated for
projecting the service life, inspection intervals, and maintenance timing. The
FSMP, IAT, and L/ESS reports are updated using the analysis of this captured
usage data and experience with the aircraft. The services will then use this infor-
mation to effectively manage their fleets to ensure aircraft readiness with
minimum downtime and reduced cost for preventative maintenance and repairs.
The rigorous, disciplined structural integrity approach followed by the F-35
program allowed the tri-company team led by Lockheed Martin to efficiently
develop and provide certification evidence for three variants of the F-35 Lightning
II. The continued application of the ASIP approach will allow the services to
operate and maintain the structure of these aircraft to ensure their effectiveness
and the safe operation throughout the 30-year/8000-hour service life of the F-35.

Fig. 22 Force life management execution.

F-35 STRUCTURAL DESIGN, DEVELOPMENT, AND VERIFICATION 285

VII. CONCLUSION
The development of the F-35 airframe structure and its lifecycle management plan
has been unique in the annals of fighter airframe development history. The concur-
rent development of three variants with widely differing requirements, to be used
by services with widely differing histories and certification philosophies, provided
unprecedented opportunity for collaboration between contractor and government
engineering organizations, as well as between worldwide industry engineering
organizations. In this environment of such complex demands and global coordi-
nation, the prescription of airframe structural integrity given by MIL-STD-1530
provided the essential framework for executing the design, verification, and force
management of the world’s first 5th-generation multirole fighter weapon system.
Furthermore, with its multinational, multiservice demands driving a unique
consideration of differing structural integrity philosophies, the F-35 program
has gained a unique perspective. Comparison of the varied approaches through
the verification phase of the program has allowed for an unparalleled understand-
ing of the structures’ capabilities, not only for its intended service usage, but also
for whatever variations of service the aircraft may encounter.
With its expansive degree of ground and flight testing, along with develop-
ment of state-of-the-art data-gathering and usage analytics, the F-35 program is
positioned to provide decades of safe and effective use of a 5th-generation high-
performance fighter platform for its many worldwide customers.

REFERENCES
[1] MIL-STD-1530C, “Department of Defense Standard Practice, Aircraft Structural
Integrity Program (ASIP),” Wright-Patterson AFB, 1 Nov. 2005.
[2] Counts, M., Kiger, B., Hoffschwelle, J., Houtman, A., and Henderson, G., “F-35
Air Vehicle Configuration Development,” AIAA Aviation Forum, June 2018
(to be published).
[3] Chichester, R. H. L., and Norwood, D., “Full Scale Drop Test Program for the F-35C
Carrier Variant,” AIAA Paper 2015-0459, 2015.
[4] Hudson, M., Glass, M., Hamilton, T., Somers, C., and Caldwell, R., “F-35 SDD Flight
Testing at Edwards Air Force Base and Naval Air Station Patuxent River,”
AIAA Aviation Forum, June 2018 (to be published).
 CHAPTER 8

F-35 Flight Control Law Design,

Development, and Verification
Jeffrey J. Harris and James Richard Stanford†
Lockheed Martin Aeronautics Company, Fort Worth, TX

The F-35 program will replace a number of aging fighter inventories

and enhance others for the U.S. Air Force, Marine Corps, and
Navy, as well as the inventories of partner/allied countries. Although
the F-35A, F-35B, and F-35C have similar planforms and mission
systems and share many common components, from a flight control
system development perspective, they are quite unique. A novel
approach to both control law design and software development was
required to efficiently develop the control laws for three versions and
ensure that the program met its demanding requirements. The use of
nonlinear dynamic inversion as a viable control law methodology has
been demonstrated on various flight control research platforms;
however, the F-35 is the first production fighter to incorporate a
dynamic inversion control law that operates from zero airspeed to
supersonic speeds, controls the jet well beyond stall angle of attack,
and provides exceptional flying qualities for operational tasks, ranging
from short takeoff and vertical landing operations on U.S. Navy amphi-
bious assault ships and large-deck carrier operations to basic fighter
maneuvers. In addition, the F-35 control law software was developed
based on a graphical model and auto-generated code. This was a key
enabler to support software development for three F-35 versions, but
it required a paradigm shift in the software development process at
Lockheed Martin. This paper explores the in-depth effort of implement-
ing a model-based control law methodology using dynamic inversion. It
also provides an overview of the software development process used for
F-35 flight control laws.

 Senior
Manager, F-35 Control Law and Air Data Design.
†
Principal Engineer, F-35 Control Law and Air Data Design.

287
288 J. J. HARRIS AND J. R. STANFORD

I. NOMENCLATURE
A aircraft dynamics matrix
B control effectiveness matrix
Bcond conditioned control effectiveness matrix
CV control variable (vector)
CVdot control variable rate (vector)
dcmd desired acceleration command
Ddavail available change in effector command
Ddcmd change in effector command
Ddlim limit on change in effector command
Ddpref change in preferred effector command (scheduled)
Ddscale scaled change in effector command
u control vector
Wd weighting matrix, desired accelerations
Wpref weighting matrix, preferred value
x state vector
ẋ state rate vector
ẋdes desired state rate vector

II. INTRODUCTION
The F-35 Lightning II Joint Strike Fighter (JSF) program was designed to provide
the next-generation family of strike fighters. By combining stealth, advanced
avionics, increased range with internal weapons carriage, and state-of-the-art
prognostics and health management, the F-35 provides a lethal, survivable,
connected, supportable, and affordable fighter platform. Although all three F-35
versions share the same general planform, core engine, avionics, and weapons,
there are fundamental differences in the aircraft from a control law development
perspective. As shown in Fig. 1, the aerodynamic differences in wing area and
horizontal tail surface area among the versions, and the differences in basic stab-
ility and control power, are not trivial. Also, the uniqueness of the F-35B configur-
ation for conducting Short Takeoff and Vertical Landing (STOVL) operations
presented a challenge for the control law engineers to achieve an optimal blend
between aerodynamic and propulsion system control effectors. All these aspects
of the F-35 versions had an impact on the task of establishing a preferred
control law structure.
Designing control laws for a new aircraft is a significant task. Designing
control laws for three versions at the same time, without standing up independent
development teams, required creative approaches in order to meet program sche-
dules. A model-based, dynamic inversion approach using auto-generated code
seemed like a viable option. One of the advantages of dynamic inversion is that
the control law designer would not be required to linearize the system at set
F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 289

Fig. 1 F-35A/B/C general characteristics.

flight conditions and develop gain schedules to provide a desired closed-loop

response, which can be very time consuming. Also, auto-generating code
instead of relying on a separate software development team could result in a sig-
nificant reduction in the software development/release cycle time. However, like
any new methodology, it was imperative to demonstrate that the advantages
would outweigh any risks, and that the new methodology could map directly
to historical flight clearance/airworthiness products. The X-35 competition
program provided an excellent opportunity to pursue this development approach.
The X-35 Concept Demonstrator Aircraft program required an efficient control
law design effort and a quick path to releasing flightworthy software. Program
management was receptive to the approach because it allowed aggressive develop-
ment schedules to be met without tripling the staff. Further, the customer was very
open to exploring new control strategies that enabled all three versions to have
similar flight characteristics.
Although early explorations of using dynamic inversion in flight control
applications began appearing in research literature as far back as the early
1980s [1], there are practical limitations to implementing this method in
flyable software, especially for highly nonlinear systems. If it were easy, we
would not still see feedback control dominating the control laws in modern
fighter aircraft. The most obvious challenge is the computing horsepower
required to execute the algorithm. However, other challenges associated with
invertibility of the system, numerical stability, and sensitivity to unmodeled
dynamics also need to be addressed. The Lockheed Martin Control Law Design
290 J. J. HARRIS AND J. R. STANFORD

Group has overcome all of those challenges to deliver a control system that pro-
vides exceptional flying qualities for all three production versions throughout
their flight envelopes. This paper explores some of those implementation chal-
lenges and discusses the development process that was used for the F-35 flight
control system software.

III. CONTROL LAW DESIGN METHODOLOGY

There have been dramatic changes in flight control system design over the past
few decades. We have seen the expansion from analog systems with limited auth-
ority augmentation on stable platforms to full-authority digital flight control
systems that are essential for allowing highly unstable jets to be safely flown.
With the development of more powerful flight control computer processors,
more advanced control methods that put a significant demand on computer
memory and throughput are being developed and flown.
Traditionally, flight control laws were developed as linear controllers with
gain scheduling used to provide desired closed-loop dynamics throughout the
flight envelope. Gains would be developed based on a linear approximation
of the aircraft and systems (aerodynamics, actuators, sensors, engine, and
other subsystems) at various flight conditions under which the jet was expected
to operate. This process can be complex because conventional fighter aircraft
have expanded into highly nonlinear flight regimes, such as high-Angle-of-
Attack (AOA) flight conditions. For the F-35, increased complexity was
added with the STOVL capability on the F-35B, which involves operating
down to zero airspeed, including a transition from aerodynamic controls to
propulsion system controls, and a transition from pneumatic air data sensors
to inertial sensors.
Faced with the challenge of designing control laws for three different versions
operating conventionally, and a version with a STOVL capability, the F-35
Control Law team explored more direct control law methodologies with which
a single control law structure could support all versions, directly accommodate
system nonlinearities, and avoid the need for complex gain scheduling. Nonlinear
Dynamic Inversion (NDI) was a very attractive option to pursue.

A. NONLINEAR DYNAMIC INVERSION

NDI, or feedback linearization, is quite simple in theory and is based on an under-
standing of the equations of motion of the aircraft. If one can predict the aircraft
response to a control effector input (e.g., the aircraft pitch acceleration response to
a horizontal tail deflection), then it should be possible to derive the required
control effector input in order to achieve a desired aircraft response. This
would be achieved by canceling (inverting) the original aircraft dynamics and
defining the desired aircraft response. In linear, state-space form, the dynamic
F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 291

equations of motion can be expressed in Eq. (1) as

ẋ ¼ Ax þ Bu (1)
Where x is the system state vector (pitch rate, roll rate, etc.), A is the aircraft
dynamics matrix (stability derivatives), B is the control effectiveness matrix
(control power), and u is the control vector (horizontal tails, rudders, engine
nozzles, etc.). Solving for the control vector u, using Eq. (2) yields

u ¼ B1 ½ẋ  Ax (2)

Put in terms of deriving the required control vector to achieve the desired state
rates, Eq. (2) becomes Eq. (3)

u ¼ B1 ½ẋdes  Ax (3)

Again, keeping with the simple theory, defining the control vector in this
manner reduces the system dynamics to a simple integrator, as shown in Fig. 2,
eliminating the need for gain scheduling.
The NDI aspect of the control laws and the plant model shown in Fig. 2
represent how the aircraft actually behaves, with all nonlinearities included. The
desired state rates reflect how the control law designer wants the aircraft to fly
in response to pilot inputs. The commands module translates the pilot’s stick,
throttle, and pedal inputs into desired aircraft movement. The regulator directly
sets good low-order-equivalent system parameters to ensure classical handling
qualities for the pilot while achieving this motion. This is generally defined in
terms of traditional flying qualities parameters, such as short-period frequency
and damping, or Tier II flying qualities metrics, such as pitch attitude bandwidth,
Neal-Smith criteria, and so forth. Those flying qualities goals are embedded
directly in the F-35 regulator because the remaining control architecture creates
the inversion, as shown in Fig. 3.
An integral part of the NDI structure shown below is the Effector Blender
(EB), which performs the control effectiveness matrix inversion and replaces
the traditional control surface mixer in the control laws [2]. Whereas a traditional
mixer uses a set of gain schedules to map an axis command to the control surfaces,
the matrix inversion typically involves a real-time, pseudo-inverse iterative calcu-
lation and logic to account for surface rate and position limits and other perform-
ance constraints, as shown in Fig. 4.

B. IMPLEMENTATION CHALLENGES
As mentioned earlier, NDI is not a new concept. It has been successfully demon-
strated working on a number of research aircraft, ranging from the F/A-18
Hornet fighter to helicopters [3, 4]. However, these research efforts were generally
focused on very limited flight envelopes to demonstrate a concept implemen-
tation. The F-35 program adopted NDI as an approach for the full operating
 292

Fig. 2 Incremental nonlinear dynamic inversion decomposition.

J. J. HARRIS AND J. R. STANFORD
F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 293

F-35 control law structure.

Fig. 3
294 J. J. HARRIS AND J. R. STANFORD

Fig. 4 Effector blender solution algorithm.

envelope, including high-AOA and STOVL operations, carrier landings, refueling,

and similar operations. The design had to be robust to the aerodynamic effects of
external store carriage and the significant range of a longitudinal and lateral
Center of Gravity (CG) that was possible with the stores that the aircraft was
designed to carry (Fig. 5). In addition, on-ground modes, reconfiguration for
system failures, and out-of-control recovery modes would also need to be accom-
modated by the control law architecture, while addressing the practical aspects of
real-time computer utilization for computationally intensive algorithms.
The decomposition of the NDI controller wrapped around the airframe
(plant) model into a simple integrator relies on the precise modeling of the aero-
dynamics of the aircraft, perfect sensors, and the ability to invert the control
effectiveness matrix B. The reality is that onboard models of the aerodynamics
and propulsion system are never perfect, and higher-order effects can be difficult
to accurately model. These effects will result in not achieving a perfect cancella-
tion of plant dynamics, and this mismatch can have a significant impact on
closed-loop dynamics, as discussed later in this paper. In addition, there are
numerical challenges associated with an algorithm that requires the inversion
of the control distribution matrix. These challenges would have to be overcome
to successfully implement an NDI-based control law structure in a production
aircraft. In addition, taking this design approach from the research environ-
ment to the production environment also has implications on system test
F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 295

requirements (both on-ground tests and flight tests), as well as airworthiness

assessment impacts.

C. NUMERICAL CHALLENGES
Early in the development of the F-35 control system (prior to flight testing), chat-
tering in the actuator commands was observed as toggling between solutions
occurred. This was generally the result of ill-conditioned control power data
(large changes in control effectiveness, local maxima/minima generated due to
nonmonotonic changes in control effectiveness, singularities, etc.). Nonmono-
tonic control power data can result in zero control-effectiveness derivatives
being generated by the onboard model, with the result being an effector not
moving from a local maxima or minima, with potentially significant effects on air-
craft flying qualities. Figure 6 shows a case in the simulator in which a local
maxima in horizontal tail control power resulted in a pitch departure.
Control effectors with comparable effectiveness may result in singularities
that lead to undesired control surface usage, such as two surfaces opposing
each other and impacting maneuvering capability, as shown in the simulation
data in Fig. 7.

Fig. 5 The F-35 is designed to carry a large complement of internal and external weapons.
296 J. J. HARRIS AND J. R. STANFORD

Fig. 6 Control effectiveness monotonicity.

All of these challenges required protections being put in place to preserve

algorithm stability and achieve the desired flying qualities for the F-35 prior to
starting the flight test program.

D. AIRCRAFT LOADS AND THE ELECTRO-HYDROSTATIC ACTUATION SYSTEM

The EB architecture allows the designer to utilize all of the available control power
from the air vehicle. The designer no longer has to define a priori the optimal way
to move the surfaces at each flight condition, but still has the freedom to define
preferences of one surface over another based on other constraints, such as aircraft
maneuver loads or aircraft performance. Integrating controller preferences into
the EB architecture provides the control law designer with an easy means to
reconfigure for system failures that result in the loss of surfaces by simply remov-
ing them from the solution. However, with that freedom comes a potential

Fig. 7 Ill-conditioned control-effectiveness matrix.

F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 297

challenge. For example, fighter aircraft have traditionally used their wing and tail
surfaces in a fixed ratio to stay within airframe loads limits. The EB does not
inherently obey that rule and had to be constrained to a fixed ratio in order to
stay within airframe loads limitations, especially at high dynamic pressures where
there is generally a delicate balance between maximizing maneuvering capability
and remaining within structural envelopes. This constraint was achieved by
simply forming virtual effectors, such as combining all the rolling surfaces (asym-
metric flaperon, asymmetric aileron, asymmetric horizontal tail) into a single
virtual rolling surface. This reduces the degrees of freedom in the solution and
keeps the asymmetric tails, flaps, and ailerons in fixed ratios to each other. As
loads increased on different surfaces in various portions of the envelope, the
ratios could simply be changed to stay within allowable loads limits while preser-
ving roll capability.
The F-35 is also the first production aircraft to incorporate an Electro-Hydro-
static Actuation System (EHAS) to replace conventional, hydraulically driven
actuators, as shown in Fig. 8 for the F-35A version. This actuation system elimi-
nated the need for a large, centralized hydraulic system by incorporating electri-
cally driven, hydraulic power that is packaged as part of each individual actuator.
This technology was a key enabler for the current aircraft configuration, and it
yielded significant aircraft reliability and vulnerability improvements, as well as
weight and cross-section reductions. However, from a control law perspective,
the actuator load capability was significantly decreased compared to conventional

Fig. 8 Electro-hydrostatic actuation system (F-35A).

298 J. J. HARRIS AND J. R. STANFORD

actuators, as shown in Fig. 9. In earlier-generation aircraft, actuators were typi-

cally sized to meet stiffness requirements, resulting in an excess load capability
with less than 50% of the total stall capability used for general maneuvering.
This provided large margins to actuator stall. The F-35 EHAS was designed to
balance stiffness and load capability, the result being an actuator operating at
closer to 90% of the capability for that same load, with a rapid loss of rate capa-
bility above 95% load. To accommodate these actuator characteristics, the F-35
control laws were tuned to use no more than 95% of the stall capability to
achieve desired maneuvering performance. This was an additional constraint
that required optimization of the EB to stay within actuator load restrictions.
The hinge moment loads were monitored during flight test, with updates to
predicted hinge moments based on flight test results. Tailoring the response to
only use up to 95% of the stall capability was only manageable due to the fact
that the Electro-Hydrostatic Actuator (EHA) retained significantly more rate
capability at 95% stall compared to a legacy actuator, as shown in Fig. 10. In
some cases, surface command limits had to be tailored to stay within hinge
moment limits as the flight test envelope was expanded during the course of
the flight test program.
The EHAs also have significantly different characteristics when failed, com-
pared to conventional actuators. In dual-redundant actuators, the imbalance in
piston area generally causes the surface to drift up when failed until it balances
with the external hinge moment loads. On the trailing edge flaps, the total
upward deflection was manageable, but not on the horizontal tails. On the
F-35A and F-35C versions, the failed tail would drift up more than what the
unfailed tail could overcome, which could result in a departure. This situation
was overcome by the addition of a small centering actuator that the control

Fig. 9 Typical load-rate characteristics for an EHA vs a conventional actuator.

F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 299

Fig. 10 Percentage of no-load rate vs percentage of stall for an EHA vs a

conventional actuator.

laws could engage that would push the failed tail down to an optimal location at
low speeds and then hold it in place. The F-35B version did not require a centering
actuator because it can convert to STOVL mode and use the engine/lift fan thrust
split to compensate for any pitching moment from the failed surface, eliminating
the weight penalty of having to carry an additional actuator on the aircraft.

E. STRUCTURAL COUPLING CHALLENGES

One of the greatest challenges with the EB and onboard model architecture was
preventing the control system from adversely responding to feedback sensor
output resulting from structural deflection, called structural coupling. This is his-
torically tested by setting the aircraft in the highest gain flight condition and per-
turbing the control surfaces/propulsive effectors to excite the structural modes,
then measuring the loop characteristics of the path through the structure,
sensors, and control system. In Fig. 11, random noise is injected into the actuators,
and the effects on the aircraft sensors are measured as well as the total control law
output. These measurements are used to assess stability when the loop is closed
and verify the structural mode margin requirements in MIL-STD-9490.
For conventional control systems, the path gain from a sensor through the
control laws is relatively easy to calculate. Control system gains are usually
fixed at any given set of flight conditions, and determining the flight condition
for maximum gain can be as simple as examining control law gain tables. For
the F-35 control laws, the EB is iterating each frame to determine an optimized
solution based on control surface effectiveness, as well as other constraints,
such as effector rate/position limits and system failures, as shown previously in
Fig. 4. At low airspeeds, where the control gains are usually the greatest, the EB
300 J. J. HARRIS AND J. R. STANFORD

Fig. 11 Structural coupling testing.

has the freedom to change the surfaces it uses to control the aircraft and to
increase the command on one surface when other surfaces hit a rate/position
limit. This makes it nearly impossible to pick a single flight condition to evaluate
during structural coupling tests. In order to ensure that the aircraft was free from
structural coupling, actuator-to-sensor transfer functions were measured on the
aircraft (like the one shown in Fig. 12 for pitch rate-to-horizontal tail) during
testing and then linearly combined with possible control systems gains through-
out the flight envelope posttest to verify structural margin. Very detailed linear
models were generated and compared to on-aircraft test results to ensure accuracy
of the airframe model. Figure 12 includes a comparison of the open-loop response
measured during a structural coupling test to results achieved analytically by com-
bining the derived airframe response with the linear model of the control system
at the tested flight condition. With an excellent correlation between the analytical
response and the measured response during testing, the Control Law team

Fig. 12 Open-loop structural response.

F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 301

successfully derived the required structural filters to achieve adequate stability

margins. No instances of aero-servo elastic instabilities were observed during
the extensive F-35 flight test program.

F. ACTIVE STICK AND THROTTLES

All three F-35 versions have unique mission requirements. Satisfying these mis-
sions with common hardware is an obvious benefit from a system architecture
perspective, and also reduces maintenance and supply chain demands. In the
cockpit, one key enabler to achieve this commonality was the use of active incep-
tors for the stick and throttle. The Active Inceptor System (AIS), shown in Fig. 13,
allows the control law designer to program various features of the stick and throt-
tle (e.g., setting force vs deflection, damping, soft stops, gates, and even ramps) as a
function of flight phase in order to satisfy the unique missions of each version. The
throttle motor is also used to back-drive the throttle to a commanded position
when automatic command modes are engaged.
These features are used extensively on the F-35B for STOVL mode landing.
Translational Rate Command (TRC) is an augmented STOVL command mode
that uses the throttle inputs out of a center detent to command a change in
forward velocity. With the throttle in the detent position, the control system is
holding a reference velocity. This mode, along with vertical/lateral TRC modes
on the stick, allows the pilot to easily fly STOVL approaches to a ship or shore-
based landing pad. When the axial TRC mode is engaged, the throttle is pro-
grammed to back-drive to a center location, and ramps are programmed on
each side. During approaches to a ship, the pilot enters the ship’s speed as the

Fig. 13 Active inceptor system.

302 J. J. HARRIS AND J. R. STANFORD

Fig. 14 Programmable throttle features for STOVL landing.

reference velocity for the center detent position. If the pilot pushes the throttle out
of the detent and onto the forward ramp, an incremental increase in forward vel-
ocity is commanded. Pulling back on the throttle decreases the velocity. When the
pilot gets to the desired position alongside the ship, simply letting go of the throt-
tle will hold that position co-speed with the ship. The roll stick similarly holds
lateral position when released and commands a change in lateral velocity when
deflected. Fore–aft deflections of the pitch stick command changes in height
rate. When the pitch stick is released, the control laws hold the current height.
When touchdown occurs, the throttle automatically back-drives to the idle
position where the control laws command the engine to ground-idle thrust.
The programmable features of the throttle for STOVL landing operations are
shown in Fig. 14.
Although the programmable features of the stick and throttle allow the control
law designer a significant amount of flexibility to tailor the characteristics for
various flight phases and modes, they also increase the complexity of the reconfi-
guration logic within the control laws to accommodate system failures. The F-35
control laws were designed to accommodate the stick or throttle going into passive
mode, where all the programmable features are lost and the inceptors operate at a
basic default spring rate. The control laws also have to accommodate instances in
which the stick or throttle may become jammed and the pilot must be able to
control the aircraft simply by applying force.

G. AERODYNAMIC MODELING CHALLENGES

As mentioned earlier, dynamic inversion relies on the cancellation of the plant
dynamics, including system nonlinearities. At its core, the NDI controller contains
a model of the aerodynamics of the aircraft (including entrained aerodynamics
F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 303

from the propulsion system) and a model of the aerodynamic and propulsion
effectors. In order to minimize performance errors resulting directly from poor
modeling, it was important to have a very detailed onboard model of the F-35
across the full operating envelope, and to account for discrete configuration
changes associated with gear doors, STOVL doors, and weapon bay doors, as
well as interactions between control surfaces. Parts of the flight envelope where
aerodynamic characteristics can change rapidly (e.g., the transonic flight
regime) require higher-fidelity modeling, with denser Mach and AoA breakpoints.
Of course, even with improvements in the memory and throughput capacity of
current flight control computers there are practical limitations on the size of
models that can be embedded in the software. The current F-35 onboard
models consist of approximately 3 million data points. The original models
were derived from wind tunnel testing.
Although the tunnel testing for the F-35 was probably the most extensive
testing conducted on a fighter program, there were still differences identified
between the aerodynamic model derived from wind tunnel testing and the
actual aircraft aerodynamics. One example of this was discovered during early
envelope expansion testing in the transonic flight regime. While conducting an
elevated g roll, a large sideslip excursion developed that led to a roll-reversal
(Fig. 15), a maneuver response that was not predicted by simulation.
The primary cause of the excursion was traced to errors in the aerodynamic
modeling of roll damping and control surface effectiveness in this part of the
flight envelope, and updates to the truth aerodynamic models used in the simu-
lation and the onboard model in the control laws were required to correct this
response. Figure 16 shows the key aerodynamic model parameter changes that
were made and the flight test response following the control law update compared

Fig. 15 Sideslip/roll excursion during an elevated g roll.

304 J. J. HARRIS AND J. R. STANFORD

Fig. 16 Complex model updates were required to correct mispredictions.

to the original response, demonstrating that the roll-reversal was eliminated and
the sideslip response was well-behaved following the software update.
Another potential challenge to an NDI implementation was associated with
the practical limitations of modeling the complex aerodynamics that are typically
encountered by fighter aircraft in the transonic, elevated G/AOA flight regime.
This part of the envelope is associated with very complex flow fields that can
lead to unsteady aerodynamics resulting in an asymmetric flow separation
capable of inducing uncommanded yaw and roll motions. Abrupt Wing Stall
(AWS) has been a challenge for fighter aircraft for decades, with documented
examples from the 1950s-era F-84 Thunderjet to modern-day fighters [5]. This
flow separation is typically characterized by discontinuous rolling/yawing
moment coefficients at discrete AOA and/or sideslip, and hysteresis in those dis-
continuities across a sweep of AOA and/or sideslip. It would be impossible to
capture these types of nonlinearities and discrete jumps in coefficients in the
onboard aerodynamics model. If left uncorrected, however, the resultant transi-
ents could be quite noticeable to pilots, as shown in Fig. 17.
It is not possible to directly model these nonlinearities, so an alternate
approach to improving the robustness of the algorithm was incorporated into
the F-35 control laws. Approximations of the model error based on measured par-
ameters were used to derive an additional acceleration error signal as an input to
the EB, as shown in Fig. 18. This implementation was originally developed to
accommodate modeling errors that were observed in the transonic flight regime
during initial envelope expansion flight testing, but it has been shown to be effec-
tive at compensating for other areas of the flight envelope where modeling errors
F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 305

can be expected, such as high AOA, where airflow separation can challenge the
effectiveness of an NDI-only control law.
Although the transonic flight regime can be extremely challenging because
of rapidly changing aerodynamics as shocks migrate across the aircraft, the
F-35 Control Law team had to deal with even more complex issues on the road
to clearing the F-35B for STOVL operations. The conversion between conven-
tional and STOVL modes involves opening/closing 11 different doors, some of
which have an obvious impact on the aerodynamics of the aircraft, as Fig. 19
demonstrates.
Prior to engaging the lift fan system for the first time, it was important to
demonstrate that the aircraft could be safely landed in conventional mode with
the doors failed open. Initial flight testing with STOVL doors open revealed
several modeling issues, including the basic directional stability of the jet, as
shown in Fig. 20.
The level of model discrepency that was identified during this initial phase of
testing had a significant impact on the planned flight test program and impacted
the ability of the test team to proceed to engaging the STOVL propulsion system;
however, it also highlighted just how complex the flow fields are around the
aircraft in this configuration and how challenging the modeling task would be.
An extensive effort was necessary to conduct the flight testing required to first
capture the data that could be used to update the aerodynamic models and
then resolve the issue. This required multiple fly-fix-fly cycles to eventually
clear the full operating envelope for all required STOVL door configurations
while allowing incremental envelope expansion to proceed.

Fig. 17 Uncommanded roll/yaw as a result of unsteady aerodynamics.

 306

Fig. 18 Augmented NDI control law structure.

J. J. HARRIS AND J. R. STANFORD
F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 307

Fig. 19 F-35B STOVL doors.

H. SOFTWARE DEVELOPMENT PARADIGM

With the incorporation of the NDI controller structure into the F-35 control laws,
the role of the control law engineer was fundamentally changed. The team was no
longer focused on developing gain schedules based on linear models, which was a
traditional control law engineer task. Now the team was much more deeply
involved in understanding the aerodynamics of the aircraft, working closely
with stability and control engineers to develop the onboard model, which is an

Fig. 20 Modeling challenge associated with STOVL doors.

308 J. J. HARRIS AND J. R. STANFORD

integral part of the control laws. In order to streamline the control law software
development process, the F-35 team also chose to implement a graphical, model-
based approach to control law development and used MATLABw Autocode to
auto-generate code. The traditional approach of control law engineers developing
S-plane diagrams and turning those over to a separate mechanization team to
recode them into the Operational Flight Program (OFP) code loaded onto the
flight control computers (Fig. 21) has been replaced with control law engineers
who are also filling the role of the embedded software engineer, responsible for
software development in accordance with established industry standards for
safety-critical systems (Fig. 22).
In order to be successful in this approach, an unprecedented level of com-
mitment and coordination by the Control Law, Stability and Control, and Soft-
ware Development/Test teams was required. A rigorous approach to large-scale
model development and configuration management was implemented, including
the development of modeling standards, so that auto-generated code would
comply with traditional software development guidelines (e.g., MISRA-C) and
satisfy established software safety standards for airborne software (e.g., DO-178B,
Defence Standard 00-56, MIL-STD-882) [6]. The overarching goal was to
develop reliable, safe, and maintainable software.
The advantages of using a model-based software development process
extended beyond just the cost savings associated with reducing the size of a sep-
arate team of software coders. Automatic code generation significantly reduced
the frequency of coding defects that are typically introduced when manually trans-
lating requirements into design and code in a traditional software development

Fig. 21 Traditional software development process.

F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 309

Fig. 22 F-35 control law software development process.

process. Also, the cycle time for developing code is reduced. Because the same
code that is generated for the OFP is integrated into the offline and piloted simu-
lations, verification of the control laws can be conducted much earlier in the devel-
opment cycle and in multiple simulation environments.
These benefits were critically important during the JSF concept demonstration
phase, when the Lockheed Martin team successfully executed a flight test demon-
stration program with the X-35 using a common control law structure for three
aircraft versions using a single OFP. However, as critical as this development
approach was to meeting the aggressive timelines associated with an X-plane
program, it proved to be an even more significant enabler to meeting the demand-
ing requirements associated with a production development flight test program
for three versions.

IV. VERIFICATION
Verification of the F-35 control law development approach has been accom-
plished through a number of methods. Although flight testing of the F-35 is the
most visible form of verification evidence, the original verification product is
the desktop graphical model. As shown in Fig. 23, the validation of the design
and testing at the desktop level forms the backbone of the OFP product that is
then tested in simulators and hot-bench (hardware-in-the-loop) test facilities
prior to being loaded onto a test aircraft and flown. All these venues have
played an important role in validating the NDI structure of the F-35 control
310 J. J. HARRIS AND J. R. STANFORD

Fig. 23 F-35 software testing.

laws and the software development approach that includes auto-generation of the
flight control OFP code from a graphical model of the control laws.
The elements shown in Fig. 23 span more than 20 years of development, either
directly supporting the X-35/F-35 programs or supporting various technology
demonstration programs and company research pursuits. The result of that
effort is an air vehicle that meets all fighter requirements of all U.S. services,
partner countries, and participating allied countries, while delivering significantly
more capability than legacy aircraft.

V. CONCLUSION
The F-35 is a fighter program that consists of three versions, each with distinct
mission requirements and aerodynamic characteristics. The STOVL version, in par-
ticular, has unique propulsion system characteristics as well. The Lockheed Martin
Control Law design team implemented a novel and cost-effective approach to meet
the demanding requirements of this program. The team was able to overcome a
number of early design challenges to implement a robust, model-based dynamic
inversion control law that provides exceptional flying qualities throughout the
flight envelope, including high-AOA operations, supersonic flight, shipboard oper-
ations, and the demanding, low-speed envelope associated with STOVL operations.

ACKNOWLEDGMENTS
The authors would like to thank the entire Lockheed Martin Flying Qualities team
for the work of its members over the years leading up to the delivery of the F-35
F-35 FLIGHT CONTROL LAW DESIGN, DEVELOPMENT, AND VERIFICATION 311

aircraft that is in operation around the world today. This team includes both
stability and control engineers and control law design engineers. The stability
and control engineers developed the aerodynamic models from wind tunnel
tests and the onboard aerodynamic models that are such an integral part of the
F-35 control laws. The control law design engineers were willing to embrace chal-
lenging new technologies for aircraft control and make them work! A special
acknowledgment to members of the team whom we have lost over the years:
John McCune and Donald Inman.

REFERENCES

[1] Enns, D. F., et al., “Control Design and Flight Hardware Implementation Experience
with Nonlinear, Dynamic Inversion Control (NASA-2) for the F/A-18 HARV,”
High-Angle-of-Attack Technology Accomplishments, Lessons Learned, and Future
Directions, NASA/CP-1998-207676/PT1 (may be released to U.S. persons with
NASA approval).
[2] Bordignon, K., and Bessolo, J., “Control Allocation for the X-35B,” AIAA Paper
2002-6020, 2002.
[3] Miller, C. J., “Nonlinear Dynamic Inversion Baseline Control Law: Flight Test
Results,” AIAA Paper 2011-6467, 2011.
[4] Bosworth, J., and Williams-Hayes, P., “Flight Test Results from the NF-15B Intelligent
Flight Control System (IFCS) Project with Adaptation to a Simulated Stabilator
Failure,” NASA/TM-2007-214629, December 2007.
[5] Chambers, J., and Hall, R., “Historical Review of Uncommanded Lateral-Directional
Motions at Transonic Conditions,” Journal of Aircraft, Vol. 41, No. 3, May/June 2004.
[6] Bridges, M., “JSF Software Safety Process: Providing Developmental Assurance,”
Systems and Software Technology Conference, Tampa Bay, FL, 2007.
 CHAPTER 9

F-35 Propulsion System Integration,

Development, and Verification
Steven P. Wurth and Mark S. Smith†
Lockheed Martin Aeronautics Company, Fort Worth, TX

At the heart of every successful aircraft is a successful propulsion system

design. The X-35 and the F-35 aircraft developments are at the top in the
history of aviation. The concept of one aircraft design to meet the
requirements of the Air Force, Marines, and Navy warfighter is not a
new concept. What distinguishes the X-35 and F-35 programs is their
success in developing a propulsion system design that meets
performance, reliability, and affordability requirements. The success of
the F-35 propulsion system can be attributed to developing verifiable
requirements, conducting cohesive systems engineering, and quickly
overcoming development challenges. The successful ground and flight
testing of the unique propulsion system led to groundbreaking flight
test programs. The F-35 propulsion system continues to successfully
power the F-35 fleet well past its current 135,000 flight-hour mark.

I. INTRODUCTION
This paper presents the development of the F-35 propulsion system following a
three-phase development life cycle. The development is similar to the design of
any major engineering effort, with conceptual design, preliminary design, and
detailed design phases. The conceptual design phase centered on the test and ver-
ification of the concept’s feasibility. It included Advanced Short Takeoff and Ver-
tical Landing (ASTOVL) and a Large-Scale Powered Model (LSPM). The
preliminary design phase also entailed test and verification but was focused on
assessing functional feasibility. For the X-35, this also included a Concept Demon-
strator Phase (CDP). For the detailed design phase, test and verification reflected
operational feasibility and included the System Development and Demonstration
(SDD) Phase for the F-35. The phases proceeded chronologically from early
ASTOVL through the F-35 SDD program.

 Lockheed
Martin Technical Fellow, Integrated Flight Propulsion Control, AIAA Sr. Member.
†
Lockheed Martin F-35 STOVL Chief Engineer (Ret.).

313
314 S. P. WURTH AND M. S. SMITH

II. F-35 PROPULSION SYSTEM OVERVIEW

Developing a propulsion system that supports all three F-35 variants is a monu-
mental task. The F-35A requires a Conventional Takeoff and Landing (CTOL)
capability, the F-35B requires a Short Takeoff and Vertical Landing (STOVL)
capability, and the F-35C Carrier Variant (CV) must withstand the rigors of
naval carrier operation. Each of these supports a branch of the military: F-35A
for the Air Force, F-35B for the Marines, and F-35C for the Navy. In order to
test the concept of one propulsion system that could satisfy all propulsion require-
ments for the three services, the STOVL system needed a special focus. The pro-
pulsion system was designed as a tri-variant solution to maximize commonality
and meet affordability goals.
The CTOL and CV propulsion systems are identical, whereas the STOVL pro-
pulsion system uses a CTOL/CV turbomachine coupled with STOVL-unique
components. These components are a LiftFanw, clutch, driveshaft, roll post
system, three-bearing swivel duct (3BSD), and STOVL exhaust nozzle. Previous
efforts under the Defense Advanced Research Projects Agency (DARPA)
allowed collaboration among Lockheed Martin, Pratt & Whitney, and Allison
Advanced Development Co. (now LibertyWorks, a Rolls-Royce division). This
team developed a STOVL propulsion system with a main engine that powers a
dual-stage LiftFan via a shaft connected to the low-pressure spool of the main
engine. Figure 1 shows the layout of the F-35 propulsion systems.

Fig. 1 F-35 CTOL/CV and STOVL propulsion system features.

F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 315

Fig. 2 F-35B STOVL mode effectors overview.

Figure 1 illustrates the basic differences between the CTOL/CV and STOVL
propulsion systems. Among these is a shaft-driven LiftFan for thrust augmenta-
tion in STOVL mode, driven via a clutch/driveshaft from the main engine. The
STOVL variant also has a unique LiftFan exhaust nozzle: the Variable Area
Vane Box Nozzle (VAVBN). In addition, it has an auxiliary inlet for improved
main engine airflow for STOVL operation and a pitch and yaw vectoring 3BSD.
Its STOVL main engine nozzle is slightly shorter than that of the CTOL/CV
variant. A similarity between the two propulsion systems is the main engine
Diverterless Supersonic Inlet (DSI). In it, the diffuser section for the STOVL
variant is modified by the presence of the auxiliary inlet.
A critical feature of the STOVL propulsion system is its control of the air
vehicle during transition and conversion, low-speed operation, and hover. In tran-
sition, the aircraft transforms from conventional flight mode to STOVL mode,
including flight down to a stationary hover. During conversion, the propulsion
system is reconfigured from a conventional CTOL/CV propulsion system to a
STOVL propulsion system with an operating LiftFan. Figure 2 shows the
316 S. P. WURTH AND M. S. SMITH

various STOVL effectors and their physical range of operation. STOVL control of
the vehicle during conversion and transition, hover, and vertical landing was
developed using Control Laws (CLAW). The CLAW govern the engine’s Full-
Authority Digital Engine Control (FADEC) and the aircraft’s vehicle management
computer [1].
Figure 2 also introduces an additional characteristic of the STOVL variant: its
unique doors and inlets that allow airflow into the STOVL propulsion system. The
STOVL variant has a three-position LiftFan inlet door that transitions based on
vehicle speed, a LiftFan exhaust door for the VAVBN, roll post doors, and
3BSD doors. These doors were an integral part of the propulsion integration of
the STOVL variant.
The following sections detail the basic design progression of the F-35 propul-
sion system. They illustrate challenges, discuss successes, and provide the history
of the design, development, and verification of the F-35 propulsion system.

III. CONCEPTUAL DESIGN, TEST, AND VERIFICATION FOR CONCEPT

FEASIBILITY: ASTOVL/LSPM
The ASTOVL program, described further in Ref. [2], included the construction of
a 91%-scale LSPM. This model (Fig. 3) established concept feasibility by assessing
such factors as:
. Low-speed and powered lift aerodynamics
. Hover and low-speed control power
. Hot Gas Ingestion (HGI) characteristics
. Propulsion system layout and general integration
The result was a very successful 160-hour ground test program executed at
NASA’s Ames Research Center. The LSPM had a representative propulsion
system and was evaluated at
both static and low-speed con-
ditions [3].

A. KEY CAPABILITIES AND

DESIGN FEATURES
One objective of the LSPM was
to show the feasibility of harnes-
sing the low spool of a main

Fig. 3 Lockheed Martin ASTOVL

LSPM ground demonstrator.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 317

engine to a bevel gearbox driving a fan. Another was to explore the STOVL aero-
dynamics of the air vehicle. To minimize costs, various components from F100
and legacy engines were used to develop this unique propulsion system. The
LSPM had no clutch or actuating front nozzle, but had an actuating aft nozzle.
The LSPM LiftFan was the first-stage fan and Inlet Guide Vanes (IGVs) from a
legacy engine. The LSPM LiftFan had only a single stage, compared to the X-35
and F-35B’s two-stage LiftFan. The performance of the assembled LiftFan,
gearbox, and driveshaft were at a power level representative of a two-stage
design. The main engine combined the fan and high-pressure core of an
F100-PW-220 and the low-pressure turbine stage from an F100-PW-229. The
engine main fan rotor was modified to attach the driveshaft to the LiftFan. The
engine case was modified so that the bypass air could be diverted to the ducts
that supply the roll control jets [4].
A bifurcated F-22-like inlet was used, along with the LiftFan inlet and auxiliary
door inlets that had been developed for the LSPM. A two-dimensional variable
area thrust-deflecting nozzle was used for the rear main nozzle. The FADEC soft-
ware was modified to control fuel flow and nozzle area as a function of the STOVL
operating line of the turbine map [4].
The propulsion system was demonstrated at Pratt & Whitney’s facility in
West Palm Beach, Florida. The performance of the main engine’s Air Induction
System (AIS) was demonstrated under hover and low-speed conditions at
NASA’s Ames Research Center. Those demonstrations also showed the LiftFan
AIS’s performance under hover and low-speed conditions. Further, they investi-
gated the propulsion effects on external aerodynamics, hover and low-speed aero-
dynamics, and HGI characteristics.

B. VERIFICATION TESTING
The demonstration of the first shaft-driven LiftFan and main engine operation
occurred at Pratt & Whitney’s outdoor engine test facility in West Palm Beach.
More than 40 hours of static testing were accomplished with very few problems.
The testing demonstrated the feasibility of changing the cycle of the cruise engine
to provide shaft horsepower for the LiftFan [4]. It also demonstrated the capability
of thrust split control. That refers to the ratio of main engine thrust to LiftFan
thrust by modulation of the legacy first-stage fan and IGVs used as the LiftFan.
The next phase of testing involved suspending the LSPM at the Outdoor
Aerodynamic Research Facility (OARF) at NASA’s Ames Research Center.
This testing investigated external aerodynamics under hover conditions, as
shown in Fig. 4.
With the LSPM suspended on the OARF, free-air hover external measure-
ments were taken while operating the STOVL propulsion system. These
showed that the jet suckdown (i.e., the tendency of STOVL jet effects to lower
the static pressure underneath the aircraft out-of-ground effects [OGE]) was
less than 3% of the total lift. The model was also tested In Ground Effect (IGE),
318 S. P. WURTH AND M. S. SMITH

Fig. 4 LSPM mounted at the OARF.

as shown in Fig. 5. This demon-

strated that the jet fountain and
lift improvement devices limited
suckdown to less than 7% of the
total lift [4].
As part of the IGE testing,
HGI was also investigated and
verified. Prior subscale HGI
testing was questionable, so the
LSPM was an ideal test platform
to see how the shaft-driven
LiftFan configuration would per-
form. The pressures and temp-
eratures of the jet flow field
around the aircraft showed that
the jet footprint of the LiftFan system was more benign. This was applicable at
the same thrust levels as those of a Harrier II AV-8B [4]. The relatively cool
exhaust from the LiftFan blocked the hot engine exhaust from moving forward
into the main engine inlets. This prevented detrimental HGI that would lower
overall thrust due to higher temperatures. Figure 6 shows the use of ground
flow visualization paint to understand the flow fields.
Another phase of the testing involved taking the LSPM to the National Full-
scale Aerodynamics Complex (NFAC) at NASA’s Ames Research Center. There,
we used the wind tunnel at low speeds to investigate propulsion control effects
on external aerodynamics, as shown in Fig. 7. Performance data were collected
on Short Takeoff (STO) performance, transition (from jetborne to wingborne
speeds) aerodynamic performance, and control power, including yaw control
power for crosswinds [4]. Data showed excellent performance characteristics
and proved the feasibility of a
shaft-driven LiftFan and its inte-
gration into an ASTOVL air-
frame configuration.
The ASTOVL program estab-
lished concept feasibility by asses-
sing multiple factors, including
low-speed and powered-lift aero-
dynamics, hover and low-speed
control power, and the HGI of a
shaft-driven LiftFan connected

Fig. 5 LSPM IGE testing.

F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 319

Fig. 6 LSPM flow field

investigation.

to the main engine of a tradi-

tional propulsion system. With
the success of this program, the
Lockheed Martin team was well-
prepared to enter the CDP. This
next section covers the pro-
gression of the design, develop-
ment, and testing of the X-35
propulsion system.

IV. PRELIMINARY DESIGN, TEST, AND VERIFICATION OF FUNCTIONAL

FEASIBILITY: CDP/X-35
In December 1996 the Joint Strike Fighter (JSF) Program Office awarded Lock-
heed Martin and Boeing CDP contracts. Our objectives were to conduct prelimi-
nary design studies for an operational aircraft and build a Concept Demonstrator
Aircraft (CDA). Our CDA would demonstrate that similar airframes and propul-
sion systems could be designed and produced to meet Air Force CTOL, Navy CV,
and Marine Corps STOVL requirements. Key requirements were to demonstrate
vehicle characteristics that were either new (STOVL operations) or unsuitable for
modeling or subscale demonstration (carrier approach handling qualities). We
were not required to demonstrate legacy capabilities, such as stealth, avionics,
and weapons carriage in the demonstrator; however, we were required to do so
for the operational air vehicle design on which the demonstrator would be based.
The Lockheed Martin team was composed
of Lockheed Martin, Northrop Grumman,
British Aerospace (now BAE Systems), Pratt &
Whitney, Rolls-Royce, and Allison Advanced
Development Co. (now LibertyWorks). We
were tasked with developing and integrating
the propulsion system into the CDA, designated
the X-35. The X-35 propulsion system consisted
of a conventional afterburning gas turbine for
the CTOL and CV variants. The STOVL pro-
pulsion system added the LiftFan for thrust aug-
mentation for vertical lift while using the same
basic gas turbine as the CTOL and CV variants.

Fig. 7 LSPM at the NFAC wind tunnel.

320 S. P. WURTH AND M. S. SMITH

Fig. 8 X-35 CTOL/CV and STOVL propulsion systems.

Other components, such as roll post nozzles and a 3BSD, were also added to
provide propulsive attitude control without needing a bleed-air reaction control
system. This approach allowed us to use a common gas generator while mitigating
STOVL impacts on the CTOL/CV propulsion system. Figure 8 illustrates the
X-35 CTOL/CV and STOVL propulsion systems.
Lockheed Martin’s demonstrator approach was to build two airframes capable
of being configured as a Marine Corps STOVL variant. Aircraft 301 was to be
initially configured as an Air Force CTOL variant to conduct CTOL flight tests.
It would then be reconfigured as a Marine Corps STOVL variant for further
flight testing. Aircraft 300 would be configured as a Navy CV variant with a
larger wing area and control surfaces. This would demonstrate the low-speed
handling qualities required for a carrier-based operation.
Pratt & Whitney proposed an engine based on the legacy design, designated
the JSF SE611, with a larger fan and a redesigned turbine for all variants. In the
STOVL variant, the engine control system used the unique turbine design of
the JSF SE611. It provided the power to drive the LiftFan via a driveshaft from
the engine’s low spool. The development of the LiftFan, pitch vectoring LiftFan
nozzle, clutch, and driveshaft was the responsibility of Allison Advanced Develop-
ment Co. (now LibertyWorks). Pratt & Whitney and Rolls-Royce developed the
roll post system and the 3BSD.
The STOVL propulsion system was designed to provide attitude control while
in jetborne or semi-jetborne mode. The roll post nozzles provided roll control by
regulating engine bypass air to nozzles located in the wings. The 3BSD provided
both yaw and pitch control supplemented by a pitch vectoring nozzle on the
LiftFan. Fore/aft balance of the aircraft, as well as pitch control, was provided
by varying the thrust split (the ratio of 3BSD thrust to LiftFan thrust).
The development and integration of a traditional propulsion system for
CTOL and CV aircraft is not a trivial task. The addition of a STOVL variant
and the need to maximize commonality across all three variants increased the
magnitude of the task. The primary propulsion focus for the X-35 was to demon-
strate a tri-variant propulsion system that supported the two key objectives of the
CDP effort: STOVL operations and carrier approach handling qualities. Propul-
sion carrier approach requirements were known from legacy operations, leaving
STOVL operations as the primary propulsion challenge for the CDP.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 321

The layout of the LiftFan STOVL propulsion system had been demonstrated
in Lockheed Martin’s LSPM under the DARPA ASTOVL program (1993 to 1995).
However, it was not representative of an operational propulsion system. Although
the LSPM gathered valuable data on propulsion-induced aerodynamics, the pro-
pulsion system featured a single-stage LiftFan without a clutch. This prevented
conversion into a CTOL-configured propulsion system and lacked the control
system necessary to execute a conversion. The engine exhaust system also featured
a heavy two-dimensional nozzle rather than the 3BSD proposed for the X-35.
Starting with the proven legacy core, all remaining elements of the STOVL pro-
pulsion system had to be developed under the CDP and flown in the X-35.
Developing the key STOVL propulsion system elements was a major technical
challenge for the CDP. It required a STOVL demonstrator aircraft that could
prove the viability of the operational aircraft design. To minimize cost and sche-
dule risks, only the necessary elements were included. The advanced integrated
systems that were part of the operational aircraft were not necessary to meet con-
tract requirements. Conventional subsystems were to be employed whenever
possible; however, for the STOVL propulsion system, virtually all elements that
would be required in an operational aircraft were necessary for a successful
demonstrator.
Other requirements were:
. Performance in both CTOL and STOVL modes
. Jetborne and semi-jetborne controllability
. The ability to convert between CTOL and STOVL modes
Other characteristics, such as system life, weight, and full envelope capability, were
not as critical. This was due to the limited flight test program planned and the
absence of full-up avionics and weapons carriage requirements.
As a target for the design team, Lockheed Martin defined a flight test mission,
Mission X, that encompassed all contract requirements in a single flight. The
mission made a strong case for the X-35 being a stepping stone for an operational
aircraft. Mission X would accomplish what no STOVL aircraft had ever accom-
plished in a single flight:
. A short field STOVL takeoff
. Conversion into CTOL mode
. A level flight acceleration to supersonic speeds
. A conversion back into STOVL mode, followed by a vertical landing
Mission X truly represented what an operational aircraft would be called upon to
accomplish. In order to accomplish it, the X-35 would be required to possess mul-
tiple capabilities. Its thrust and fuel burn characteristics had to meet the
up-and-away requirements of an operational combat aircraft. Its performance
had to meet the STOVL requirements of an operational combat aircraft. It
322 S. P. WURTH AND M. S. SMITH

needed to have a fault-tolerant Integrated Flight Propulsion Control (IFPC)

system that seamlessly interfaced with the aircraft CLAW and minimized pilot
workload. Also, it needed to be able to convert from STOVL into CTOL mode,
and vice versa.
Developing and maturing these capabilities for the X-35 in the time available
was the program’s key challenge. Risk mitigation was integral and required iden-
tifying areas where contingency plans were required to allow a rapid response to
development problems. Detailed schedules were prepared showing development
milestones, presenting a methodical buildup approach leading to flight test.

A. PROPULSION SYSTEM REQUIREMENTS

1. PERFORMANCE
Performance of the X-35 propulsion system had to satisfy both CTOL/CV and
STOVL requirements using the same basic gas generator. Extensive modeling
was performed leading up to the CDP contract by all parties. We defined the pro-
pulsion cycle that would best meet the needs of the operational air vehicle. Pratt &
Whitney generated engine models, and Allison Advanced Development Co. (now
LibertyWorks) generated LiftFan performance models. Lockheed Martin com-
bined these models with aircraft installation effects to predict operational aircraft
performance considering range, combat performance, and payload. The result was
the JSF SE611 engine cycle coupled with a LiftFan. This configuration balanced
CTOL/CV requirements with STOVL requirements to produce a truly tri-variant
propulsion system.
For the CTOL and CV variants, what remained was to provide an adequate
engine response to ensure acceptable handling qualities. For the STOVL
variant, having sufficient power to accomplish a mission was just the beginning.
Controlling the aircraft while jetborne required not only thrust, but also thrust
through the appropriate effectors. These had to be coupled with a control
system that could keep the aircraft in balance and execute the pilot’s commands.
A new level of controls integration was required to couple the propulsion system
with the aircraft flight controls, provided in the IFPC.
CTOL/CV performance requirements centered on mission range, dry power
thrust (military power), and afterburner thrust levels (maximum power). The
STOVL requirements were more extensive, including all the CTOL/CV require-
ments plus the need to provide STO thrust and propulsive lift for a vertical
landing. Maintaining and controlling thrust during the transitions between wing-
borne and jetborne flight were also key requirements.
During the development period, the most stressing condition for the STOVL
propulsion system was either STO thrust at deck-edge rotation or vertical thrust at
the end of a mission. These two conditions were captured as part of the STOVL
variant’s overall mission requirements. STO thrust was dictated by mission
payload and fuel load at the deck edge necessary to meet the design mission.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 323

Vertical thrust was driven by the end of the mission where vertical lift bring-back
dictated minimum landing weight. Vertical lift bring-back comprises a specified
weapons load and a minimum fuel load that includes sufficient fuel for a
go-around pass. These competing requirements confirmed that the correct
engine cycle had been selected to balance performance requirements. The final
result was a propulsion system in the 40,000 lbf class. STOVL thrust is distributed
among the four lift posts: the LiftFan, two roll posts, and engine primary nozzle.

2. AIR INDUCTION SYSTEM AND HOT GAS INGESTION

Designing the AIS was a key element in meeting performance requirements. Lock-
heed Martin proposed a diverterless bump inlet, a design that had both weight and
survivability benefits. Without any moving parts and no need for boundary layer
bleed, the inlet minimized weight and provided good supersonic performance;
however, like all supersonic inlets, performance was inadequate at the low
speeds that would be encountered during STOVL operations. An auxiliary inlet
was required to improve static inlet recovery and meet performance goals.
Similar performance goals were required for the LiftFan inlet. The inlet was
shaped as much like a bell mouth as possible, subject to integration constraints.
Multiple wind tunnel entries, Computational Fluid Dynamics (CFD) modeling,
and static tests were conducted to maximize the performance of the AIS, including
modeling the doors. Both the auxiliary inlet and the LiftFan required doors that
would not impact inlet recovery while closing for up-and-away flight as part of
a stealth configuration.
A critical factor in the performance of all STOVL aircraft is the potential for
HGI. When a STOVL aircraft performs a vertical landing, hot exhaust from the
aircraft itself strikes the ground and reflects upward. It can potentially enter
the engine inlet and result in a dramatic decline in performance. The layout of
the X-35 STOVL propulsion system relied on the relatively cool exhaust of the
LiftFan. The fan provided an aerodynamic barrier that would prevent the
engine exhaust gases from making their way to the inlet. Quantifying how effec-
tively this prevented HGI was a critical step in defining the safe envelope for a
vertical landing.

3. STOVL INTEGRATED FLIGHT PROPULSION CONTROL AND CONVERSION DEVELOPMENT

The development of the IFPC system required a rigorous and thorough appli-
cation of an established Fly-By-Wire (FBW) flight control standard. This standard
was modified to adapt a legacy twin-engine propulsion system into a single-
engine, flight-critical application. The JSF SE611 engine was adapted from a
twin-engine legacy application with a control system and redundancy architecture
appropriate for a twin-engine application. There were clear deficiencies in the
redundancy architecture when simply carried over to a single-engine airframe.
324 S. P. WURTH AND M. S. SMITH

Addressing this issue with hardware and software redesign efforts was the major
focus of IFPC activities.
The legacy engine used an active/standby hydraulic configuration, but an
active/active electrical configuration for actuator control. The hydraulic active/
standby feature posed the most concern due to the inherent transient that
occurred as the system switched from active to standby after a failure. Consider-
able effort had been spent to minimize these transients to reduce the aircraft-level
effects. The legacy engine lacked fail-safe positioning after second failures. This
deficiency led to efforts implementing features that would drive some actuation
systems to aircraft-safe positions. These concerns were the impetus to establish
fault-tolerance requirements for the JSF X-35B lift system [5].
The X-35B’s flight test success was critical to winning the program. As such,
select members from the Flight Control Integrated Product Team (IPT) and the
Propulsion IPT formed the IFPC IPT to apply flight control standards to the
legacy propulsion system. The IFPC IPT was tasked with developing the inte-
grated CLAW that integrated aircraft, main engine, and LiftFan hardware and
software for STOVL mode flight. Additionally, the IFPC IPT had to apply the
flight control standards and requirements for fault tolerance and failure mode
handling to the development of the new hardware STOVL effectors. The IPT’s
objectives were to [5]:

. Rigorously and thoroughly apply an established FBW flight control standard.

. Develop the standard into an achievable process for the X-35B development.
. Ensure that the effects of considered failures were well-understood, simulated,
and documented.
. Design the X-35B propulsion system such that, under the effects of these con-
sidered failures, the X-35B would complete a successful and safe flight
test program.

Converting the aircraft from CTOL mode into STOVL mode and vice versa
required the development of a clutch system that would be new to the CDP.
This development required the integration of clutch CLAW with the engine
and aircraft CLAW so that the transition between modes was seamless to the
pilot. Pilot/vehicle interface design was also part of the IFPC development for
this new STOVL X-35 development so that controlling the propulsion system
was an easy task. The synergy of aircraft and engine CLAW development inte-
grated with the STOVL effectors became the genesis of the IFPC IPT.
Requirements for the hardware were developed based on previous Lockheed
Martin flight control FBW experience and applied to the new STOVL effectors.
Propulsion requirements for redundancy and control characteristics were
flowed to Pratt & Whitney and Allison Advanced Development Co. (now Liberty-
Works)/Rolls-Royce for the main engine, LiftFan, and clutch system. These
requirements were defined similarly to actuation requirements, but with an
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 325

added complication: the actuation requirements had to be related to thrust and

pitch control, roll control, and yaw control.
Pratt & Whitney provided the General Actuator Model (GAM) to help Lock-
heed Martin establish the effector requirements to provide the needed aircraft
desired dynamics. The handling quality requirements were then decomposed to
identify which effector requirements were necessary to meet those requirements.
The redundancy requirements were established using a CDP-developed fault tol-
erance process [5]. This process involved reviewing legacy propulsion systems and
adding redundancy to manage the X-35B STOVL flight test risk. The results
exposed areas that needed added redundancy. For areas to which redundancy
could not be added, minimizing exposure to these risks could be performed by
flight procedures.
Conversion requirements drove redundancy and engagement characteristics
while engaging and disengaging the LiftFan. This ensured that the aircraft
CLAW could deal with the introduction of LiftFan thrust during inbound conver-
sions and the removal of LiftFan thrust for outbound conversions. The conversion
modes are shown in Fig. 9, along with a photograph of the X-35B completing a
conversion.
Because engaging and disengaging the clutch was so critical to the success of
the X-35B program, special considerations for the requirements of the clutch and
lube system were established. No possible single or dual failure(s) of the clutch
system were permitted to engage the LiftFan when not commanded to do so.
Also prohibited were any possible single or dual failure(s) of the clutch system
that disengaged the LiftFan when not commanded to do so. Failures of the
clutch system during conversion were to be detected and safely accommodated.
These requirements ensured that the design of the clutch control system
had mitigated any failures that would cause an inadvertent disengagement or
engagement of the LiftFan. The
establishment of a specialized
built-in test was the result of
these requirements. This test
checked all conversion-related
actuation and software systems
immediately prior to engaging
or disengaging the clutch. If
that test failed, the pilot could
land in CTOL or STOVL mode
and prevent the aircraft from
being in the precarious state of
being in between modes.

Fig. 9 X-35B conversion

propulsion requirements and
conversion modes.
326 S. P. WURTH AND M. S. SMITH

The conversion process and the choreography of all the various propulsion
systems, LiftFan, clutch, and main engine actuation were established by the
IFPC IPT. This involved the systems integration of not only propulsion and
flight controls, but also hydraulics and electrical power systems. This ensured
that all of them operated in concert to convert the X-35B into/from conventional
flight into/from jetborne flight.
The IFPC requirements for this supersonic STOVL prototype required a
specialized effort. We defined flight control and fault tolerance requirements
that would meet aircraft STOVL flight control handling quality and safety
requirements to ensure a successful X-35B flight test program. Adapting a
legacy propulsion main engine and developing new LiftFan and clutch hardware
to meet these requirements would ensure this success. The IFPC IPT established
these propulsion system STOVL effector requirements and incorporated changes
to the system. This placed the aircraft in a fail-safe configuration in the event of a
catastrophic failure of the propulsion system. As a result of establishing hard-
ware and software requirements, extensive modeling and planning during the
flight test program were performed. This ensured that pilots could safely
recover the aircraft in the event of any improbable failures in the STOVL
control system.

4. AIR VEHICLE INTEGRATION

Air vehicle integration of the propulsion system in the X-35 followed a conven-
tional approach to minimize technical, cost, and schedule risk. Employing con-
ventional subsystems meant that off-the-shelf components could be used where
suitable. An Airframe-Mounted Accessory Drive (AMAD) was connected to
the engine gearbox via a power takeoff shaft and used to drive the aircraft subsys-
tems. The Auxiliary Power Unit (APU) was taken from the F-22, and the hydrau-
lic pumps were taken from the YF-23. The air turbine starter, generators, and
Environmental Control System (ECS) were all obtained from legacy aircraft. By
using this approach, subsystem components were readily available to support
development and flight testing with a supply of spares. This minimized cost
and avoided Safety-of-Flight (SoF) component qualification testing. This philos-
ophy was followed throughout the X-35 design, allowing the team to focus on
the unique capabilities required.
Because the X-35 was a demonstrator and not an operational prototype, some
capabilities required in an operational aircraft could be omitted, simplifying the
design task. The capabilities to fly at night in clear or adverse weather, and to a
full flight envelope, were not required. Limiting the maximum altitude relaxed
emergency power requirements in terms of duration and simplified life-support
system design. Similarly, although carrier approach handling qualities were a con-
tract requirement, performing an arrested landing was not, eliminating the need
for a carrier tail hook.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 327

B. DEVELOPMENTAL, GROUND, AND FLIGHT TESTING

The X-35 propulsion system development effort followed a methodical approach.
It began with component-level testing and proceeded through system-level
ground tests, culminating in air vehicle ground tests of the installed propulsion
system. Ground testing for the X-35A mimicked legacy practices, whereas exten-
sive STOVL ground testing was required for the X-35B. This section provides an
overview of that testing.

1. PROPULSION SYSTEM DEVELOPMENTAL TESTING

The development of the propulsion system followed multiple component-level
paths. It culminated in an integrated STOVL propulsion system test at Pratt &
Whitney’s test facilities in West Palm Beach. Pratt & Whitney initially began
testing the CTOL/CV propulsion system, proving out the JSF SE611 engine in
a conventional environment. Parallel efforts were underway to develop STOVL
components and construct test facilities. Conventional mode engine testing was
accomplished in a run station before sending the engine to the Arnold Engineer-
ing Development Complex (AEDC) for altitude testing. The JSF SE611 was
common between the CTOL/CV and STOVL systems, the so altitude testing
applied to all variants and included performance, throttle transients, air starts,
and inlet stability. Several engines were produced by P&W for engine system ver-
ification and validation testing. Allison Advanced Development Co. (now Liberty-
Works) developed several LiftFans, clutches, and driveshafts for similar testing.
Rolls-Royce developed the roll post system (nozzles, actuators, and ducting)
and the 3BSD module for development testing. Allison used various rigs to exer-
cise the gearbox and the clutch. A full-up test of the LiftFan would occur only
when coupled with a JSF SE611 engine during the system-level test in West
Palm Beach. The driveshaft was tested separately on rigs at the supplier’s facility.
In addition to the torque and speed requirements, key characteristics of the drive-
shaft were its ability to tolerate misalignment between the engine and LiftFan, air-
craft deflections, and axial growth of the engine due to thermals. Due to the high
rotational speed of the driveshaft, the rapid accumulation of load cycles also
required it be designed for infinite life.
The 3BSD and roll post system were designed and underwent component
testing at Rolls-Royce’s facilities in Filton, United Kingdom, just north of
Bristol. The 3BSD was a special challenge, featuring large-diameter flexible bear-
ings that would allow the rotation of the exhaust duct segments to provide the
desired vectoring capability. Actuators mounted on the duct would engage
gears on the outside of the bearings to provide the motive force for vectoring.
The 3BSD in both unvectored and vectored positions is shown in Fig. 10.
A key test objective was to collect data on the engine’s Low-Pressure Turbine
(LPT) during both CTOL and STOVL operation. The design of the LPT was the
most challenging aspect of the JSF SE611 due to its wide operating range. The
328 S. P. WURTH AND M. S. SMITH

Fig. 10 3BSD in CTOL and STOVL positions.

turbine was designed to meet all up-and-away requirements while also being able
to provide the horsepower necessary to drive the LiftFan. Pratt & Whitney modi-
fied two of its West Palm Beach test cells to accomplish these key objectives and
demonstrate STOVL system operation. One test stand was reconfigured to allow
the installation of the STOVL propulsion system. This included ducting to carry
exhaust gases from the LiftFan and roll posts outside the test cell. The exhaust col-
lector for the engine exhaust was modified to allow vectoring of the 3BSD. Initial
testing occurred in this test cell with a highly instrumented engine and LiftFan.
The second test stand (Fig. 11), which had been reconfigured by Pratt &
Whitney for STOVL testing, was an outdoor commercial engine test stand.
This allowed mounting the STOVL propulsion system at a height that permitted
STOVL mode operation without using exhaust ducts or collectors. This thrust bed
allowed the measurement of system performance to confirm modeling of the inte-
grated system. It also enabled demonstrating that the system would meet the
requirements for a successful flight test program. Other test facility modifications
included placing high-temperature concrete beneath the test stand to permit
extended STOVL test runs and deflectors to minimize any HGI.
This initial testing was the first time all STOVL propulsion components
had been assembled as a system. It was also the first time the LiftFan was
operated at high power. In order to reduce risk and allow for control system
development, initial runs were accomplished without engaging or disengaging
the clutch. Engine starts were accomplished with the clutch preset to the
desired configuration based on the intended test program. Initial test runs were
at low power as confidence in the system grew and the control system matured.
High-power runs were first accomplished with the LiftFan disengaged, followed
by engaged runs. Once high-power runs had been accomplished, testing pro-
ceeded with individually exercising each propulsion effector before proceeding
with integrated system operation. The final step was the execution of clutch
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 329

engagements, proving the capability of the system to convert into/from CTOL

and STOVL modes.
Testing on outdoor STOVL test stands confirmed that system performance
would meet the needs of the flight test program. It also verified that the propulsion
effectors could produce the required control authority for acceptable handling
qualities. The thrust measurement system provided steady-state measurements.
From these, data from the test stand were used to calibrate system models to
predict how the control system would perform under dynamic conditions.
Rapid thrust modulation between control effectors also excited the structure of
the test stand. This resulted in modifications during SDD to increase the overall
stiffness of the test stand.
Once the STOVL propulsion system had sufficiently matured to fly a simu-
lated mission on the test stand, an endurance test or Accelerated Mission Test
(AMT) was designed. The AMT would demonstrate sufficient life and durability
to meet the needs of the flight test program. It would also demonstrate twice the
needs of the flight test program, exercising the propulsion system in both CTOL
and STOVL modes. This would show that the system was safe and ready to fly.
Developing the AMT program required condensing the planned flight test
program and removing any nondamaging time (e.g., ground idle). It also required
exercising the system through a series of missions that simulated the stressing por-
tions of the flight test program. A single AMT would be performed to encompass
the flight test programs of all three variants.
The CDP AMT program had some interesting differences from what an oper-
ational AMT program would require. An operational aircraft would spend a con-
siderable portion of its life at a cruise condition. By contrast, the X-35 test
program would stress the propulsion system constantly while executing test
points. This was especially true for the lift system, because STOVL operations

Fig. 11 STOVL propulsion system in full afterburner on test stand.

330 S. P. WURTH AND M. S. SMITH

during the demonstrator flight test program were expected to approach the total
design life of the lift system for an operational aircraft. Whereas the engine would
have to demonstrate only a portion of a full-life operational system, the lift system
was required to demonstrate a nearly full design life capability.
The objectives of an AMT are to demonstrate readiness for flight and highlight
areas requiring special monitoring or components that might require replacement
during the flight test program. The AMT of the X-35 propulsion system went
smoothly, with only minor issues identified, such as fuel pump leaks partway
through the program. No major issues were found with the propulsion system,
clearing the way to execute the planned flight test program.

2. IFPC DEVELOPMENT TESTING

The IFPC system used an extensive process of testing and simulating all the
various STOVL and CTOL systems to ensure that the X-35 had a very robust
IFPC system. This included performing component testing for every actuation
system, to include:
. Functional and environmental testing
. Integrated propulsion testing
. Final vehicle integration testing
Integrated propulsion testing included fault detection and accommodation
testing, and the air vehicle and IFPC integrated verification of software- and
hardware-represented models in vehicle integration laboratories. This approach
to testing components, integrated systems, and vehicle integrated modeling of
the IFPC system was very successful in mitigating X-35 flight test program risk.
Simulation of IFPC failures established flight test procedures and allowed pilots
to practice emergency procedures for these failures. The Lockheed Martin/Pratt
& Whitney/BAE Systems/Rolls-Royce/Allison Advanced Development Co.
(now LibertyWorks) team understood all failures and flight characteristics. This
was as a result of this process, through which we acquired both a CTOL and
STOVL flight clearance to proceed to first flights of all three variants.

3. AIR INDUCTION SYSTEM DEVELOPMENT TESTING

The development of the AIS for the engine and LiftFan occurred in parallel with
that of the propulsion system. It proceeded from modeling and subscale testing
(both statically and in wind tunnels) to the construction of the demonstrator air-
craft. The X-35 incorporated the Lockheed Martin –developed and – patented
DSI [6]. As Fig. 12 shows, the development of the DSI started in 1995 for the
X-35. The DSI provides high aeroperformance across the flight envelope in a
fixed-geometry design without using a boundary layer diverter, bleed system, or
bypass system. Fixed-geometry, diverterless apertures are integrated on each side of
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 331

Fig. 12 DSI development timeline for the X-35.

the forward fuselage. The integration of the bump and the forward-swept aperture
lips combine to force the boundary layer air out the aft close-off point of the aper-
ture along the fuselage sidewalls.
Testing of the propulsion system with a representative AIS was not possible
until the installation of the propulsion system in the X-35. All ground and altitude
testing of the propulsion system was accomplished with ideal inlets (i.e., bell
mouths). By contrast, distortion screens based on inlet modeling were used to
demonstrate propulsion system operability. Designing the AIS for the engine
and LiftFan began with CFD computer modeling, followed by testing wind
tunnel models.
Both X-35 demonstrators had to be STOVL capable, requiring a single AIS
design; however, this still required the development of three inlets as opposed
to a single AIS for a conventional aircraft. These were the primary inlet in
CTOL mode, primary inlet with the auxiliary inlet open, and the LiftFan inlet.
The flight regime also required operation from static conditions all the way to
supersonic speeds, with the ability to convert into/from CTOL and STOVL
inlet configurations. Subscale testing through the speed regimes was conducted
with both CTOL and STOVL inlets, as Fig. 13 shows. Subscale testing was also
conducted to quantify the effects of HGI during a vertical landing.
A lesson learned while developing the LiftFan resulted in transient limitations
on STOVL operations during the X-35 flight test program and a design change for
SDD. Flow into the LiftFan exhibited higher than acceptable distortion at certain
332 S. P. WURTH AND M. S. SMITH

Fig. 13 CTOL and STOVL AIS wind tunnel testing.

airspeed/LiftFan rotational speed combinations. The result was higher and

unacceptable aeromechanic stress levels in the blades of the LiftFan due to the
distortion pattern shown in Fig. 14. This level of distortion limited the operating
time at those conditions and thus required constant and real-time control room
monitoring. Although manageable for a flight test program, a solution was
required for an operational aircraft.
BAE Systems characterized HGI during a vertical landing at a unique test
facility in Wharton, United Kingdom. The company used a subscale model of
the X-35, complete with flowing inlets and temperature instrumentation, to
perform dynamic testing. It did so by translating the model vertically into a
ground plane, thereby simulating a vertical landing. More than 10,000 test runs
were accomplished, investigating descent rates, aircraft attitude, ground slope,
nozzle vector angles, and ground speed/headwinds. Analyses of the results pro-
vided a set of conditions under which the X-35 could land without risking
HGI. Hover flow field IGE wind tunnel testing using oil flow indicated that, in
a calm hover IGE, the jet of air from the LiftFan created a wall. This wall prevented

Fig. 14 X-35 LiftFan inlet distortion pattern due to side-hinged LiftFan inlet doors.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 333

Fig. 15 Comparison of X-35B IGE wind tunnel oil flow and flight test demonstration.

the main engine exhaust from moving forward. During X-35B vertical landings, a
comparison of the IGE oil flow results with flight test results, using night vision
goggles (Fig. 15), shows that the LiftFan prevents main engine hot gases from
moving forward as the X-35B approaches the ground during a vertical landing [7].

4. AIRCRAFT PROPULSION GROUND AND FLIGHT TESTING

The final verification of the X-35 propulsion system was accomplished with the
ground testing of the installed propulsion system in the X-35A, X-35B, and
X-35C aircraft. X-35A and X-35C propulsion testing involved restrained full-
power engine runs comparable to those of legacy systems. These ensured that
the aircraft and subsystems were ready for flight. These aircraft then completed
very successful flight test programs, proving that the propulsion systems met
X-35 requirements for the Air Force and Navy. X-35B STOVL propulsion
system ground testing was considerably more involved and was the primary
focus in preparation for the X-35B flight test program. After completing
334 S. P. WURTH AND M. S. SMITH

ground tests, the X-35B progressed to STOVL press-ups on the hover pit to
demonstrate vertical capability and control. The X-35B then moved to Edwards
Air Force Base (EAFB), where the flight test program exercised the STOVL pro-
pulsion system in flight. It then progressed from wingborne flight to semi-jetborne
flight, to jetborne flight, and culminated in a vertical landing.

X-35A
As the first of the variants to enter flight test, the X-35A was charged with accom-
plishing all the basic SoF testing expected of any new aircraft configuration.
Among other things, these tests included:
. Aircraft handling qualities checks
. Subsystem operations checks
. Takeoff and landing characteristics
. Engine throttle transients
. Turns at elevated gs
The flight test program was highly successful at meeting or exceeding all flight test
objectives, clearing the way for the first flights of the X-35B and X-35C. The first
flight of the X-35A occurred on 24 October 2000. A key milestone accomplished
near the end of the flight test period was the first supersonic flight of the X-35 on
21 November 2000, less than one month from first flight. In order to increase
redundancy, the X-35 was designed for continuous operation of the APU in
flight, except for the supersonic portion of the flight envelope. This required
demonstrating the ability to shut down and restart the APU in flight before per-
forming a supersonic flight. Flight test missions were planned and executed that
exercised the APU in flight. These determined which flight conditions would best
support an APU restart.
The X-35A test program encompassed 27 flights, totaling 27.4 flight hours,
with six pilots and an unprecedented flight rate of approximately seven flights
per week.

X-35B
After the successful X-35A flight test program, Aircraft 301 was converted into an
X-35B. The forward fuel tank was removed, and the LiftFan and roll post system
were installed (Fig. 16). The 3BSD was part of the basic engine assembly, so no
special installation was required. The process of converting Aircraft 301 from
the CTOL X-35A configuration into the STOVL X-35B configuration spanned
approximately four months.
Ground testing of the X-35B began with subsystem checks to verify the instal-
lation of the STOVL propulsion system. All engine-mounted STOVL com-
ponents, such as the 3BSD, were driven by engine fueldraulics and could be
tested during an engine ground run; however, the remainder of the STOVL
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 335

Fig. 16 First LiftFan installation in an X-35B.

propulsion effectors was powered by aircraft hydraulics. The variable geometry in

the LiftFan, including the vectoring nozzle, clutch, and roll post system, was
driven by the aircraft’s hydraulic systems.
A purpose-built test facility was necessary to conduct all the necessary check-
outs on the X-35B. To this end, Lockheed Martin constructed a hover pit at its
Palmdale, California, facility specifically for testing the X-35. An additional
hover pit was built in Fort Worth, Texas, for STOVL checkouts of the F-35B.
The purpose of the hover pit was to allow restrained testing of the X-35B in
STOVL mode. During testing, the exhaust from the STOVL system was directed
away from the aircraft, aerodynamically simulating a free-air hover while avoiding
HGI and exhaust gas impingement on the aircraft. This permitted extended test
runs that were not representative of an operational environment.
Two systems were used for restraining the X-35B on the hover pit to support
different test objectives. A force-and-moment hard-mounted restraint system was
used to measure forces on the aircraft in STOVL mode, confirming performance
predictions and control power. This was the first time the STOVL propulsion
system was operated with representative inlets and other installation effects, as
Fig. 17 shows.
All previous performance predictions had analytically combined STOVL pro-
pulsion system testing results with inlet performance predictions from CFD and
subscale tests. Results confirmed predictions, with performance being as good as
or better than predicted. An interesting aspect of the testing was the first demon-
stration of full thrust split range. Previous system testing had included ideal inlets,
so the testing in the X-35 demonstrated the full thrust split range achievable with
representative inlets. The testing highlighted the need for improvements in
336 S. P. WURTH AND M. S. SMITH

Fig. 17 X-35B force-and-moment ground testing using hard support.

instrumentation, control sensor placement, and seals around STOVL door hinges
that were implemented prior to flight. Control room experience for the flight test
program was also gained by using a full control room during all ground testing.
This reduced test risk while training personnel.
The second restraint system
used on the hover pit was a soft-
mounted support to complete
STOVL mode structural coup-
ling testing. Structural coupling
tests are routine with today’s
FBW flight control systems. They
allow tuning of the CLAW to
avoid structural modes that could
be driven by the flight control
system. The soft system on the
nose gear, shown in Fig. 18, was
composed of air springs. These
isolated the vehicle from the rigid
ground while exercising the
STOVL propulsion system with
the CLAW loops closed and
monitoring the aircraft structure
with accelerometers. Results of
the tests were used to adjust the

Fig. 18 X-35B nose gear

soft support.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 337

Fig. 19 X-35B structural coupling test soft support system.

STOVL CLAW to avoid any interactions between the flight control system and the
aircraft’s structure. Figure 19 shows the X-35B over the hover pit in preparation
for engine-on STOVL structural coupling testing.
Completion of the STOVL ground tests on the X-35B cleared the way for flight
operations. Lockheed Martin’s plans for clearing the X-35B for STOVL operations
called for a build-down approach from conventional flight to semi-jetborne flight,
to jetborne operations. Before this could be accomplished, we needed to verify that
the STOVL propulsion system could support the aircraft in a jetborne condition
with sufficient control power. This verification would be accomplished in a series
of short hops (press-ups) from the hover pit. The first hops would demonstrate
performance necessary to support the aircraft, whereas later press-ups would
investigate control authority in hover.
The first hop of the X-35B (Fig. 20) was piloted by Simon Hargreaves on 24
June 2000 and was a complete success. Simon found the aircraft rising higher
than intended due to increasing thrust of the propulsion system; however, he
did not want to move the throttle abruptly on his first flight. As such, he slowly
retarded throttle as the aircraft climbed until the ascent was arrested and the air-
craft descended back onto the pit. Required performance had been demonstrated,
with the IFPC providing a stable and controllable platform. Posttest analyses
showed that thrust had increased as engine thermals stabilized, requiring more
throttle control than anticipated to retard the climb.
Press-up testing on the hover pit continued with accomplishing two short
hops the next day and a stable hover lasting longer than two minutes on the
third day of press-ups. Testing continued for a week with three test pilots accom-
plishing a total of 14 press-ups. The final press-up test included control system
checks and was the longest hover accomplished, lasting more than three and a
half minutes.
Completing the press-ups from the hover pit cleared the X-35B for the move
to EAFB and the continuation of the flight test program. Over the next two and a
half weeks, the X-35B exercised the STOVL propulsion system in flight, steadily
338 S. P. WURTH AND M. S. SMITH

Fig. 20 Two views of X-35B’s first press-up from hover pit.

progressing through the test program to lower and lower airspeeds. Testing cul-
minated in a vertical landing on the EAFB hover pad.
The X-35B flight test program was extremely successful, conducting 39 flights
totaling 21.5 flight hours with four pilots in an intensive six-week test program.
With its unique propulsion system, the X-35B accomplished:
. 22 hovers
. 18 STOs
. 27 vertical landings
. 21 airborne conversions
. 6 supersonic flights
The aircraft also executed Mission X twice, reaching the milestones of STO, super-
sonic dash, and vertical landing in a single flight for the first time in history.
c. X-35C
Aircraft 300 commenced its flight test program configured as a Navy variant in
late 2000, the same time when the X-35A completed its flight test program. The
X-35C accomplished 73 flights totaling 58 flight hours with eight pilots, including
the first transcontinental ferry by an X airplane. After initial flights at EAFB, the
X-35C continued its flight test program at a Naval Air Station Patuxent River test
facility. There it accomplished 250 field carrier landing practices to demonstrate
carrier approach handling qualities. The propulsion system provided the necess-
ary throttle response and performance required to meet all criteria with the flight
CLAW. It performed well enough to prompt a pilot to state, “We are ready to go to
the boat with this airplane.”
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 339

The three variants, using two airframes and two propulsion systems (CTOL/
CV and STOVL), successfully met all objectives. No flight control or engine soft-
ware control changes were required, and the propulsion systems worked flawlessly
through a very rigorous flight schedule. The flight test program concluded on 31
July 2001. The X-35’s flight test program successes were key to Lockheed Martin
being awarded the JSF F-35 SDD program on 26 October 2001. The success of the
propulsion system during the flight test program in particular was instrumental
in this.

V. DETAILED DESIGN, TEST, AND VERIFICATION FOR OPERATIONAL

DEPLOYMENT: SDD/F-35
The X-35 flight test program was a success, proving the functionality of the pro-
pulsion system and meeting the demonstration requirements of STOVL operation
and a carrier approach. The Lockheed Martin JSF team was then tasked with com-
bining everything learned from the X-35 demonstrators with the operational air-
craft preliminary design and producing an operational warfighting machine.
Christened the F-35, the aircraft incorporated the very successful propulsion
system design into an airframe design that was lethal, stealthy, and survivable.
Those requirements alone were difficult engineering challenges for the Lockheed
Martin/BAE Systems/Northrop Grumman/Pratt & Whitney/Rolls-Royce F-35
team. Adding to them, the team had to make everything work with a propulsion
system design that met Air Force, Marine Corps, and Navy tri-variant require-
ments for performance, reliability, maintainability, supportability, lethality, and
survivability. The production propulsion system was designated the F135, depart-
ing from the demonstrator JSF SE611 nomenclature.
Beginning with the X-35 propulsion system, major changes were needed to
address the CDP lessons learned and produce an operationally suitable full-life
propulsion system. The most significant changes were made to the LiftFan and
its integration into the air vehicle. An inlet redesign was necessary to address
the aeromechanical problems encountered during the flight test program, and a
complete LiftFan nozzle redesign made the nozzle part of the airframe structure.
This reduced weight and provided area control that could be used to enhance
operability. Other changes to the propulsion system and its integration are dis-
cussed in the following sections.

A. PROPULSION SYSTEM DESIGN REQUIREMENTS

1. PERFORMANCE
The X-35 flight test program demonstrated slightly better STOVL performance
than had been expected due to the performance of the AIS and the propulsion
system itself. This highlighted an opportunity to increase the up-and-away
340 S. P. WURTH AND M. S. SMITH

performance of the air vehicle and the vertical landing capability of the STOVL
aircraft. Slightly increasing the size of the primary inlet of the operational aircraft
took advantage of a new engine fan design being implemented for producibility.
This allowed higher airflows with reduced distortion and promised even better
inlet recoveries than had been experienced with the X-35.
By far the biggest challenge for the operational propulsion system was the
requirement for full life. The STOVL system life requirements were also a chal-
lenge for the hot parts due to the more strenuous duty cycle of the STOVL aircraft.
The goal was to use common parts across all variants and minimize any redesigns
due to STOVL requirements. A key element in meeting full life was the specialized
FADEC control system employed. Conventional engine control systems run to
schedules, such as fan speed. Accordingly, as an engine deteriorates, thrust actu-
ally increases due to increasing turbine temperatures. With the new F135 control,
engine performance remains constant throughout its life. Turbine temperatures
are controlled to meet thrust targets rather than holding the engine on a
control schedule as it deteriorates.
The major challenges for the engine, as might be expected, were in the turbine
section. The high-pressure turbine was refined to better utilize cooling air,
improving both life and efficiency over the JSF SE611. The LPT was a major chal-
lenge as well, with the JSF SE611 featuring a design that could not meet full-life
requirements. New blade damping techniques were needed to meet operational
life requirements.
Some of the changes to the propulsion system were refinements based on
experience gained from the CDP. Variable geometry in the turbine exhaust case
included for STOVL operation was found to be unnecessary and was removed
from all variants. The aggregate vector angle of the 3BSD was found to be more
than required. This allowed a reduction in bearing angles that produced benefits
in overall tailpipe length, weight, and ground clearance. The offtake valves included
as part of the roll post system were deleted because they did not add redundancy
and were heavy due to their stiffness requirements. These types of changes were
instrumental in maintaining performance targets while reducing weight and cost.
Operational design changes for the air system also drove changes to the
propulsion system. The F-35 does not have an AMAD, which requires aircraft
generators, hydraulic pumps, and lubrication pumps to be mounted directly on
the engine gearbox. Heat rejection, a major problem for 5th-Generation aircraft,
required a thermally efficient fuel pumping system on the engine and the ability to
handle fuel temperatures as high as possible. Heat exchangers were mounted in
the engine gas path to cool bleed air from the engine and provide a heat sink
for the aircraft Power and Thermal Management System (PTMS). At the same
time, every effort was made to improve maintainability by tailoring engine exter-
nals to line up with airframe access points as much as possible. Propulsion system
access was more limited, relative to the X-35. This was due to the presence of
weapons bays and the need to minimize the number of access panels in the aircraft
outer surface.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 341

2. AIR INDUCTION SYSTEM

The redesign of the primary inlet for higher airflows was only a portion of the inlet
redesign. The inlet was also shortened as a weight-saving measure. It entailed
moving the aperture aft and changing the cowl from the four-edge design on
the X-35 to a simpler three-edge design for the F-35. The redesigned inlet aperture
also provided improved inlet performance at high angle of attack. The aft portion
of the inlet diffuser was made variant-unique, with the CTOL and CV variants
sharing one design and the STOVL variant’s auxiliary inlet having its own con-
figuration. The STOVL inlet also featured a tunnel for the LiftFan driveshaft,
extending forward from the engine face to the inlet bifurcation.
Common requirements across all variants included a source of nacelle venti-
lation air, the addition of an ice detector, and the introduction of an inlet debris
monitoring system. The location of the ice detector was to be common across all
variants, if possible, but was subject to the results of ice accretion tests that would
be performed. The inlet debris monitoring system was a developmental prognos-
tics and health monitoring system that promised to reduce or eliminate the need
for inlet and engine face inspections between flights.
The refinement of the auxiliary inlet design was an evolutionary step from the
system, which had performed well on the X-35B (Fig. 21). Recovery improved
significantly, and the doors were lighter than the original X-35B doors. The com-
bined thrust increase with the recovery improvement and the lighter weight added
500 lb of STOVL thrust over the X-35B configuration. Additional F-35B changes
focused on lessons learned from the X-35, such as the following improved
features:
. Seals on the door hinges
. Locations of inlet sensors on the LiftFan
. Smooth contours with any air vehicle outer mold line changes

The LiftFan inlet was the biggest challenge facing the Lockheed Martin
internal aerodynamics team. The X-35B inlet had performed adequately, but
the uneven flow field had unintended consequences on LiftFan aeromechanics.

Fig. 21 F-35B improved auxiliary inlet.

342 S. P. WURTH AND M. S. SMITH

Fig. 22 X-35B LiftFan flow field causing LiftFan aeromechanic issues during forward flight.

Figure 22 shows the X-35B LiftFan inlet configuration, CFD streamlines into the
inlet, and the resultant adverse pressure flow field that created the aeromechanics
issue. The existence of transient-only zones was not acceptable for an operational
aircraft.
The stream tube was a special inlet test fixture designed from CFD streamlines
that replicated the LiftFan flow field for a given forward velocity while testing at
static conditions. This test technique was developed during SDD and was vali-
dated through comparisons with X-35 flight test data and CFD. The technique
proved invaluable and was used on subscale tests, full-scale LiftFan tests at
Rolls-Royce, and Pratt & Whitney test stands in West Palm Beach. Figure 23
shows flight test total pressure (recovery) and a comparison of CFD and a test
using a LiftFan stream tube to reproduce the adverse flow field [8]. The stream
tube test technique proved its value by supporting the development of a LiftFan
inlet that had acceptable levels of distortion [8]. With the technique, the
inlet also eliminated the transient-only zones experienced during the X-35 flight
test program [8].

Fig. 23 X-35B LiftFan aeromechanic stream tube test rig results.

F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 343

The level of inlet distortion was known and easily understood. The challenge
was how to improve the inlet such that the LiftFan would tolerate the resulting
flow field while having acceptable performance. The aft hinge configuration was
subsequently chosen for future development, and work began on characterizing
the resultant flow field. Distortion estimates were generated and provided to
Rolls-Royce to assess the aeromechanics effects. It was also necessary to redefine
the interface between the LiftFan and the inlet stream to more accurately model
the performance of the system. Once a preferred configuration had been
defined, testing was planned to verify that the system worked as predicted [8].
The aft hinge door had an inlet bell mouth lip modification and LiftFan case
flare. It was the only concept to meet the distortion and high-cycle fatigue goals
while maintaining an acceptable flight envelope. The F-35 propulsion team
used CFD, subscale inlet testing, and stream tube testing to solve the X-35B
LiftFan aeromechanics issue [8].

3. F-35 INTEGRATED FLIGHT PROPULSION CONTROL DEVELOPMENT

AND FAILURE IMMUNITY
The F-35 presents one of the most complex integrated control problems for
aircraft and propulsion to date through its tight interaction of numerous subsys-
tems. These interactions span the obvious flight controls and propulsion systems
and, as Fig. 24 shows, extend into utilities and subsystems, power and thermal
management, and fuel systems. Utilities and subsystems include STOVL door
control, hydraulics, fuel, electrical power systems, landing gear, weapons bay

Fig. 24 F-35 integrated systems layout.

344 S. P. WURTH AND M. S. SMITH

door systems, and pilot ejection systems. Power and thermal management
includes an Integrated Power Package (IPP) and thermal cooling.
The synergy between aircraft CLAW and engine software was drastically
improved over the X-35. IFPC requirements were inherent from the first day of
the program, in contrast with the X-35 process of adapting IFPC requirements
to existing propulsion control design. The IFPC CLAW architecture is a
complex series of multivariable loop closures involving [1]:

. Six-degrees-of-freedom flight controls

. Five-degrees-of-freedom propulsion control
. Fuel system control
. Lift system door control

The Lockheed Martin nonlinear dynamic inversion flight control methodology

has evolved from the early X-35 implementation to the present-day F-35
system. The architecture enables a single CLAW structure supporting all three air-
craft variants. The STOVL design is the most complex, with the direct control of
all six degrees of freedom. Higher levels of augmentation have been incorporated
to reduce pilot workload and increase safety. The primary goal of the CLAW
modes was to make STOVL takeoffs and landings as routine as they are for con-
ventional aircraft. This also benefits fleet usage by reducing the training burden for
STOVL mode-unique portions of the syllabus [1].
The F-35 CLAW improved from the X-35 in that more modes of augmenta-
tion were incorporated to ease pilot workload. This augmentation drove the need
for the F135 engine control to drastically improve its capability with higher pre-
cision control and faster engine dynamics. These changes were necessary for the
challenging STOVL and Navy requirements. From the beginning, the JSF
program set out to develop performance-based specifications in which only the
performance capability and desired operational mission requirements were put
on contract. The specification levied against the aircraft contractor required
Level 1 flying qualities for all three variants. This, in turn, caused the flow
down of propulsion system dynamic performance requirements to constitute
the most technically challenging requirement of the entire program. It required
the STOVL propulsion system to deliver commanded total thrust (sum of all
posts) to within 0.5% accuracy, with a bandwidth requirement that was up to
several times faster than any legacy designs.
The most technically challenging of all requirements was either the
thrust-to-pitch and pitch-to-thrust decoupling or the ability to stay wings level
as the aircraft heaved in the vertical axis. Additionally, the aircraft could not
have uncommanded movement in the pitch axis while performing a roll maneu-
ver. To achieve these requirements, a quantum leap in integrated flight and pro-
pulsion control technology had to be realized. During the X-35B concept
demonstrator phase, the operability of the integrated STOVL propulsion system
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 345

and the variant’s flying qualities from hover through supersonic flight were
demonstrated. Several advanced flight and engine control concepts were being
explored. Among them was a government-sponsored research program on the
United Kingdom’s Vectored thrust Aircraft Advanced Control (VAAC)
Harrier. Many revolutionary areas of STOVL advanced control were explored,
from pilot inceptor types to aircraft response mapping to those inceptors. This
was demonstrated in large-amplitude, motion-based simulators and flight demon-
strations, both ashore and during embarked operations at sea. Studies from VAAC
changed the pilot control concept used in the X-35 to the F-35’s unified flight
control strategy, discussed later in this paper. This strategy was made possible,
in part, with the successful development and deployment of Pratt & Whitney’s
F135 Advanced Multi-Variable Control (AMVC) design [9].
The AMVC is a multivariable engine control using a unique control technol-
ogy that allows the engine software to decouple the highly coupled engine system
composed of a main engine connected to the LiftFan via a driveshaft. The resul-
tant aircraft performance was wings-level hover and vertical landing Level 1 flying
qualities. The AMVC team met the immense challenge to develop and certify the
control architecture concurrent with the JSF F135 development program. It did
not have the benefit of a dedicated technology development program, however.
Notwithstanding, the team successfully matured the state-of-the-art control
system as part of the F135 development program, culminating in the operation
of the F-35B in powered lift mode.
Pratt & Whitney’s AMVC uncouples complex, cross-coupled, nonlinear pro-
pulsion system dynamics and makes them appear similar to a set of uncoupled
virtual flight effectors. In contrast to normal propulsion system control, this
one controls five degrees of freedom of thrust at four nozzles. These responses
must be linear and consistent with relatively high bandwidths and with relatively
little cross-coupling among the degrees of freedom. The virtual effector command
response bandwidth required is greater than the propulsion feedback loop band-
width and approaches the propulsion actuator bandwidths. Also, similar to most
engine controls, the CLAW must run the engine close to and onto numerous
limits without coupling into the virtual effector responses.
The failure management strategy for a single-engine FBW STOVL 5th-
Generation aircraft has extensively evolved from its early implementation in the
X-35B aircraft to the current F-35B system [5]. Based on modern triple- and
quadruple-redundant FBW flight controls, the program requirements drove
redundant channel control and feedback for propulsion effectors. In this way,
they ensured a fail-operational/fail-degrade capability in the event of a failure
of primary flight-critical flight control systems in STOVL flight. During STOVL
mode, the propulsion system becomes the primary flight control system for
control and, as such, it drove this redundancy into the design.
Achieving a fail-operational/fail-safe philosophy for dual mechanical failures
for all the propulsion effectors was deemed impractical for a weight-sensitive
STOVL aircraft. The performance-based requirements for the F-35B IFPC
346 S. P. WURTH AND M. S. SMITH

Fig. 25 (a) X-35B LiftFan hood nozzle and (b) F-35B VAVBN.

failure management drove a philosophy into the design that was different from
that of the X-35B. This is not to say, however, that the aircraft could not
achieve fail operational/fail safe for a wide range of failures. Nor was it the case
that the aircraft could achieve fail operational in all cases for dual electrical fail-
ures. Rather, the basic difference in philosophy focused on how faults affected
flying qualities and their probability of occurrence. Design success was achieved
by adhering to a basic tenet giving a carrier pilot the utmost confidence of aircraft
recovery under a wide range of failures and meteorological conditions. This tenet
became known as “failure immunity in blue water ops.” As a result, the F-35B
IFPC system and its fault detection and accommodation strategy vastly improved
over the X-35B and achieved a lighter-weight solution. IFPC redundancy has
been designed into all the propulsion effectors and meets the failure immunity
requirements. The FADEC and aircraft CLAW failure accommodation allow
the pilot to achieve a safe recovery to a ship in blue water operations. During
this, the pilot has at least Level 2 flying qualities for any single or dual combination
of failures that have a greater than 1027 probability of occurrence.
Many hardware redundancy improvements were incorporated into all engine
and LiftFan systems as a result of the new failure immunity requirements. Specifi-
cally, the X-35B LiftFan hood nozzle was removed and replaced with the VAVBN.
This greatly improved the capability to vector in the event of a failure. The
VAVBN also provided nearly 40% variation in exit area, allowing for additional
LiftFan vector and thrust control in the event of failures. The VAVBN system
also represented a 35% reduction in weight over the X-35 LiftFan hood system
[10]. Comparative pictures of both systems are shown in Figs. 25a and 25b.
The LiftFan clutch, gearbox, driveshaft, and lube system were vastly improved
over those of the X-35. The F-35 clutch system used aircraft dual hydraulics to
actuate and lock the clutch with improved control and feedback systems. A
new lube pump was driven off the driveshaft via a gear drive that only supplied
oil to the gearbox and clutch bearings. It did not have the task of engaging the
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 347

clutch like the X-35B lube system. The driveshaft was an all-new design that
.

improved producibility while maintaining an angular misalignment capability.

It also addressed compression/tension tolerances between the engine and
LiftFan commensurate with an operational F-35 system. The number of clutch
plates was reduced, allowing thicker plates that increased the number of engage-
ments before replacement.
To avoid using twice as many actuators to support a dual hydraulic clamp
and lock system, an ingenious system was implemented to detect leaks in the
clutch system. If the system detected a leak, it would alert the pilot and allow
him or her to preserve the remaining hydraulic system to ensure a successful
landing in STOVL mode if required. In 2015, the F-35B clutch system was incur-
ring an inordinate amount of clutch failures attributed to nuisance issues and a
nonrobust design. The Lockheed Martin/Pratt & Whitney/Rolls-Royce IFPC
team redesigned the clutch, engine, and hydraulic redundancy algorithms. The
team’s modifications allowed the removal of certain clutch components while
providing the necessary redundancy tolerance to meet failure immunity
requirements.
F-35B conversion and aircraft flight control during conversion was also dra-
matically improved. In the X-35, the conversion was initiated when the pilot
moved the Thrust Vector Lever (TVL) out of the detent, which initiated the con-
version process. The TVL was a legacy device from AV-8B to control the vector
angle of the nozzles while the pilot controlled height using the throttle. The F-35B
had an improved pilot inceptor design, named Unified Control Strategy (Fig. 26).
It used traditional stick, called the Right-Hand Inceptor (RHI), yaw pedals, and
throttle controls, called the Left-Hand Inceptor (LHI), to reduce the amount of
pilot workload significantly. The strategy was to decouple the aircraft control
axis and provide the pilot with the most desired response for his or her inputs.
It also simplified the cockpit by removing the third TVL inceptor, providing
the same inceptor functionality
whether the aircraft was station-
ary or flying at 500 kt.
Conversion was now handled
by the identical method with
which the F-35C (CV) would
extend the carrier hook, a simple
button in the upper-left region of
the cockpit. The difference with
the F-35B, however, was that
it was labeled HOOK/STOVL.
Conversion in the F-35B was

Fig. 26 F-35B Unified Control

Strategy inceptor configuration.
348 S. P. WURTH AND M. S. SMITH

designed to be as simple as the extension of landing gear or the carrier arresting hook.
The pilot simply flies the aircraft below 250 kt and pushes the HOOK/STOVL con-
version button, which initiates a complex transformation of hardware and software
in the aircraft. The conversion is done completely automatically. The aircraft
CLAW initiate commands to the aircraft to open doors, as shown in Fig. 27. Com-
mands are also sent to the FADEC to engage the LiftFan and prepare the engine to
drive the lift system. Commands then move the 3BSD and VAVBN vectors into
position for STOVL flight—all while maintaining the pilot’s desired flight path
and speed.
The improved redundancy architecture and associated Lockheed Martin/
Rolls-Royce/Pratt & Whitney fault detection logic provided acceptable fault
tolerance for all failures meeting the failure immunity requirement. The fault
detection logic was driven by the new failure immunity requirements to meet
affordability goals. Improvements in all these STOVL effectors improved redun-
dancy and capability to allow STOVL to be as simple as CTOL.
The success of this philosophy has been observed in flight test and fleet
pilot confidence. If a failure were to occur when a pilot entered the slow
STOVL regime, the aircraft would accommodate it. The aircraft’s response
would keep a transient from forcing a pilot to make the decision to egress (via
ejection) the aircraft.

4. AIR VEHICLE INTEGRATION

The integration of the F135 into the F-35 was a complete rework, compared to the
X-35, as Fig. 28 shows [11]. The entire suite of subsystems was changed from
off-the-shelf systems pulled from legacy platforms to an innovative, highly inte-
grated set of subsystems. One immediately obvious change is the lack of an air
vehicle AMAD. Eliminating the AMAD reduced the overall length of the aircraft

Fig. 27 F-35B lift system doors.

F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 349

Fig. 28 Comparison of X-35 federated components with F-35 integrated PTMS.

by providing more flexibility in the placement of the weapons bays. In the case of
the STOVL variant, it created more room for the LiftFan nozzle. Without an
AMAD, this required all mechanically driven aircraft systems to be mounted
on the engine gearbox. One of the systems that made this approach possible
was the use of electrical power for flight controls rather than hydraulics.
Electrical power for flight controls reduced hydraulic demands to the level of
a utility system. An Engine Starter/Generator (ESG) would be mounted on the
gearbox, serving as both the engine starter and the prime generator for the entire
aircraft. The internal architecture of the ESG made it functionally a triplex gen-
erator, allowing redundancy on a single gearbox pad. The electrical system
architecture was configured to shed a load in the event of a failure, prioritizing
power based on the criticality of the need. In the event of a total ESG failure,
backup power would be provided by batteries and the IPP, another highly inte-
grated subsystem providing multiple functions. The integration trade-off was
that multiple power feeders would have to be connected to the generator once
the engine was installed (generator mounts on engine prior to engine install).
In addition, the lubrication system would be partially mounted on the
gearbox (pump), with the remainder on the aircraft (heat exchanger and oil
reservoir).
The use of electrically powered flight control actuators avoids much of the heat
that would be generated by conventional hydraulic systems. Hydraulic systems
operate continuously at full pressure, whereas electrically powered actuators
350 S. P. WURTH AND M. S. SMITH

operate on an on-demand basis. This reduction was a key enabler in meeting

thermal requirements, as discussed later in this section.
The PTMS combined the functions of an APU and an ECS. The PTMS pro-
vides conditioned air for cockpit and equipment cooling, refrigerated liquid for
equipment cooling, and electrical power for ground maintenance, main engine
starting, and in-flight emergencies. The cooling function is performed by a
closed-loop air refrigeration cycle powered by expanding engine bleed air
across a power turbine. Waste heat from the refrigeration cycle is transferred to
fuel and to engine fan bypass air. A polyalphaolephin liquid loop and two heat
exchangers are used to transport heat between the air cycle and the fuel. Heat
is transferred to the engine fan air through air-to-air HXs mounted in the
engine Fan Duct HXs (FDHXs).
APU functions were facilitated by reconfiguring valves so that the PTMS tur-
bomachine could operate like a traditional APU. In this way, it provided power for
self-start, ground maintenance, main engine start, and in-flight emergency power.
Effectively combining two functions into one system, the PTMS provided all
necessary functionality in a much smaller package. The use of FDHXs required
more ducts to be connected to the engine, but it reduced the need for ram air
inlets and heat exchangers.
As previously mentioned, F-35 hydraulic heat loads were greatly reduced by
the use of electric flight control actuators; however, hydraulic waste heat is just
one of several heat sources transferred to fuel. The PTMS, electrical generator,
engine lubrication system, and engine fuel system also dissipate waste heat to
fuel. The mechanically driven, continuously running engine fuel pumps are a
major source of engine waste heat. In addition to boosting fuel pressure to
engine combustor pressure, the pumps also provide the power for engine
nozzle and other actuation. Minimizing the heat generated by the engine
fuel pumping system reduced the total cooling load on the aircraft. It was
accomplished by using a thermally efficient fuel pump for the main engine
while optimizing pump configuration when not required. When coupled
with the electric flight control actuators, these changes made dramatic differ-
ences in the total heat load and allowed the air vehicle to meet its mission
thermal requirements.
One major source of heat that could not be completely mitigated was the
heat generated by the LiftFan during STOVL operations. When engaged, the
gearbox generated large quantities of heat that were removed by the lubrication
system. The most challenging condition was at the end of a mission, when the
fuel heat sink was smallest (low fuel warmed during the mission) and the
quantity of heat generated could not be entirely removed. Consequently,
when a vertical landing was to be performed at the end of a mission, the air-
craft would operate at a deficit, heating up during STOVL operations. Due to
the short durations of vertical landings, coupled with the thermal mass of the
system, this deficit proved to be manageable, as confirmed by flight test
operations.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 351

B. DEVELOPMENTAL INLET TESTING AND PROPULSION INTEGRATION

GROUND TESTING
The development test program for the F-35 propulsion system was much more
extensive than employed by the CDP. During the CDP, testing was accomplished
using existing facilities with minimal modifications plus a few purpose-built facili-
ties (e.g., the 3BSD rig in Bristol). For SDD, dedicated test facilities were con-
structed and existing facilities were modified to provide support for the life of
the test program. Legacy test facilities already existed for CTOL/CV propulsion
systems, so most facilities were targeted at STOVL testing. One example is the
dedicated LiftFan test facility at Rolls-Royce Indianapolis. Another example is
Pratt & Whitney’s modified C-12 test stand in West Palm Beach (Fig. 29),
which provided a STOVL test capability. A third example is Lockheed Martin’s
hover pit in Fort Worth, Texas, for force-and-moment testing and acceptance
testing of production STOVL aircraft.
The ground test program started in 2004 and was completed with more than
10,000 engine test hours. Modeling of the propulsion system was critical to the
success of the F-35. The program developed detailed models that were integrated
with the aircraft models to develop the aircraft and propulsion CLAW. The engine
and LiftFan ground hardware and software components underwent extensive
testing [1]. This included:
. Bench and rig testing
. LiftFan testing
. Engine testing
. Integrated propulsion system testing
. AMT

Fig. 29 STOVL propulsion system on the C-12 test stand.

352 S. P. WURTH AND M. S. SMITH

. Vehicle Integration Facility testing

. Vehicle Systems Integration Facility testing
. Air vehicle testing on the F-35 Hover Run Facility

1. DEVELOPMENTAL INLET TESTING

An expansive F-35 inlet test plan was launched at the program start in 2001. This
plan called for using subscale wind tunnel models to improve the X-35 inlet
system for an operational F-35 aircraft. CFD and subscale and full-scale inlet
testing were added to the test program as the SDD propulsion testing phase con-
tinued [11]. This testing supported all milestones for the program and for lines
freeze and validation.
Beginning with the main inlet, CFD and model testing were used to determine
the performance of the primary inlet. Changes in aperture location, number of
edges, and diffuser shape drove the need to treat the inlet development as a
new configuration relative to the X-35. In addition, lessons learned from the
CDP were incorporated. Subscale testing, from subsonic to supersonic, was con-
ducted in multiple wind tunnels, including wind tunnel facilities at AEDC and
NASA Glenn, shown in Figs. 30a and 30b, respectively. Test results were used
to calibrate and refine CFD models, generating an iterative process that continued
to refine the overall configuration while responding to integration considerations.
An ice accretion test program was included as part of the SDD program to
investigate the buildup of ice and identify potential locations for an ice detector
in the inlet. Testing was conducted at Italy’s Centro Italiano Ricerche Aerospaziali
(CIRA) facility in Capua, Italy north of Naples. The test featured a large half-
model of the primary inlet that was subjected to various combinations of freezing
moisture to determine where and how fast ice would accumulate. The data were
also used to design the engine front-frame ice protection system.
Results from the test were used to calibrate an analytical ice prediction model
for the inlet, which allowed later simulations for locating the ice detector in the

Fig. 30 (a) Main inlet development subscale testing at AEDC 16T and (b) NASA Glenn.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 353

Fig. 31 Main inlet development full-scale testing.

inlet. Although the test identified the optimum location for an ice detector, inte-
gration considerations required compromising its placement. (Such consider-
ations included the backup structure and the auxiliary inlet for the STOVL
variant.) Later design changes made to the auxiliary inlet doors for structural
reasons, discussed later in this paper, drove a subsequent relocation of the detector
for the STOVL variant, requiring additional analyses.
The primary inlet development test program followed the same basic
approach as that used for recent legacy systems. It relied on CFD and model
testing to arrive at a final configuration. The addition of STOVL requirements
drove additional testing, relative to legacy systems. This, in turn, required con-
ducting full-scale tests at both Rolls-Royce’s and Pratt & Whitney’s test facilities.
By contrast, STOVL requirements drove stream tube testing for the LiftFan at
Rolls-Royce’s test facility. Figures 31 and 32 show full-scale inlet testing at Pratt
& Whitney’s West Palm Beach test facilities [8].

Fig. 32 STOVL propulsion system inlet development full-scale testing.

354 S. P. WURTH AND M. S. SMITH

The most unique and extensive inlet test program addressed the LiftFan inlet
and the aeromechanics issues encountered during the CDP. The development test
program included low-speed wind tunnel testing and static testing using stream
tubes. Figure 33 shows multiple views of a fully instrumented 20%-scale model
tested at German-Dutch Wind Tunnels’ (DNW’s) low-speed wind tunnel in Mar-
knesse, Netherlands. The model was equipped with an instrumented LiftFan,
including variable IGVs and an auxiliary inlet with adjustable doors. The simu-
lated inlet face included a standard inlet rake to map distortion of the inlet
with the auxiliary inlet open. Suction was applied to the LiftFan exhaust and
the simulated engine to generate representative flows through the AIS [8].
Further testing of the LiftFan inlet was accomplished at static conditions
through the use of stream tubes under both subscale and full-scale conditions.
A drawback of the stream tube approach is that the shape of the stream tube is
unique for a given LiftFan airflow/airspeed combination. Building multiple
stream tubes at subscale is manageable, but constructing full-scale stream tubes
is much more difficult and costly. This is due to the sheer size of the stream
tube itself and the loads imposed by an operating LiftFan. A compromise was
reached for the approach using multiple subscale assets. First, the most stressing
conditions were identified, and then full-scale stream tubes were constructed to
reflect these conditions.
Figure 34 shows the streamlines from a CFD simulation that define the shape
of a stream tube. Combined with a model of the upper surface of the aircraft, the

Fig. 33 Twenty-percent STOVL inlet model at DNW’s Marknesse facility.

F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 355

Fig. 34 Representative stream tube lines for a

given inlet flow/airspeed condition.

blue surface would be constructed to generate

the flow field representative of the given flight
condition [8].
The low loads generated during subscale
static testing allowed the use of stereolithogra-
phy to fabricate the stream tubes, reducing
cost and minimizing turnaround time. CFD solutions were required to generate
the lines to fabricate the stream tube. Accordingly, a CFD solution was available
to directly compare against subscale test results, confirming that the desired flow
field had been generated. Figure 35 shows a group of 20%-scale stream tubes used
during the test program [8].
A full-scale stream tube was constructed for use at Rolls-Royce’s LiftFan test
facility in Indianapolis, Indiana. This was the final test of the LiftFan inlet design.
With a flow field representing a stressing condition, a strain-gaged LiftFan was
operated with the stream tube in place. This test collected the data that proved
that the aeromechanics issue encountered during the CDP had been successfully
mitigated. Figure 36 shows a full-scale stream tube used at the Rolls-Royce test
facility [8].
Multiple STOVL jet effects aerodynamics wind tunnel test programs were
conducted to characterize all aspects of the STOVL air vehicle during jetborne
and semi-jetborne modes. These test programs were closely related to the charac-
terization of the STOVL propulsion system. Many of the forces on the air vehicle

Fig. 35 Subscale LiftFan stream tube models (20% scale).

356 S. P. WURTH AND M. S. SMITH

Fig. 36 Full-scale LiftFan stream

tube prior to test stand installation.

during STOVL operations

depended on system thrust and
vector angles. It was necessary
to characterize how the forces
would be accounted for in
analytical models of the aircraft.
We also needed to identify how
they fed into the overall air
vehicle flight control system.
Accurate STOVL jet effects
data are an integral part of the
high-fidelity, STOVL-mode flight simulations required to execute the F-35
program. In jetborne mode, the STOVL jet effects can account for a reduction
in thrust of approximately 8% OGE. This OGE value depends heavily on multiple
variables, including wind speed and direction. When the F-35 is IGE, the OGE
level can be magnified many times, depending on such variables as relative jet
angles and thrust values, and height above the ground. For STO-critical con-
ditions, the STOVL jet effects can reduce vertical thrust by approximately 20%
and axial thrust by approximately 2%. One force affecting the pitching moment
on the aircraft that must be accounted for is the inlet suction on the various
inlets of the STOVL propulsion system. Another force comes from the suckdown
effects from the various nozzles. Figure 37 shows an F-35 STOVL 12%-subscale jet
effects model upon entry at DNW’s Large Low-Speed Facility. Ten separate test
entries were conducted, covering all STOVL flight regimes from pure hover,
IGE, through transition to wingborne flight [12].
The particle image velocimetry technique is used to examine the flow field
generated by the exhaust system installed in the aircraft. The technique requires
the flow field to be seeded with a vegetable-based oil. A laser mounted aft of
the model is used to illuminate and capture the particles at a given moment in
time at a given frequency (approximately 3 Hz). Two cameras mounted to a tra-
versing rig on the side of the wind tunnel are used to capture the image produced
by the laser. These images are combined and used to generate a complete picture
of the flow field. In the case of STOVL jet effects, both the LiftFan and core nozzle
plumes are seeded.
Figure 38 shows a flow field center-line cut example with forward airspeed. A
horseshoe vortex is formed as the wall jet spreads forward and is re-entrained into
the LiftFan plume. The location of the horseshoe vortex is heavily dependent on
the strength of the forward wall jet and forward aircraft velocity. At these higher
forward velocities, the fountain is tilted aft and re-entrained into the core nozzle
plume [12].
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 357

Fig. 37 F-35B STOVL jet effects subscale testing.

Similar to what was accomplished during the CDP, HGI testing was con-
ducted at the BAE Systems facility in Wharton, United Kingdom. The test
setup was updated to reflect the configuration of the F-35 and a similar but
expanded test program conducted to establish a suitable operational envelope.

Fig. 38 Particle image velocimetry example (hover IGE).

358 S. P. WURTH AND M. S. SMITH

2. PROPULSION INTEGRATION AND INTEGRATED FLIGHT PROPULSION CONTROL

GROUND TESTING

The degree of integration between the air vehicle and the propulsion system
resulted in test programs not found in legacy systems. Two key areas were inves-
tigated at Pratt & Whitney’s West Palm Beach test facilities: the integrated sub-
system architecture and the IFPC during STOVL operation.
The level of systems integration between the air vehicle and the propulsion
system dictated that risk reduction testing be accomplished before all systems
came together in the air vehicle. An enhanced integrated systems test was con-
ducted that included installing major aircraft subsystems on a test stand in
West Palm Beach. This test included the IPP, the ESG, hydraulic pumps, the
ESG lubrication system, and all associated control hardware. The IPP was put
through its paces, beginning with operation as an APU, then providing power
to start the engine. It then transitioned to cooling mode, operating off engine
bleed air and using the FDHX as a heat sink.
Failure scenarios were also simulated, such as engine flameouts in which the
IPP would convert from ECS mode to an emergency power combustion mode.
The ESG was used to both start the engine and provide power as if operating
in the aircraft. The hydraulic pumps operated as if installed, generating
minimal load on the engine gearbox during starts, then pressurizing once the
engine had achieved a stable idle. Load banks to simulate aircraft systems were
included on the test stand to represent aircraft operation. The testing verified
the design architecture while gaining experience connecting all the systems and
observing their interactions before an actual aircraft installation took place.
The early design of the AMVC was initially tested in support of a 2003 F135
ground test. This initial architecture was matured and qualified to support flight
operations of the initial CTOL variant aircraft AA-1. First flight was achieved in
December 2006, and the aircraft has been flying flawlessly since, earning the praise
of the pilot community. F-35 Chief Test Pilot Jon Beesly stated [8]: “The engine
performed flawlessly. The performance was actually better than I expected,
which is interesting because I expected a lot.”
Building on the initial architecture, improvements necessary to meet the strict
pitch control and limit control requirements of STOVL operation were
implemented in 2007. Following the completion of the ground test program for
STOVL, powered-lift flight clearance was granted for the AMVC control architec-
ture in January 2009. This was a major milestone for Pratt & Whitney and the
F135 program. Meeting the pitch coupling and limit control requirements with
this software represented a key milestone as the program headed toward powered-
lift operation. A key area of STOVL propulsion system operation is engaging a
physical rotor speed redline limit. In this situation, a traditional control system
would automatically throttle back fuel flow to the engine, cutting back thrust.
This would be an unacceptable and undesirable outcome for a STOVL aircraft
operating in vertical mode. As described previously, the AMVC automatically
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 359

takes the appropriate action to adjust not only fuel flow, but also nozzle areas and
LiftFan actuators such that thrust and pitch control are maintained [8].

C. AIRCRAFT PROPULSION GROUND AND FLIGHT TESTING

Combined, the 18 flight test aircraft completed more than 16,000 flight hours and
64,000 test points in the program, which started in 2008. The validation and ver-
ification programs of the F-35 propulsion system had to ensure that the propul-
sion system could withstand the extreme rigors of warfighter usage at all
maneuver, climatic, and battlefield conditions. On 15 December 2006, the first
flight of the CTOL variant (F-35A), under Jon Beesly, was completed and initiated
one of the largest flight test programs in the history of aviation. On 11 June 2008,
the first F-35B (BF-1), piloted by Graham Tomlinson, took to the skies in CTOL
mode flight for its first flight milestone. Shortly after its first flight, BF-1 was
moved to the hover run facility to exercise the propulsion system in STOVL
mode. No high-power runs were conducted initially, but the successful operation
cleared the use of STOVL mode, if required, for an aircraft recovery. STOVL clear-
ance would follow later, after full-power operations on the hover pit. By April
2009, the F-35 was ready for full-power STOVL testing on the hover pit.
Similar to the X-35 program, the aircraft was mounted on a force-and-
moment system to measure performance and control power, as shown in
Fig. 39. Unlike the X-35, no press-ups from the hover pit were to be performed,
because an in-flight build-down approach was selected as a lower-risk approach
than conducting press-ups. STOVL performance was confirmed, and control
authority of the various effectors was verified. The aircraft completed an aggres-
sive series of full-up powered-lift engine runs to exercise and evaluate the robust-
ness of the F135 control systems and the performance of the AMVC. Hover pit

Fig. 39 Mounted on force-and-moment system at hover run facility.

360 S. P. WURTH AND M. S. SMITH

ground tests validated the propulsion system and aircraft response, and the first
F-35B was ready for its first STOVL flight.
With the successful completion of the hover pit test, the F-35 was cleared to
begin powered-lift operations. On 17 October 2009, BF-1 completed its first non-
tethered test of the STOVL system during taxi tests. On 15 November 2009, BF-1
arrived at Naval Air Station Patuxent River, Maryland, in preparation for hovers
and the first vertical landing. Finally, on 7 January 2010, BF-1 engaged its STOVL
propulsion system in flight for the first time.
On 18 March 2010, BF-1, under the control of Graham Tomlinson again, con-
verted, slowed to a hover, and completed the program’s first vertical landing. The
STOVL flight test program continued, and challenges were uncovered. One was
that the Auxiliary Air Inlet Door (AAID) design was not stiff enough to handle
turbulence generated by the Upper Liftfan Door (ULFD) at certain speeds.
During envelope expansion of the F-35B, it was discovered that under certain
flight conditions during semi-jetborne flight with the ULFD opened to its mid-
position, there were increased oscillatory loads acting on the AAID. The
primary culprit for these increased loads was buffeting from the ULFD.
However, the physical mechanism by which the buffeting was occurring was
not fully understood. CFD was used to provide that understanding. A summary
of the steady-state results is presented in Fig. 40 for the mid-position ULFD con-
figuration. The streamlines tend to flow around the AAID at the higher Mach

Fig. 40 Summary of steady-state AAID CFD solutions.

F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 361

numbers, whereas at the lower Mach numbers the yellow streamlines are much
lower and impinging on the AAID. The effect of the engine power setting can
also be seen in the sequence of pictures. As the engine power setting increases,
the streamlines are pulled into the AAID opening [13].
After extensive CFD and analysis of the stiffness of the door and its actuation
system, a redesign of the actuation system and stiffening of doors was
implemented in 2011. This returned the door design to a full-life capability
[13]. The F135 propulsion system has performed acceptably throughout its
flight test program and, as a result, has led the F-35A/B/C flight test programs
to a successful SDD closure.

VI. CONCLUSION
The integration of the F135 propulsion system with the F-35 spans more than 20
years of propulsion design, development, and testing. Both STOVL and CTOL/
CV propulsion systems meet all requirements. The LiftFan-based system, one
of the higher-risk technologies for the program, works and integrates well with
the F-35, with many challenges overcome during its development.
The ASTOVL program established the concept feasibility of a LiftFan-based
STOVL system. The lift system architecture was proven using the LSPM and
testing at facilities at NASA’s Ames Research Center. These successfully showed
that low-speed, powered-lift aerodynamics, hover control power, and HGI of a
shaft-driven LiftFan were acceptable and an improvement over legacy STOVL
systems.
The demonstration of a flightworthy propulsion system in the X-35 proved
that a well-integrated team is required for success. The formation of a highly inte-
grated IFPC IPT allowed for quick solutions to the formidable control challenges
the team faced. The adaptation of a legacy conventional main engine with an
unconventional LiftFan was a success. The IFPC IPT integrated X-35B aircraft
and engine software and related hardware that led to a successful flight test
program. The innovative DSI inlet design allowed for a short, lightweight inlet
system that had excellent performance characteristics. Also, HGI, a concern for
decades with legacy STOVL aircraft, was proven to be manageable and essentially
a nonissue with the LiftFan arrangement.
The F-35’s successful development of the propulsion system took the
lessons learned from the X-35 and applied them across all the propulsion
systems. The highly integrated cross-company team of Lockheed Martin, North-
rop Grumman, BAE Systems, Pratt & Whitney, Rolls-Royce, and the F-35 Joint
Program Office worked very well together. With a great combination of CFD, sub-
scale wind tunnel testing, full-scale testing, stream tubes, and engine and LiftFan
testing, the team used design techniques that maximized the benefit of each. The
LiftFan inlet was improved to solve the serious LiftFan aeromechanical problem
seen on X-35B with a new door design. The auxiliary inlet benefited from CFD
362 S. P. WURTH AND M. S. SMITH

analysis to solve an unforeseen door vibration issue. The full-scale testing, stream
tubes, and extensive subscale inlet testing gave the propulsion system unprece-
dented performance for a single-engine 5th-Generation fighter. The VAVBN
improvement over the X-35B’s LiftFan hood provided much more flexibility in
LiftFan operation. All these examples show how the team evolved these designs
from concepts to operational reality. The degree of IFPC integration was absol-
utely unprecedented, with the propulsion and flight control systems in complete
control from conversion to STOVL mode, slowing to a hover, and performing a
vertical landing. The F-35 propulsion integration exemplified the importance of
power and thermal management in the development of an integrated solution.
A testament to the F-35’s propulsion design success was captured with a quote
from a Royal Air Force F-35B test pilot in an interview on 26 January 2018. Sqn.
Ldr. Andy Edgell stated:
I will never forget my first hover in the Harrier and I will never forget my first
hover in the F-35B. My first hover in the Harrier is akin to trying to stay alive
on a unicycle and I don’t think I made much of a conscious thought at the time
in the hover. It was only once I managed to get her down on deck safely, and
then retrospectively I thought, “How on earth did I manage that?” Whereas in
my first hover in the F-35B I sat there at that moment, and looked around, rea-
lized, I was stationary in every axis, and thought, “Goodness gracious me, the
guys who designed this are absolute geniuses.” [14]

Fig. 41 Propulsion systems at work in the F-35A, F-35B (CTOL and STOVL modes), and F-35C.
F-35 PROPULSION SYSTEM INTEGRATION, DEVELOPMENT, AND VERIFICATION 363

The fleet continues to grow, with more than 290 F-35As, F-35Bs, and F-35Cs,
totaling more than 135,000 flight hours in the field. As such, the F-35 and its
unique and successful propulsion system will continue to meet the warfighter’s
every challenge, as shown in Fig. 41. In the final analysis, the F-35’s propulsion
integration, design, and development were a success.

ACKNOWLEDGMENTS
Steven P. Wurth and Mark S. Smith would like to thank their colleagues
throughout the multicompany propulsion enterprise whose efforts brought
this propulsion system to life. The ASTOVL, X-35, and F-35 propulsion teams
constituted one of the best collections of subject-matter experts in the fields of
propulsion aerodynamics, integration, and IFPC. The authors would also like to
thank the executive review team for providing guidance in the development of
this paper.

REFERENCES

[1] Wurth, S., Walker, G., and Fuller, Dr. J., “F-35B IFPC Development,” AIAA Paper
2013-4243, August 2013.
[2] Maddock, I. A., and Hirschberg, M. J., “The Quest for Stable Jet Borne Vertical Lift:
ASTOVL to F-35 STOVL,” AIAA Centennial of Naval Aviation Forum, 11th AIAA
ATIO Conference, Virginia Beach, VA, Sept. 2011.
[3] Bevilaqua, P. M., “Joint Strike Fighter Dual-Cycle Propulsion System,” Journal of
Propulsion and Power, Vol. 21, No. 5, Sept./Oct. 2005.
[4] Bevilaqua, P. M., “Future Applications of the JSF Variable Propulsion Cycle,”
AIAA Paper 2003-2614, July 2003.
[5] Wurth, S., Mahone, T., Hart, J., and Baxter, J., “X-35B Integrated Flight Propulsion
Control Fault Tolerance Development,” AIAA Paper 2002-6019, Nov. 2002.
[6] Weigand, C., et al., “F-35 Air Vehicle Technology Overview,” AIAA Aviation 2018
Conference, Atlanta, GA, May 2018 (submitted for publication).
[7] Buchholz, M. D., “Highlights of the X-35 STOVL Jet Effects Test Effort,”
AIAA Paper 2002-59622002, Nov. 2002.
[8] Sylvester, T. G., Brown, R. J., and O’Conner, C. F., “F-35B LiftFan Inlet
Development,” AIAA Paper 2011-69402011, Sept. 2011.
[9] Fuller, J., “Advanced Multi Variable Control (AMVC) and Its Application in
Turbomachinery,” AIAA Paper 2010-01-1737, Nov. 2010.
[10] Lo Gatto, E., “Overview of the JSF LiftSystemTM,” International Powered Lift
Conference, Hartford, CT, Sept. 2016.
[11] Smith, M. S., “Joint Strike Fighter: X-35 and F-35 Propulsion System Integration,”
39th AIAA/ASME/SAE/ASEE Joint Propulsion Conference, Huntsville, AL,
July 2003.
364 S. P. WURTH AND M. S. SMITH

[12] Mange, R., and Hoggarth, R., “Highlights of the Lockheed Martin F-35 STOVL
Jet Effects Program,” 08IPLC-0022, Royal Aeronautical Society International Lift
Conference, London, UK, July 2008.
[13] Cox, Dr. C. F., “F-35B Auxiliary Air Inlet Analysis and Design,” AIAA Paper
2013-218, Jan. 2013.
[14] Robinson, T., “Inside F-35B Flight Test,” Royal Aeronautical Society [interview with
Sqn. Ldr. Andy Edgell], https://www.aerosociety.com/news/inside-f-35b-flight-
test/ [retrieved 26 Jan. 2018].
 CHAPTER 10

F-35 Subsystems Design,

Development, and Verification
Drew Robbins , John Bobalik

†
, David De Stena‡, Ken Plag§,
Keith Rail} and Ken Wall
Lockheed Martin Aeronautics Company, Fort Worth, TX

F-35 vehicle subsystems were developed, qualified, and fielded as part of

an overall integrated systems architecture. Because of the unique
requirements derived from the common tri-variant integrated systems
architecture, several key technologies were developed as part of the
overall subsystems development. These include electric primary and
secondary flight control actuation, electrical starter/generator systems,
lithium-ion batteries, integrated power and thermal management
systems, and expanded ejection system performance envelopes. With
any development and qualification of new technologies, several
discoveries were observed and resolved during qualification. With the
successful implementation of a fully integrated subsystem architecture,
F-35 vehicle subsystems are fully qualified and operational, providing a
high-performing, integrated architecture. These subsystems continue to
successfully support the F-35 fleet well past its current 100,000 flight-
hour mark.

I. INTRODUCTION
The vehicle subsystems on the F-35 were developed, qualified, and fielded as part
of the F-35 System Development and Demonstration (SDD) program. Continuing
from previous efforts in both configuration assessments and technology develop-
ment [1], these systems were developed as part of an overall integrated air vehicle
philosophy [2]. This paper discusses five systems: the Electro-Hydrostatic Actua-
tion System (EHAS), the Power and Thermal Management System (PTMS), the
Electrical Power System (EPS), the Hydraulic and Utility Actuation (HUA)

 SeniorManager, F-35 Utilities and Subsystems, MZ 6500.

†
Aeronautical Engineer Principal, Advanced Development Programs, MZ 2353.
‡
Principal Systems Engineer, F-35 Utilities and Subsystems, MZ 6500.
§
Principal Systems Engineer, F-35 Electrical Power Systems, MZ 6500.
}
Deputy Manager, F-35 Vehicle Management Hardware, MZ 6468.
 Manager, F-35 Power and Thermal Management System and Pilot Systems, MZ 6500.

365
366 D. ROBBINS ET AL.

system, and the ejection system. As part of the development of these systems, the
key technologies discussed are electric primary and secondary flight control actua-
tion, electrical Starter/Generator (S/G) systems, lithium-ion batteries, integrated
PTMSs, and expanded ejection system performance capabilities (Fig. 1). With any
development program, integration challenges occur. The challenges this paper
reviews are the increased priority of modeling and simulation, improvements
implemented because of the early AA-1 inflight emergency, development
of lithium-ion batteries, regenerative power impacts on both the EHAS and
EPS, engine start functionality, emergency power transitions and qualification,
and hydraulic system integration. Because of the integrated effort by design
engineers across multiple companies, the F-35 flight controls and utilities and
subsystem hardware are fully fielded and operational, supporting the growing
F-35 fleet.
As discussed in Ref. [2], the final aircraft design required an integrated
architecture vs a more traditional federated or distributed system. A federated
system relies on several independently designed subsystems, whereas an inte-
grated architecture attempts to achieve equivalent or greater capability with less
hardware. As a result, these architectures require more integrated development
and testing. Integrated architectures allow for increased performance, improved
affordability, and improved weight due to dramatically reduced amounts of hard-
ware. Multiple development programs formulated the final configuration, as dis-
cussed in Ref. [3]. Figure 1 compares a federated system to the F-35 integrated
system. Of note, the integrated architecture drove new hardware development
and had an impact on the overall system management (including software),
qualification, and flight test approach. Rather than relying on individual
systems to operate to a predefined set of criteria, an integrated architecture
relied heavily on a robust system engineering approach. This applied throughout
the process, from requirement derivation and allocation through integrated qua-
lification and flight testing.

Fig. 1 Federated vs integrated vehicle subsystems.

F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 367

Fig. 2 Basic schematic of an F-35 EHAS.

II. SYSTEMS OVERVIEW

This section provides a brief overview of the five systems comprising the vehicle
subsystems, including a brief history of how the systems evolved under the
pressures to integrate functions into more efficient products. It provides a very
basic understanding of how the federated functions combined into integrated
systems. This is important in understanding how and why the aircraft-level
requirements drove the flow down of requirements, the key technologies, and
the integration and development issues discussed in later sections.

A. ELECTRO-HYDROSTATIC ACTUATION SYSTEM

The F-35 is the first manned aircraft and first fighter aircraft to use electric
primary and secondary flight control actuation. As the primary actuation
system, the EHAS (Fig. 2) incorporates a new type of self-contained actuator
with an integral motor and pump driven solely by electric power (Fig. 3). These
components, along with a reservoir, reside directly on the manifold of the actuator
and replace the control valve used on a conventional actuator. The pump directly
pushes fluid into the actuator, and the direction of actuator movement follows the
direction of the pump rotation. Each Electro-Hydrostatic Actuation (EHA) motor
is controlled by an Electronic Unit (EU) channel that contains a low-power (28
VDC) side and a high-power (270 VDC) side. The low-power side is responsible
for digital loop closure and fault monitoring, and the high-power side is respon-
sible for three-phase power switching.
Due to this new technology, the development effort included significant design
maturation and challenges with scaling the design for the unique needs of the
368 D. ROBBINS ET AL.

Fig. 3 EHAS installations.

three aircraft variants. The aircraft size, performance, and safety requirements
drove the system to be larger and more integrated than previous electric actuators
used in drone applications or for concept demonstration. One example of the
latter case is the Joint Strike Fighter (JSF) Integrated Subsystems Technology
(J/IST) program [2]. The interfaces between the EHAS and PTMS had to go
beyond basic connections and performance criteria. They had to include modes
of operation and how to collectively transition between modes in case of a
failure in any of the systems. All these considerations played into iterative
improvements as the EHAS matured through design, integration, and flight test.

B. POWER AND THERMAL MANAGEMENT SYSTEM

The PTMS is a multifunctional system designed by Honeywell International Inc.
combining traditionally separate environmental control and auxiliary power
functions into one integrated system. It provides conditioned air for cockpit
and equipment cooling, refrigerated liquid for equipment cooling, and electrical
power for ground maintenance, main engine starting, and inflight emergencies.
Figure 4 provides a simplified schematic of the system.
The cooling function is performed by a closed-loop air cycle powered by
expanding bleed air across a Power Turbine (PT). The closed-loop air cycle
includes a compressor (C in the figure), Cooling Turbine (CT), and S/G
mounted on a single shaft. High-temperature, high-pressure air exiting the com-
pressor dissipates heat to a Fan Duct Heat Exchanger (FDHX) installed in the
engine fan air duct. The engine FDHX is unique with respect to keeping exchanger
waste heat output within the propulsion cycle rather than overboard. The air then
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 369

passes through a polyalphaolephin (PAO) to air Heat Exchanger (HX) to sup-

plement air cycle waste heat removal when fan duct temperatures are too high
to be used as an effective heat sink. The air is cooled further in a recuperator
HX before being expanded through a CT. The cold, low-pressure air exiting the
CT passes through the load HX.
The load HX includes an air-to-air core to cool preconditioned bleed air to cool
the cockpit, equipment, and Onboard Oxygen Generation System (OBOGS). It also
has a liquid-to-air core to provide liquid cooling to the majority of the mission
systems equipment. Air exiting the load HX passes through the recuperator HX
before returning to the compressor inlet. In addition to providing preconditioned
air to the load HX, the engine precooler supplies the Onboard Inert Gas Generation
System (OBIGGS). The OBIGGS removes oxygen from air and provides nitrogen-
enriched air to the fuel tanks for vulnerability and lightning protection.
In addition to the cooling and air conditioning mode, the PTMS has four
other modes: ground maintenance, main engine start, and self-powered and
bleed-driven emergency electrical power. Figure 5 summarizes the PTMS operat-
ing modes and depicts the turbomachine. This integrated approach to cooling
relied on unprecedented integration among the propulsion, power, and environ-
mental control system cycle.
Similar to the way in which it operates in the bleed-driven cooling mode, the
PTMS can operate as an emergency generator in flight by reducing cooling and

Fig. 4 Simplified PTMS schematic.

370 D. ROBBINS ET AL.

Fig. 5 The PTMS provides both environmental control and auxiliary power functions.

diverting shaft power to the S/G. The PTMS can provide cockpit conditioning,
flight-critical equipment air cooling, and electrical power in this mode.
The PTMS can be reconfigured by valves to operate like a traditional Auxiliary
Power Unit (APU) for self-start, ground maintenance, main engine start, and
inflight emergency power. The compressor-combustor-turbine section of the
PTMS turbomachine is effectively an APU that can produce shaft power to
drive an integrated S/G and/or an open-loop air cycle. The S/G mounted on
the same shaft system as the turbomachine provides electrical power for checkout
of the aircraft’s electrical, avionics, and flight control systems, as well as inflight
emergency power. When provided with 270-VDC aircraft battery power, the
S/G also acts as a start motor, providing the initial rotation and torque required
to start turbomachine operation in the combusted mode.
Once the PTMS has been reconfigured to the combusted mode, the propulsion
engine can be started electrically by providing PTMS-generated 270-VDC electri-
cal power to an S/G mounted on the engine gearbox. In this same configuration,
the PTMS can also provide air and liquid cooling and electrical power to perform
ground maintenance without using ground support power or cooling. Because the
FDHX is not available in this mode, cooling cycle waste heat is transferred to fuel
from the PAO/air HX. An onboard fan is used to enhance heat transfer from the
heated fuel to the atmosphere.
The PTMS also can reconfigure to the combusted mode in flight. In the event
of an engine flameout, the PTMS transitions from the normal cooling mode to the
combusted mode throughout the flight envelope. The 270-VDC aircraft battery
provides flight-critical electrical power at high altitudes until the aircraft descends
and the PTMS can provide sufficient power to the flight control actuation system.
After mode transition is complete, the PTMS provides nonessential electrical
power at high altitudes and flight-critical electrical power at lower altitudes,
including enough power to facilitate a safe landing. Along with providing flight
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 371

control power, the PTMS simultaneously provides power to assist with propulsion
engine air-start within the engine start-assist flight envelope.

C. ELECTRICAL POWER SYSTEM

The EPS consists of two subsystems: the Electrical Power Generating System
(EPGS) and the Electrical Power Management System (EPMS). The overall
control and monitoring of the system is supplied by redundant software
running in the Vehicle Management Computers (VMCs).
The EPGS consists of a single Engine Starter/Generator (ES/G) that provides
two independent outputs of 80-kW power each during normal operation. During
main engine start, the ES/G uses electrical power from the PTMS via a pair of
Inverter/Converter/Controllers (ICCs) to create mechanical shaft power to
motor the engine. After engine start, the ES/G automatically switches to generate
mode, in which the ICCs convert the unconditioned power from the ES/G into
regulated 270 VDC. Two voltage converters, each rated at 160A of power,
provide 28 VDC for low-power, critical loads. The 270-VDC-to-115-VAC inver-
ter provides the aircraft with 5.4 kVA of 115-VAC power for weapons stations and
wing-fold on the F-35C aircraft. The three Converter/Regulators (C/Rs) provide
redundant 28-VDC power to the flight-critical loads. These C/Rs accept power
from the 28-V distribution system, 28-VDC battery system, and two sets of per-
manent magnet generators: one located within the ES/G and the other within the
PTMS turbomachine. These inputs provide highly reliable, break-free power. Each
C/R provides power to one branch of the triple-redundant flight control system.
The EPMS consists of the power distribution equipment, providing the on/off
switch control of the loads and short-circuit protection for the wiring. The equip-
ment uses a combination of traditional mechanical circuit breakers and relays,
smart contactors, and solid-state switches to distribute power from the various
sources to the using loads. (Smart contactors combine overcurrent monitoring
and switching into a single unit.) The 28-VDC battery system consists of an eight-
cell lithium-ion battery and charger/controller. The 270-VDC battery system is
made up of an 84-cell lithium-ion battery and a charger/controller. This system
provides fill-in power to 270-V flight-critical loads and provides electrical
power for PTMS self-start.
The F-35 EPS has expanded on the level of integration incorporated compared
to the F-22 aircraft. All load switching is now performed by the EPS via software
commands from the using system. When the PTMS turbomachine is the
power source, a hand-off between the PTMS and EPS provides a power-available
signal. This depends on the instantaneous capacity of the turbomachine, but
allows the EPS to actively load manage to meet critical needs within the available
capacity. A similar relationship is present between the EPS and the stores manage-
ment system. In it, the instantaneous power capacity is provided so they can
manage which weapons are powered based on the available power and mission
needs.
372 D. ROBBINS ET AL.

Fig. 6 EPS architecture.

An overview of the system architecture is shown in Fig. 6, with the part

locations depicted in Fig. 7. The system is nearly common across all variants.
The exception is the size of the output rating for the inverter, which is slightly
larger to accommodate the wing-fold function of the F-35C aircraft.

D. HYDRAULIC AND UTILITY ACTUATION

The hydraulic system was designed to provide maximum power to F-35 utility
actuation systems and F-35B lift system functions. It uses approximately one-
fourth of the hydraulic power capacity that would have been required for conven-
tional subsystems. The relatively small capacity of the hydraulic system and large
power demands of the utility systems led to a unique balance of actuation technol-
ogies, power sources, and system architecture. Although the hydraulic power
system architecture is common across the F-35 aircraft variants, the unique actua-
tion systems across the aircraft variants resulted in approximately three times the
number of utility actuation functions, compared to legacy aircraft. A block
diagram of the F-35 utility actuation systems is shown in Fig. 8.
Characteristics of the hydraulic system include two independent 4000-psi
hydraulic power systems that are common across all three variants. Each hydrau-
lic system consists of:
. A 29.8-gallon per minute (gpm) engine-driven variable-displacement pump
. A boot-strap reservoir (F-35A/C at 640 in3, F-35B at 945 in3)
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 373

Fig. 7 EPS installations.

Fig. 8 HUA block diagram.

374 D. ROBBINS ET AL.

. Maintenance-free accumulators for reservoir pressurization and parking

brake functionality
. An electric-driven pump for ground maintenance and inflight emergency
operations
. Five-micron filtration

Redundancy was incorporated for safety-critical actuation applications, including

redundant lines to the actuator applications and shuttle valves at the actuators.
Additional features include:
. Inflight stowage and latching to support aircraft stealth requirements
. Quick disconnects on selected components to facilitate rapid aircraft turn
times
. Isolation of landing gear and F-35B Short Takeoff and Vertical Landing
(STOVL) doors systems when not in use
. Switched high/low-side electrical controls to preclude unintended actuation
. Computerized reservoir-level sensing
. An electronic filter replacement indication

E. EJECTION SYSTEM
The F-35 ejection seat (Fig. 9) provides a
means for rapid egress from the cockpit
under emergency conditions. To initiate ejec-
tion, the pilot pulls the ejection handle,
which allows the transparency removal
system to cut the canopy, fires the rocket
motor, and activates the leg/arm restraints.
When the seat separates from the aircraft,
the drogue chute deploys to stabilize and
decelerate the seat/occupant mass. The para-
chute deployment box fires to initiate seat/
occupant separation and deploy the main
parachute at the calculated delay provided
by the sequencer to ensure deceleration to a
non-injurious parachute opening load. The
parachute harness features a head support
panel to control the head from excessive aft
extension.

Fig. 9 Production ejection seat.

F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 375

The auto-escape system is configured in the F-35B only for the event of
catastrophic failure of the LiftFanw. Such a failure can produce a rapid onset
of forward pitch accelerations, perhaps exceeding the pilot’s ability to react and
manually eject. The system is designed to provide detection of inertial attitudes
and rates in the pitch-down plane within tolerances to filter out normal flight
control input and response. The detection occurs in the control law software
application of each VMC during conditions in which the system is armed.

III. REQUIREMENTS DERIVATION

This section discusses the flow down of requirements driven by the contract spe-
cification and airframe design decisions. Several major design decisions drove the
requirement for several key technologies to be discussed in the next section. These
technologies were critical to the success of the overall platform. In addition, the
weight and volume growth obstacle discussed in Ref. [1] introduced, in some
cases, an additional design cycle midway through the development.

A. ELECTRO-HYDROSTATIC ACTUATION SYSTEM

The EHAS is much more complex than a conventional hydraulic actuation system
and does not follow the classical growth trades well when compared to a
stand-alone system. This is especially true because the EHAS is scaled up to the
size required for the F-35C. The increased F-35C capability and, therefore, size
are driven by the unique and stringent performance requirements for landing
the F-35C on Navy carriers. The reasons for choosing the EHAS as the flight
control actuation for the F-35 go beyond the individual system and look at the
benefits to the overall aircraft configuration [2]. Large components like hydraulic
pumps, which normally reside on the accessory gearbox in the body of the aircraft,
can be removed and distributed in the wings and aft portions of the aircraft. This
provides available fuel volume and improves the aircraft cross-section, benefitting
aircraft performance. As a result, the EHAS became the system of choice for
the F-35.
The architecture chosen for the EHAS and vehicle systems in general relies on
the distributed processing of certain lower-level functions based on commands
from the VMC. This design, along with greater interaction among the EHAS,
EPS, and PTMS, increased the complexity of the interface requirements used to
maintain coordination and the timing needed to safely control the systems. The
additional requirements were derived and managed in the subsystem specifica-
tions and the Vehicle Systems Network interface control document.

B. POWER AND THERMAL MANAGEMENT SYSTEM

The F-35 PTMS requirements were derived from legacy federated subsystem
architecture requirements for both auxiliary power and environmental control.
376 D. ROBBINS ET AL.

The PTMS turbomachine performs legacy air cycle machine functions to provide
equipment cooling and cockpit environmental conditioning. It also performs
legacy APU functions to start the propulsion engine, provide ground maintenance
power and cooling, and provide inflight emergency power. The integration of
these functions required turbomachinery with a wide operating range to cover
open-loop, APU-like operation from sea level to 50,000 ft and closed-loop,
refrigeration mode operation throughout the flight envelope. Propulsion engine
start requirements sized the turbomachine PT and the compressor required to
feed it. Sizing the compressor for engine start resulted in a compressor that was
a good match for the F-35 environmental control functions, although not opti-
mized for those functions. Propulsion engine start requirements also sized the
generator integrated into the turbomachine. Sizing the generator for engine start
provided ample power for inflight emergency modes and ground maintenance.

C. ELECTRICAL POWER SYSTEM

Following on from prior designs, the F-35 EPS was designed around a 270-VDC
system that provides a significant weight savings and enables the system to
provide essentially break-free power transfers under many conditions. Break-free
power transfers ensure that the power-to-aircraft loads are not interrupted when a
transfer to another power source occurs. The ability to provide break-free power
at very high load levels enabled the flight control system to implement the EHAS,
which benefitted the overall aircraft configuration extensively.
The advent of high-powered flight controls created the need for redundant
sources, and these are provided by the pair of 80-kW generators housed in
a single chassis and spinning on the same shaft. This provides sufficient electrical
isolation and redundancy to meet the flight-critical power needs, but also mini-
mizes weight. The original configuration of these generators was a switched-
reluctance topology, with the idea that the units would be implemented deep
within the engine. This dictated a unit capable of operating in that high-
temperature environment. Other issues prevented the generators from being
located within the engine, so the need for this specialized topology was never rea-
lized. As part of the continued evaluation of the system, it was determined that
a weight savings of more than 100 lb could be realized by converting from a
switched-reluctance system to a more conventional synchronous system. This
system was implemented midway through the flight test program.
The next step aligned with the more-electric aircraft technology thrusts of
the Joint Advanced Strike Technology and J/IST programs [4, 5] and the drive
toward more integrated systems. The generating system was implemented as
the electric start system to provide significant weight and complexity savings to
the overall aircraft through the removal of the more traditional engine starting
system. Additionally, and more important to the overall aircraft configuration,
this allowed for the removal of the accessory drive gearbox. This allowed for
a reduction in the cross-section of the airframe. Through the creation of the
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 377

ICCs, the generators are operated as motors during the start mode to provide
torque to the engine.
To provide nearly instantaneous fill-in power to the EHAS while recovering
from a generator or engine fault, a lithium-ion battery was tied to the 270-VDC
generation bus. It is generally referred to as a 270-VDC battery because of the
bus voltage, although its actual operating voltage exceeds 300 V.

D. HYDRAULIC AND UTILITY ACTUATION

The transition from legacy hydraulic flight-control actuation systems to electro-
hydrostatic flight control actuation led to unique challenges for the actuation of
the numerous utility-actuation functions. The quantity of utility functions and
wide range of power demands led to a combination of hydraulic and electrical
power sources for the utilities. Extensive studies with numerous hydraulic, pneu-
matic, and electric actuation suppliers with a focus on weight and cost efficiencies
led to the utility system architecture shown in Fig. 7. Although there was extensive
use of common utility functions across the F-35 aircraft variants, unique aspects of
each aircraft resulted in a wide range of utility actuation systems (Table 1).
The electric architecture of the F-35 flight control actuation systems, high
quantity of utility system functions, and stealth aircraft requirements resulted
in several differences in the utility system architecture, compared to legacy
utility actuation systems. These differences included the elimination of the
airframe-mounted accessory drive gearbox because the hydraulic pumps were
mounted directly to the main engine gearbox. A dual hydraulic system architec-
ture was used to provide fault tolerance for the F-35B lift system actuation and

TABLE 1 HUA SYSTEM COMMONALITY

Common Partial F-35A Unique F-35B Unique F-35C

Commonality Unique
. Hydraulic . Weapon Bay . Refuel . Lift-Fan . Wing-
Power Door Drive Receptable Inlet Guide fold
Generation . Refuel . Arresting Vanes, . CV
. Landing Probe Gear Vane-Box Arresting
Gear (F-35B/C) Nozzle and Hook
. Internal Roll-Post
Architecture . Missionized Gun and . Ailerons
Nozzles
. PTMS and Gun Door . Cat-
Utility (F-35B/C)
. Lift-System
Doors Launch
Function Bar
Door . LiftFanw
Clutch
. Gear
Shrink
378 D. ROBBINS ET AL.

backup actuation for essential landing, tail hook, and aerial refueling functions.
Utilization of the redundant hydraulic system for safety-critical applications
eliminated the legacy pneumatic bottles and associated servicing. In addition,
advanced hydraulic motors were incorporated onto the Weapon Bay Door
Drive (WBDD) systems based on their power efficiency and weight optimization.
Electric motor pumps were implemented to support maintenance at austere
sites, aircraft tow braking, and dead-stick landings. There was also extensive use
of door sequencing functions for the F-35B STOVL doors, weapon bay doors,
and landing gear doors. Intermediate door positions for the LiftFan inlet door
actuation were incorporated to accommodate the LiftFan inlet recovery. The
weapon bay doors are driven to a mid-stroke position and held to support the
lift improvement device’s F-35B vertical landing function. Finally, an advanced
wing-fold system was implemented on the F-35C for weight optimization and
volume reductions.

E. EJECTION SYSTEM
The F-35 Martin-Baker US16E ejection seat was built on legacy ejection seat
designs. The design was modified to incorporate performance features to
comply with the increasingly challenging anthropometric and pilot safety require-
ments. The ejection seat needed to be able to accommodate a wide range of
pilot anthropometry, from Case 1 to Case 8, with a nude mass range of 103
lbm to 245 lbm. At the same time, it had to balance this with the advanced aircraft
operating envelope and strict terrain clearance requirements defined for the F-35.
The US16E design also integrates components of the F-35 life support systems,
such as the breathing regulator and the backup oxygen supply. Further, it inte-
grates into the unique helmet-mounted display system.

IV. KEY TECHNOLOGY DEVELOPMENT

Several key technologies were critical to the development program to enable the
weight, volume, and performance promises of the integrated systems. These
technologies included the EHAS, an FDHX, the integration of a permanent
magnet S/G on the same high-speed shaft as the turbomachine, a 270-V
battery, the ES/G, the WBDD system, specialized door sequencing, an advanced
wing-fold system, and the integration of several features of the seat, along with a
sensing of the environment and management of parachute deployment.

A. ELECTRO-HYDROSTATIC ACTUATION SYSTEM

A number of key technologies went into developing the EHAS. One area that most
affected integration into the aircraft was power control and regenerative energy
handling. New hardware and software were developed to safely manage the
EHAS’s power draw and dissipate the regenerative energy that occurs during
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 379

motor deceleration so there are no effects on the power system. Another area
affecting development was the management of cold temperature performance.
The EHAS is a large user of electrical power on the aircraft. How this power is
managed is critical to the successful operation of the EHAS, as well as the other
components that share the power buses. As the aircraft and EHAS matured, the
losses and operating needs increased in fidelity and the required power draw
went up. This increase was driven by the need to meet performance at cold
temperatures. In addition to the power increase, design features were added to
help mitigate temperature effects.
When the cold temperature issue was identified, it looked inevitable that the
EHAS would need to live with added weight, cost, and complexity due to the
addition of heaters in the manifolds to warm the system in cold environments.
As the issue matured, it was determined that the software could be used to aid
in actuator warming by allowing the EHAS to use losses in the motor as a heat
source. The software was altered so that it could modify the commutation effi-
ciency to increase the losses and, therefore, the heat generated in the motor.
The heat then dissipates into the pump and manifold. This helps to keep the
fluid warm and the fluid properties within the desired range.
The F-35 is a dynamic, naturally unstable aircraft, and controlling it
requires high-bandwidth actuators that can stop and reverse directions quickly.
To accomplish this, the EU uses regenerative energy to slow down the
motor. The regenerative energy must be used or dissipated to prevent voltage
transients that can damage electrical components. Early in the program the
decision was made to keep the regenerative energy in the EU instead of allowing
it back onto the power bus. This was intended to avoid transients on the bus but
required additional provisions within each EU channel to dissipate the regenera-
tive energy.
As the program developed, it was determined that the magnitude of this
regenerative energy exceeded the capability of the EUs to dissipate it. The
ceramic resistors would crack under thermal stress and the solder used to
connect them would reflow and cause the circuits to open. Through multiple
design iterations, additional resistor capacity was incorporated and the thermal
bonding between the resistor and the lid of the EU was improved to better dissi-
pate the heat to the environment. This included the use of steel plate resistors and
high-temperature solder to increase the operating temperature range.

B. POWER AND THERMAL MANAGEMENT SYSTEM

Two key technologies were developed to maximize PTMS integration benefits at
the aircraft level. The FDHX minimized the fuel displacement and ram drag of a
typical ram air circuit and ram air/bleed air HX. Integrating a permanent magnet
S/G on the same high-speed shaft as the PTMS turbomachine facilitated the elim-
ination of a gearbox on the PTMS turbomachine. It also reduced the weight penal-
ties of a more conventional low-speed generator.
380 D. ROBBINS ET AL.

The FDHX consists of three titanium cores and one INCONELw core
arranged in parallel in the F135 fan air duct. The FDHX is critical to dissipating
heat generated by the aircraft electronics into the propulsion system. Hot side flow
for the three closed-loop HXs is provided in parallel. The titanium cores transfer
heat from the closed-loop air refrigeration cycle to fan air. The INCONEL core
transfers heat from bleed air to fan air before it is conditioned by the closed-loop
air refrigeration cycle and provided to the cockpit and forced air –cooled
equipment.
A typical HX core is shown in Fig. 10. Each HX includes two separate cores
welded together at an angle to allow for installation in the F135 annular fan
duct area while maximizing the frontal HX fan air capture area. Hot closed-loop
air enters the center header and splits between the left and right cores. The
internal fin geometry turns the hot flow as shown in the figure to facilitate a
counter-flow design, thereby maximizing performance and minimizing weight
and volume. The cooled bleed or closed-loop air exits the cores through the left
and right exit headers and is collected in a manifold installed on the outside of
the F135 engine.
The PTMS turbomachine includes a compressor, a CT, a PT, and an S/G
mounted on a single shaft. The turbomachine is shown in Fig. 11 with callouts
indicating locations of the major components. To reduce weight and volume,
the turbomachine was designed to operate at the highest speed possible. The
high shaft speed, up to 59,000 rpm, drove the generator design to a permanent
magnet rotor design. The S/G produces approximately 5 hp when used as a
motor to start the turbomachine. The motor is powered by a 270-VDC battery
that also provides inflight emergency power for the flight control system. In gen-
erator mode, the electrical power generation system is capable of producing 80

Fig. 10 Typical fan duct heat exchanger.

F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 381

Fig. 11 PTMS turbomachine mounted on a single shaft.

kW continuous and 120 kW peak 270-VDC power. The generator provides power
for main engine start and inflight emergency power. The F-35 flight controls are
electrically actuated, so there is no need for an emergency hydraulic pump on the
turbomachine. In addition, electrically driving the turbomachine fuel pump and
lubrication system, along with the single-shaft design approach, facilitated remov-
ing the gearbox typically used on aircraft APUs.

C. ELECTRICAL POWER SYSTEM

Many of the technologies used in the F-35 EPS were developed during the devel-
opment of the F-22, greatly reducing risk during the development phase. The
overall bus architectures are very similar, and the solid-state power controllers
are almost identical. The use of a central processing system to manage the
overall control of the system demonstrated many benefits on the previous
program. In addition to raising the level of integration among the various sub-
systems, two new key technologies were developed: a 270-VDC battery and the
ES/G. These met the requirements that had been developed during the trade
study phase.
To support the need for nearly instantaneous fill-in power to the EHAS,
a battery was required. The battery provided fill-in power on the 270-VDC
main bus while load and source management tasks were performed. The
382 D. ROBBINS ET AL.

driving requirement for this battery was the need to provide 6.5-kW of power to
start the PTMS turbomachine at 408F. Then, following the start load, it would
provide 8.6 kW with 40-kW peaks to support the EHAS load for an inflight emer-
gency fill-in. The maximum charge voltage was limited to 350 VDC, and the
voltage at the terminal was required to remain above 207 VDC throughout the
event. Weight and volume are always a driver for aerospace applications, so an
84-cell lithium-ion battery system was defined. Saft Batteries was selected by
General Electric as the supplier with the best chance of meeting this requirement,
and its cells proved to have excellent performance after extensive development
efforts.
The generating system consisting of the ES/G, ICCs, and a dedicated oil
system for cooling the ES/G was developed by Hamilton Sundstrand (now
UTC Aerospace Systems). In addition to providing the prime generator power,
this system replaced the traditional engine-starting system (i.e., providing
torque to the engine gearbox for engine starting). As mentioned previously, two
complete systems were developed: the switched-reluctance system for the early
flight test program, and its replacement, a more traditional synchronous system
for later jets. The electric start system was the driving requirement for the ICC,
while the continuous and transient loads drove the size of the generators.
Much of the technological development work had been accomplished during
the Power Management and Distribution System for a More Electric Aircraft
program. However, many issues were uncovered and surmounted through these
development efforts. One feature of the More Electric Aircraft concept that was
not obvious initially was the effect of having transient loads representing a very
large percentage of the total load. With the F-35C configuration, 65% of the tran-
sient capability of the generating system is allocated to the EHAS. This drove a
requirement for specifying the generating system’s capability at 50 ms in addition
to the more traditional steady-state, two-minute and five-second periods.
Additionally, unique control functions are required to differentiate normal tran-
sient loads from short-circuit conditions.

D. HYDRAULIC AND UTILITY ACTUATION

Advanced developments were included in the weapon bay door drive system to
meet the large hydraulic power demands. The F-35 WBDD system incorporated
an advanced over-center motor that optimized the drive system performance to
align with the hydraulic system’s power capacity. Legacy pressure-compensated
hydraulic motors experience hydraulic power losses across the control valves
upstream of the motor. The use of the advanced over-center motor maximized
the hydraulic pressure available to the motor, resulting in optimum power effi-
ciency of the drive system. It also facilitated integration with the F-35 hydraulic
power system. The decision to build a WBDD system integration facility at the
supplier’s site was instrumental in developing the drive system, controls, and soft-
ware to facilitate successful and timely integration into the aircraft.
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 383

The utility actuation system also used door sequencing extensively to meet the
needs of the air vehicle requirements. Overlapping doors prompted extensive
use of left-to-right door sequencing. A variety of methods were used for the over-
lapping door sequencing, all tailored to the applications and software controlled to
ensure reliability. The majority of the air vehicle doors incorporated door latches
to support the stealth aircraft requirements, all requiring methods to integrate and
sequence the latches with the door system applications.
The F-35C variant included an advanced wing-fold system for aircraft weight
and volume efficiencies. The wing-fold system included new technologies devel-
oped and patented by Moog Aircraft Group that reduced the wing thickness
requirements by approximately 1.5 in. These reduced weight reductions signifi-
cantly, compared to legacy wing-fold systems. Extensive development, buildup
testing, and system qualification led to the successful incorporation of this
advanced technology into the F-35C aircraft. Figure 12 provides a view of the
uninstalled wing-fold system.

E. EJECTION SYSTEM
The US16E ejection seat system incorporates several key technologies to support a
safe ejection environment across a wide range of pilot body types (weight and
height). Several features of the seat, such as head position control and arm and

Fig. 12 Wing-fold actuation system.

384 D. ROBBINS ET AL.

leg restraints, are integrated. There is also sensing of environment and manage-
ment of parachute deployment. With this combination, the seat successfully
demonstrated a full envelope capability that minimizes the risk of injury.
The seat is qualified for safe ejection at up to 550 kt Calibrated Air Speed
(KCAS) and provides superior terrain clearance performance over other fighter
aircraft. Some of the design features that enable this performance are a five-mode
electronic sequencer, active arm and leg restraint system, and neck protection
device. The five-mode electronic sequencer energizes upon ejection and senses
environmental data, such as pressure and acceleration. With this information, it
determines the sequencing of drogue parachute deployment, main parachute
deployment, and seat/pilot separation. A unique feature of the sequencer is its
manual selection of main parachute deployment based on the pilot-weight
range. To minimize accelerations during main parachute deployment, pilots
weighing less than 135 lbm select a longer delay, minimizing acceleration and
allowing the US16E to minimize head and neck loads. The selection is made by
toggling a switch mounted on the left side of the seat, which is easily accomplished
and verified during aircraft ingress. As a result of this design, the US16E ejection
system achieves a greater terrain clearance and escape injury-protection envelope,
compared to legacy systems (Fig. 13).
A second feature of the US16E ejection seat is the active arm and leg restraint.
To minimize injury due to arm and leg flail during ejection, the US16E actively
restrains the pilot’s arms and legs during ejection. It uses fabric lines that are
integrated into the pilot’s flight ensemble. During ejection, the legs are drawn
back and the arms are drawn to the lap and held until the lines have been
severed. Figure 14 shows the US16E arm and leg restraint system. A third
feature of the US16E is the neck protection device. This is an inflatable air
beam that deploys during ejection to stabilize the pilot’s head and reduce head
and neck loading.

Fig. 13 F-35 US16E ejection envelope comparison.

F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 385

Fig. 14 US16E arm and leg restraints.

V. INTEGRATION QUALIFICATION EFFORT/CHALLENGES

There was a need for integrated systems and the development and qualification of
several key technologies. Based on this, integrated testing and qualification
resulted in several challenges during the development phase, including:
. The development of integrated modeling and testing
. Early flight test electrical system failures
. Lithium-ion battery development
. Integrated engine start and emergency power
. Hydraulic system integration
. Full envelope ejection seat qualification
Along with the individual system development at the suppliers, there was a need
for a full-up integration facility. The facility had to be capable of testing the
386 D. ROBBINS ET AL.

integrated system architecture not only for normal operation, but also—and
perhaps more importantly—under failure conditions.
The Vehicle System Integration Facility (VSIF) supports both hardware-in-
the-loop and pilot-in-the-loop capabilities for all three F-35 variants. It includes
the following:

. EHAS
. EPS
. HUA systems
. PTMS and main engine drive stands
. Cockpit and visual display
. Bus architecture (including aircraft hardware)
. VMCs

Each of those systems can be integrated into the laboratory or emulated with soft-
ware models, depending on the test needs. A six-degrees-of-freedom simulation
provides the necessary plant models to simulate the aircraft environment.
Figure 15 provides the layout for the VSIF.

A. ELECTRO-HYDROSTATIC ACTUATION SYSTEM

Aircraft-level integration played a significant role in the development of the
EHAS. The complexity of the system and the time-critical interactions among
the EHAS, VMC, EPS, and PTMS made integration a necessity. It was integral
in verifying that the systems worked in coordination with each other both in

Fig. 15 Vehicle System Integration Facility layout.

F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 387

Fig. 16 EHA loading fixtures.

normal conditions and as failures were inserted. Within the VSIF, an EHA loading
fixture was developed to house the EHAS, along with loading fixtures (Fig. 16).
Load cells were required to correctly mimic the EHAS environment specifically
as it relates to thermal characteristics and electrical regenerative performance.
The load cells’ forces were computed by the aircraft simulation aerodynamic
model. This allowed for varying hinge moments as a function of the aircraft envel-
ope, rates, and accelerations, along with surface positions.
The PTMS provides the cooling air for the EHAS’s EUs. The cooling air
primarily cools the high-power switches that drive the EHA motors, but it also
has indirect effects on all components in the box. The software solution to the
cold temperature issue noted in the key technology development section was
implemented. From this it was found that although increasing the motor losses
could improve the performance of the EHA, it also increased the temperatures
of the switches. To account for this, a new interface was created to request
increased cooling flow from the PTMS. Additionally, the PTMS could not
provide the increased flow under all conditions, which drove interlocks into the
design to limit when the motor heating could be applied and prevent the switches
from overheating.
Further complicating the cooling interface, the tubing carrying cooling air to
the EUs on the right side of the aircraft was running around the engine and
causing heat impingement. It was also elevating cooling air temperatures when
388 D. ROBBINS ET AL.

compared to the left side. During normal operations the warmer airflow does not
impact the ability of the EUs to operate. However, when the EUs are being
stressed, the internal temperatures can rise and reach the component limits.
This made it necessary to add a level of cooling flow that could be requested
based on actual and predicted component temperatures.
The VMC is connected to the EPS, PTMS, and EHAS through a communi-
cations bus. Accordingly, commands and status messages must travel back and
forth going from, to, and through the VMC. This allows for a more integrated air-
craft and generates latency, timing issues, and the need for increased coordination
among the systems.
The EUs control electric motors that commutate at a high frequency and
can fail and burn out in milliseconds. As such, they must operate at internal
frame rates that significantly exceed the bus. This means that the EU must be
able to manage its state somewhat autonomously and address failures as they
occur; however, it must do so predictably and within the expectations of the
VMC running the aircraft control laws and providing position commands. The
commands from the VMC had to be orchestrated such that it maintained
ultimate control but did not limit the ability of the aircraft to self-protect
during a failure. Reset processing and a built-in test capability also had to be
sequenced to accommodate latencies and system interdependencies. This was
most evident on the F-35B with the additional components required for
STOVL operations.
All these interfaces were tested with hardware in the loop at the VSIF. That
facility includes a complete triplex VMC and remote input/output setup, along
with a full set of EHAs and EUs connected to a load system. It can also
connect the EHAS to the EPS in the same manner as how it would be in an
aircraft. This allowed interconnected failure modes to be tested so that the
EHAS and EPS were both robust in responding to the failure events that can
occur in a high-power system.
Early in flight test, the systems experienced one of these events. A high-power
short in one channel of a dual-channel EU propagated to the opposite channel,
causing a short across both power buses and a temporary loss of power. In
spite of this, the aircraft was still able to recover, reconfigure, and land safely.
Overcoming this problem illustrated that the design and integration testing
performed had adequately prepared the systems for such an issue; however, it
also brought into question other scenarios that required improvements. In part,
the EUs needed to allow the EPS the opportunity to fully recover and address
the variability of the timing of that recovery. As a result, new criteria were estab-
lished to limit power draw and improve failure hold-off timing during a power
loss event. The criteria also needed to allow for critical surfaces to automatically
reset after a total power loss. The approach had to be balanced between the
EPS response time and the time before the actuator would lose stiffness and
cause surface flutter. New integration test cases showed improved handling of
shorts and power losses and a significant increase in the robustness of the system.
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 389

B. POWER AND THERMAL MANAGEMENT SYSTEM

The PTMS normally operates in the closed-loop refrigeration mode, providing

cockpit conditioning and equipment cooling during flight. In the event of an
engine flameout, the PTMS transitions to an open-loop combusted power cycle
and produces electrical power to drive flight control actuation and engine-start
assist. Transition from the cooling mode to the emergency power mode requires
deploying a ram air scoop to feed the compressor. In the process, this manipulates
several valves to convert the closed-loop refrigeration cycle into an open power
cycle, while also lighting the combustor.
The maturation of this transition started by performing a high-fidelity transi-
ent simulation of the process. A detailed transient model of the PTMS was
built using MATLABw/Simulinkw. The model was used to develop the sequence
of valve openings and closing and combustor light-off required for successful
mode transition. Software was developed based on the transient modeling results,
and both software and hardware were tested statically in a ground test stand and
in an altitude chamber. The rigorous modeling and laboratory testing proved
their value in the very successful flight test sequences for transition from bleed-
cooling to self-powered emergency, in-flight propulsion engine start-assist, and
in-flight PTMS self-start.
Legacy fighter aircraft programs mitigated the risk of propulsion system
loss by running the APU during early flights. This was done until confidence in
the propulsion system was gained through successful flight envelope expansion.
The F-35 EPS contains a 270-VDC battery capable of providing emergency
flight control power. However, actually landing the aircraft with only the available
battery power is considered risky due to the unknown battery state of charge
during the high current draw expected during landing. In addition, the PTMS’s
capability to produce electrical power depends on the recovery of the system’s
turbomachine inlet. Prior to flight, inlet recovery had only been evaluated by com-
putational fluid dynamics analysis. In addition, the ability to transition from the
closed-loop refrigeration mode to the open-loop power mode had only been
analytically modeled and tested statically in an altitude chamber. Statistical
failure analysis was used to show that the probability of the engine failure scenario,
especially at benign flight conditions, was low enough to begin the flight test
program with confidence.
Additional flight testing evaluated PTMS turbomachine inlet recovery by
installing a total and static pressure rake in the turbomachine inlet throat. The
PTMS was also operated in the open-loop, emergency power mode with the
propulsion system operating. The process of transitioning from the closed-loop
refrigeration mode to the open-loop emergency power mode was also demon-
strated. It was tested in flight with the propulsion system operating while being
facilitated by a flight test aid to simulate engine failure conditions.
Once the PTMS’s inflight operation had successfully been demonstrated,
engine inflight spool-down restarts were demonstrated. These had the PTMS
390 D. ROBBINS ET AL.

preset in the open-loop emergency power mode, followed by engine-start

assist testing with the PTMS preset in the open-loop emergency power mode.
Ultimately the complete engine flameout scenario with the PTMS providing
engine-start assist was demonstrated by cutting fuel flow to the engine. This simu-
lated an engine flameout and allowed the PTMS to transition from the closed-loop
refrigeration mode to the open-loop emergency power mode, followed by a
successful engine-start assist.
The aircraft operating environment is critical to the PTMS and main engine,
so the VSIF laboratory did not incorporate the PTMS or main engine systems.
Instead, drive stands were incorporated to mimic the necessary inputs in
support of the overall aircraft. The two drive stands are shown in Fig. 17.

C. ELECTRICAL POWER SYSTEM

As with any development program, changes were implemented as a result of dis-
coveries from planned testing and unplanned events. Early in the design cycle,
Northrop Grumman assembled a dedicated team to construct and run computer-
based models. With these, the team analyzed various system interactions and
helped to guide the requirements development and implementation. Models

Fig. 17 Main engine and PTMS generator drive stands.

F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 391

Fig. 18 ES/G testing at Northrop Grumman.

were used to predict power quality during main engine starts and high-power
operations, and the effects of EHA in-rush currents during normal operations
and failure modes. The models were extremely useful and were maintained
throughout the entire qualification effort. Northrop Grumman also conducted
extensive development testing in its facilities, particularly on the generating
system and the batteries (Fig. 18).
Each component was subjected to extensive environmental, life, and perform-
ance testing as an individual component. In some cases, such as the battery
systems, the components were tested as a sub-subsystem to ensure that the final
products would be robust in the specified environments. As is typical, the electro-
magnetic environment and vibration environment were the most difficult.
Extensive testing was performed on the EPS as a stand-alone system and
with several other key systems to ensure that the high level of systems-level inte-
gration was successful. A System Integration Laboratory (SIL) was constructed at
a Hamilton Sundstrand (now UTC Aerospace Systems) facility where all the com-
ponents of the EPS were assembled. The controlling software was also assembled
there to ensure that the EPS components were well-integrated with each other.
Lockheed Martin’s VSIF (Fig. 19) was used extensively to expand on the efforts
in the SIL. This included integration with the EHAS and key mission systems
components, ultimately including a pilot in the loop as well. The latter was accom-
plished via a flight simulator in which end-to-end integration and an exhaustive
list of failure modes were extensively examined.
392 D. ROBBINS ET AL.

Fig. 19 EPS test stand in Lockheed Martin’s VSIF.

During laboratory testing of the overall EPS, an inadvertent short circuit

was applied to the output of the 270-VDC battery. As expected, the extremely
low impedance of the battery allowed extremely high currents into the fault.
The distribution system was designed to tolerate the current, but the incident
revealed a problem with the cell design: the internal connections were not sized
to tolerate the extremely high currents, which resulted in the destruction of the
battery. As a result, a fuse was added into the battery case to prevent external
short circuits from cascading damage to the battery.
Similarly, an early failure during flight testing showed that the logic
implemented to protect and isolate shorts in the electrical distribution unit
loads was inadequate. Although protection was in place to ensure that both
generators would not be susceptible to a short circuit, the battery system was
not adequately considered. Based on the short circuit capability of the battery,
it was decided that it would be prudent to postpone connecting the battery as
long as possible. Waiting would allow the fault to clear before bringing online
the remaining power source for flight control actuation.
In spite of our modelling and integration efforts, some problems were still
beyond our ability to analytically duplicate even after they were discovered.
One example was how the generating (starting) system reacts with the power
source (PTMS turbomachine) during engine starting. An interaction between
the load (ICC and ES/G), source (turbomachine), and interconnect wiring was
discovered during the early production program. In it, the turbomachine would
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 393

occasionally trip for instantaneous overcurrent conditions during specific

frequency-dependent portions of the start profile. This anomaly was not discov-
ered during development testing and, despite numerous attempts to duplicate
the phenomenon through simulations, the team was unable to duplicate it;
however, additional inductance between the turbomachine and ICC resolved
the issue.
Because of the highly integrated systems implemented in the subsystems,
the EPS was also influenced by failures in other systems. The EHAS inherently
regenerates energy that must be dissipated. Early testing of the circuitry for con-
suming this energy within the EHAS controller showed that much more energy
was being generated than could be dissipated by the circuitry. This was resulting
in early failures of the circuitry.
Since volume in the EHA controller was limited, increasing the dissipation
capacity was difficult, so alternate methods of consuming the regenerative
energy were investigated. An early alternative was to allow the energy to flow
back onto the electrical bus, where (in theory) it would either be consumed by
other loads or back-drive the engine through the generating system.
Implementing a circuit to manage the regenerative energy within the E/SG
ICC was not especially difficult; however, implementing the overall system was
not as easy, particularly after trying to accommodate failure modes. It was
finally determined that there was no feasible way to handle failure modes of
the regenerative circuitry (open circuits) without risking the destruction of all
the other equipment on the bus. This was due to the high voltage that would
be created when a regenerative pulse was unable to be consumed by the regenera-
tive circuitry. The first solution tested was the addition of burden resistors within
the ICCs to ensure that there was always a sink for the regenerative energy. This
then led to the final solution of localized regenerative consumption within the
EHAS.
In spite of the team’s efforts to integrate with the various stakeholders,
a problem was discovered just prior to the start of the flight test program. This
was the sharp rise-time of power demand from the EHAS, particularly during
an F-35C carrier-based landing with numerous, rapid movements of the large
horizontal tails. The affected power demand generated concern over how the
electrical transients would be handled by the drive train through the generating
system, gearbox, and engine. This was particularly concerning during the
conversion of the electrical energy draw from the mechanical system. Extensive
testing in the VSIF with multiple pilots in the loop established more precise
usage profiles throughout multiple flight profiles, particularly the landing profiles
of the F-35C. Analyses and modeling of this new data was beneficial, with no
changes required to any of the units; however, the life of the ES/G stub shaft
became a life-limited part as a result of the dynamic stresses imparted on the
shear section.
As a final lesson learned, an incident during early flight testing revealed the
need to ensure that the supplier base abides by the intricacies of high-voltage
394 D. ROBBINS ET AL.

power. Proper isolation of 270-VDC power is absolutely necessary when

designing units, particularly those containing both high- and low-voltage
inputs. Such isolation ensures that the high voltage cannot short to the low-voltage
inputs, or between isolated high-voltage inputs, during other catastrophic failure
modes.
Through this development effort the EPS has proven to be a very robust
and reliable system. The team spent considerable effort integrating the EPS
with the PTMS and engine to minimize issues between/among these systems.
Extensive testing during the Enhanced Integrated Systems Test [6] proved to be a
worthwhile endeavor as well. No significant problems were discovered between/
among these systems throughout the development and flight test programs.
Looking to the future, the EPS is well-situated for growth, with margin avail-
able on the 270-VDC system. By contrast, the 28-VDC system has little growth
available as currently implemented. Implementing active, real-time load manage-
ment (planned for a future software release) will allow the current system to
meet all the projected needs for the foreseeable future. It is expected that future
improvements in switching and conversion technologies will allow extending
the capacity of the current 28-VDC system without increasing weight or volume.

D. HYDRAULIC AND UTILITY ACTUATION

The complexities of the hydraulic system requirements and associated design
implementation led to the development of integration facilities to support the
systems development. The development of a Hydraulic Integration Facility
(HIF), or iron bird facility, was originally planned to take place at a supplier
facility. Early planning transitioned this test facility to Fort Worth, Texas, to
incorporate the test facility as part of the overall F-35 VSIF. Transitioning the
HIF (Fig. 20) to the VSIF in Fort Worth offered cost efficiencies and supported
a higher level of system integration. The HIF was instrumental in the HUA hard-
ware and software development, and supported the rapid resolution of system
anomalies during the development phase of the F-35 program.
Another test facility that was instrumental in the successful development
of the F-35 utility actuation systems was the WBDD system test rig (Fig. 21).
It was developed by UTC Aerospace Systems in Wolverhampton, United
Kingdom. The weapon bay doors are designed to operate within the full flight
envelope of the aircraft. The WBDD system includes various operational modes
for weapon deployment, vertical landing, and ground operations. In addition,
various safety features were designed into the system to protect maintenance per-
sonnel. The WBDD test rig included the complete actuation system, weapon
doors, locks, and aircraft door lands. Overall system operation and performance
were thoroughly tested throughout the operational loads and temperature
extremes. The test facility was instrumental in developing the system software.
In addition, aircraft rigging procedures were developed at the test facility, support-
ing successful aircraft integration and initial operations.
F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 395

Fig. 20 Hydraulic integration facility.

E. EJECTION SYSTEM
The successful performance of the US16E ejection seat was proven through a
rigorous qualification test program. The program included traditional com-
ponent testing and incorporated a series of catapult tests, sled track ejection
tests, and ejection testing. The latter was performed at altitudes requiring
the seat to be ejected out of the cockpit of a Martin-Baker Aircraft Co.
Ltd. – owned Gloster Meteor airplane. The catapult testing was carried out at
Martin-Baker’s facility in Chalgrove, United Kingdom. The catapult testing
demonstrated the performance of the seat through the firing of the primary
and secondary cartridges and the deployment of the neck protection device.
Catapult testing is an efficient, cost-effective means to ensure that the
impulse provided by the main cartridges meets the head and neck injury cri-
teria without the complexities of a full-up sled test. Figure 22 shows a picture
of a catapult test.
The US16E also incorporated an extensive sled test program to ensure that the
entire escape system operated as designed across the required speed range. Two
ground test tracks were used during the qualification program: Martin-Baker’s
396 D. ROBBINS ET AL.

Fig. 21 F-35 WBDD system test facility.

Fig. 22 Catapult test at Martin-Baker’s Chalgrove facility.

F-35 SUBSYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 397

Fig. 23 Sled test at Martin-Baker’s Langford Lodge facility.

facility in Langford Lodge, Ireland, and the test facility at Holloman Air Force
Base, New Mexico. The Langford Lodge facility was used for most of the sled
testing, and the Holloman facility was used for the 550-to-600 KCAS testing.
Figure 23 shows a picture of a sled test at the Martin-Baker facility. Martin-Baker
modified a Gloster Meteor aircraft to enable an ejection seat to be ejected at alti-
tude. The altitude testing was executed at Cazaux Air Base, France (Fig. 24).

Fig. 24 US16E altitude ejection testing.

398 D. ROBBINS ET AL.

VI. CONCLUSION
The F-35 Vehicle Subsystems team successfully integrated previously federated
systems into a cohesive system of systems that reliably and safely meets very strin-
gent aircraft requirements. The team’s rigorous, structured process of systems
engineering included extensive modeling, simulation, integration, and testing.
Through it, the resultant systems demonstrated significant reductions in weight,
volume, and complexity across a wide range of operating conditions. The F-35
Vehicle Systems constitute the first production deployment of an integrated sub-
systems architecture for a combat aircraft. These systems and their designers over-
came several developmental problems. The systems are well-positioned to meet
future needs. Lessons learned from the team’s efforts provide a firm grasp of
the processes and techniques required to progress the level of integration to
meet future requirements.

ACKNOWLEDGMENTS
The authors thank all their colleagues on the F-35 Vehicle Subsystems team who
were instrumental in developing, qualifying, and fielding these systems on the
F-35. Composing this paper was only made possible by their dedication and
expertise. The authors would also like to thank graphic artists Daniel Buck and
Dennis Soultaire for their assistance. In addition, we extend our gratitude to the
executive review team whose members provided guidance in the development
of this paper.

REFERENCES
[1] Sheridan, A., and Rapp, D., “F-35 Program History—From JAST to IOC,” AIAA
Aviation Forum, June 2018 (to be published).
[2] Counts, M., Kiger, B., Hoffschwelle, J., and Houtman, A., “F-35 Air Vehicle
Configuration Development,” AIAA Aviation Forum, June 2018 (to be published).
[3] Wiegand, C., et al., “F-35 Air Vehicle Technology Overview,” AIAA Aviation Forum,
June 2018 (to be published).
[4] Burkhard, A., and Deitrich, R., “Joint Strike Fighter Integrated Subsystems
Technology, A Demonstration for Industry, by Industry,” Journal of Aircraft, Vol. 40,
No. 5, Sept./Oct. 2003.
[5] Weimer, J., “Past, Present & Future of Aircraft Electrical Power Systems,” AIAA Paper
2001-1147, Jan. 2001.
[6] Wurth, S., Smith, M., Celberti, L., and Walters, J., “F-35 Propulsion System Design,
Development, and Verification,” AIAA Aviation Forum, June 2018 (to be published).
 CHAPTER 11

F-35 Mission Systems Design,

Development, and Verification
Greg Lemons , Karen Carrington†, Thomas Frey‡ and
John Ledyard§
Lockheed Martin Aeronautics Company, Fort Worth, TX

F-35 mission systems allow pilots to execute traditional advanced tactical

missions critical to allied and partner nations. The systems include the
most advanced sensor management and data fusion of any current
fighter aircraft. These capabilities offer unmatched situational
awareness to the pilot and provide decision aids that allow the pilot to
make critical timely decisions. The development of these capabilities
was accomplished utilizing use cases derived from mission vignettes to
perform capability and model-based development. This approach was
extended to the test and evaluation phase to develop mission-level
scenarios used for validating the models and capabilities.

I. INTRODUCTION
The F-35 comprises a set of highly common aircraft for the Air Force, Marine
Corps, and Navy, as well as the 12 current F-35 partner nations. Although the air-
frames themselves possess slight differences unique to each variant’s operating
environment requirements, the mission systems’ hardware and software are
common. In addition, product-line engineering tags have been integrated into
the mission systems’ software requirements baseline. This enables repeatable
and affordable country-unique builds for production off of a U.S. baseline.
The F-35 sensor suite includes the following:
. AN/APG-81 Active Electronically Scanned Array (AESA) radar
. AN/ASQ-239 Electronic Warfare (EW)/Countermeasures (CM) system
. AN/AAQ-40 Electro-Optical Targeting System (EOTS)

 Systems Engineer Sr. Mgr., F-35 Mission Systems Design.

†
Systems Engineer Dir., F-35 Mission Systems Design.
‡
Senior Technical Fellow, Software Engineering.
§
Technical Fellow, F-35 Mission Systems.

399
400 G. LEMONS ET AL.

. AN/AAQ-37 Electro-Optical (EO) Distributed Aperture System (DAS)

. AN/ASQ-242 Communications, Navigation, and Identification (CNI) avionics
suite
These five sensors provide F-35 fusion with object detection and measurements in
the Radio Frequency (RF) and Infrared (IR) spectrums. This compilation of data
gathering disseminates more information about the environment than what has
ever been available on a fighter aircraft.
In addition to receiving information from the onboard sensors, the F-35
receives off-board tracks and measurements from the Link 16 datalink and the
Multifunction Advanced Data Link (MADL). Designed for 5th-Generation
aircraft, MADL provides fusion-quality data on all air and surface tracks to
other members of the flight group. These data include the track state, track covari-
ance, identification features, and passive RF data.
The amount and fidelity of the off-board information provided by MADL was
one of the largest challenges for the fusion design. The capability of the sensors
and information sharing across MADL presented a challenge for sensor fusion.
The challenge was to ensure that the tracks displayed were real and not duplicated,
which would result in display clutter. The last few software builds in the System
Development and Demonstration (SDD) phase of the F-35 program were
aimed at tackling display clutter problems. The objective was to ensure that the
pilot had accurate and timely information to make real-time tactical decisions
in the cockpit.
This paper discusses the design, development, and verification of each
of these systems, as well as the system of systems integrated into the F-35
aircraft.

II. THE VISION

A. AIRCRAFT CONCEPT
The F-35 was born of the need for an affordable multirole, multiservice, multina-
tional (U.S. Air Force, Marine Corps, and Navy, and international operators)
fighter aircraft to replace an aging fleet of fighter and attack aircraft. The aircraft
being replaced were battle tested by their operators. The challenge of this
concept was rooted in the varied missions and operating environments that these
platforms were satisfying. In order to replace their collective capabilities with a
single fighter platform, a new way of thinking about single-seat fighter avionics
design was needed.
The concept for the F-35 developed by Lockheed Martin centered on return-
ing the pilot to the role of tactician. This principle was the driving force behind
the avionics development plan. One of the Lockheed Martin Mission Systems
team’s design goals was to develop a set of sensors that could collect information
across multiple spectrums. Another goal was to develop a sensor control scheme
F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 401

Fig. 1 Returning the pilot to the role of tactician.

of autonomous sensor management. This, along with a next-generation cockpit,

would provide the pilot with an unprecedented amount of information distilled
down to an easily consumable format. Prior to this, the pilot spent precious
minutes setting up radar scans and adjusting tilt, gain, and refresh rates, while
also monitoring multiple displays to run an intercept. With this new suite, the
pilot is able to view a picture of the multispectral battlespace in a consolidated
format, as depicted in Fig. 1.
The strategy for developing the avionics system was based on a block
buildup strategy. It was founded on basic warfighting capabilities and then
built into the most advanced fighter weapons system currently in service.
The most fundamental elements were developed first, and then the design
moved to higher levels of capability. In this way, the team reduced the risk
of software development produced by having a single large software release.
This strategy was adapted from previous tactical fighter programs, such as
the F-16, and more recently, the F-22. Those programs demonstrated that it
is essential to break the software development into manageable blocks to
reduce the complexity and cost of testing. Incremental releases also provide
more of an opportunity to manage requirements creep and incorporate technol-
ogy changes that directly benefit the warfighting capability. This was evident in
the F-35 program’s ability to implement additional weapons (e.g., GBU-39) and
capabilities (e.g., operational test support changes) to support emerging require-
ments. The approach is structured as three development blocks that establish
the basic flight control systems and essential mission systems before building
up mission capability. The approach to building up the block plan in detail
is depicted in Fig. 2.
402 G. LEMONS ET AL.

Fig. 2 Block development buildup.

B. SYSTEM ARCHITECTURE
It was recognized early in the concept development phase that the architecture of
the mission systems would be key to the program’s success. To succeed, many
challenges had to be faced to develop the right architecture. One was that the com-
puting resources needed for the full set of capabilities could not fit within the
power, weight, volume, and thermal limitations of the air vehicle using available
technology. Another was that the long development cycle and initial low-rate pro-
duction was expected to result in Diminishing Manufacturing Source (DMS) pro-
blems. In addition, the aircraft needed to be easily adaptable to support the unique
needs of multiple countries. Further, it needed to be unclassified on the pro-
duction line and on the ramp to avoid increased production and sustainment
costs. Beyond this, it also needed to operate in future battlespaces where the move-
ment of data at multiple levels would be key to interoperability.
The plan for overcoming computing resource and DMS challenges was to
execute multiple technology refreshes of the computers during development.
The processing update would allow Moore’s law to take effect, providing increased
processing capability over time that would fit within the limitations of the air
vehicle. The updates would also allow updates to mitigate DMS and validate
that the application software was independent of the underlying processor.
To achieve the goal of making the application software independent of the
processor changes, three design approaches were used. The first approach
(Fig. 3) was to layer the software on top of Commercial, Off-the-Shelf (COTS)
operating systems, under the assumption that the virtual platform would
not change.
The second approach was to use rate-based processing for all threads when
timing and latency were critical. This approach would achieve constant system-
level timing, even with faster processing. Also, it would enable analyzing
the system and proving that it was schedulable using rate-monotonic theory.
F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 403

Fig. 3 Layered software design.

Both benefits supported easier integration, reduced regression testing, and sup-
ported airworthiness and safety certification.
The third approach was to use messages for communication among all appli-
cation software components, as well as components and subsystems [1]. This
created controlled interfaces among the components and enabled moving appli-
cations to different processors without impacting the software. The approach con-
tributed to solving interoperability and adapting the software for multiple
countries. With the clearly defined interfaces and communication paths in the
system, it was possible to control data paths using the trusted computing base.
This enabled isolating data access for specific address spaces and ensured that
the application remained at the designed security level for a specific datalink.
Further, combining the messaging and access control with a COTS operating
system with a high assurance level made it possible to design write-down appli-
cations for each datalink.
The capability was then advanced to interoperate with multiple participants in
the battlespace at different operational levels and message formats. From this, the
messaging and access control was partitioned to the external communications
domain. This provided a broker for data on and off the aircraft, ensuring the
correct classification level and translating the external data into formats consistent
with internal data. It also transformed internal data into the message formats and
needs of the external links.

III. F-35 SENSOR SUITE

The F-35 sensor suite includes the AN/APG-81 AESA radar, AN/ASQ-239 EW/
CM system, AN/AAQ-40 EOTS, AN/AAQ-37 EO DAS, and AN/ASQ-242 CNI
system. This collection of advanced multispectral sensors, shown in Fig. 4, pro-
vides the F-35 with a next-generation ability to see the battlespace.
404 G. LEMONS ET AL.

Fig. 4 F-35 installed sensor locations.

A. AN/APG-81 RADAR
Northrop Grumman’s Electronic Systems sector’s AN/APG-81 radar was devel-
oped as a next-generation version of its AN/APG-77 AESA radar, first fielded
on the F-22A Raptor. The design was further refined with the AN/APG-80
fielded on the Block 60 F-16. This lineage of AESA radar designs allowed for a
rapid development and insertion of previously fielded common waveforms in
early software blocks. It also paved the way for more complex functions during
later deliveries.
AN/APG-81 testing was approached in a buildup fashion of increasingly
complex integration into the rest of the avionics system. The integration began
with stand-alone laboratory testing isolated from the rest of the avionics
system. It then progressed through the Northrup Grumman flying testbed,
where dynamic stand-alone open-air testing was performed. From there the
AN/APG-81 was integrated into the F-35 avionics suite to continue both labora-
tory testing and dynamic open-air testing on the Lockheed Martin Cooperative
Avionics Test Bed (CATB). After the system was proven on the CATB it
progressed to full airborne testing on the F-35. The buildup timeline for the
AN/APG-81 is shown in Fig. 5.
The F-35 radar system has an active, electronically scanned Multifunction
Array (MFA) and the RF support electronics necessary to support a fully func-
tional radar. It also has integrated radar software modes that are hosted on the
integrated core processor. The radar operates through the nose radome, which
F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 405

Fig. 5 AN/APG-81 integration buildup.

has a wide bandwidth, enabling high-power transmissions over a large frequency

range (Fig. 6).
The AN/APG-81 is designed to operate as a radar, Electronics Support
Measures (ESM) receiver, and jammer. It includes active and passive Air-to-Air
(A/A) and Air-to-Surface (A/S) target detection, track, and identification capa-
bilities. In addition, it allows many of these to be interleaved, providing both
A/A and A/S functionality. The sensor also supports the Advanced Medium-
Range Air-to-Air Missile (AMRAAMw) and synthetic aperture radar mapping,
ground and sea moving target detection and track, and A/S ranging. Radar func-
tions include electronic protection for operation in jamming environments and
low probability of intercept features to minimize the likelihood of emissions
being usefully detected by airborne or surface-based receivers. Radar functions
also support system health determination and calibration. Figure 7 depicts the
various radar functions.

Fig. 6 Installed radome and radar locations.

406 G. LEMONS ET AL.

Fig. 7 Radar system air-to-air and air-to-surface operation.

B. AN/ASQ-239 ELECTRONIC WARFARE/COUNTERMEASURE SYSTEM

The AN/ASQ-239 EW/CM system is an integrated suite of hardware and soft-
ware. It is optimized and designed to provide the F-35 with a high level of A/A
and A/S threat detection and self-protection. It can search, detect, identify,
locate, and counter RF and IR threats. The EW system supports the application
of Electronic Support Measures (ESM) through such functions as:

. Radar warning
. Emitter geolocation
. Multiship emitter location (including high-sensitivity states)
. High-Gain (HG) ESM
. HG electronic CM
. HG Electronic Attack (EA) via radar MFA utilization

The EW functions are designed for:

. Wide frequency coverage

. Quick reaction time
. High sensitivity and probability of intercept
. Accurate direction finding and emitter geolocation
F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 407

. Multiship geolocation
. Self-protection countermeasures and jamming

The countermeasure subsystem provides multiple self-defense responses, including

preemptive and reactive techniques, based on available expendable payload and/
or threat-specific self-protection plans. The EW/CM system provides emitter
tracks to the sensor fusion function, which fuses EW track reports and other
sensors (e.g., radar and DAS, off-board sensors) and displays the information
to the pilot. The EW/CM system consists of the following primary elements:
. Band 3/4 apertures
. Band 3/4 aperture electronics
. Centralized EW electronics (Racks 2A and 2B)
. CM controller unit
. CM dispensers
. RF and digital interfaces with the MFA
. Digital clock reference interfaces with the CNI system

The installation locations of the EW/CM system-related equipment are depicted

in Fig. 8. The EW system is common among the three F-35 variants, except for
the forward Band 3/4 arrays, which employ longer elements for the F-35C

Fig. 8 EW equipment location.

408 G. LEMONS ET AL.

Carrier Variant (CV). Also differing, the distance between the inboard and out-
board arrays is less on the CV variant due to the wing fold. In addition to the
EW Band 3/4 apertures, the radar MFA is employed to support EW functions.
There are growth provisions allocated for Band 5 radar warning such that Band
5 apertures, aperture electronics, and the Band 5 switch can be incorporated
into the EW subsystem architecture.
The EW apertures comprise six multielement antenna array sets covering por-
tions of the Band 3 and Band 4 frequency spectrum, along with both vertical and
horizontal polarization. All the arrays have Azimuth (AZ)-only designs that do
not rely on the use of Elevation (EL) arrays. The passive array assemblies use a
traveling wave-notch element approach designed to balance gain, polarization,
Field of View (FOV), and radar cross-section features. Each Band 3/4 aperture
feeds an aperture electronics module that amplifies and passes the detected RF
signals from the apertures. It does this through a switch matrix and tuners that
distribute the RF to a set of wideband EW Receivers (EWRs). The switch
matrix also receives RF signals from the radar MFA when tasked to support
HG modes.
The EWR suite consists of 12 wideband receivers grouped into three inte-
grated sets of four channel receivers. The wideband receivers take in RF
energy and convert the data into digital information via a set of high-speed
analog-to-digital converters for processing. Each EWR performs initial data pro-
cessing and generates pulse parameter reports that are sent to the EW controller/
preprocessor. The processor then provides further signal processing and algor-
ithms to support all EW activities. Various intelligence products combine to
produce an Electronic Intelligence (ELINT) database in a preplanned mission
data file. This data file provides the system with the necessary parametric descrip-
tions of the emitters of interest for emitter identification and scan schedule
operations.
EW system testing was performed in a buildup fashion of increasingly
complex integration into the rest of the avionics system. The integration began
with stand-alone laboratory testing isolated from the rest of the avionics system
and progressed through a Sabreliner T-39-based flying testbed. From there the
EW system was integrated into the F-35 avionics suite to continue both laboratory
testing and dynamic open-air testing on the Lockheed Martin CATB. After the
system was proven on the CATB it progressed to full airborne testing on the
F-35. The buildup timeline for the EW system is shown in Fig. 9.

Fig. 9 Electronic warfare system integration buildup.

F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 409

C. AN/AAQ-40 ELECTRO-OPTICAL TARGETING SYSTEM

The F-35 requirement for a Low Observable (LO) combat configuration did not
allow for the traditional Targeting Forward-Looking Infrared (TFLIR) solution.
The legacy pod systems could not be both missionized and easily concealed for
LO operations. The solution was to integrate the targeting pod system into the
outer mold line of the aircraft. The EOTS, built by Lockheed Martin Missiles
and Fire Control, was built specifically for the F-35 to provide the jet with an
LO IR targeting capability. Its integration was approached in a buildup fashion,
as shown in Fig. 10. The initial open-air testing was performed with a modified
Sabreliner T-39 jet to test the EOTS as a stand-alone sensor to verify sensor-level
behavior. The EOTS was integrated into the rest of the avionics system on
the Lockheed Martin CATB flying testbed to test the interactions between the
sensor and the full avionics system with a pilot in the loop. The final testing
and verification came with the full integration of the EOTS into the F-35.
The EOTS is an internally mounted advanced Midwave Infrared (MWIR)
targeting system with a faceted window having LO characteristics, designed for
A/A and A/S targeting support. By using the midwave portion of the IR spectrum
the EOTS provides a sharper image and less susceptibility to target obscuration
from smoke and haze. The EOTS may be used in the imaging mode in A/A
and A/S, or in the IR Search and Track (IRST) mode in A/A. Design consider-
ation has been given to achieving the following:
. A good receiver signal-to-noise ratio
. Effective FOVs for A/A and A/S performance
. A broad field of regard
. Auto-search pattern modes
. Low false alarm rates

The EOTS uses low-profile gimbals with an optical system that maintains bore-
sight accuracy between the Forward-Looking Infrared (FLIR) and laser functions.
Precise stabilization of the EOTS’s line of sight is achieved by gyro-controlled AZ
and EL gimbals, and fine stabilization is achieved through a fast-steering mirror.
Equipped with a staring 1024-by-1024-element MWIR focal plane array, the

Fig. 10 EOTS integration buildup.

410 G. LEMONS ET AL.

Fig. 11 EOTS capabilities.

EOTS is a dual-FOV system. The narrow FOV is optimized for targeting func-
tions, and the wide FOV is developed to maximize search performance.
The EOTS’s functionality consists of a TFLIR image, laser range finder/
designator, laser spot tracker, and IRST, as shown in Fig. 11.

D. AN/AAQ-37 ELECTRO-OPTICAL DISTRIBUTED APERTURE SYSTEM

The program required a 360-deg spherical coverage missile warning system. The
EO DAS consists of six identical MWIR sensors distributed on the aircraft, each
with a corresponding airframe window panel. The sensors are installed such
that their respective FOVs (95-deg AZ and EL) overlap to provide total spherical
coverage. This EO DAS subsystem provides the pilot with both an MWIR tracking
capability and FLIR visual scene, but its FLIR is more comprehensive. In legacy
FLIR systems, the pilot’s visual scene was limited to the forward sector. With
the F-35’s EO DAS, the pilot has a 360-deg spherical view of the environment.
This allows for a true synthetic vision system, with the image displayed on the
pilot’s Helmet-Mounted Display (HMD).
The EO DAS integration began with a single sensor installation within a pod.
This pod was mounted on an F-16 to support initial testing and data collection for
image processing algorithm validation. This podded system was also mounted on
a QF-4 drone to enable testing of the missile warning function. The next step in
integration was to mount a sensor in an integration-representative fashion on the
Northrop Grumman –owned BAC 1-11 flying testbed. The first introduction of
F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 411

Fig. 12 DAS buildup integration.

multiple EO DAS cameras into the integrated avionics system was performed on
the Lockheed Martin CATB platform. This marked the beginning of integrating
the EO DAS sensors into the Lockheed Martin – developed fusion algorithms.
The final step to fully incorporate the EO DAS into the integrated avionics
system came in March 2011, with the first flight testing on an F-35. The EO
DAS integration timeline is depicted in Fig. 12.
Key EO DAS operational functions in Block 1 of Flight Test Update B are
Navigation Forward-Looking Infrared (NAVFLIR) and missile warning. Block 2
of Flight Test Update B added Surface-to-Air Missile (SAM) launch point report-
ing and situational awareness IRST. These EO DAS functions are available simul-
taneously and serve to enhance situational awareness and defensive response.
Figure 13 illustrates these capabilities.

Fig. 13 DAS capabilities.

412 G. LEMONS ET AL.

E. AN/ASQ-242 COMMUNICATIONS, NAVIGATION, AND IDENTIFICATION SYSTEM

The CNI system (Fig. 14) is an integrated subsystem designed to provide a broad
spectrum of:

. Secure/anti-jam/covert voice and data communications

. Precise radio navigation and landing capability
. Self-identification, beyond-visual-range target identification
. Connectivity with off-board sources of information

In support of the stealthy operation and design goals of the F-35, the CNI subsys-
tem includes techniques to reduce the probability of detection, interception, and
exploitation, and can deploy electronic CM. These techniques include frequency
agility, spread spectrum, emission control, antenna directivity, and low probability
of intercept design capabilities. The CNI system provides interoperability with
existing (legacy) military and civilian communication, RF navigation, and Identi-
fication Friend Foe (IFF)/surveillance systems. It is also interoperable with appro-
priate civilian systems for U.S. and European airspace operations. The CNI system
provides an inherent growth capability and the flexibility to incorporate additional
functionality through software upgrades. It also provides for hardware upgrades
driven by parts obsolescence and enables manufacturing cost reduction and/or
performance improvement.

Fig. 14 CNI system components.

F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 413

The CNI-specific data, signal, and cryptographic processing are performed in

unique CNI processors, and the integrated core processor as required. The CNI
system includes all functionality related to audio generation and distribution
for the aircraft. This includes the pilot intercom; integrated caution, advisory,
and warning messaging; pilot audio alerts; and support for the voice recognition
function.
The CNI system includes an all-attitude Inertial Navigation System (INS) and
anti-jam GPS. These provide outputs of linear and angular acceleration, velocity,
body angular rates, position, attitude (roll, pitch, and platform AZ), magnetic and
true heading, altitude, time tags, and time. The INS and GPS provide navigation
data to the ownship kinematic model, which produces the navigation solution for
the aircraft. The baseline system provides high-rate motion compensation data to
the radar and EOTS.

IV. FUSING THE DATA INTO INFORMATION

The F-35 fusion engine is the software module at the heart of the integrated
mission systems capability on the aircraft. Fusion involves constructing an inte-
grated description and interpretation of the tactical situation surrounding
ownship [2]. It draws from onboard, cooperative, and off-board data sources to
enhance situational awareness, lethality, and survivability [2]. The fusion func-
tionality is divided into two major subfunctions: Air Target Management
(ATM) and Surface Target Management (STM). The purposes of these functions
are to optimize the quality of air and surface target information, respectively.
Their functionality is implemented in three primarily software modules: the
A/A Tactical Situation Model (AATSM), the A/S Tactical Situation Model
(ASTSM), and the Sensor Schedule (SS).
The AATSM software module receives data from onboard and off-board
sources about air objects in the environment. It then integrates this information
into kinematic and identification estimates for each air object. Similarly, the
ASTSM software module receives data from onboard and off-board sources
about surface objects in the environment. It then integrates this information
into kinematic and identification estimates for each surface object.
Objects that are ambiguous between air and surface are sent to both Tactical
Situation Models (TSMs). Each TSM assesses the quality of its tracks to identify
any information needs. The System Track Information Needs (STINs) are sent
from the TSMs to the SS software module. The SS prioritizes the information
needs by track and selects the appropriate sensor mode command to issue in
order to satisfy the information need. The SS provides the autonomous control
of the tactical sensors to balance the track information need and the background
volume search needs.
Measurement and track data are sent to fusion from the onboard sensors
(e.g., radar, EW, CNI, EOTS, DAS) and off-board sources (e.g., MADL, Link 16).
414 G. LEMONS ET AL.

Fig. 15 Closed-loop sensor data fusion.

When this information is received at the TSM, the data enter the data associ-
ation process. This process determines whether the new data constitute an
update for an existing system (fusion) track or potentially new tracks. After
being associated with a new or existing track, data are sent to the state esti-
mation to update the kinematic, identification, and Rules of Engagement
(ROE) states of the object.
Kinematic estimation refers to the position and velocity estimate of an object.
It can also include an acceleration estimate for maneuvering air track. The kin-
ematic estimate also includes the covariance for the track, an estimate of the
track accuracy. Identification estimation provides an estimate and confidence of
the affiliation, class, and type (platform) of the object. The identification
process also evaluates the pilot-programmable ROE assistant rule to determine
when the sensing states and confidences have been met for declaration. Estimation
publishes the updated track state (kinematic, identification, and ROE statuses) to
the system track file. At a periodic rate (about once a second), each track is prior-
itized and then evaluated to determine whether the kinematic and identification
content meets the required accuracy and completeness. Any shortfall for a
given track becomes STINs. The STIN message for the air and surface tracks
are sent to the SS to make future tasking decisions for the onboard sensor
resource. The process continues in a closed-loop fashion with new pieces of
data from the sensors or datalinks. Figure 15 illustrates this process.
F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 415

The results of this fusion of information are provided to the other elements
in mission systems. They are provided to the Pilot/Vehicle Interface (PVI)
for display, fire control and stores for weapon support, and EW for CM
support. This allows these elements to perform their related mission functions
to provide:
. A clearer tactical picture
. Improved spatial and temporal coverage
. Improved kinematic accuracy and identification confidence
. Enhanced operational robustness.

For a clearer tactical picture, multiple detections of an entity are combined into
a single track instead of multiple tracks. For improved spatial and temporal
coverage, a target can be continuously tracked across multiple sensors and
FOVs. This is made possible by the extended spatial and temporal coverage
of the onboard sensors, as well as the off-board contributor. Improved kin-
ematic accuracy and identification confidence requires the effective integration
of independent measurements of the track from multiple sensors or aircraft.
This integration is what improves the detection, tracking, positional accuracy,
and identification confidence. Enhanced operational robustness requires the
abilities to fuse observations from different sensors and hand off targets
between sensors. This leads to increased track resilience to sensor outages or
countermeasures. Increased dimensionality of the measurement space (i.e.,
different sensors measuring various portions of the electromagnetic spectrum)
then reduces vulnerability to denial of any single portion of the measurement
space.

V. NEXT-GENERATION
COCKPIT
The visible product of this infor-
mation gathering is ultimately
the fighter pilot’s office. The
F-35 cockpit (Fig. 16) was
designed to accommodate the
unprecedented amount of data
available to the pilot of a single-
seat fighter. The F-35 display
suite and associated PVI were

Fig. 16 F-35 cockpit.

416 G. LEMONS ET AL.

Fig. 17 Gen III helmet-mounted display system.

conceptualized through a multiyear immersive simulation and evaluation, along

with flight demonstrations on surrogate aircraft.

A. HELMET-MOUNTED DISPLAY
The HMD, included in Fig. 17, was developed to provide the pilot with both exist-
ing and evolutionary tactical advantages. In it, legacy systems, such as the Joint
Helmet-Mounted Cueing System (JHMCS), were supplemented by an integrated
digital Night Vision Camera (NVC). The HMD also provides integrated flight
reference information that allowed for the removal of the traditional Head-Up
Display (HUD). This improved upon the legacy tactical and navigation display,
with an integration surpassing that of the HUD/JHMCS combination. The
HMD design evolved from concepts demonstrated using the Viper-II HMD on
the Variable-stability In-flight Simulator Test Aircraft (VISTA) F-16 program.
This was the basis for proving the concept of a HUD-less cockpit. The HMD
design and integration then matured through several iterations during the SDD
phase. Ultimately, it met ejection safety limits (Gen I to Gen II), program require-
ments for line-of-sight accuracy (Gen II to Gen III), and (later) full ejection envel-
ope (Gen III to Gen III Lite).
The HMD is a monochromatic 1280-by-1024-resolution bi-ocular display
with a 30-by-40-deg FOV. The helmet position is determined by a hybrid
(magnetic, inertial, and optical) tracker system that allows for low-latency
symbol positioning. The HMD can display either the MWIR image provided by
the DAS or a near-IR image provided by the embedded NVC. When the pilot
is looking forward, the NVC image is blended with the image provided by a
F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 417

fixed camera mounted on the glare shield. This is done to eliminate the obstruc-
tion caused by the canopy bow frame.

B. PANORAMIC COCKPIT DISPLAY

A striking difference between the F-35 and other legacy fighters is the incor-
poration of a large-format touchscreen display. The Panoramic Cockpit
Display (PCD) provides the pilot with a programmable and reconfigurable
display that allows for customized views to satisfy mission needs. The PCD
was the replacement selected for the original rear-projection Multifunction
Display Suite (MFDS) [3]. The MFDS rear-projection display technology was
abandoned due to several factors. The lack of available contrast and insuffi-
cient resolution for tactical displays within the F-35 cockpit environment
caused the symbols to be difficult to read. Pilots required the system to be
at full brightness for all daytime flights, which shortened the projection lamp
lifespan. This resulted in frequent maintenance being needed for the develop-
mental test aircraft. Due to these and other factors, the program chose to
recompete the display and associated processing elements early in the SDD
contract.
The PCD consists of an 8-by-20 active matrix liquid crystal display and an
Electronics Unit (EU), as shown in Fig. 18. The display resolution is 2560-by-
1024 color pixels. The PCD uses a portal window concept to support multiple
formats on a single display surface. A touchscreen interface and/or hands-on
throttle and stick provide pilot input to the PCD. The EU has two separate
power supplies and independent display management computers for each half
(left and right) of the display. This, along with the ability to display any format
on either display, provides redundancy to support mission operations and air-
worthiness. The EU accepts multiple video inputs for display on the PCD and
is also responsible for outputting an MPEG-2 standard compressed video
stream to a recording function to facilitate pilot debrief. The PCD unit is compa-
tible with the night vision imagery system.

Fig. 18 PCD display unit and electronics unit.

418 G. LEMONS ET AL.

VI. VERIFICATION
A challenge for the F-35 program was to reduce the cost of integration and ver-
ification. The concept developed was based on verification of performance to
specific missions depicted in Fig. 19. The focus of the plan was to break down
the missions into verifiable pieces that could be developed into a buildup plan
for test and verification. This allowed a view of the problem space that did
not center on any one specific scenario. Instead, it took a broader look at how
the aircraft performed in representative test scenarios.
To rely less on flight testing for developmental buildup, the F-35 Mission
Systems team leveraged mission-representative vignettes that had been developed
for verification. The team applied those same scenarios and criteria to the lower-
level testing. The subsystem and laboratory testing venues allowed for a rapid
turn of changes once a testable product was provided. The laboratories constructed,
along with their associated capabilities listed in Table 1, provided unique venues
to add or subtract levels of complexity to grow the system and isolate issues.
The F-35 Mission Systems team used mission-based scenarios and a pyramid-
type test approach (Fig. 20). With these, the team could progressively add
complexity to the integrated system and quickly compare system performance
to uncover problems early in integration. This also allowed for releases to flight
test that may have a limitation in one area but verification-capable performance
in another. By including the developmental test pilots in the final laboratory
testing sessions, pilots were able to get an idea of performance and learn where
they may experience shortfalls. This allowed them to more efficiently perform

Fig. 19 F-35 mission types.

F-35 MISSION SYSTEMS DESIGN, DEVELOPMENT, AND VERIFICATION 419

TABLE 1 VERIFICATION LABORATORIES

Laboratories Capabilities
Simulated Systems 1. Core avionics hardware and PVI
Integration Station 2. Supplier-provided models for sensors and
weapons
3. Ability to connect to vehicle system and other
integrated avionics laboratories

Open-Air Systems 1. Open-air environment with representative

Integration Station apertures and associated hardware
2. Full avionics hardware suite
3. Capable of CATB cooperative testing

Stimulator-Based Systems 1. RF and IR stimulation capability

Integration Station 2. Representative sensor hardware
minus apertures
3. Multiaircraft cooperative simulation utilizing
using connections to multiple laboratories

Fig. 20 Test buildup methodology.

flight tests to gather the

required data. The partnership
and communication between the
developmental test pilots and the
development team proved essen-
tial to maintaining the rapid pace
required to achieve full system
verification.

VII. CONCLUSION
The F-35 mission systems suite of sensors, displays, and advanced fusion algor-
ithms was developed to satisfy the need to provide tomorrow’s fighter pilot
with an unprecedented level of information. At the same time, the suite allows
pilots to still perform the tactics mandated by the mission. The F-35’s perform-
ance has been demonstrated in more than 100,000 flight hours covering develop-
mental test, operational test buildup, and training flights. The F-35 also showcased
its capability during Red Flag 17-1, where the platform achieved a 20:1 kill ratio in
simulated combat exercises [4].
420 G. LEMONS ET AL.

F-35 mission systems design, development, and verification were achieved

with an object-oriented, multilevel COTS-based architecture and a suite of power-
ful multispectral sensors. As successful products of the F-35 Mission Systems
team’s efforts, the advanced fusion algorithms and next-generation cockpit
enable the pilot to return to the role of tactician.

REFERENCES
[1] Levis, J., Sutterfield, B., and Stevens, R., “Fiber Optic Communication within the F-35
Mission Systems,” IEEE Conference Avionics Fiber-Optics and Photonics, Annapolis,
MD, 2006, pp. 12 – 13. doi: 10.1109/AVFOP.2006.1707476
[2] Frey, T., “F-35 Information Fusion,” AIAA Paper, June 2018 (to be published).
[3] Kalmanash, M. H., “Status of Development of LCOS Projection Displays for F-22A,
F/A-18E/F, and JSF Cockpits,” Proc. SPIE 4362, Cockpit Displays VIII: Displays for
Defense Applications, 7 Sept. 2001. doi: 10.1117/12.439116
[4] Demerly, T., “‘Red Flag Confirmed F-35 Dominance with a 20:1 Kill Ratio’
U.S. Air Force Says,” The Aviationist, https://theaviationist.com/2017/02/28/
red-flag-confirmed-f-35-dominance-with-a-201-kill-ratio-u-s-air-force-says/
[retrieved 3 May 2018].
 CHAPTER 12

F-35 Information Fusion

Thomas L. Frey , J. Chris Aguilar†, Kent R. Engebretson‡,
David K. Faulk§ and Layne G. Lenning}
Lockheed Martin Aeronautics Company, Fort Worth, TX

Information fusion is a set of algorithms that combines data from

all sources to create an integrated view of the environment to provide
situational awareness. Fusion is a core attribute of the F-35, designed
into the mission systems from initial conception. The F-35 Information
Fusion development leveraged experience from past fusion projects
across the corporation and industry; however, there were some funda-
mental architectural decisions and algorithmic solutions that are
unique to the F-35 concept of operation. This paper discusses some of
the key design decisions and features that shaped the final F-35 Infor-
mation Fusion solution.

I. NOMENCLATURE
~
D vector difference between Observer 1 and Observer 2
Kk Kalman filter gain
k Dempster-Shafer measure of conflict between two masses
mðAÞ Dempster-Shafer probability mass of proposition A
Pkjk1 previous covariance estimate propagated to time k
Rk measurement covariance for Zk
R̂1 range estimate from Observer 1 to the target
R̂2 range estimate from Observer 2 to the target
S1 covariance matrix for the target from Observer 1
S2 covariance matrix for the target from Observer 2
Sint covariance matrix for the intersection point

 Senior Fellow, F-35 Information Fusion, AIAA Associate Fellow.

†
Software Engineer Sr. Staff, F-35 Information Fusion.
‡
Fellow, F-35 Information Fusion, AIAA Senior Member.
§
Embedded S/W Engineer Staff, F-35 Information Fusion, AIAA Senior Member.
}
Principal Research Scientist, F-35 Information Fusion.

421
422 T. L. FREY ET AL.

X̂ kjk resultant state estimate at time k

X̂ kjk1 previous state estimate propagated to time k
û1 unit direction vector from Observer 1 to the target
û2 unit direction vector from Observer 2 to the target
Zk measurement vector at time k

II. INTRODUCTION
In his 1982 book Megatrends, John Naisbitt predicted that in the Information Age
we would find ourselves “. . . drowning in information but starved for knowledge”
[1]. For many aviation or military applications, as the amount of information
increases, the sheer volume of data becomes overwhelming and results in the
loss of situational awareness [2]. As the amount of data increases and control
choices multiply, the pilot workload increases exponentially to a point where
eventually even the most able pilots begin to miss important information or fail
to recognize critical situations [3]. The loss of situational awareness degrades
the reaction time of the user [4, 5]. Shenk coined the term data smog to describe
the data overload caused by the dramatic increase in information without a means
to readily incorporate this information or its significance [6]. In Consilience: The
Unity of Knowledge, E. O. Wilson predicted the need for a synthesizer that could
combine relevant information at the right time to support critical decisions [7].
Kline identified fusion algorithms as this synthesizer to reduce information over-
load, improve situational awareness, and reduce user reaction time [8].
Fifth-Generation fighters, with their diverse sensor suites and multirole
mission sets, require some form of information fusion to support the pilot’s situa-
tional awareness (Fig. 1). The F-35 avionics suite comprises several complemen-
tary sensors and off-board datalinks but is a single-seat fighter with no weapon
systems operator. Without some form of information fusion, the pilot would be
left to manually correlate sensor and datalink tracks together, while executing tac-
tical air, land, and sea missions while also trying to fly the aircraft. This combi-
nation of factors can increase pilot workload and quickly lead to an
overwhelming amount of displayed information. Information fusion algorithms
aggregate the onboard and off-board sensor information to provide a complete
and accurate representation of the environment, resulting in an increase in situa-
tional awareness, which is the ultimate goal of fusion [9, 10].
The terms data fusion, sensor fusion, and information fusion are often used
interchangeably, and yet these terms have subtle distinctive connotations within
the community. The Joint Directors of Laboratories (JDL) Data Fusion Model
defines a useful categorization of fusion algorithms and techniques used in the sol-
ution of many general fusion problems [11]. It defines data fusion as the combin-
ing of information to estimate or predict the current or future state of the
environment. Level 1 fusion is focused on object assessment. Level 1 fusion algor-
ithms include: 1) data association algorithms, which determine whether
 F-35 INFORMATION FUSION

Fig. 1 Information fusion is a set of algorithms that combines data from all sources to create an integrated view of the environment.
423
424 T. L. FREY ET AL.

information from multiple sources describes the same object; and 2) state esti-
mation algorithms, which estimate the current (and, in some cases, future) state
of the physical object in the environment. The estimate includes both the kin-
ematic state (e.g., position, velocity) and an estimate of the object’s Identification
(ID). Level 2 fusion focuses on aggregating the Level 1 objects, inferring relation-
ships between/among the objects and corresponding events, and assessing the
unfolding situation. Level 3 fusion assesses the impact of perceived, anticipated,
or planned actions in the context of the unfolding situation, for instance, in
terms of lethality and survivability. Level 4 fusion is focused on process refine-
ment, including sensor resource management or sensor feedback to modify
sensor actions and refine the overall situational picture.
There are many published algorithms that offer methods to address each of
these fusion functions; however, in the field the imperfect nature of data and
the varying fidelity of disparate data sources make the fusion problem much
more challenging to solve. To quote Yogi Berra, “In theory there is no difference
between theory and practice. In practice, there is” [12]. For example, object refine-
ment is relatively easy when there are a few well-separated objects, but as the
number of objects increases and their spacing decreases, the data association
problem can become much more difficult or even unsolvable. The potential for
the object to perform maneuvers, coupled with a highly dynamic environment
involving multipath, interference, signal blockage, and weak signals, can compli-
cate the ability of fusion to make sense of the environment. Information fusion
must be robust enough to provide reliable situational awareness, even in this
challenging environment.

A. THE EVOLUTION OF FUSION TECHNOLOGY

The first radar was introduced to a fighter in the mid-1940s. As confidence grew in
this technology, pilots came to rely on the radar to provide situational awareness
in the environment. A radar warning receiver was later added to give a coarse
indication of the general direction of a hostile emitter. In the 1970s and 1980s,
new sensors and datalinks added secondary sources of ranged tracks. This
ushered in the initial generation of fusion technology, which used data association
or correlation algorithms to identify tracks that most likely described the same
object in space and then suppressed all except the most accurate copy of the
track from the display (Fig. 2). This is sometimes referred to as correlation or
display fusion. The accuracy of this technique was equal to the accuracy of the
best track.
The next step in fusion technology was to combine the output of multiple
sensor tracks into a blended system solution. By blending the tracks from two
or more sensors, the resultant system track accuracy approached the accuracy
of the best parameter of the contributing sensors. For example, blending tracks
from a radar and an Infrared Search and Track (IRST) sensor could have the
range and range rate accuracy of the radar along with the angle and angle rate
F-35 INFORMATION FUSION 425

Fig. 2 The evolution of fusion technology.

accuracy of the IRST; however, the accuracy of the resultant track remains
limited by the track’s update rate. If the track’s update rate (fusion rate) is
larger than the measurement’s rate, then there is a loss of accuracy, even with
optimal algorithms [13].
Fifth-Generation aircraft are designed to process the sensor measurements
rather than the sensor tracks, resulting in an integrated system track containing
the most precise track accuracy and enabling cooperative sensing across aircraft.
Measurement-level processing can provide earlier discovery of objects in the
environment that are hard to detect. By processing the measurement-level data,
the system can use detections from any sensor (or aircraft) to confirm a track
before any single sensor can make the declaration. The focus on the measurement
data rather than track data also means that combat ID information from a sensor is
retained by the system track, even when the track is no longer in the sensor’s field of
view because the system track can be maintained by other sensors or aircraft.
In addition to improved accuracy and detection performance, the introduc-
tion of an Autonomous Sensor Management capability provided the ability to
react and refine objects in the environment much faster than any human could
respond [14]. The addition of the Autonomous Sensor Manager is referred to
as closed loop fusion. This capability provides the fusion process a feedback
loop to coordinate the actions of the sensors in a complementary way to detect,
refine, and maintain tracks based on system priorities [15]. The sensor manage-
ment capability evaluates each system track, determines any kinematic or ID
needs, assesses those needs according to system track prioritization, and cues
the sensors to collect the required information. Analogous to John Boyd’s
Observe, Orient, Decide, and Act (OODA) loop [16], which expressed the engage-
ment advantage related to the pilot’s ability to understand and react to an adver-
sary, closed loop fusion accelerates the ability of the pilot to understand and
respond to an object in space faster and often at a much greater range than
legacy systems.

III. FUSION ARCHITECTURE

The F-35 is not one but three highly common aircraft for the U.S. Air Force, U.S.
Marine Corps, and U.S. Navy, designed to avoid the higher costs of developing,
426 T. L. FREY ET AL.

Fig. 3 The F-35 sensor suite.

procuring, operating, and supporting three separate tactical aircraft designs to

meet the services’ similar but not identical operational needs. Numerous
partner countries are also involved, and their needs are incorporated into one
or more of these three basic variants. The mission system sensors and software,
on the other hand, are common among all three variants. The F-35 sensor suite
(Fig. 3) includes the APG-81 Active Electronically Scanned Array (AESA)
radar, the ASQ-239 Electronic Warfare (EW)/Counter Measures (CM) suite,
the AAQ-40 Electro-Optical Targeting System (EOTS), the AAQ-37 Electro-
Optical Distributed Aperture System (DAS) system, and the ASQ-242 Communi-
cation, Navigation, and Identification (CNI) system. These five sensors provide
object detection and measurements in the Radiofrequency (RF) and Electro-
Optical/Infrared (EO/IR) spectrum to F-35 Information Fusion, resulting in
more information about the objects in the environment around the aircraft
than has been available to a fighter aircraft before.
In addition to the onboard sensors, the F-35 transmits and receives tracks via
the Link 16 datalink and the Multifunction Advanced Data Link (MADL). Link 16
provides connectivity with legacy aircraft and command and control systems,
which allows the aircraft to transmit information about selected tracks and
receive surveillance information from command and control centers. Designed
primarily for cueing, information about the quality and timeliness of the track
is limited.
The F-35 MADL communication link was designed explicitly for the F-35 to
support the sharing of information among the flight group. Unlike legacy data-
links, MADL bandwidth supports passing fusion-quality information on all air
and surface tracks to other members of the flight group. These data include a
locally derived track state, track covariance, ID measurement history, and RF
history, as well as other metadata associated with each track. The amount of
F-35 INFORMATION FUSION 427

potentially redundant off-board information provided by MADL was one of the

largest challenges for fusion design. It was now possible to have many copies
(in some cases greater than 10) of the same track from onboard and off-board
sources simultaneously, creating a high potential for clutter. The fusion of this
spatially diverse data offered a significant improvement in situational awareness
and cooperative sensing.

IV. THE F-35 INFORMATION FUSION APPROACH

Prior to the introduction of the 5th-Generation fusion systems, fusion historically
referred to only the data association and estimation processes. The earliest parti-
tioning of the F-35 fusion capability envisioned the sensor management capability
to be independent of the fusion process; however, there was already strong evi-
dence that the autonomous sensor manager was fundamental to efficient fusion
performance and sensor optimization. During the early stages of design, the
sensor manager was repartitioned to the F-35 fusion design. Figure 4 shows the
top-level functional architecture of the F-35 fusion design, highlighting the data
association, estimation (both kinematic and ID), and sensor management
functionality.
The F-35 Information Fusion design isolates fusion algorithms from both the
sensor and datalink inputs, as well as any consumers of fused data. Essentially, the
fusion algorithms comprise a black box, known internally as the fusion engine,
and sensor inputs and data consumers are encapsulated in external software
objects known as Virtual Interface Models (VIMs). For incoming data, the sensor-
specific or datalink-specific VIMs fill in missing data (e.g., navigation state, sensor
bias values), preprocess the information, and translate it into a standard form for
the fusion process. For data leaving fusion, the outgoing VIM, known internally as

Fig. 4 F-35 fusion architecture.

428 T. L. FREY ET AL.

the fusion server, provides data to the various consumers of fused information,
both onboard and off-board. The fusion server isolates users of the fused infor-
mation from both the fusion process and data sources. Legacy fusion implemen-
tations reported fusion tracks as a monolithic block (i.e., one size fits all) where all
data consumers received the same message. Any propagation of the data or con-
version was the responsibility of the recipient. This created a coupled interface
between fusion and the data consumers. When a new data source was introduced
to fusion, the interface changes to make these data available impacted all consu-
mers of that message, whether the data were used or not, making changes to
fusion very costly. The fusion server sends each information consumer a tailored
message that contains only the information required to support that consumer.
This isolates that consumer from changes to any data source or to the fusion
algorithm. The use of VIMs enables the fusion architecture to be extensible to
new sensors and data sources, as well as new data consumers, over its lifetime.

V. INFORMATION TIERS
Sensor fusion can result in poor performance if incorrect information about
sensor performance is used:

A common failure in data fusion is to characterize the sensor performance in

an ad hoc or convenient way. Failure to accurately model sensor performance
will result in corruption of the fused results. [17]

One of the key architecture decisions for F-35 fusion is how to share information
among aircraft. Independent data can be incorporated optimally into a filter for
the highest accuracy; however, if dependent data are incorporated under the
assumption of independence, the result will be track instability and, eventually,
track loss [18]. Data consumers on the F-35, including the pilot, receive the kin-
ematic and ID estimate of each track based on all available data sources, both
onboard and off-board. This is referred to as the Tier 3 solution. However,
when sharing information with other aircraft, each F-35 shares the information
describing a track based solely on measurements from onboard sensors. This is
referred to as the Tier 1 solution. By ensuring that the information received
from MADL is independent, the track information can be converted into equiv-
alent measurements [19] by the recipient supporting both track-to-track
and measurement-to-track of the information. The sharing of Tier 1 data
ensures that the information is not coupled to any specific fusion algorithm
and provides a method for dissimilar fusion platforms to share optimal fusion
data in the future (Fig. 5). In late 2016, Lockheed Martin and the U.S. government
used this technique to share an F-35 fused track of a target drone across MADL to
a surface-based weapons system that had no line of sight to the drone. The
surface-based weapons system converted the F-35 MADL Tier 1 information
into equivalent measurements that were consumed by the native engagement
F-35 INFORMATION FUSION 429

Fig. 5 Information tiers.

tracker. Together, the networked systems achieved a successful acquisition, gui-

dance, and kinematic intercept of the track using a surface-to-air missile.
To incorporate dependent information or unknown pedigree information into
the F-35 estimate, fusion includes a covariance intersection update that is more
tolerant of rebroadcasted data [20]. The covariance intersection algorithm
reduces the error of the estimate when novel information is introduced but
does not improve the error with redundant data. This technique provides flexi-
bility for incorporating data from many different datalinks in the future, where
the pedigree of the data is not known.
The integration of legacy datalinks uncovered other challenges to optimal
fusion. When system tracks are sent off-board over a datalink, it is important
to send the most precise error characterization available for the system track.
The covariance matrix provides this sensor characterization to other users of
the data; however, in many legacy datalinks, quality factors are used to describe
track errors rather than multidimensional covariance matrices. These quality
factors denote a maximum area or volume uncertainty value for the data. This
lack of specificity in dimensional accuracy requires the recipient to assume the
worst possible accuracy for the data in all dimensions to ensure stable tracking
behavior under all conditions. From an algorithmic perspective, the pessimistic
error characterization leads to a higher incidence of false correlation and a
deweighting of the off-board data [21].
Given the off-board data have successfully correlated to the correct system
track, the reported error of the remote track is used to weight the influence of
the data. Recall from the Kalman filter that the optimal gain Kk is used to deter-
mine the blending of the new measurement [22, 23]. Equation (1) is an alternative
form of the Kalman filter, showing how the optimal filter gain blends the
430 T. L. FREY ET AL.

propagated system estimate and the observation data. As the gain increases, the
observation data Zk has a larger influence on the new estimate.
X̂ kjk ¼ Kk Zk þ ð1  Kk Þ X̂ kjk1 (1)
The gain for a Kalman filter is determined by the estimated system track errors
and the measurement errors, as shown in Eq. (2).
h i1
Kk ¼ Pkjk1 Pkjk1 þ Rk (2)

If the measurement errors are pessimistic, the filter gain decreases and results
in a natural deweighting of the measurement influence on the new estimate. Con-
versely, if the measurement errors are optimistic, the filter gain increases, and the
measurement has too large an effect on the track. The use of quality factors on
legacy datalinks leads to smaller gain values and reduces the potential accuracy
of the resultant track. The F-35 MADL datalink reports the track covariance for
each track, allowing for accurate weighting of the off-board contributions.

VI. EVIDENCE-BASED COMBAT IDENTIFICATION

Combat ID is the process of attaining an accurate characterization of detected
objects in the joint battlespace [24]. Characterizations may include affiliation,
class, type, nationality, and mission configuration. The F-35 employs a probabil-
istic, evidential reasoning approach to ID in a fundamental departure from
heuristic algorithms employed on other tactical platforms. The decision to
switch to a probabilistic formulation was driven by lessons learned from those
platforms to eliminate misidentification due to forced declarations and the
inability to handle ambiguous information sources. The transition to a probabil-
istic framework forced design changes throughout the rest of the information
fusion design, including sensors, pilot displays, and sensor tasking.
Early in the F-35 design, a trade was performed between Bayesian and
Dempster-Shafer inferencing algorithms. Ultimately, Dempster-Shafer was
selected due to greater robustness to conflicting data, less reliance on a priori
target distributions, and a more natural formulation of relationships between
sensor attributes and corresponding platform sets. Algorithms based on Bayes’s
rule provide intuitive probabilistic confidence values; however, Bayesian inference
places a higher dependence on a priori probabilities, the PðAÞ term in Eq. (3), to
seed the probability of a given target ID declaration. Accurate a priori information
can be difficult to obtain in practice.
PðBjAÞPðAÞ
PðAjBÞ ¼ (3)
P ðBÞ
In contrast, Dempster-Shafer belief theory does not require explicit a priori
information. Further, Fixsen and Mahler have shown that Dempster-Shafer
F-35 INFORMATION FUSION 431

probability masses can approach Bayesian probabilities [25] when probability

masses are chosen appropriately. In our Dempster-Shafer formulation, each
sensor declaration maps to a set of platform disjunctions, or a list of disjunctions
in the case of ambiguities. These sensor disjunctions are combined using Demp-
ster’s rule of combination, shown in Eq. (4). Inconsistent propositions, designated
by k in Eq. (5), are removed and their confidence is distributed across the consist-
ent terms through normalization.

1 X
mð AÞ ¼ mðBÞmðCÞ (4)
1k
B>C=;
X
k¼ mðBÞmðCÞ (5)
B>C¼;

Implementation of the Dempster-Shafer algorithm requires first defining the

frame of discernment: the set of objects over which the algorithm reasons. Pre-
vious approaches to combat ID have been forced to limit their platform set due
to processing and memory constraints. By contrast, the F-35 design took a holistic
approach to include all tactically relevant platforms in the theater, minimizing
misidentification of platforms that are not in library. Platforms are interconnected
in a strict hierarchical taxonomy with levels in the taxonomy labeled subtype, type,
class, and affiliation, as shown in Fig. 6. The algorithm creates ambiguity lists and
associated confidences for each level of the taxonomy. The type ambiguity list is
inferred from the subtype ambiguity list; the class ambiguity list is inferred from
the type ambiguity list; and the affiliation ambiguity list is inferred from the class
ambiguity list. The strictly enforced hierarchy allows for processing simplifica-
tions and ensures that the confidence at a higher level in the taxonomy is
always greater than or equal to the confidence at lower levels [26].

Fig. 6 Platform taxonomy levels.

432 T. L. FREY ET AL.

The platform taxonomy also provides the framework for mapping sensor
attributes to a platform set. This mapping, contained in an onboard relational
threat database, encodes the transitional probability of observing a feature
given a particular platform subtype. Each sensor declaration must also have an
associated probability or confidence so the input to fusion is weighted based on
its relative merit in the same way a kinematic report is weighted based on its
covariance matrix. At the beginning of the F-35 program, none of the sensor
modes produced confidence values, so these needed to be generated. Called the
Basic Probability Assignment (BPA), the generation of confidence values
proved to be one of the more challenging aspects of the design. It should go
without saying, the generated error must also correctly bound the true error dis-
tribution. Continuing with the kinematic tracking analogy, it does no good to
report a track range error as having a standard deviation of 1 m when the true
error statistics are 100 m. Similarly, the ID probabilities must accurately reflect
the true target uncertainty.

Fig. 7 F-35 expanded data window.

F-35 INFORMATION FUSION 433

Although the soft decisions developed by fusion provide a more accurate rep-
resentation of what is known and not known about a given target’s ID state, the
pilot requires actionable information, which requires that the soft decision be con-
verted into a hard declaration. The Dempster-Shafer algorithm produces support-
plausibility intervals that bound the estimate of probability but do not directly
provide platform declarations. The F-35 design converts the probability masses
into pignistic probabilities that effectively distribute the platform disjunction con-
fidence across all elements of the disjunction [27]. The system makes a hard ID
declaration when the translated confidence exceeds a user-defined threshold.
The taxonomy is traversed from lower nodes to higher nodes until the confidence
threshold is exceeded. The F-35 combat ID output, shown in the second row of
Fig. 7, is flexible enough to allow a display of information from any level in the
taxonomy. The output also contains ID declarations from other sensors and off-
board sources, which helped to develop trust in the fused outputs and transition
from legacy platforms. In this example, a Link 16 declaration of fighter was com-
bined with the MADL declaration of F-35 to produce a high-confidence type
declaration of F-35 and a friendly confidence of 1.

VII. AUTONOMOUS SENSOR MANAGEMENT

Technical advances in the capability of modern sensors, increases in the number
of multiple sensor systems, and the migration to increased connectivity have led to
a sensor network that has exceeded the ability of a human to efficiently control
them [28]. This performance gap led to the development of an automatic feedback
mechanism in the closed-loop fusion model that autonomously modifies the
actions of the sensor suite to achieve a mission-level objective or behavior. The
use of an algorithm to task or modify the behavior of the information sources is
referred to as process refinement and is categorized as JDL Fusion Level 4 [29].
The goals of the sensor manager are to: 1) reduce pilot workload by automating
sensor actions and selection; 2) prioritize information requests (including
pilot requests, background volume searches, and fusion information needs); and
3) reconfigure sensor assets to compensate for individual sensor loss or unavail-
ability [30].
Sensor management strategies focus generally on the control of heterogeneous
collocated or spatially diverse sensors to achieve a desired goal (e.g., track accu-
racy). An autonomous sensor management system requires an objective function
to optimize sensor utilization to select a preferred action over many possible
alternatives. Many early strategies [31 – 33] were ad hoc or rule-based, having a
strong coupling to sensor requirements. Others [34 – 38] explored information-
theoretic approaches that sought to minimize uncertainty in the track covariance,
thus decoupling the objective function from a specific sensor suite. The advantage
of information-theoretic solutions is their adaptability to many problems;
however, they are focused on optimizing the information of a given track while
434 T. L. FREY ET AL.

neglecting the value of that information to the mission objectives [39]. For a
fighter aircraft, the primary objective is to provide situational awareness to the
operator during a complex and changing environment to support critical
mission decisions [40].
Endsley [41] defines situational awareness from the perspective of the fusion
customer (pilot), which could then be decomposed further into times or distances
where key decisions must be made. The necessary information to support these
key decisions are independent of the sensor’s ability to meet these needs. The
information boundaries and associated information needs define a dimensional
space that can be used to derive a global objective function to support auton-
omous sensor management [42] and can be used to define the sensor and
fusion capabilities necessary to support these decisions. A benefit from this
mapping is that the mission goals can be directly related to the sensor capabilities
in terms of range, accuracy, and latency. This allows the designer to trace system-
level fusion situational awareness requirements to individual sensor performance
requirements.
For the F-35, the automated sensor management function is responsible for
efficiently managing the sensor suite to provide critical information about the
objects in the environment to the pilot, supporting critical decisions and
actions. It does this by prioritizing the systems tracks, autonomously directing
system resources to maintain existing tracks, gathering mission ID and Rules of
Engagement (ROE) information, and balancing track maintenance with new
track discovery through searching. System priorities, track information needs,
and track accuracies are based on the track type, pilot emphasis (if any), and infor-
mation boundaries around the aircraft. The autonomous sensor manager also
provides methods for the pilot to collaborate with the fusion system, both for
refinement or reprioritization of existing tracks and to cue the system to search
for new objects in the environment. The pilot can designate a line of sight (air)
or area (surface) and command a cued search. The sensor manager will direct a
series of active and passive scans focused on the detection of new air or surface
objects. If the pilot selects an existing track, sensor management will raise the pri-
ority of the track and will cue sensors to meet the information needs associated
with the tactical zone (full state).
The autonomous sensor management algorithm focuses on providing the
information needs for every track based on priority. The goal is not to drive
each track to the best accuracy, but to instead drive it to sufficient accuracy
and information content. In practice, for situational awareness, there is a level
of component accuracy (e.g., range, angle) where the information is sufficient
to support the pilot’s understanding of the environment to decide. Additional
accuracy beyond this point does not significantly improve the pilot’s awareness
or decision-making ability. Therefore, information gain above a sufficient level
does not support the mission objective and should be directed toward other
objectives in the environment, such as searching for new objects. This concept
of sufficiency can be used to define a new constrained objective function that
F-35 INFORMATION FUSION 435

incorporates the concept of sufficient information and targets specific infor-

mation needs. The addition of the sufficiency constraint removes the dependency
of the objective function on the accuracy of the sensor making the update, as long
as the sensor can provide the required accuracy. This objective function encapsu-
lates the mission goals in terms of situational awareness while incorporating the
constraint of sufficiency. The definition of sufficiency changes at each boundary
based on the pilot’s needs to make critical decisions. The closed-loop nature of
the autonomous sensor manager enables the system to respond more rapidly
to changes in the environment and to optimize the sensor behavior, freeing the
pilot from the role of sensor manager and returning him or her to the role
of tactician.

VIII. COOPERATIVE SENSING

The F-35 MADL was designed to support full sharing of information among
aircraft. MADL bandwidth supports the exchange of all air and surface tracks
between/among participants within the flight group. Given that each F-35 has
multiple sensors detecting multiple targets—and sometimes spurious signals—
this can lead to the exchange of numerous, potentially duplicative tracks over
MADL. Therefore, the F-35 places limits on the kinds of tracks and associated
information that can be transmitted over the link.
For MADL distribution, a single F-35 system track is divided into three mess-
ages: the basic MADL surveillance track, extended combat ID (XID), and RF para-
metric extensions. The basic surveillance track provides the independent
kinematic state estimate and track covariance at the time of the last measurement
update. It is important to note that the kinematic estimate for a sent track can be
either ranged or angle-only (no observed range). This distinction becomes impor-
tant for advanced multiship tracking techniques, such as angle/angle ranging or
Time Difference of Arrival (TDOA), described later. The MADL surveillance
track also includes a list of sensors contributing to this track, as well as ID
summary data. The XID message contains a higher-fidelity ID ambiguity list, in
addition to ID measurements (e.g., IFF). The RF parametric message contains
the Electronic Signal Measurement (ESM) data correlated to this track. The
sharing of this detailed information allows each aircraft to leverage the spatial
diversity of the flight group.
One of the initial multiship capabilities of the F-35 was the ability to coopera-
tively range airborne emitters by finding the intersection (or point of nearest
approach) for angle-only tracks on two or more different aircraft (Fig. 8).
On the receipt of a MADL angle-only track, the receiving fusion system deter-
mines possible intersection points with its own onboard angle-only tracks. It is
possible for a given MADL angle-only track to intersect with multiple ownship
tracks. In truth, only one of the intersections is correct. These alternate angle/
angle candidates are referred to colloquially as ghosts. Once all ghosts have
been eliminated, the range of the track from each participant can be calculated
436 T. L. FREY ET AL.

Fig. 8 Passive angle/angle ranging of airborne emitters.

in Eqs. (6) and (7) as follow [43]:

 
  ~ D ~u2 
~ 
R̂1 ¼ R1 ¼ (6)
j~
u1  ~
u2 j
 
  ~ D ~u1 

R̂2 ¼ ~ 
R2 ¼ (7)
u1  ~
j~ u2 j
Rotating the covariance of each observer into a common reference frame, the
range error in the common reference frame is approximately the intersection of

Fig. 9 Passive ranging of surface emitters using TDOA techniques.

F-35 INFORMATION FUSION 437

the two error covariance matrices, which can be expressed in Eq. (8) as
 
1 1
Sint ¼ S1 1 þ S2 (8)

For surface emitters, fusion incorporates a TDOA capability for precision

location. The TDOA capability provides the ability for multiple aircraft to syn-
chronize ESM dwells in time and frequency. Upon initiation, the autonomous
sensor manager configures the dwells across the network, and then all aircraft
send the time of arrival of any received pulses to the initiating aircraft. Fusion
processes these pulse streams from the cooperating participants to detect
common pulse pairs. Common pulse pairs form a surface of constant delta-time
that, when intersected with the Earth, produce a hyperbola of constant time differ-
ence called an isochrone. Multiple pulse pairs represent the intersection of one or
more isochrones and produce range estimate (Fig. 9).

IX. SUMMARY
The F-35 Information Fusion software combines information from both onboard
and off-board data sources, providing the pilot with advanced capabilities not
available on legacy aircraft. Further, this extensible approach to information
fusion leverages the spatial and spectral diversity among multiple F-35
wingmen, creating an innovative tactical network where data are shared instan-
taneously with other F-35s and legacy aircraft. The F-35 Information Fusion
implementation of data association, state estimation, and combat ID ensures
that the pilot has accurate situational awareness, allowing for advanced target
detection, tracking, and tactical employment. The autonomous sensor manager
provides timely reaction to a changing environment and ensures that all tracks
are refined to a prespecified quality based on priority, allowing the pilot to
return to the role of tactician. The F-35 MADL provides sufficient bandwidth
for complete sharing of detailed fusion solutions and accuracies for all air and
surface targets, resulting in improved situational awareness for all pilots in the
MADL network. Using data-sharing methods to ensure that the data pedigree
is maintained, the MADL information can be processed like a remote sensor,
resulting in improved accuracies and new capabilities.

REFERENCES

[1] Naisbitt, J., Megatrends: Ten New Directions Transforming Our Lives, Warner Books,
New York, 1982.
[2] Chalmers, B. A., “On the Design of a Decision Support System for Data Fusion and
Resource Management in a Modern Frigate,” NATO System Concepts and
Integration Panel Symposium, 1998, pp. 2-1– 2-13.
438 T. L. FREY ET AL.

[3] Musick, S., and Malhorta, R., “Chasing the Elusive Sensor Manager,”
Proceedings of the IEEE 1994 National Aerospace and Electronics Conference,
NAECON 1994 (23– 27 May 1994, Dayton, OH), Vol. 1, IEEE, New York, 1994,
pp. 606 – 613.
[4] Endsley, M. R., “Situation Awareness in an Advanced Strategic Mission,” NOR DOC
89-32, Northrop Corporation, Hawthorne, CA, 1989.
[5] Endsley, M. R., Farley, T. C., Jones, W. M., Midkiff, A. H., and Hansman, R. J.,
“Situation Awareness Information Requirements for Commercial Airline Pilots,”
ICAT-98-1, International Center for Air Transportation, Massachusetts Institute of
Technology, Cambridge, MA, 1998.
[6] Shenk, D., Data Smog: Surviving the Information Glut, Harper, San Francisco, 1997.
[7] Wilson, E. O., Consilience: The Unity of Knowledge, Vantage Books, New York, 1998.
[8] Kline, G. B., Moon, B., and Hoffman, R. R., “Making Sense of Sensemaking 1:
Alternative Perspectives,” IEEE Intelligent Systems, Vol. 21, No. 4, July/Aug. 2006,
pp. 70 – 73.
[9] Canavan, R., Ganesh, C., and Matuszak, B., “Evolution of Fusion in Navy Tactical
Systems,” 17th International Conference on Information Fusion (FUSION), 2014,
pp. 1 – 7.
[10] Endsley, M. R., and Garland D. J (eds.), Situation Awareness Analysis and
Measurement, Lawrence Erlbaum, Mahwah, NJ, 2000.
[11] Steinberg, A., Bowman, C., and White, F., “Revisions to the JDF Data Fusion Model,”
Proceedings of AIAA Missile Sciences Conference, Naval Postgraduate School,
Monterey, CA, 17 – 19 Nov. 1998.
[12] Yogi Berra Quotes, BrainyQuote.com, Xplore Inc. [online database], https://www.
brainyquote.com/quotes/yogi_berra_141506 [retrieved 22 Jan. 2018].
[13] Chong, C., Mori, S., Govaers, F., and Koch, W., “Comparison of Tracklet Fusion and
Distributed Kalman Filter for Track Fusion,” Proceedings of 17th International
Conference on Information Fusion, Salamanca, Spain, 2014.
[14] Steinberg, A. N., and Bowman, C. L., “Revisions to the JDL Data Fusion Model,”
Chapter 3 in Handbook of Multisensor Data Fusion, Liggins, M. E., Hall, D. L.,
and Llinas, J. (eds.), CRC Press, London, 2009.
[15] Steinberg, A., and Snidaro, L., “Levels?,” Proceedings of the 18th International
Conference on Information Fusion, Washington, DC, July 2015.
[16] Schechtman, G. M., “Manipulating the OODA Loop: The Overlooked Role of
Information Resource Management in Information Warfare,” master’s thesis, AFIT/
GIR/LAL/96D-10, Air Force Institute of Technology, Wright-Patterson Air Force
Base, OH, 1996, p. 28.
[17] Hall, D. L., and Steinberg, A. N., “Dirty Secrets in Multisensor Data Fusion,” Chapter
21 in Handbook of Multisensor Data Fusion, Hall, D. L., and Llinas, J. (eds.), CRC
Press, London, 2001.
[18] Blasch, E., and Israel, S., “Situation/Threat Context Assessment,” Proceedings of the
18th International Conference on Information Fusion, Washington, DC, 6 – 9 July
2015.
[19] Chong, C., Mori, S., Barker, W. H., and Chang, K., “Architectures and Algorithms
for Track Association and Fusion,” IEEE Aerospace and Electronic Systems
Magazine, 2000, Vol. 15, No. 1, pp. 5 – 13.
F-35 INFORMATION FUSION 439

[20] Julier, S. J., and Uhlmann, J. K., “A Non-divergent Estimation Algorithm in the
Presence of Unknown Correlations,” Proceeding of the American Control Conference,
1997, pp. 2369– 2373.
[21] Frey, T., Rasmussen, N., and Engebretson, K., “Information Exchange
Considerations for Effective Fusion among Heterogeneous Network Participants,”
Proceedings of the AIAA Infotech@Aerospace Conference, Jan. 2018.
[22] Kalman, R. E., “ A New Approach to Linear Filtering and Prediction Problems,”
Journal of Basic Engineering, Vol. 82, 1960, p. 35.
[23] Kalman, R. E., and Bucy, R. S., “New Results in Linear Filtering and Prediction
Theory,” Journal of Basic Engineering, Vol. 83, No. 1, 1961, pp. 95 – 108.
[24] Capstone Requirements Document for Combat Identification, Joint Requirements
Oversight Council, 19 Mar. 2001, pp. I – 1.
[25] Fixsen, D., and Mahler, R. P. S., “The Modified Dempster-Shafer Approach to
Classification,” IEEE Transactions on Systems, Man, and Cybernetics—Part A:
Systems and Humans, Vol. 27, No. 1, Jan. 1997, pp. 96 – 104.
[26] Shafer, G., and Logan, R., “Implementing Dempster’s Rule for Hierarchical
Evidence,” Artificial Intelligence, Vol. 33, 1987, pp. 271– 298.
[27] Smets, P., “Data Fusion in the Transferable Belief Model,” Proceedings of the 3rd
International Conference on Information Fusion, Paris, France, 2000, pp. 21– 33.
[28] Freedberg, S. “F-22, F-35 Outsmart Test Ranges, AWACS,” Breaking Defense, 7 Nov.
2016, https://breakingdefense.com/2016/11/f-22-f-35-outsmart-test-ranges-
awacs/ [retrieved 23 Jan. 2018].
[29] McIntyre, G., “A Comprehensive Approach to Sensor Management and
Scheduling,” Ph.D. dissertation, George Mason University, 1998.
[30] Steinberg, A., and Bowman, C. L., “Revisions to the JDL Data Fusion Model,”
in Handbook of Multisensor Data Fusion, Hall, D. L. and Llinas, J. (eds.), Boca Raton,
FL, CRC Press, 2001, pp. 2-1–2-18.
[31] Adrian, R., “Sensor Management,” Proceedings of the 1993 AIAA/IEEE Digital
Avionics System Conference, Fort Worth, TX, 25 – 28 Oct. 1993, pp. 32 – 37.
[32] Bier, P., Rothman, P., and Manske, R., “Intelligent Sensor Management for Beyond
Visual Range Air-to-Air Combat,” Proceedings of the IEEE 1988 National Aerospace
and Electronics Conference NAECON 1988, Vol. 1, Dayton, OH, May 1990.
[33] Rothman, P., and Bier, S., “Intelligent Sensor Management Systems for Tactical
Aircraft,” Second National Symposium on Sensors and Sensor Fusion, Chicago, IL,
March 1989.
[34] Buede, D., Martin, J., and Sipos, J., “AIFSARA: A Test Bed for Fusion, Situation
Assessment and Response,” Proceedings of the 4th AAAIC 1988 Conference, Dayton,
OH, Oct. 1988.
[35] Greenway, P., and Deaves, R., “Sensor Management Using the Decentralized
Kalman Filter,” Sensor Fusion VII, Proceedings of the SPIE—The International
Society for Optical Engineering, Vol. 2355, Boston, MA, 31 Oct –1 Nov. 2004,
pp. 216 –225.
[36] Greenway, P., and Daves, R., “An Information Filter for Decentralized Sensor
Management,” Signal Processing, Sensor Fusion, and Target Recognition III,
Proceedings of the SPIE—The International Society for Optical Engineering, Vol.
2232, Orlando, FL, 4 – 6 April 1994, pp. 70 – 78.
440 T. L. FREY ET AL.

[37] Jenison, T., “Improved Data Structures, Multiple Targets/Target Types and Multiple
Sensors in Discrimination Based Sensor Management,” master’s thesis, University of
Minnesota, July 1996.
[38] Hintz, K., “A Measure of the Information Gain Attributable to Cueing,” IEEE
Transactions on Systems, Man, and Cybernetics, Vol. 21, No. 2, March/April 1991,
pp. 237 –244.
[39] McIntyre, G. A., and Hintz, K., “Sensor Management Simulation and Comparative
Study,” Signal Processing, Sensor Fusion, and Target Recognition VI, Proceedings of
the SPIE—The International Society for Optical Engineering, Vol. 3068, Orlando, FL,
20 – 25 April 1997.
[40] Avasarala, V., Mullen, T., and Hall, D., “A Market-Based Approach to Sensor
Management,” Journal of Advances in Information Fusion, Vol. 4, No. 1, June 2009,
pp. 52 –53.
[41] Endsley, M. R., “Designing for Situation Awareness in Complex Systems,”
Proceeding of the Second International Workshop on Symbiosis of Humans, Artifacts
and Environment, Kyoto, Japan, 2001.
[42] Frey, T., “A Constrained, Information Needs Based Sensor Fusion Resource
Management Design Methodology,” Proceedings of AIAA Infotech@Aerospace
Conference, St. Louis, MO, March 2011.
[43] Wadley, J., “Important Considerations for Cooperative Passive Ranging of Aircraft,”
Proceedings of the AIAA Infotech@Aerospace Conference, Jan. 2018.
 CHAPTER 13

F-35 Carrier Suitability Testing

Tony Wilson
Lockheed Martin Aeronautics Company, Fort Worth, TX

Carrier suitability is a multidiscipline specialization of aircraft test

and evaluation. The discipline combines theories from aircraft loads,
flying qualities, and performance in a system of systems approach to
assess the suitability of an aircraft to operate from ships and austere
sites. Additionally, navigation and guidance, sensor integration, data
link interoperability, pilot – vehicle interface, supportability, and
maintainability are evaluated to ensure the aircraft is capable of
operating as a system within a system. This paper will provide an
introduction to carrier suitability flight test of the F-35C, a
carrier-based, multirole, 5th-Generation stealth fighter to be used by the
U.S. Navy and U.S. Marine Corps. Topics will include a discussion of
the shore-based testing prerequisites and results, an overview of the
challenges experienced with the original tailhook and its redesign, and
the use of autopilot functionality in shore-based testing. This will be
followed by an examination of the necessity to test in the shipboard
environment, a discussion of shipboard testing, shipboard catapult and
arrested landing methods of test and results. Throughout, an
investigation on how advanced approach mode control laws used for
shipboard landings were implemented, the results from three ship trials,
and the implications to future operations will be analyzed.

I. NOMENCLATURE
a angle of attack
d off-center distance
H/E hook to pilot eye distance
i initial condition
L pilot eye to hook point length

 Test Pilot Senior Staff, Fighter Flight Operations.

441
442 T. WILSON

PT total probability
tx time increments
VA approach speed
VE engaging speed
VV sink speed
d difference between flight path angle and glideslope angle
g flight path angle
F pilot eye to fuselage reference line angle
w roll angle
c yaw angle
Q glideslope angle
u pitch angle

II. INTRODUCTION
The idea of naval aviation, the launching and recovering of aircraft from a sea-
going vessel, is neither new nor novel; its development started less than a
decade after the Wright brothers’ first flight. U.S. naval aviation can trace its
inception back to 14 November 1910, when Eugene Ely took off in a Curtiss
Pusher from a temporary platform erected over the bow of the light cruiser
USS Birmingham. Upon rolling off the deck, the airplane sank 83 ft downward
and did not start to climb until the aircraft’s wheels had dipped into the water.
Just two months later, Ely was at it again, this time landing his aircraft on a
120-ft platform attached to the deck of the USS Pennsylvania, an armored
cruiser anchored in San Francisco Bay. A series of ropes attached to sandbags
were stretched across the landing platform designed to be snatched by a hook
attached to the aircraft and stop it prior to rolling into a canvas awning attached
to the ship [1]. With these historic events, Eugene Ely not only proved the concept
of naval aviation, but also started the discipline of ship suitability flight test.
Ship suitability flight test is a multidiscipline approach to aircraft test and
evaluation used to determine how well an aircraft system either integrates
within a ship system or operates from an austere site. Multiple nations operate
varied aircraft (fixed wing, rotary wing, and powered lift aircraft) from the
decks of military and civilian ships. To try to address all combinations would
be a herculean effort. Therefore, this paper will provide an introduction to ship
suitability flight test of the F-35C, shown in Fig. 1, a carrier-based, multirole,
5th-Generation stealth fighter to be used by the U.S. Navy and U.S. Marine
Corps both shore-based and aboard U.S. aircraft carriers (CVN). (This
acronym is the U.S. Navy’s designation for nuclear-powered aircraft carriers.
The C designates the ship as a carrier, the V is a designation for fixed wing oper-
ations, and the N represents the nuclear power aspect of the ship).
Carrier Suitability (CVS), a subset of the ship suitability discipline, is com-
posed of both shore-based testing and ship-based trials during which a
F-35 CARRIER SUITABILITY TESTING 443

Fig. 1 The author conducting the first shipboard landing of the F-35C.

system-of-systems approach is used to assess the suitability of an aircraft to

operate from the deck of an aircraft carrier. This assessment begins ashore with
an assessment of deck handling evaluations, catapult spotting, steam ingestion,
Jet Blast Deflector (JBD) compatibility, approach speed survey, Powered
Approach (PA) throttle transients, bolter and wave-off performance, approach
handling qualities, and a structural survey or demonstration. The difference
between a survey and a demonstration is the degree to which envelope exploration
or expansion is required: a demonstration requires a higher degree of expansion
whereas a survey is used to validate models and predictions. For the F-35C, a
structural survey was completed; this is a series of shore-based catapults and
arrestments designed to test the structural integrity of the landing gear, arresting
hook, aircraft structure, weapons, and weapon suspension equipment.
Shore-based testing of the F-35C commenced in the spring of 2011. It was
quickly realized that the original tailhook design would need alterations due to geo-
metric constraints. This resulted in a prolonged pause in testing while the tailhook
was redesigned. The new tailhook was installed in December 2013 and testing
recommenced. The delay in testing acted as a forcing function to develop new tech-
niques to increase testing efficiency. Prior to the F-35, structural survey test points
were flown mostly, if not completely, with the pilot in the loop, making test point
444 T. WILSON

completion highly dependent upon pilot experience and technique, resulting in

multiple attempts at a single test point. The advanced control laws of the F-35C
were used in a manner never attempted before to attain test point parameters
and improve repeatability. This resulted in a rapid completion of the clean wing
aircraft structural survey, a prerequisite for shipboard testing, and paved the way
for F-35C ship trials. Benefits to future carrier aviation programs will be realized
through the examination of how the advanced control laws were used, the advan-
tages and disadvantages to each mode, and how F-35C flight test benefited from
this breakthrough.
Once shore-based prerequisites are complete, ship trials commence. Ship-
board testing is a necessity due to the ship’s flight deck being approximately 60
ft above the ocean’s surface. This elevation results in the aircraft rotating about
its Center of Gravity (CG) as it leaves the deck during a catapult launch vs rotating
about its main gear during a conventional takeoff. Additionally, the wind interacts
with the structure of the ship and produces disturbances and perturbations both
behind and in front of the ship that cannot be reproduced or simulated on shore.
The F-35C conducted its initial ship trial (DT-I) in November 2014, aboard USS
Nimitz (CVN 68) where initial evaluation of the clean wing aircraft catapult
launch and recoveries began. DT-I consisted of 124 catapult launches, 124
arrested landings, and 222 Touch-and-Go (T&G) landings with two test aircraft
during day and night Visual Meteorological Conditions (VMC). DT-II was com-
pleted in October 2015 aboard USS Dwight D. Eisenhower (CVN 69), where
increased weight and forward CG were evaluated with 66 catapult launches, 66
arrested landings, and 40 T&G landings during day and night VMC. The third
ship trial (DT-III) was conducted aboard USS George Washington (CVN 73) in
August 2016, consisting of 121 catapult launches, 121 arrested landing, and 67
T&G in which shipboard operations with full internal weapons bays and external
weapons were evaluated. With respect to catapult testing, the largest technical
challenge and most dangerous portion of the evolution is Minimum End Airspeed
(MEAS) catapults. During these particular launches, the fly-away airspeed of the
aircraft is slowly decreased, intentionally allowing the aircraft to sink below deck
level to a targeted height above the sea surface, allowing the minimum safe
fly-away airspeed to be anchored for a given aircraft gross weight. This section
will provide a detailed discussion on the theory of MEAS testing and other cata-
pult launches, the results of the shipboard evaluation, and lessons learned while
deployed during the three ship trials.
Once an aircraft is launched, it must be recovered via an arrested landing.
Almost 600 approaches were made during ship trials to various aircraft carriers
in a multitude of wind conditions ranging from minimum recovery headwinds
to winds in excess of 40 kt. During these approaches, three different approach
modes were tested. The basic mode, also known as manual, is flown on the backside
of the power curve in which the pilot controls rate of descent with thrust and air-
speed with aircraft attitude. The next mode is Approach Power Compensation
(APC), in which the pilot commands the flight path rate via longitudinal control
F-35 CARRIER SUITABILITY TESTING 445

stick inputs while the control laws contained in the vehicle management computers
maintain Angle of Attack (AOA) via engine thrust request and throttle backdrive.
The final mode, Delta Flight Path (DFP), allows precision control of flight path
(gamma) by integrating APC with Integrated Direct Lift Control (IDLC). These
advanced approach modes are synthesized via the nonlinear dynamic inversion
control paradigm that is shared will all other F-35 variants. The remainder of the
paper will focus on the design and implementation of these advanced control
modes, the methods used to evaluate the approach handling qualities in the
various wind cells using all three approach modes, and the results of these tests.
In addition to catapults and arrested landings, several other areas were evalu-
ated including navigation and guidance, sensor integration, data link interoper-
ability, pilot –vehicle interface, supportability, and maintainability—all of which
will be touched upon in this paper. In the end, the reader will learn about the the-
ories and applications of the multidisciplinary field of carrier suitability flight test,
how they were applied to the F-35C, and the implications to future operations and
test programs.

III. BACKGROUND
This section is intended to provide the reader the necessary information to under-
stand shore-based and ship-based carrier suitability testing. It will begin with an
overview of the F-35C with a focus on the Control Laws (CLAW) and Arresting
Hook System (AHS). This will be followed by a description of the landing aids,
arresting gear, and catapult. The background section will be concluded with a
brief discussion of the pilot’s tasks while working with a Landing Signal Officer
(LSO). If the reader is already familiar with carrier suitability, this section may
be skipped or serve as a refresher on the basics.

A. F-35C LIGHTNING II
The F-35C, shown in Fig. 2, is a carrier based, multirole, 5th-Generation stealth
fighter intended for use by the U.S. Navy as a replacement for its aging
4th-Generation fighters. The F-35 is a single-seat, single-engine, all-weather
fighter with low observable features designed for ground attack and air superiority
missions. It is powered by the Pratt & Whitney F135 engine that generates
approximately 40,000 lb of thrust. The F-35C shares its basic structure and
primary load paths with the F-35A and F-35B variants but incorporates special-
ized design features and strengthened internal structure to perform catapult
launches and arrested landings. The F-35C was designed to be compatible with
and operationally supported from the Nimitz class (CVN 68) aircraft carrier.
The F-35C design incorporates 45% more wing area, larger horizontal stabi-
lators, increased control surface size, and outer wing ailerons, which contribute
to precise slow-speed handling characteristics required for carrier approaches
446 T. WILSON

Fig. 2 Unique features of the F-35C Lightning II.

and landings. Unique features to the F-35C include folding wings to reduce the
amount of deck space needed while aboard the carrier, a launch bar for catapult
operations, and a recessed stowable tailhook for shipboard arrestments. The air-
craft has 11 weapons stations, 4 internal and 7 external. The internal weapon bay
stations are designed for a combination of Advanced Medium-Range Air-to-Air
Missiles (AMRAAMs) and two Air-to-Surface (A/S) weapons up to 2000 lb
each. The F-35C has three weapons stations under each wing, allowing it to
carry a combination of Air-to-Air (A/A) and A/S weapons externally. The center-
line fuselage station is designed for carriage of a low-observable gun pod. Table 1
depicts all the configurations that were tested shipboard [2].

B. F-35C CONTROL LAWS

The F-35 uses nonlinear dynamic inversion control architecture that is operated
in three redundant Vehicle Management Computers (VMCs) to provide
fly-by-wire control. The CLAW that augment aircraft dynamics and provide stab-
ility are common among the three variants; however, each variant’s CLAW has
been optimized and contains advanced features to reduce pilot workload and
 TABLE 1 SHIPBOARD FLIGHT TEST WEAPONS CONFIGURATIONS

Loading Station
1 2 3 4 5 6 7 8 9 10 11
F-35 CARRIER SUITABILITY TESTING

Clean Wing DART-1k Suspension Suspension Suspension

Empty Bays Equip Equip Equip
Clean Wing DART-1k AIM-120 AIM-120 GBU-31
(Most Fwd CG)
Pylons Only LAU Pylon Pylon DART-1k Suspension Suspension Suspension Pylon Pylon LAU
Equip Equip Equip
Heavy Symmetric LAU GBU-12 GBU-12 DART-1k AIM-120 GUN AIM-120 GBU-31 GBU-12 GBU-12 LAU
Heavy Symmetric LAU GBU-12 GBU-12 DART-1k Suspension Suspension GBU-31 GBU-12 GBU-12 LAU
No Gun Equip Equip
Symmetric (Most LAU GBU-12 GBU-12 DART-1k Suspension Suspension Suspension GBU-12 GBU-12 LAU
Aft CG) Equip Equip Equip
Light Symmetric LAU GBU-12 Pylon DART-1k Suspension Suspension Suspension Pylon GBU-12 LAU
Equip Equip Equip
Asymmetric LAU GBU-12 Pylon DART-1k Suspension Suspension GBU-12 Pylon Pylon LAU
(10,000 ft/lb) Equip Equip
447
448 T. WILSON

provide carefree handling for specific tasks (e.g., hovering in an F-35B). For the
F-35C, these optimizations include, but are not limited to, shipboard launches
and recoveries.
A shipboard catapult launch can be divided into three distinct phases: accel-
eration, rotation, and flyaway. The launch starts with an acceleration. The air-
craft is attached to a shuttle on the deck by a launch bar on the nose strut.
The shuttle, attached to a steam-powered piston under the deck, provides the
motive force to accelerate the aircraft to a safe flyaway speed. During this
phase, the CLAW pre-positions the control surfaces to provide aircraft rotation
upon shuttle release. The horizontal tail is scheduled to achieve roughly
12.5 deg/second of pitch rate at operationally representative aircraft weights,
Wind Over the Deck (WOD), and energy imparted on the aircraft by the cata-
pult. Next, the launch bar is released from the shuttle at the end of the catapult
stroke; stored energy in the nose strut and aerodynamic moments generated by
pro-rotation control surfaces (symmetric horizontal tail trim, symmetric ailer-
ons, and toe-in [trailing edge inboard] rudders) combine to provide a smooth
and quick aircraft rotation. Finally, during the flyaway phase, the CLAW tran-
sitions from pro-rotation to pro-lift to optimize lift and minimize sink once air-
borne. The pro-lift control surfaces are symmetric Trailing Edge Flaps (TEFs)
and symmetric ailerons.
The catapult launch control law logic is evoked when the launch bar is com-
manded down while the aircraft is on deck, positioning the control surfaces for
pro-rotation based on gross weight and the aircraft’s CG. The pro-rotation con-
figuration is designed to achieve a positive flight path as quickly as possible follow-
ing catapult launch in order to minimize sink. Additionally, catapult launch logic
causes a 30-lb throttle detent to engage at the 100 Engine Thrust Request (ETR)
MIL gate, Nose Wheel Steering (NWS) is disengaged, and engine AB Limit
(ABLIM) is activated. The ABLIM feature is designed to reduce thermal heating
of the tail boom, empennage structure, JBD panels, and flight deck. If throttle pos-
ition is set for ETR . 122% during the catapult launch sequence, ABLIM control
logic will automatically limit ETR to 122% power prior to launch. When a catapult
launch is declared or longitudinal acceleration (Nx) . 0.5 g, the ABLIM feature
will clear, allowing thrust to increase up to the commanded power setting
during aircraft acceleration down the catapult track. The Flight Control System
(FLCS) uses multiple sensor to provide feedback to the CLAW to declare a catapult
launch:
. ETR  95% and longitudinal acceleration (Nx) . 2 g or calibrated airspeed
(VCAS) . 100 kt or
. ETR , 95%, wheel speed . 20 kt, and Nx . 2 g or VCAS . 100 kt

As the launch bar leaves the shuttle, the pro-rotation control surfaces begin to be
removed as Angle of Attack (AOA) or pitch rate increase to protect from over-
rotation. The control system commands 20 deg AOA initially but backs off to
F-35 CARRIER SUITABILITY TESTING 449

15 deg as AOA passes through 14 deg AOA. This command is then blended back
to zero once a 1-deg flight path (gamma) is achieved. If AOA and pitch rate do not
increase prior to the aircraft rolling off the deck edge, the pro-rotation aids are
removed quickly once weight-off-wheels is sensed and replaced with pro-lift
aids. Once airborne, catapult launch mode is disabled. The control law reschedules
surfaces from pro-rotation to pro-lift positions in order to minimize sink and to
enhance fly-away characteristics.
The CLAW also augments the roll axis during catapult launches. Roll rates
and bank angles induced by asymmetries and crosswinds at the deck edge are
removed via a launch roll trim mode that commands differential flaps and ailerons
during a catapult launch. Launch roll trim activates when VCAS exceeds 100 kt
and is reset as the aircraft leaves the deck. Additionally, a bank assist was
added to hold wings level on flyaway. Bank assist activates at weight off wheels
(WoffW) and attempts to maintain a wings-level attitude (0 deg bank angle)
during flyaway for 10 seconds. The pilot may override bank assist at any time
with a lateral stick input.
The F-35C has three approach modes: manual, Approach Power Compen-
sation (APC), and DFP. All three modes are designed to maintain a precise
AOA throughout the approach, providing the pilot fine control of touchdown
location on the deck at the optimum pitch angle to facilitate a successful arrest-
ment. Manual mode allows the pilot to fly a traditional backside approach to a
ship, using thrust via the throttle to control rate of descent and longitudinal
stick inputs to control airspeed. APC automates engine response and AOA
such that the pilot controls flight path, and thus rate of descent, with longitudinal
stick inputs. DFP uses APC logic and automates flight path maintenance by cap-
turing and maintaining a designated flight path.
All approach modes use IDLC, a pilot command TEF/aileron interconnect
that provides improved wave-off, bolter, and approach flying qualities by enhan-
cing the flight path control of the nominal PA control system. This dynamic sym-
metric deflection command is added to the nominal PA schedules. Pilot input is
passed through a first-order high pass filter to create a dynamic IDLC command.
This dynamic command deflects the symmetric TEFs/ailerons for rapid direct lift
capability to offset the slow response of the engine and provide a more immediate
flight path response. IDLC is only used while in the AOA portion of the PA system
where the aircraft is designed to be “back-sided” (i.e., throttle changes control rate
of descent). IDLC response is initiated via the throttle or longitudinal stick inputs
when in manual or APC, respectively; however, TEF/aileron command gains are
increased 300% in APC over that of manual.
Delta Flight Path engages APC and commands an automated gamma capture
and maintenance. As perturbations move the aircraft away from the desired gli-
deslope, the pilot inputs and holds a longitudinal stick displacement to generate
a change in gamma. As the aircraft approaches the desired glideslope, the pilot
leads the correction and releases the stick, thus commanding the aircraft to recap-
ture the programmed gamma.
450 T. WILSON

C. F-35C ARRESTING HOOK SYSTEM

The Arresting Hook System (AHS) consists of a tailhook, hydraulic actuators and
dampers, and a set of clamshell doors to conceal the tailhook in a bay for Low
Observability (LO) considerations. It is designed to achieve rapid deceleration
during shipboard landings by engaging an arresting wire on the deck of an aircraft
carrier, or during shore-based aborted takeoffs or emergency landings. The arrest-
ing hook is electrically controlled, hydraulically actuated, and attached to the air-
craft structure via a pitch pivot pin. Figure 3 shows the redesigned arresting hook,
and Fig. 4 shows the AHS components.
Testing on the original AHS began in 2011, and it was quickly realized that the
AHS was not functioning per design: the tailhook engagement rate (number of
times the hook engaged and maintained contact with the arresting wire) was sig-
nificantly low and the hook shank did not have enough load carrying capacity. An
engineering investigation identified three root causes for the low engagement rate:

1. Aircraft geometry: The distance between the tailhook point and the main
landing gear is short, 7.1 ft, as shown in Fig. 5 [3]. The short distance was
the result of geometric constraints imposed by low observability require-
ments. To reduce radar cross-section, the AHS is contained in a recessed
bay that is covered with
doors similar to the
weapons bay. The size of
the bay, limited by aircraft
size, dictates the allowable
length of the AHS.
2. Tailhook point design: The
hook point’s geometry was
too rounded, thus lowering
its scooping efficiency, as
shown in Fig. 6 [3].
3. Tailhook hold-down per-
formance: The hold-down
damper provided an insuffi-
cient hold-down force, thus
allowing excessive tailhook
bounce [3].

These three factors culmi-

nated in the main landing gear
tires rolling over the arresting

Fig. 3 Redesigned F-35C

arresting hook.
F-35 CARRIER SUITABILITY TESTING 451

Fig. 4 F-35C arresting hook components.

wire, initiating a transverse or “kink” wave. This wave propagated away from the
point of tire rollover, both outwards and inwards. The wave propagation speed
and aircraft forward velocity were such that the arresting wire was lying flat on
the ground as the rounded tailhook point arrived at the wire, causing the hook
to skip over the wire. These dynamics were further exacerbated by the low hold-
down force. When the hook point initially contacted the deck or was perturbed by
an imperfection on the landing surface, the hook would bounce [4], leaving the
surface and either glancing off or missing the arresting wire completely. These
dynamics can be seen in the series of photos shown in Figs. 7 –10.
In 2013, the redesigned AHS was completed, delivered, installed, and tested. The
new design offered several improvements. First, the hook point geometry was
altered from a rounded design to a wedge design (Fig. 11), ensuring that the apex
of the wedge was lower than the centroid of the arresting wire, as shown in Fig. 12.
Additional improvements included increased structure to the y-frame and
hook shank, thereby increasing the load carrying capacity; the hold-down

Fig. 5 U.S. Navy aircraft main landing gear to hook point geometry comparison.
452 T. WILSON

Fig. 6 Original hook point with rounded geometry.

damper size also was increased to increase the hold-down force and reduce hook
bounce. Furthermore, an end-of-stroke damper and lateral limiter were installed.
The end-of-stroke damper was added to compensate for the increased down force
of the hold-down damper. When the aircraft is on the deck and the hook is com-
manded down, the hook point will strike the deck prior to the hook shank achiev-
ing full range of motion; however, when airborne, the hook shank will travel its
fully allowable arclength, striking the end-of-stroke damper. The end-of-stroke
damper absorbs the shock downwards force imparted on the hook by the hold-
down damper, preventing overload conditions. The lateral limiter acts to
prevent excessive movement of the hook shank. It is attached to the hinge
point where the shank meets the y-frame. If the aircraft rolls or touches down
in a bank angle and imparts an angular velocity on the hook shank, the lateral
limiter acts as a bumper, absorbing the energy. If too much energy is imparted,

Fig. 7 During roll-in testing, nosewheel initiates kink wave.

F-35 CARRIER SUITABILITY TESTING 453

Fig. 8 Kink wave propagates away from the source of excitation.

the elastic deformation region of the material is exceeded, resulting in plastic

deformation of the lateral limiter, providing a telltale sign to maintainers and
pilots that an overload condition has been prevented. The resulting maintenance
requires the removal and replacement of the lateral limiter, but provides a signifi-
cant savings over damaging the hook shank.
The result of the redesigned AHS was tested for the first time in December
2013, where it successfully engaged the arresting wire on the first attempt, as

Fig. 9 Tailhook arrives at arresting wire flush with ground.

454 T. WILSON

Fig. 10 Tailhook bounces over flush arresting wire.

seen in Fig. 13. Subsequent flight tests have provided a statistically significant
sample size to validate the effectiveness of the F-35C’s AHS.

D. NIMITZ-CLASS AIRCRAFT CARRIER

Testing was conducted aboard USS Nimitz (CVN 68), USS Dwight D. Eisenhower
(CVN 69), and USS George Washington (CVN 73), all Nimitz-class aircraft car-
riers. Nimitz-class aircraft carriers are 1092 ft long, 135 ft wide at the waterline,
252 ft wide at the widest point on the flight deck, have a 37-ft draft, and displace
approximately 100,000 long tons fully loaded. Each Nimitz-class ship has an
angled flight deck, four C-13 steam catapults, four Mk 7 arresting gear (three
Mk 7 arresting gear CVN 76 and later), an IFLOLS, drop lights, laser line-up
system, flight deck lighting, a large below-decks aircraft hangar, and four aircraft
elevators. Each ship has an approximately 788-ft angled deck (9 deg to port from

Fig. 11 Wedge hook point geometry.

F-35 CARRIER SUITABILITY TESTING 455

Fig. 12 Geometric comparison of rounded to wedge hook point.

centerline) to facilitate simultaneous launch and recovery of aircraft. Aircraft are

lowered to and raised from the hangar deck via the four elevators: three located on
the starboard side and the other located to port near the stern. A flight deck
diagram is presented in Fig. 14.

E. CATAPULT
The catapult provides the motive force for shipboard launches of aircraft. The
main components of a catapult are the power cylinders, piston assemblies,
shuttle assembly, and a water brake, illustrated in Fig. 15. Steam, provided by
the ship’s boilers, enters the aft end of the cylinders through the launching
valve. This steam pressure then acts on the piston assemblies and accelerates

Fig. 13 The first successful arrested landing with the redesigned arresting hook system.
456 T. WILSON

Fig. 14 Nimitz-class aircraft carrier deck diagram with catapults in red and arresting wires
in blue.

the shuttle assembly to a safe launching velocity. The aircraft is connected to the
shuttle assembly via the launch bar attached to the aircraft’s nosewheel, shown in
Fig. 16. At the end of the catapult stroke (which is the forward end of the ship), the
tapered spears, one mounted to the front of each piston, enters the water brake
assembly, which absorbs the energy of the piston assemblies and rapidly brings
them to a stop. At this time the aircraft is released from the shuttle with sufficient
speed to become airborne.
Proper heating and temperature maintenance are a critical factor for the safe
operation of the catapult. Finned tub heat exchangers installed around the power
cylinders, also supplied with steam from the ship’s boilers, provide heat to main-
tain the cylinders at a near constant temperature. The catapults are preheated to
achieve the proper thermal expansion. Once the proper temperature is reached,
the flow of steam is varied to maintain a constant elongation of both power cylin-
ders. The maintenance of proper cylinder elongation is critical in that all launch
bulletins (documents that provide the settings for the Capacity Selector Valve
[CSV]) are based on a constant elongation, thus providing for reliable and repea-
table catapult performance between launches.
The water brake assembly provides the catapults braking capability via a 9-ft
spear on the forward end of the piston assembly. At the end of the catapult stroke,
the spear is guided into the water brake cylinder, which is kept full of water by a
centrifugal vortex induced by jets of water [5].
Two types of catapults were used during testing: a shore-based C-7 catapult
and a C-13 aboard the Nimitz-class aircraft carriers. The primary difference
between the two catapults is length. The C-7 was designed for an earlier class

Fig. 15 Simplified diagram of steam catapult components.

F-35 CARRIER SUITABILITY TESTING 457

Fig. 16 F-35C launch bar connects to shuttle forward of nose strut.

aircraft carrier, thus limiting its length to 276 ft with a 253-ft stroke. Conversely,
the C-13 has a length of 325 ft with a 310-ft stroke [6]. Due to this difference, the
jerk (time rate of change of acceleration), or jolt, is greater on the C-7 catapult.
Despite this difference, the C-7 provides operationally representative launches,
mirroring accelerations and both aircraft and catapult tow loads experienced
during a shipboard launch. A typical longitudinal acceleration and launch bar
axial loading as a function of time during a catapult launch are shown in Fig. 17.

F. VISUAL LANDING AIDS

The U.S. Navy has two shore-based visual landing aids: the Mk 8 Fresnel Lens
Optical Landing System (FLOLS) and the Mk 14 Improved Fresnel Lens
Optical Landing System (IFLOLS) to display glideslope and trend information
to pilots approaching the runway. Each system consists of a yellow source light
assembly, green datum lights, red wave-off lights, and green “cut” lights. The
lights are useable at a range of 1.0 n mile. The yellow source light is projected
through a series of linear Fresnel lenses to focus the light into narrow horizontal
bands. A lenticular lens is attached in front of the Fresnel lenses to provide the
desired source light color, increase the azimuth visibility arc to +20 deg, and dis-
perse sun reflection from the lens face. The result is a light source that appears to
the pilot as the “meatball” or a “ball” that moves vertically across the face of the
lens as the pilot’s eye moves through the beam’s arc length. The ball provides the
pilot a visual means to determine the aircraft’s position relative to the designated
glideslope. A jackscrew and hand-crank on the front of the trailer base are used to
adjust the system to achieve desired glideslope. The task of the pilot is to capture
and maintain the ball between the datum lights. If the aircraft is flying on the
desired glideslope, the pilot will see the ball in line with the green datum lights.
458 T. WILSON

Fig. 17 Launch bar axial loading and longitudinal acceleration of the F-35C as a function
of time.

If the aircraft is high, the pilot will see the amber ball above the datum arms; con-
versely, if the aircraft is below glideslope, the pilot will see the ball below the datum
lights. If the pilot proceeds too far below glideslope, the amber ball will change to a
red ball. The red wave-off lights are manually activated by the LSO when danger-
ous landing conditions exist. The green cut lights were historically used to inform
the pilot of a propeller-driven aircraft when to “cut” the engines; however, today
cut lights are used to indicate a landing clearance in the event of lost communi-
cations [7].
The Mk 8’s source box, shown in Fig. 18, contains fewer cells than the Mk 14,
shown in Fig. 19, and is consequently a shorter display—a 4-ft-tall source box vs
the 6-ft-tall source box of the Mk 14. This results in the fidelity of the Mk 14
display being approximately double that of the Mk 8, as can be seen when com-
paring Tables 2 and 3. The Mk 8 FLOLS pro-
vides optical glideslope information within
approximately +34 deg of the desired glideslope.
The source light unit is composed of five cells
(each cell is designed to illuminate 20.450 of
arc) for a total of approximately 1.7 deg of verti-
cal coverage. A subsequent consequence of the
Mk 14’s larger size is that it is transversely
mounted on its trailer as compared to the
mounting orientation of the Mk 8. Although

Fig. 18 Mk 8 FLOLS light box geometry.

F-35 CARRIER SUITABILITY TESTING 459

Fig. 19 Mk 14 IFLOLS light box geometry.

the Mk 14 offers higher fidelity with respect to

changes in glideslope, it is not easily or quickly
moved when a change in touchdown point is
desired. To conduct carrier suitability testing
efficiently, there is a need to rapidly move the
desired touchdown location. The orientation of
the Mk 8 on its trailer provides this flexibility
over the Mk 14, thus making it the lens of
choice for carrier suitability testing.
Mk 13 Mod 0, shown in Fig. 20, is used for
shipboard operations. It provides glideslope
and trend information to the pilot in the same exact manner as the Mk 14 shore-
based model. It is set to a base angle of 3.5 deg to ensure a Hook-to-Ramp (H/R)
distance of at least 14 ft, as shown in Fig. 21. This will be discussed in more detail
in the “Glideslope Geometry” section. The largest difference between the two is
that the Mk 13 is dynamically stabilized to compensate for ship’s motion in
pitch, roll, and heave [5].

G. MK 7 ARRESTING ENGINE
The Mk 7 constant runout arresting gear, illustrated in Fig. 22, is a hydro-
pneumatic system composed of the engine structure, a cylinder and ram assembly,
a crosshead and fixed sheaves, a control valve system, an accumulator system, air
flasks, and a sheave and cable arrangement. An Mk 7 arresting engine is installed
in the approach end of runway 32 at Naval Air Station Patuxent River, Maryland.
The shore-based Mk 7 site is similar to shipboard arresting gear and considered
operationally representative.

TABLE 2 MK 8 FLOLS GEOMETRY BY CELL

Cell Basic Angle (deg) Hook-to-Ramp (ft)

Top 4.4 23.3
1 4.2 20.0
2 3.8 17.7
3 3.5 14.0
4 3.2 10.3
5 2.8 6.6
Bottom 2.7 4.7
460 T. WILSON

TABLE 3 MK 14 FLOLS GEOMETRY BY CELL

Cell Basic Angle (deg) Hook-to-Ramp (ft)

Top 4.3 23.6
1 4.2 22.8
2 4.1 21.2
3 4.0 19.6
4 3.8 18.0
5 3.7 16.4
6 3.6 14.8
Center 3.5 14.0
7 3.4 13.2
8 3.3 11.6
9 3.2 10.0
10 3.0 8.4
11 2.9 7.2
12 2.7 4.7
Bottom 2.6 3.4

During an arrested landing, the Mk 7 brings the aircraft to a smooth, controlled

stop in the following manner: The aircraft’s arresting hook engages the cross-deck
pendant (also called the “wire” or arresting cable) spanning the landing area. The
cross-deck pendant is a 1-3/8-in.-diameter cable with a polyester core, weighing
approximately 3 lb/ft and having a minimum static breaking strength of 205,000
lb [5]. The forces generated by the aircraft’s forward motion are transferred via
the cross-deck pendant to purchase cables that are reeved around a movable cross-
head of sheaves and the fixed sheave assembly of the arresting engine. The movable
crosshead is moved toward the fixed sheave assembly as the aircraft pulls the
purchase cables off the arresting engine, forcing a ram into the cylinder holding
pressurized hydraulic fluid (ethylene glycol). The flow of fluid is controlled by
the constant runout control valve, which in turn controls the pressure in the cylin-
der thereby providing the restraining force to absorb the aircraft’s energy.

H. LANDING SIGNAL OFFICER

The Landing Signal Officer, pictured in Fig. 23, is a pivotal member of the test
team. The LSO is a pilot who, through training and experience, is capable of
quickly correlating the effects of wind, weather, aircraft capabilities, and pilot
F-35 CARRIER SUITABILITY TESTING 461

Fig. 20 Shipboard Mk 13 IFLOLS.

Fig. 21 Glideslope geometry relationships: hook-to-ramp, hook-to-eye, hook touchdown

point, and IFLOLS glideslope projection.
462 T. WILSON

Fig. 22 Simplified diagram of arresting gear components.

performance to provide optimum control and assistance in achieving the test

point. The LSO’s primary responsibility is safety. Additional duties include
serving in a capacity similar to that of a test conductor, directing the testing
efforts of the day. The test pilot and LSO form a professional and disciplined
team founded on mutual respect and the highest degree of trust.
Shore-based, the LSO is stationed in a mobile structure on the side of the
runway abeam the intended point of landing and is able to observe the test aircraft
throughout the landing pattern. Ship-based, the LSO and team are stationed on
the port (left) side of the flight deck, approximately 50 ft from the aft end of

Fig. 23 The Landing Signal Officer (LSO).

F-35 CARRIER SUITABILITY TESTING 463

the ship. From the moment of the “ball call,” signifying that the light source on the
Fresnel lens is visible, to the arrestment, touch and go, or wave-off, the pilot is
under the control of the LSO. Analyzing trends in both pilot and aircraft perform-
ance, assessing external environmental factors, and having a thorough under-
standing of what techniques the pilot is using, the LSO transmits informative
advisory and directive calls to aid the pilot in safely achieving the test point.

I. PILOT’S TASK
A test pilot flying a structural survey arrested landing has a different task than that
of the fleet aviator executing a carrier-style landing. The fleet aviator is expected to
capture and maintain glideslope, centerline, and angle of attack from the start of
the approach all the way to touchdown. Although there are tolerances for devi-
ations, if the appropriate correction is not made in a timely manner, the fleet
aviator will be waved off. However, a carrier suitability structural survey is a
specialized loads test, thus in addition to glideslope, line up, and angle of
attack, the test pilot has the additional tasks, depending on the specific test
point, of capturing and maintaining a ground speed; setting a pitch, roll, and/
or yaw attitude prior to touchdown; or attaining a target sink speed. Another
difference is hook touchdown location. The fleet aviator, while expected to main-
tain a centered ball, has an acceptable touchdown window of 120 ft whereas the
test pilot’s acceptable touchdown window varies from 40 ft to less than a foot,
again depending on the test point.
The base glideslope is a 3.25-deg (above the horizon) fixed path determined by
the angle of the Fresnel lens. The rate of descent necessary to maintain this glide-
slope depends on the aircraft ground speed (and therefore changes slightly with
wind conditions). Proper execution of the approach requires an accurate, rapid
scan to allow the pilot to identify deviations and make corrections as soon as
one is required; if an appropriate correction is not made, the deviation will only
increase in magnitude. This is due to the constant arc length of the light the
Fresnel lens is displaying. As the pilot approaches the light source, the lens fidelity
increases, resulting in ball movement increasing at an exponential rate; for
example, if the pilot captures the desired 3.25-deg glideslope but is one ball
high and does not center the light source between the datum, the pilot will initially
see the ball moving upward at a slow rate. As the pilot’s eye approaches the source,
the rate of ball movement increases exponentially, resulting in the ball appearing
to shoot off the top of the lens.
Lineup for shipboard landings is critical; the relatively small size of the landing
area makes it imperative that the pilot lands on the centerline with no drift. What
adds to the difficulty of shipboard lineup is that the landing area is misaligned
with the ship’s centerline by 9 deg, resulting in the landing area having a sideward
velocity or drift. Conversely, shore-based testing is completed to a runway with no
drift; however, lineup is no less critical for the test pilot. For shore-based testing, an
extended landing pattern is used with the pilot capturing runway centerline
464 T. WILSON

somewhere between 2.0 and 1.5 n miles from touchdown, thus requiring the pilot to
perform the centerline maintenance task for up to one minute. Depending on the
wind gradient and magnitude of gust, this task can be challenging, requiring the
pilot to identify and correct lineup deviations as early as possible. Failure to make
lineup corrections in a timely manner will typically cause the test pilot’s scan to
break down. This becomes evident to both the pilot and LSO with a manifestation
of a deviation in glideslope or another critical test parameter. Lineup deviations as
little as 3 ft from centerline can have a profound impact on hook loading. Addition-
ally, one of the test points that will be discussed later requires the test pilot to engage
the cross-deck pendant within 2 ft. The test point calls for an off-center engagement
of at least 18 ft from centerline; however, if the pilot exceeds 20 ft, the arresting
engine requires an immediate inspection because damage may have occurred.
For both the fleet aviator and the test pilot, lineup deviations that are not corrected
in a timely manner will typically cause scan breakdown, resulting in deviations in
other parameters. Shipboard, a camera is mounted on the centerline flush with
the deck facing aft. This image is monitored by the LSO and allows him or her to
see if the pilot is failing to maintain the landing area centerline. Unfortunately,
the shore-based LSO does not have this capability. This results in the pilot having
to assess crab angle and make lateral and/or directional inputs just prior to touch-
down to precisely place the hook point within the lateral tolerances of the test point.
Historically, the pilot was also required to maintain the aircraft within a specific
AOA, or alpha, range. The range is specific to each aircraft and is determined by
two factors: speed and geometry. The lower end of the acceptable alpha range is
determined by maximum arresting gear engagement speed. If the aircraft’s velocity
is too high (low AOA), the aircraft structure may yield or the arresting gear engine
may be damaged. Conversely, if the aircraft is too slow (high AOA), the airspeed
margin above aerodynamic stall is unacceptable or the aircraft’s tailhook point is
sufficiently lowered that the threat of an in-flight engagement exist. Due to the care-
free handling offered by the F-35’s FLCS, the CLAW attempts to maintain AOA
tightly within the acceptable range, removing the pilot from controlling AOA
directly; although it is important for the pilot to mind the alpha and avoid large
AOA transients, this is now a secondary concern for the pilot.

J. GLIDESLOPE GEOMETRY
The task of executing carrier suitability testing, be it on land or sea, requires the
capability to precisely have the hook point touchdown on the targeted position. To
facilitate this, glideslope geometry, shipboard landing dynamics, and the environ-
mental effects must be thoroughly understood. The following are definitions of
commonly used terms with respect to glideslope geometry also illustrated in
Figs. 24 and 25:
. Natural wind: Existing surface wind caused by the moving air mass indepen-
dent of ship’s motion
F-35 CARRIER SUITABILITY TESTING 465

Fig. 24 Vector relationships of shipboard winds.

. Ship’s wind: Wind created by ship’s velocity, equal in magnitude and opposite
in direction
. Relative wind or Wind Over Deck (WOD): The vector sum of natural wind and
ship’s wind
. Recovery Headwind (RHW): That component of WOD parallel to the center-
line of the landing area
. Recovery crosswind: That component of WOD perpendicular to the centerline
of the landing area
. Angle of attack (a): Angle between the aircraft Fuselage Reference Line (FRL)
and flight path
. Flight path (g): Aircraft’s velocity vector through the air mass measured from
the FRL
. Glideslope angle (Q): Angle between the optical landing aid light source and
the horizon

Fig. 25 Glideslope geometry.

466 T. WILSON

. Pilot eye to FRL angle (F): Angle between FRL and a line connecting the pilot’s
eye to the hook point
. Pilot eye to hook point length (L): Straight line distance between the pilot’s eye
and the hook point
The Hook-to-Eye (H/E) distance of an aircraft is of critical importance to place
the aircraft touchdown at the proper point in the landing zone of the aircraft
carrier. The H/E distance needs to be precisely determined; any error will
result in either an early touchdown that reduces Hook-to-Ramp (H/R) clearance
or a late touchdown that increases bolter rates. A stadiametric view of glideslope
geometry is presented in Fig. 25 [6].
Applying the law of sines to triangle ABC,
. A ¼ ðQ  g Þ þ a þ F (1)

. C ¼ 90  Q (2)
H=E L
¼ (3)
sin A sin C
L sin ((Q  g) þ a þ F)
H=E ¼ (4)
sin (90  Q)
sin (90  Q) ¼ cos Q (5)
L sin ((Q  g) þ a þ F)
H=E ¼ (6)
cos Q
Small angle theorem, cos Q ¼ 1; therefore,
H=E ¼ L sin ((Q  g) þ a þ F) (7)
To determine the difference between the glideslope angle and the flight path
angle, the relative velocity between the aircraft’s true airspeed and the landing
point’s velocity relationship is evaluated. The landing point’s relative velocity is
a function of ship’s speed and natural winds, as shown in Fig. 26.
d¼Qg (8)
H=E ¼ L sin (d þ a þ F) (9)
Applying the law of sines to Fig. 26,
RHW TAS
¼ (10)
sin d sin (180  Q)
sin (180  Q) ¼ sin Q (11)
sin QRHW
sin d ¼ (12)
TAS
F-35 CARRIER SUITABILITY TESTING 467

Fig. 26 Vector relationship of true airspeed, recovery head wind, and glideslope angle.

Small angle theorem, sin Q ¼ Q; therefore,

QRHW
d¼ (13)
TAS
It is seen that the largest value of d will be obtained with increasing glideslope
angle and RHW and decreasing approach airspeed. Another way of looking at
Eq. (13) is to manipulate glideslope angle based on changes in RHW to maintain
d near constant for pilot considerations. To accomplish this, the shipboard glide-
slope is routinely increased during high recovery headwind conditions.
The flight path angle can be determined by
QRHW
g¼Qd¼Q (14)
TAS
 
RHW
g¼Q 1 (15)
TAS
An alternative way to observe effective glideslope is presented in Fig. 27. At
time one t1, the pilot’s eye is on the optical glideslope. A few moments later, as

Fig. 27 Effect of the ship’s forward motion on the effective glideslope.

468 T. WILSON

the pilot continues to approach the ship, the ship’s position is displaced by its vel-
ocity through the water. This dynamic continues all the way to touchdown at t3.
Under ideal operational conditions, the glideslope is set to 3.5 deg, but the pilot
maintains a 3-deg flight path, the effective glideslope.

IV. SHORE-BASED TESTING

Shore-based carrier suitability testing of the F-35C started on 17 March 2011. His-
torically, new seagoing aircraft were required to demonstrate compliance with
MIL-D-8708; however, for the F-35C, requirements that outlined the shore-based
tests required were derived from the Joint Contract Specification (JCS) and JSF
Operation Requirements Document (ORD). Specifically, a structural survey vs a
structural demonstration would be executed. The difference between a survey
and demonstration is the degree to which envelope exploration or expansion is
required: a demonstration requires a higher degree of expansion whereas a
survey is used to validate models and predictions. To meet the requirements of
the JCS and ORD, shore-based testing was broken into three phases: prerequisites
to support CVS (phase one), the structural survey (phase two), and initial ship
trial prerequisites (phase three).
During the first phase, prerequisites to support CVS testing were completed to
include static loading, drop test of a ground article (CG-1), and PA flying qualities.
The second phase was designed to support shore-based catapult launches and
arrested landings. To facilitate shore-based launches, JBD compatibility and
environmental testing was completed to evaluate deck heating, JBD panel
cooling, and vibro-acoustic, thermal, and Hot Gas Ingestion (HGI) environments
on and near the aircraft, shown in Figs. 28 and 29. The C13-2 catapult at Naval Air
Engineering Site (NAES) Lakehurst, New Jersey, was intentionally degraded to
represent a worst-case steam leakage condition to evaluate the effect on engine
performance and tolerance to steam ingestion. Additionally, ground and deck
handling tasks were conducted to ensure aircraft compatibility with the confined
space taxi requirements, catapult spotting procedures, and physical geometry of
the above deck equipment (launch bar and repeatable release hold back
[RRHB] bar compatibility). The final task with respect to shore-based catapult
testing was to subject the aircraft to design limit load launches from the TC-7 cat-
apult at Naval Air Station Patuxent River, Maryland (PAX). Prior to starting
shore-based arrested landings, bolter and wave-off performance test points were
flown and roll-in arrestments were completed. Roll-in arrestments, completed
at NAES Lakehurst, were a ground-based test where the aircraft was positioned
at a predetermined distance on the runway from the arresting gear so as to
reach a targeted groundspeed by the arresting gear. When the aircraft was
within the allowable groundspeed band, the tailhook was lowered, and the
F-35C “rolled” into the arresting gear. This line of testing was performed both
on and off runway centerline (to a maximum of 20 ft off centerline) and is
F-35 CARRIER SUITABILITY TESTING 469

Fig. 28 F-35C conducting deck heating and JBD panel cooling evaluations.

designed to measure both arresting gear and tailhook loads prior to executing
fly-in arrestments. The third, and final, phase consisted of testing required to
support the initial ship trials: reliability and maintenance (deck tie-down and
towing), electromagnetic spectrum capability, navigation and guidance, sensor
integration, data link interoperability, pilot– vehicle interface, and supportability

Fig. 29 JBD panel cooling evaluation with the F-35C engine in afterburner limited by RATS.
470 T. WILSON

Fig. 30 First shore-based catapult launch of the F-35C.

were evaluated to ensure the aircraft was capable of operating as a system within a
system (aircraft aboard a seagoing vessel).

A. SHORE-BASED CATAPULTS
Shore-based catapult launches of the F-35C conducted during the structural
survey, shown in Fig. 30, were used to define and/or validate engineering
models of the structural components to withstand the forces and accelerations
the aircraft would be subjected to during a shipboard launch. The extent to
which the aircraft was stressed was based on the maximum shipboard takeoff
weight and the performance characteristics of the launching catapult. An
additional objective of shore-based catapults was to ensure adequate clearance
between the F-35C (including all external stores) and the catapult launch com-
ponents (i.e., catapult shuttle, RRHB, etc.) throughout the launch cycle
(hookup, launch, shuttle release, and rotation).
The results of shore-based catapult testing are used to predict the limit con-
ditions and establish pilot technique that will be used for shipboard operations.
For the F-35C, the pilot technique used was open-loop, or hands-off, because
the CLAW is designed to capture and maintain the optimum flyaway attitude
and maintain wings level. Testing consisted of a series of launches within an allow-
able weight band that built up in Nx. The Nx was incrementally increased until the
endpoint, maximum longitudinal acceleration (Nxmax) was achieved or the launch
bar horizontal tow load limit was reached. After Nxmax was achieved, a series of
off-center catapult launches were conducted in which the longitudinal axis of
F-35 CARRIER SUITABILITY TESTING 471

the aircraft was misaligned with the longitudinal axis of the catapult to excite
directional oscillations. The off-center testing was conducted within a constant
weight band and at a constant Nx while incrementally increasing the off-center
angle until the maximum delta had been achieved [6].
Most airplanes will experience directional oscillations during catapult
launches when excited by crosswinds, asymmetric loadings, or inadvertent
lateral or directional inputs. When directional oscillations are excited, aerody-
namic forces and tire cornering forces combine as a forcing function to realign
the aircraft with the catapult; however, inertia will cause an overshoot. This
cycle continues until the forcing functions act to dampen out the oscillations. Off-
center catapult testing is used to excite directional oscillations for two reasons:

1. Confirm that forcing function dynamics will dampen oscillations during the
power stroke prior to shuttle release.
2. Measure side forces and bending moment loads on the launch bar and
nosewheel strut.

Catapult launch test points at high gross weights resulted in conditions at which
the catapult end airspeed was less than that required to become airborne. Testing
was conducted at PAX, where the centerline of the C-7 catapult intersects runway
32 at an angle of 8 deg approximately 1000 ft from the end of the catapult power
stroke, as seen in Fig. 31. For these points, significant performance calculations,
modeling, and simulation were conducted to ensure the aircraft would be airborne

Fig. 31 Relative geometry of NAS Patuxent River catapult site intersecting runway 32 at an
8-deg angle.
472 T. WILSON

prior to reaching the centerline of runway 32. The results of this effort were used
in training the test pilots in simulators where aircraft system failures were input
during the catapult power stroke to ensure pilot recognition and appropriate
actions were taken to align the aircraft with the runway for continued takeoff
or abort. If the pilot would be unable to safely realign the aircraft or was required
to eject, the test point was deferred until it could be completed at NAES Lakehurst
where the catapult is aligned with the runway.
Shore-based catapults were completed in each of the configurations shown in
Table 1. The data obtained from shore-based testing were used to construct the
allowable launch envelope and initial Aircraft Launch Bulletins (ALBs), tables that
use aircraft weight and WOD as inputs to determine the CSV setting for launch.

B. SHORE-BASED ARRESTMENTS
In carrying out the carrier suitability testing, a majority of the time, effort and
budget are spent on the shore-based arrestment portion of the structural
survey. This phase of testing can be thought of as specialized loads testing in
which the aircraft executes a series of arrested landings at various sink rates, atti-
tudes, and airspeeds to evaluate the loads on the landing gear, arresting hook, and
other critical load nodes. Prior to the F-35, limitations in computer processing
power constrained the test team’s ability to use aircraft capabilities to increase
testing efficiency; however, with computer processor power ever-increasing,
advances in flight control computers and Full Authority Digital Engine (or Elec-
tronics) Control (FADEC) systems have lessened the need to keep the pilot in the
loop, thus providing an avenue to increase testing efficiency.
The first F-35C structural survey arrestment was accomplished on 4 February
2014. This testing was filled with discoveries and engineering challenges that
resulted in delays; however, using a “blank canvas” philosophy to challenge
legacy assumptions and techniques, new methods were developed to complete
testing and recover lost time. To achieve this end, a methodical five-step iterative
process was created. The first step was to evaluate current test equipment in use
and conduct a historic review of carrier suitability testing to understand the
reasons for assumptions under which shore-based testing was conducted, thus
giving insight into the legacy techniques. The second step consisted of completing
a comprehensive examination of the geometry, control laws, and flying qualities of
the F-35C with respect to approach handling qualities and landing and arrestment
characteristics. Third, the results of the first two phases were compared and con-
trasted, the outcome of which was used to identify every aircraft capability that
could be used or exploited to achieve test points in the structural survey. Next,
as new techniques were identified, a rigid developmental process was followed
consisting of extensive simulator utilization for maneuver development, flying
qualities evaluation and maneuver safety assessment. Finally, the maneuvers
and techniques were used airborne to achieve test points during which their
merits were evaluated. The methods that worked were adopted and the ones
F-35 CARRIER SUITABILITY TESTING 473

that didn’t were refined by repeating the process. The “blank canvas” philosophi-
cal approach consisted of a systematic iterative approach that resulted in the
development of new test techniques that will be used to advance the field of
carrier suitability flight test.
A structural survey is essentially a loads envelope expansion program. Ten
specific maneuvers are flown to endpoints, to achieve 80 –100% of the com-
ponents allowable load limit. For each maneuver, multiple components’ loads
are being monitored via instrumentation, such as main gear compression and
bending moments, nose strut compression and bending moments, hook side
forces, and ordnance suspension forces. To replicate the landing attitudes the air-
craft may experience during a shipboard landing, 10 maneuvers were flown to
reach the limits of the landing design envelope. The landing design envelope
for U.S. Navy aircraft is expressed as a mathematical probability in terms of the
multivariate distribution of landing impact conditions [Eq. (16)] [6]. The total
probability of occurrence of any landing is expressed in terms of the probability
of eight variables that can be expected during landing. Sink speed is defined by
a Pearson Type III distribution function, and the other seven variables are
expressed in terms of a normal or Gaussian distribution function [6].
  
PT ¼ VA ., VAi VE ., VEi VV ., VVi fu ., ui g
 fw ., wi gfẇ ., ẇi gfc ., ci gfd ., di g (16)

PT ¼ Total probability
VA ¼ Approach speed
VE ¼ Engaging speed
VV ¼ Sink speed
u ¼ Pitch attitude
w ¼ Roll attitude
ẇ ¼ Roll rate
c ¼ Yaw attitude
d ¼ Off-center distance
i ¼ Initial condition
., ¼ Greater than and, alternately, less than
All three approach modes of the F-35C were used in pursuit of structural
survey test points. Through the iterative process of the blank canvas philosophy,
each was found to be best suited for certain test points and environmental con-
ditions. Manual with IDLC was used either as pilot preference or in cases
where the aircraft was being flown beyond the CLAW limits for APC or DFP
(i.e., high or low AOA). The challenge with manual was IDLC. Although IDLC
474 T. WILSON

provides fine glideslope control in an operational sense, the “heave” response to

power inputs complicated the pilot tasks for some test points, requiring a simul-
taneous longitudinal stick input to counter the heave or sink to maintain the
desired sink rate; however, this, in turn, would change the pitch attitude at
landing. As will be discussed, both sink rate and pitch attitude had tight tolerance
on most test points, making manual approaches a viable option only when outside
of PA control AOA bands.
APC with 300% IDLC was also used either to establish the pilot on the desired
glideslope for the test point or for high sink test points when environmental con-
ditions allowed. The same heave challenges faced with manual were also faced
with APC, but due to mechanization, the pilot was unable to counter changes
in sink rate with pitch inputs.
DFP was found to be the mode of choice for most test points, provided the
aircraft was inside of the PA AOA band. It provided a very stable sink rate
even when the pilot would make lineup corrections with lateral stick inputs.
DFP CLAW maintains tight control of gamma, and thus sink rate, while allowing
AOA to vary but still be maintained within tolerances. With the CLAW maintain-
ing gamma, sink rate, and AOA, this provided spare capacity for the pilot to main-
tain other parameters for the test point.
In addition to DFP, the pilots were provided flight test scales on their displays
to fine tune test parameters such as AOA, sink rate, and yaw angle. With the flight
test scales and DFP, the pilot was able to set sink rates accurately to 1 ft per second
(fps). The ability for the pilot to capture and maintain the glideslope to within 1
fps resulted in the realization that the FLOLS no longer provided truth data. His-
torically, the pilot did not have the tools within the cockpit to precisely set glide-
slope parameters (sink rate and gamma), and therefore relied on the FLOLS for
truth data. The challenge with the FLOLS at the test site is that it does not sit
on a surveyed pad; it resides on asphalt on the side of a crowned runway. The
imperfections in both the asphalt surface and the crowned slope that cause the
lens to have a roll angle introduce variances that change the desired glideslope.
The lens would be set in increments of 0.25 deg; however, it was noticeable to
the pilots using information in the cockpit that the lens was not set to the
proper angle. This meant that the only useful information the lens provided
was touchdown location; as long as the pilot touched down with a centered
ball, the hook point would be properly positioned for the arrestment. The take-
away at this point was that the FLOLS no longer presented truth data.
To complicate the task, for valid loads measurements the aircraft needed to
land on the nonskid, outlined in red in Fig. 32, which extends 20 ft longitudinally
in either direction from the wire. Nonskid is a surface treatment that increases the
coefficient of friction on the deck of aircraft carriers; there is a patch of nonskid at
the test site’s arresting gear to replicate the shipboard landing environment. The
landing gear is required to touch down on the nonskid for proper wheel spin-up
and landing gear spring-back. With respect to tailhook aircraft, all AHSs exhibit a
bounce on touchdown. When the aircraft and hook hit the deck, the hook bounces
F-35 CARRIER SUITABILITY TESTING 475

Fig. 32 Overhead view of NAS Patuxent River arresting site on runway 32 with nonskid
outlined in red.

off the deck for a finite distance as the main mounts begin to compress. For the
F-35C, this initial bounce is approximately 5 ft. When added to the nonskid con-
strain, that meant that the pilot had to precisely place the hook between 5 and 20 ft
in front of the wire for a successful test point.
The final aircraft characteristic that will be discussed in this section is ground
effect, the aerodynamic phenomenon in which lift increases and drag decreases
when the aircraft is within approximately a half of a wingspan off the deck. This
effect results in a shallowing of the flight path and reduction in sink rate. This pre-
sented a challenge initially as the test team observed a difference between the cal-
culated and actual sink rates, presented in Fig. 33. There was a hesitance in targeting
high sink rates by increasing the lens angle due to concerns that the ground effect
reduction would be reduced or negated at higher gammas, resulting in an overshoot
in targeted sink rates, thus damaging the aircraft. Eventually, the decision was made
to progress and monitor the ground effect. There is a point at glideslopes in excess
of 5 deg where a reduction in ground effect was observed.
High sink test points are designed to achieve 80% or more of the landing gear
allowable load limit. They are conducted to verify structural integrity of an aircraft
that has developed a high sink rate landing aboard the deck of a pitching aircraft
carrier in various sea states. The test point emulation scheme is mean pitch cor-
relates to a level deck, tail down mimics a ship bow down attitude, and nose down
equates to the ship’s bow being up.

1. MEAN PITCH HIGH SINK

Despite the title of this landing being self-explanatory, the objective was to achieve
a landing at or above 21.1 fps at a targeted pitch angle. This endpoint was required
476 T. WILSON

Fig. 33 Calculated and observed aircraft sink speed as a function of lens angle.

to be hit a minimum of three times with at least one deck obstruction (e.g., cross-
deck pendant), with rollovers at 75% of the maximum main landing gear com-
pression load or greater. One of the reasons that the shore-based structural
survey is time consuming is that regardless of what was achieved the day
before, each day began with a “nominal” arrestment using a 3.5-deg glideslope
to ensure the aircraft, instrumentation, and arresting gear were operating prop-
erly. From there, the lens angle would be increased at 0.5-deg increments until
5.0 deg, at which point the increments were decreased by 0.25 deg until the end-
point was achieved. This meant that for high sink test points (mean pitch, tail
down, and nose down), on average a minimum of approximately six arrestments
were needed to achieve an endpoint, barring any complications such as delays
with the airfield, instrumentation issues, aircraft maintenance challenges, or limit-
ations on working hours.
Delta Flight Path was the mode of choice for this test point, providing a high
degree of consistency and repeatability regardless of pilot techniques or environ-
mental conditions. The pilot would be briefed on the lens angle and target sink
rate, and then using the flight test scales, would set the sink rate and engage
DFP. The team decided that the pilots would not compensate for ground effect,
again removing variability due to pilot technique and increasing consistency
and repeatability. The importance and utility of DFP cannot be overemphasized;
F-35 CARRIER SUITABILITY TESTING 477

its use allowed the test team to achieve these test points in minimum time, making
up for delays incurred due to weather, engineering discoveries, or programmatic
issues.

2. TAIL DOWN HIGH SINK

The goal of this test point was to achieve a landing at or above 21.1 fps at the tar-
geted pitch angle plus 2.2 deg, putting the aircraft in a tail down attitude. The
buildup for this test point was the same as discussed in the mean pitch high
sink section. The largest variable with this and the nose down test point was
wind. If the aircraft maintains a constant AOA as the lens angle is increased,
the aircraft’s pitch attitude decreases. To achieve the required high pitch (tail
down) attitude, the aircraft was flown at a high AOA and thus slower TAS. A
slower TAS means a reduction in vertical velocity or sink rate. One solution to
this dilemma is to continue to increase the glideslope angle; however, this has
two adverse safety ramifications:
1. Field of view: At high AOAs, the pilot’s field of view over the aircraft’s nose is
restricted and he or she loses sight of the lens or even the runway.
2. Margin to stall and handling qualities: At high AOAs, the aircraft stall margin
is reduced and becomes a safety concern. Additionally, the handling qualities
change; the aircraft’s response is asymmetric. If the pilot is high on the glide-
slope, a forward longitudinal input results in a rapid decrease in AOA and
pitch attitude (perhaps out of the test band) and a sharp increase in rate of
descent. Conversely, a low on glideslope correction, aft longitudinal stick,
may initially result in nothing more than increased AOA and pitch atti-
tude—the aircraft squats, before shallowing the gamma and returning to
the desired glideslope. It is a disharmonious control scheme.
Therefore, the optimal solution is wind direction. For tail down high sink, the best
environmental days were either no wind or a tailwind. This would allow the air-
craft to fly at the required AOA to achieve the target pitch attitude while still
allowing the sink rate to be achieved without giving up sight of the landing
area or increasing the risk of the test point.
Again, DFP was the mode of choice for this test point, provided the AOA was
within the CLAW limit, for all of the same reasons discussed previously. At AOAs
higher than the CLAW limit, either the point was flown manually with little
success or the team targeted another type of landing.

3. NOSE DOWN HIGH SINK

The goal of this test point was to achieve a landing at or above 21.1 fps at the tar-
geted pitch angle minus 4.5 deg, putting the aircraft in a nose down attitude. The
buildup for this test point was the same as discussed in the previous high sink sec-
tions. Again, the largest variable with this and the nose down test point was wind.
478 T. WILSON

Despite the pitch attitude decreasing as the glideslope was increased, theta was still
high and out of the desired test band. To reduce theta more, AOA was reduced,
thereby increasing airspeed. The challenge faced was that at a low enough AOA to
put theta in the test band, airspeed was too high (above maximum engaging
speed) with zero to little headwind. Achieving the endpoints safely required a
headwind in excess of 5 to 10 kt, depending on the aircraft’s weight.
As with the other high sink test points, DFP was the mode of choice, provided
the AOA was within the CLAW limit, for all of the same reasons discussed
previously. Alternatively, manual or APC could be used for these points, but
they introduced variations due to pilot technique, reducing consistency and
repeatability.

4. ROLL
The roll only test point, shown in Fig. 34, simulates either a late lineup correction
initiated by the pilot or the ship having a roll attitude caused by sea motion, a
slight turn, or not being properly ballasted. It was conducted in both directions,
left roll and right roll. The objective of this test point was to achieve a sink rate
in excess of 17 fps, with the aircraft’s roll attitude statically set (zero roll rate)

Fig. 34 F-35C conducting a roll-only arrestment during the structural survey.

F-35 CARRIER SUITABILITY TESTING 479

at 6 deg or more. The test point was set up based on the wind direction such that
the maneuver was performed into the crosswind component. This was done to
reduce the magnitude of the lateral drift that would develop as a result of the air-
craft turning prior to touchdown. The test team was fortunate in that the environ-
mental conditions supported all of the roll arrestment testing into the wind
without incurring delays waiting on favorable wind directions.
To achieve the roll arrestment, the pilot would offset laterally from the runway
centerline approximately 5 to 10 ft opposite the side of the roll (i.e., offset left for
right roll), depending on the direction and magnitude of crosswind feet. The LSO
provided a “roll now” call over the radio at an appropriate height above the
runway (just prior to touchdown). With the aid of the LSO, the pilot would
capture and maintain the desired bank angle. The safety concerns were twofold.
First, if the LSO or pilot initiated the roll too early, the difference between the
runway’s longitudinal axis and the aircraft’s longitudinal axis would be significant,
resulting in a dangerous condition as the aircraft’s directional vector at touch-
down was biased towards the side of the runway vs aligned with the runway.
Second, despite having adequate displays for setting other aircraft parameters,
the Pilot-to-Vehicle Interface (PVI) is not optimized for setting precise roll
angles required in flight test. The pilot has a small bank angle indicator on the
bottom of the display consisting of an inverted triangle and an arced scale out
to 30 deg in 10-degincrements. The size of the PVI requires the pilot to divert
attention from glideslope maintenance to set a bank angle with any degree of
accuracy. Additionally, the aircraft’s roll mode time constant is relatively high
in PA, complicating the pilot’s task of capturing the required bank angle. To
capture bank angle, the typical pilot’s control inputs resulted in an initial target
overshoot followed by a reduction in bank angle. The bank angle reduction was
necessary for safety; as bank angle increased, the separation between the
wingtip and runway surface decreases. Similarly, due to the relatively high roll
mode time constant, when the pilot would attempt to reduce the roll angle, an
overshoot would occur resulting in a bank angle of less than 6 deg and an
invalid test point.
Again, DFP was the mode of choice to achieve this test point for its capability
of maintaining a constant sink rate. Another benefit of DFP is the response to
lateral stick inputs. DFP is designed to maintain a constant gamma; therefore,
as the pilot commands a roll rate with lateral stick inputs, the control surfaces
respond not only to provide the requested roll rate, but also to maintain the com-
manded gamma. This allows the pilot to divert attention away from glideslope
maintenance, and set the bank angle precisely with little to no roll rate (lateral
stick inputs removed) prior to touchdown.

5. YAW
The yaw arrestment point is a buildup to the roll/yaw arrestments and is only
conducted in one direction—into the wind. The objective of this test point was
480 T. WILSON

to achieve a sink rate in excess of 17 fps, with the aircraft’s yaw attitude statically
set (zero yaw rate) at 5 deg or more with bank angle as close to zero as possible.
The yaw angle was defined by the difference in the aircraft’s longitudinal axis and
directional velocity vector. Again, the arrested landing was accomplished primar-
ily for safety. To achieve the yaw test point, the setup was the same as the rolled
arrestment—the pilot would offset laterally from the runway centerline approxi-
mately 5 ft opposite the side of the yaw input. The offset side was chosen based on
the wind direction such that the maneuver was performed into the crosswind
component to limit lateral drift induced by the yaw. The LSO provided a “yaw
now” call over the radio at an appropriate height above the runway (just prior
to touchdown). Prior to executing the test point, the pilot would engage a
Flight Test Aid (FTA) that enabled a beta limiter. This allowed the pilot to use
a single step or ramp input to full rudder pedal deflection to command the
target beta. The safety concern was similar to those associated with the rolled
arrestment: If the LSO or pilot initiated the yaw too early, the difference
between the runway’s longitudinal axis and the aircraft’s longitudinal axis
would be significant, resulting in a dangerous condition as the aircraft’s direc-
tional vector at touchdown was biased towards the side of the runway vice
aligned with the runway.
A combination of advanced approach mode CLAW and FTA logic resulted in
this maneuver being consistent and repeatable independent of pilot technique
and with low pilot workload. APC was used to capture the desired glideslope
followed by DFP to maintain the desired sink rate. The FTA beta limiter
greatly increased the consistency and repeatability by removing all variability of
pilot technique—full rudder pedal deflection. Another design point of the
CLAW that reduced pilot workload for this test point was the decoupling of the
lateral and directional axes. As the rudders deflect, the roll induced (roll due to
yaw) is near zero, meaning that the pilot was able to accept any roll experienced
because it would be minimal and inside of the test point tolerance.

6. ROLL/YAW OPPOSITE
Roll/yaw test points simulate a late lineup correction in a crosswind landing. The
objective of this test point was to achieve a sink rate in excess of 17 fps and the
aircraft established in a steady-heading sideslip with the flight path perpendicular
to the cross-deck pendant. The test point was valid provided both the roll angle
and yaw angle were equal to or greater than 5 deg. As a general rule, when the
environmental conditions allowed, the yaw was into the relative crosswind; this
decreased the pilot’s centerline maintenance workload.
The employment of DFP and the FTA beta limiter resulted in the completion
of this line of testing in one day. DFP was the optimum approach mode for this
test point. The pilot would engage the beta limiter while setting up for the test
point. Once aligned with the runway centerline at a range of approximately 1.5
miles with the desired glideslope captured, DFP was engaged, the appropriate
F-35 CARRIER SUITABILITY TESTING 481

rudder pedal was fully deflected, and the pilot used lateral stick inputs to set an
angle of bank that maintained the desired ground track. The test pilot scans gli-
deslope (the ball), line up, bank angle, yaw angle, and rate of descent; however,
using this technique, the pilot’s workload was low and repeatability was high.
DFP maintained the desired rate of descent while the pilot used primarily
lateral stick inputs to maintain centerline. Just prior to touchdown, the LSO
would give a “roll now” call as a preemptive command to have the pilot check
the roll angle. This call was made with enough time to allow the pilot to make
one to two more lateral stick inputs to ensure the bank angle was within test tol-
erance. All of the difficulties and safety concerns discussed in the roll section are
also applicable to roll/yaw opposite with the exception of runway misalignment;
despite the unusual landing attitude, on touchdown, the aircraft was aligned with
the runway.

7. ROLL/YAW SAME
The objective of this test point was to achieve a sink rate in excess of 17 fps and the
aircraft’s attitude statically set with both a roll angle and a yaw angle equal to or
greater than 5 deg in the same direction. The maneuver was performed into the
relative crosswind to limit the lateral component of the velocity vector while
allowing the yaw angle to reach the desired target. As with the roll only and
yaw only test points, the pilot needed to be prepared for potentially high drift
rates and runway misalignment on touchdown should a bolter occur.
To achieve the test point, the pilot would engage the beta limiter while setting
up for the test point. Once aligned with the runway centerline at a range of
approximately 1.5 miles with the desired glideslope captured, DFP was engaged
to maintain the desired sink rate. As the aircraft approached the runway, the
pilot would establish a 5-ft lateral offset from the runway centerline opposite
the direction of the maneuver. Two techniques were employed in the achievement
of this test point: a steady-heading sideslip reversal and a dynamic maneuver. In
the dynamic method, the pilot would simultaneously set the roll angle with lateral
stick and input a full pedal deflection when the LSO called “roll now.” The pilot
relied on the beta limiter both to establish a yaw angle greater than 5 deg and not
to exceed yaw angle limits where high main gear side loads were predicted. With
the steady-heading sideslip reversal method, the pilot would establish the aircraft
in a steady-heading sideslip using the roll/yaw opposite technique. Upon the
LSO’s “roll now” call, the pilot would reverse the bank angle with lateral stick
while relying upon DFP to maintain the sink speed and the beta limiter to main-
tain the yaw angle. Although the dynamic maneuver was successful in completing
the test point, it was deemed too susceptible to pilot technique with insufficient
repeatability; therefore, the steady-heading sideslip reversal method was preferred.
Both techniques utilized DFP and FTA beta limiters extensively to reduce pilot
workload and increase repeatability, thereby increasing testing efficiency and
effectiveness.
482 T. WILSON

8. MAXIMUM ENGAGING SPEED ON-CENTER

The data from the maximum engaging speed arrestment were used to measure
and verify the design limit arresting hook axial load. An initial maximum enga-
ging speed had been established during the ground-based roll-in testing at
NAES Lakehurst. To evaluate the load, an arrested landing was accomplished at
a sink rate of at least 17 fps while flying a nominal approach AOA. This
allowed the pilot, control room team, and LSO to evaluate the headwind gradient
during the approach, descending from an approximate altitude of 1000 ft to sea
level. After the hook loading was evaluated, the pilot was given a ground speed
to target for the next arrestment.
The pilot used the speed hold autopilot functionality of the CLAW that cap-
tures and holds a constant calibrated airspeed. This mode was intended for
up-and-away flight. It was found that when engaged in PA, it easily excited the
phugoid mode, and the open loop response resulted in a gentle longitudinal undu-
lation about the desired glideslope. Although the magnitude of this undulation
was relatively small, if unchecked, it could have resulted in the aircraft landing
before the nonskid or boltering, both of which result in an invalid test point.
During the approach, the pilot was able to accept the undulations provided the
mean was about the desired glideslope; however, starting at approximately 0.5 n
mile from touchdown, the pilot would dampen out the undulations and maintain
the desired glideslope with longitudinal stick inputs. Not using DFP further
increased pilot workload. When making a lateral correction to maintain align-
ment with the runway centerline, the flight control system does not automatically
compensate for loss of lift or the increase in drag, nor is the engine response
coupled with control surface movement. Because the parameter of primary
concern was ground speed, if the aircraft developed a lateral drift, the pilot
would initiate a correction towards centerline with roll and then use longitudinal
stick to increase AOA momentarily to maintain the desired glideslope. A similar
series of control inputs were used to recapture runway centerline. Both the pilot
and LSO have instruments that provide the aircraft’s ground speed. As the
F-35C approaches the arresting gear, the LSO advises the pilot of required
speed changes. As target groundspeeds were increased, longitudinal stick forces
changed due to the CLAW trimming to a nominal approach AOA. The pilot
would engage Speed Hold at a calibrated airspeed that equated to the targeted
ground speed for the environmental conditions of that particular day. The pilot
manually trimmed the longitudinal stick forces to zero. Then, using a switch on
the throttle, the pilot adjusted the captured calibrated airspeed in 1-kt increments
as required to maintain the desired ground speed. The pilot and LSO worked as a
team to ensure that the aircraft engaged the arresting gear at the targeted
ground speed.
This technique of the pilot maintaining glideslope with longitudinal stick
inputs, lineup with a combination of lateral and longitudinal stick inputs, and
ground speed with multiple switch actuations results in a higher pilot workload
F-35 CARRIER SUITABILITY TESTING 483

but was very repeatable, resulting in accurate engaging speeds. The endpoint of
this series of testing was achieving an allowable hook axial load limit of 100%
while engaging the cross-deck pendant within 5 ft of the runway centerline and
the aircraft’s groundspeed within 1 kt—1 kt too slow resulted in a hook loading
less than 100% and 1 kt too fast resulted in an overstress condition requiring
testing to stop while both the aircraft and arresting engine were inspected
for damage.

9. MAXIMUM ENGAGING SPEED OFF-CENTER

This series of testing is designed to stress both the arresting hook and arresting
gear loading. Maximum engaging speed endpoints were required to be executed
to both the left and right side of the cross-deck pendant. This is due to the path
the arresting gear cables travel from the arresting engine to the cross-deck
pendant—one length is shorter than the other, resulting in slightly higher loads
on the short side. Testing was conducted using the same techniques and with
the same safety concerns discussed in “Maximum Engaging Speed On-Center”;
however, the buildup sequence was slightly different. After an on-center engage-
ment at a 17-fps sink rate had been achieved at a nominal approach AOA, the off-
center distances of 10 ft, 15 ft, and 18 ft were targeted. Once reaching the
maximum off-center distance (18 ft with a tolerance of –1 and þ2 ft), the enga-
ging speed was incrementally increased until the arresting hook loads or the
arresting gear limits were reached.
Asymmetric loading of the aircraft added another level of complexity to this
series of test points as well as the yaw and roll/yaw series. Due to the lateral
shift in CG, the aircraft’s dynamics during the arrestment were closely monitored
for wing tip – to-runway clearances, arresting gear –to-aircraft interactions,
weapon bay clearances, or adverse stores response during inertially induced rolls.

10. FREE FLIGHT OR HIGH PITCH/LOW ANGLE

This test point duplicates an approach where the pilot initiates an aggressive
nose-up pitch maneuver close to the deck, causing the arresting hook to engage
the cross-deck pendant prior to main gear touchdown or with the aircraft’s vel-
ocity vector above the horizon. The objective of this point was to achieve at
least an 85% arresting hook limit load while generating an as high as possible
nose gear limit load. This type of landing, also called an “inflight engagement,”
results in high nose-down pitch rate that causes peak loads in the nose gear.
Although not impossible, the F-35C’s geometry makes a free flight or inflight
engagement extremely unlikely because at optimum approach AOA, the
bottoms of the main gear tires are lower than the hook point. This results in
the main gear touching down on the deck or runway prior to the hook point.
The setup for this test point was to set the FLOLS to a minimum safe glide-
slope of 2.25 deg. The aircraft was flown at 15 deg AOA. Under these conditions,
484 T. WILSON

this places the hook point and the main gear in the same horizontal plane, result-
ing in simultaneous touchdown of the main gear and hook point. Although the
pilot was able to utilize DFP, at the higher AOA the response was asymmetric
and disharmonious. If the pilot was above the desired glideslope, a forward longi-
tudinal input resulted in a rapid decrease in AOA and pitch attitude, raising the
hook point and increasing the rate of descent. Conversely, to correct for being
below the desired glideslope, an aft longitudinal stick input initially resulted in
an increased AOA and pitch attitude, a lowering of the hook point, and an aircraft
squat before shallowing the gamma and returning to the desired glideslope.
Additionally, flying at high AOAs limited the pilot’s field of view over the aircraft’s
nose. Finally, due to the desire for the hook to engage the cross-deck pendant sim-
ultaneously as the main gear touched down, hook point touchdown placement
needed to be precise. These factors combined to make this series of testing one
of the most difficult for the pilot to achieve with a marginal success rate.
As a result of aircraft geometry and the test point’s lack of repeatability due to
the sensitivity in pilot technique, it was decided that the test point would be rede-
signed for future testing. The new test, high pitch/low angle, uses a low-angle gli-
deslope being flown with the aircraft at an elevated AOA. The touchdown
requirements are loosened with the objective being to generate pitch rates and
nose gear loading above calculated values that still allow the models and predic-
tions to be validated.
Shore-based ship suitability testing of the F-35C began on 17 March 2011 and
culminated in the first shipboard arrestment on 3 November 2014. It took
approximately three and a half years to complete the prerequisite testing to

Fig. 35 The F-35C conducting a structural survey arrestment with GBU-12s and a gun pod.
F-35 CARRIER SUITABILITY TESTING 485

safely conduct the initial ship trials. During that time, multiple discoveries were
made and engineering challenges faced, primarily the AHS being redesigned,
but that is to be expected in developmental flight test. The structural survey
started on 4 February 2014, after the installation of the new AHS; however, the
program continued to make discoveries that complicated the completion of the
structural survey. Between the start of the structural survey and 19 August
2014, 51 arrested landings, of which 15 were endpoint, were achieved. That is
an average of 0.26 arrestments per day. On August 19 the last remaining restric-
tion (no more than six arrested landings per day) was removed. And on Septem-
ber 22, the test team completed the structural survey, executing an additional
102 arrestments, of which 23 were endpoints, in 34 days. See Fig. 35 for an
example of a structural survey arrestment. This pace increased the average daily
arrestment-per-day over 11-fold from 0.26 to 3.0. This order of magnitude
increase in testing efficiency has been shown to be linked to three factors:

1. Challenging the assumptions that previous carrier suitability projects were

conducted under, the F-35C was approached with a blank canvas philosophy
that paved the way for the use of automation in the pursuit of CVS testing that
increased efficiency.
2. The advanced approach modes and IDLC provide the pilot with an unprece-
dented ability to control the aircraft during shipboard approaches. This tech-
nology, specifically DFP, and FTAs provided tools that allowed challenging
test points to be achieved with a high level of repeatability without relying
on a test pilot’s skill (or lack thereof) to finesse aircraft performance.
3. The blank canvas philosophy approach enabled the use of the advanced
approach modes and FTAs in ways that had been previously thought unusa-
ble or unsafe.

V. SHIP-BASED TESTING
Flying qualities and performance testing are first conducted shore-based, then
evaluated at the ship. The shore-based tests are used to define the minimum
acceptable approach airspeed and give the testers an initial look at airplane
flying qualities in the more benign shore-based environment. Once satisfied
with the shore-based results from a structural suitability, flying qualities, and per-
formance standpoint, the airplane is ready to proceed to the ship. Some of the
highest risk shipboard tests are the minimum catapult end airspeed tests where
the slowest safe speed for catapult flyaway is determined. Other shipboard
launch tests include establishing crosswind limits, evaluating launches off waist
catapults, and evaluating the effects of low energy launches (low gross weights,
high winds) on flyaway. Recovery testing includes establishing crosswind limits,
evaluating bolter and wave-off performance, and evaluating handling qualities
at various wind-over-deck conditions. Additional testing includes evaluating
486 T. WILSON

Fig. 36 Example of ship attitudes routinely attainable while at sea.

compatibility with shipboard facilities and support equipment: heavy weather

tie-down (as seen in Fig. 36), canopy opening under high wind conditions,
dynamic tipback following an arrestment, and hangar bay towing and spotting.
Over the course of two years, three ship trials were completed: the initial ship
trial, DT-I, aboard the USS Nimitz (CVN 68) in November 2014, DT-II aboard the
USS Eisenhower (CVN 69) in October 2015, and DT-III aboard the USS George
Washington (CVN 73) in August 2016. While embarked, the F-35C flew in
several different configurations, shown in Table 1, building up in weight, shifting
center of gravity, and increasing asymmetry.

A. WHY TEST SHIPBOARD?

Shipboard testing is conducted for two reasons: first and foremost is to evaluate
the aircraft in the environment in which it will be operated, and second, it is
not currently possible to re-create or simulate the environmental conditions
experienced at sea. The environmental conditions include launching from a
deck approximately 60 ft above sea level, the aircraft carrier air wake or
“burble,” and deck motion.
The dynamics of a shipboard launch are very different than taking off from a
runway. When an aircraft takes off from a runway, the pilot applies power and
the aircraft accelerates down the runway. When the aircraft’s velocity has
reached rotation speed, the pilot applies an aft longitudinal input on the
control stick to deflect the horizontal stabilators, thus producing an aerodynamic
pitching moment. The aircraft rotates about the main wheels, increasing the AOA
F-35 CARRIER SUITABILITY TESTING 487

of the wing and thereby lift. As the aircraft continues to accelerate down the
runway, lift increases and the aircraft flies off the runway. This process is illus-
trated in Fig. 37. Conversely, for a shipboard launch, the aircraft is accelerated
to flyaway speed in just under three seconds with the horizontal stabilators pre-
positioned to generate a pitching moment. At shuttle release, stored energy in the
nose strut is released and aided by the aerodynamic pitching moment, a pitch rate
is started; however, a moment later (under one second) the aircraft leaves the
deck. The aircraft now rotates about its CG vs the main wheels. This process is
illustrated in Fig. 38. The dynamics and flyaway characteristics of a shipboard
launch are more easily affected by crosswinds, pitch angle of the aircraft carrier’s
deck, and the carrier wake turbulence, and therefore preclude shore-based
re-creation.

Fig. 37 Conventional takeoff dynamics rotating about the aircraft’s main landing gear.
488 T. WILSON

Fig. 38 Catapult launch dynamics rotating about the aircraft’s center of gravity.

As a U.S. aircraft carrier moves through the water, it has a profound effect on
the air mass through which it moves. Starting at the bow of the ship, two vortices
are generated aligned with the longitudinal axis of the ship, similar to wingtip vor-
tices. The vortex generated on the starboard side rotates counterclockwise as
viewed from the deck of the ship looking forward but has negligible effects on
the launch and recovery of aircraft. The vortices of concern with respect to reco-
vering aircraft are generated off of the port bow and the outboard leading edge of
the landing area; both are aligned with the longitudinal axis of the ship, rotate
counterclockwise, and flow through the landing area. Next, the aircraft carrier’s
superstructure creates an area of turbulent air. Due to the angled landing area,
F-35 CARRIER SUITABILITY TESTING 489

Fig. 39 Overhead view of ship airwake effects on the deck of an aircraft carrier.

aircraft transit through the turbulent air during the approach. All of these effects
are shown in Fig. 39 with the approach path of the aircraft depicted by the dashed
black arrow.
The airwake off the end of the aircraft carrier, called the “burble,” is charac-
terized by a downdraft as the air flows off of the deck followed by an area of
updraft aft of the downdraft, as seen in Fig. 40. The magnitude and position
of the downdraft and updraft are a function of natural wind and the ship’s vel-
ocity. As an aircraft on approach interacts with the burble it will first experience
an uplift; the open looped response would be a reduction in rate of descent
followed by a greater than normal increase in rate of descent. With compu-
tational fluid dynamics, the fidelity of computer-modeled airwake effects is
increasing; however, outside of a simulator, the shore-based re-creation to
date is impossible.
The deck of the aircraft carrier is subject to the motions of the ship: pitch, roll,
heave, drift, and Dutch roll. These dynamics result in complicating launches and
recoveries. For launches, any bow-down attitude results in an initial negative
gamma (downward flight path) for the aircraft. With respect to recoveries,
ship’s motion results in landing area and optical lens motion, increasing the dif-
ficulty of a shipboard landing. In addition to complicating launches and recov-
eries, the ship’s dynamics also induce airwake effects. As the ship pitches and
heaves up and down, a vortex aligned with the ship’s lateral axis is formed at
the bow. The magnitude of this vortex is a function of pitch and/or heave
rates, growing as the bow pitches down or the ship settles after a heave. This
lateral vortex has been found to affect catapult launches. Finally, the ship’s
dynamics prevent deterministic loads calculations, requiring stochastic analysis
of the multivariate distribution of landing impact conditions.

Fig. 40 Airflow profile over the flight deck and aircraft’s approach path.
490 T. WILSON

B. CATAPULT LAUNCHES
Shipboard catapult launches provided data used to verify the launch envelope,
similar to that seen in Fig. 41 [6], construct ALBs, and develop sections of the
flight manual. The launch envelope is bounded by the aircraft’s maximum allow-
able gross weight, launch bar axial load limit, and longitudinal acceleration limit.
Other envelope-bounding considerations include catapult limits and the
minimum airspeed required to safely fly away. The launch envelope and ALBs
provide limitations and procedures to ship personnel to safely catapult the
F-35C; these documents contain information such as wind envelopes and catapult
settings. Conversely, the flight manual provides the pilot limitations such as
maximum gross launch weight, maximum asymmetric loading, and maximum
gross arrestment weight (also known as “max trap”).
To collect the required data, four distinct types of catapult launches were con-
ducted from both the bow and waist catapults aboard the ship: Minimum End
Airspeed (MEAS), high excess, low energy, and crosswind. To better understand
shipboard catapult testing, the following terms and phrases are defined:
. Digital End Speed Indicator (DESI): A chronograph that is integrated with the
catapult and uses a series of magnetic sensors to provide a catapult end speed.
. Catapult end speed: The speed of the aircraft relative to the deck (catapult
shuttle speed) at the end of the catapult power stroke, normally measured
by the DESI.

Fig. 41 Example catapult envelope illustrating the factors that are evaluated and verified
during testing.
F-35 CARRIER SUITABILITY TESTING 491

. Minimum End Airspeed (MEAS): The end airspeed at which the aircraft can
safely fly away at a given weight but not without the aircraft’s CG sinking a pre-
determined amount below the deck of the aircraft carrier.
. Recommended launch airspeed: A launching airspeed in excess of MEAS that
provides a margin of safety for variations in launch parameters. It is normally
expressed in terms of excess airspeed above MEAS (e.g., an aircraft launched at
an end speed 20 kt above MEAS would be phrased “min plus twenty” and
would be written min þ20). Under normal operational conditions, aircraft
are launched at min þ15.

1. MINIMUM END AIRSPEED

All catapult launches are based on the minimum end airspeed values. Through
examination of the figure and intuition, it can be deduced that if the catapult
launches the aircraft at higher speeds or the ship produces a higher WOD, then
margins are increased; however, launching aircraft at the lowest safe airspeed
possible has many advantages, including:
. Lower WOD requirements reduce the speeds at which the ship must move
through the water to produce winds, thereby reducing the rate at which the
nuclear fuel is consumed.
. Lower WOD requirements reduce the environmental stress on the deck crew.
(Exposure to sustained high winds results in fatigue.)
. Lower launch airspeeds result in lower catapult acceleration profiles, thereby
decreasing loads imposed on the airframe and increasing fatigue life.
. Lower launch airspeeds reduce the amount of energy that the catapult must
impart to the aircraft, thereby conserving the ship’s fresh water supply (the cat-
apults are steam-driven) and nuclear fuel.
. Lower launch airspeeds increase the operational capability of both the ship and
the aircraft.
Minimum end airspeed launches were conducted at various weights, CGs, and
power settings (military power [MIL], which is 100% of rated thrust, and
maximum afterburner [MAX] at 150% of rated thrust) (see Fig. 42). The MEAS
is calculated based on the aircraft’s CG sink below the flight deck or “sink off
bow.” Several factors are considered when determining the targeted sink off
bow including thrust required, thrust available, proximity to stall warning, aircraft
flying qualities, longitudinal acceleration over vertical acceleration (a/g), flight
control response, and pilot comfort [6].
The build-down process began at the predicted minimum end airspeed plus 15
kt excess using CSV, then 10 kt of excess, and finally 6 kt excess, all while keeping
WOD constant. CSV was then held constant to remove catapult variability, and
492 T. WILSON

Fig. 42 Minimum end airspeed testing was conducted in multiple configurations, including
heavy weight external loadings, as shown here.

WOD was reduced by slowing the ship’s speed for the next two launches targeting
3 kt and 0 kt excess (minimum end airspeed). Environmental conditions needed
to be ideal each time MEAS testing was conducted. Winds needed to be steady
with little to no gust for consistency and safety. Additionally, the winds needed
to be high enough such that testing could continue once the ship’s speed
reached minimum steerage (the minimum ship’s speed to maintain rudder effec-
tiveness). The targeted CG sink off bow for MIL and MAX was 13 ft and 18
ft, respectively.

2. HIGH EXCESS END AIRSPEED

Launching with high excess end airspeeds is an operational capability requirement
in which the aircraft is catapulted with up to 40 kt of excess. During these
launches, the aircraft may “kite” or become airborne prior to the end of the cat-
apult power stroke [6] or overrotate upon shuttle release. Although these undesir-
able flying qualities were not encountered, had they been, the testing would have
been halted and a maximum excess end airspeed limit would have been defined,
bounding the catapult launch envelope. High excess end airspeed launches uti-
lized a buildup from minimum end airspeed plus 30 kt to plus 45 kt. CSV was
the primary driver to provide the excess; however, a minimum of 15 kt of
WOD was required. The F-35C exhibited neither kiting nor overrotation, but
instead, high excess end airspeeds resulted in the aircraft flying off of the deck
in a nonobjectionable flat attitude, shown in Fig. 43, vs the typical rotation
about the CG.
F-35 CARRIER SUITABILITY TESTING 493

3. LOW CATAPULT ENERGY

During pilot carrier qualification, aircraft are routinely launched at lighter
weights, slightly above max trap, to facilitate an immediate recovery. When
lighter weight aircraft are launched in high WOD conditions, CSV settings are
lowered to reduce fatigue stresses on the airframe, resulting in lower catapult
end speeds and longitudinal accelerations. Historically, low catapult energy
launches reduced the inertial loads action of control surfaces, resulting in small
control surface deflections during open-looped launches; unintentional sink off
bows were a potential due to degraded rotation characteristics [6]. However,
the F-35C is designed to be launched stick free; the CLAW schedules the
control surfaces as a function of air data, longitudinal acceleration, and wheel
speed. Low catapult energy launches were conducted to assess the flyaway per-
formance and rotation characteristics in high WOD conditions; low launch bar
loads result in lower nose strut squat at RRHB release, imparting less stored
energy at shuttle release. This series of catapult launches started at minimum
plus 15 kt excess. Using ship’s speed, WOD was then incrementally increased
above 45 kt and until a minimum CSV setting or  5% launch bar axial tow
force was achieved. The safety concern with low axial tow forces was that the
shuttle could spit the launch bar.

4. CROSSWIND CATAPULT
During crosswind catapult launches, an up-wash condition is created on the wind-
ward side of the ship’s hull, effectively increasing the AOA of the upwind wing,
resulting in high torque loads on the launch bar. Upon shuttle release, the air-
craft’s open loop response is to yaw into and roll away from the crosswind

Fig. 43 F-35C flying away in a flat attitude after a high excess end airspeed launch.
494 T. WILSON

while developing a lateral drift aligned with the crosswind. The objective of the
stick-free response was to characterize the aircraft’s stability and evaluate the
flying qualities and FLCS to return the F-35C to balanced flight.
Catapults in crosswinds were conducted to evaluate and establish the aircraft’s
operational limits. Launches were accomplished from the bow and waist catapults
with the aircraft in two configurations: the maximum allowable shipboard asym-
metry and the most forward CG operationally possible. There is a correlation
between excess end speed and crosswind effects: crosswinds under 10 kt have neg-
ligible effects on the launching characteristics of an aircraft [6]; therefore, testing
began at 10 kt and expanded outwards to a maximum of 15-kt crosswind. Testing
was achieved first by building up in crosswind from 10 to 15 kt, then by building
down in excess end airspeed from 15 to 10 kt. Additionally, the crosswinds were
evaluated from both the port and starboard side of the ship. During testing, launch
bar torqueing loads were well within structural tolerances. Additionally, aircraft
dynamic response in yaw, roll, and drift at shuttle release were deemed nonobjec-
tionable by the test pilots. Both of these factors allowed the crosswind envelope to
be cleared out to 15 kt in either direction down to a minimum excess of 10 kt.
(With the exception of catapult 4, 15 kt excess is the minimum due to a longer
deck roll than catapults 1, 2, and 3).

5. AIRCRAFT LAUNCH BULLETINS

The results of shipboard catapult testing were used to develop a series of ALBs, an
example of which is shown in Table 4 [8], that contain the catapult settings and
WOD required to launch the aircraft throughout the full range of gross weights as
a function of aircraft configuration and power setting. The bulletins also contain
information on WOD limitations, corrections for atmospheric conditions, correc-
tions for catapult expansion due to heat, JBD limitations, and CSV settings.

TABLE 4 EXAMPLE OF AN ALB THAT PROVIDES THE UNITLESS CSV SETTING FOR CATAPULT
LAUNCHES AS A FUNCTION OF AIRCRAFT WEIGHT AND HEADWIND

Headwind for 15 Kt Excess

20 22 24 26 28
Aircraft Weight CSV Setting
(1000 lb)
45 114 110 106 102 98
46 118 114 110 106 102
47 124 120 116 112 108
48 128 124 120 118 116
F-35 CARRIER SUITABILITY TESTING 495

Fig. 44 Carrier approach segment nomenclature and approach handling qualities testing
envelope bounded by the limits of the IFLOLS (bottom red and top yellow lines) and
proximity to the ship (from the start [X] to in the middle [IM]).

C. ARRESTED LANDINGS
Shipboard recovery handling qualities were evaluated using both nominal and off-
nominal (intentional error) approaches in environmental conditions emulating
those to be expected by fleet pilots. In the course of the three ship trials, evalu-
ations were conducted in multiple configurations including forward CG,
maximum allowable lateral asymmetries, and full external stores. In addition to
the many configurations, all three approach modes were evaluated in four differ-
ent wind cells. Although handling qualities were not evaluated with the aircraft in
simulated degraded states (control surface failures or degraded modes) during
ship trials, degraded and failure state shipboard recovery handling qualities
were extensively evaluated in the simulator to develop emergency procedures
and recommended pilot recovery techniques. The results of the simulator trials
and shipboard testing were used to establish procedures and publish pilot tech-
niques in the flight manual.
The final approach to an aircraft carrier starts with the “ball call” and ends
with an arrestment, touch and go, bolter, or wave-off. The ball call is a radio trans-
mission made at 34 n mile from the aft end of the ship by the pilot to inform the
LSO that the lights of the OLS and ships landing area are in sight. The approach is
segmented into five parts, four of which are shown in Fig. 44, and are divided by
distance in nautical miles from the aft of the ship. The ball call coincides with the
start. The times shown below the distances are based on a fighter aircraft flying the
optimum approach airspeed and correlate to time in seconds from touchdown.
The test pilots would intentionally set up a deviation in glideslope, lineup, or
airspeed to evaluate the handling qualities. Testing began with single deviations at
the start such as a “high start” or lined up left start, also known as an angling
approach (see Fig. 45). The pilot would then correct to be on glideslope, on center-
line, and on-speed no later than the in close position. As a safety measure, if the
pilot had failed to make the appropriate corrections by the in close position, he or
she would have been waved off.
496 T. WILSON

Fig. 45 Carrier approach nomenclature for lateral deviations.

After single deviations were evaluated, testing progressed to multiple devi-

ations, ultimately evaluating multiaxis and speed deviations such as low on glide-
slope, slow airspeed, high bank angle, and laterally displaced to the right of the
landing area centerline at the start—known to naval aviators as a low slow
wrapped up overshooting start. All of these deviations were evaluated in the
three approach modes: manual, APC, and DFP. After start deviations were eval-
uated, the test pilots would hold the deviation to the in the middle position with a
requirement to correct the deviation by the in close position.
The initial rounds of evaluations were completed in the nominal wind cell (25
kt+5 kt WOD). The burble effects aft of the ship increase as a function of WOD
and may result in special pilot techniques due to increased ship airwake disturb-
ances. After the nominal wind cell was complete, low wind (16 kt þ4/ –0 kt
WOD), medium wind (35 kt+5 kt WOD), and high wind (40 kt þ0/ – 5 kt
WOD) cells were evaluated, again using all three approach modes. Handling
Qualities Ratings (HQRs), using the Cooper-Harper rating scale, were assigned
for both the capture and maintenance tasks associated with glideslope, centerline,
and airspeed. The results of all three ship trials are shown in Figs. 46 and 47. Air-
speed, or AOA, ratings were not evaluated for APC and DFP because the CLAW is
controlling AOA and the pilot has no direct control.
All three approach modes exhibited level one handling qualities, but DFP was
the preferred approach mode for both test pilots and operational pilots. As a result
of the handling qualities, the tasks associated with shipboard landings not only
were deemed satisfactory in achieving desired performance, but DFP also
reduced pilot workload. The general control scheme/philosophy is that the
pilot engages DFP as the glideslope is captured. As perturbations move the aircraft
away from the desired glideslope, the pilot inputs and holds a longitudinal stick
displacement to generate a change in gamma. As the aircraft returns to the
desired glideslope, the pilot leads the correction and releases the stick, thus com-
manding the aircraft to recapture the programmed gamma and desired glideslope.
F-35 CARRIER SUITABILITY TESTING 497

Fig. 46 DT-I and II approach handling qualities results for the clean wing configuration (no
external stores) evaluation. (Numbers to the right in red correspond to the number of
deviations and approaches attempted for that specific task with breakouts according to
handling quality levels.)

Additionally, as pilots are making lateral inputs to capture and maintain center-
line, the DFP functionality attempts to maintain the commanded gamma such
that once the pilot’s scan returns to the OLS, the ball is still between the datums.
The F-35C was designed to conduct carrier approaches with the TEFs posi-
tioned at 30 deg down. This was done to minimize the approach airspeed.
However, due to aircraft geometry and hook shank length, the driving design
objective was to minimize hook touchdown point dispersion to ensure succes-
sful arresting gear engagement. It was determined during testing that the task
of glideslope capture and maintenance was easier when the TEFs were positioned
to 15 deg and IDLC was increased 300%, as seen in Fig. 48. A flight path (gamma)
hold feature was added to the CLAW, in addition to Approach Power

Fig. 47 DT-III approach handling quality evaluation results with external stores, both
symmetrically and asymmetrically, loaded.
498 T. WILSON

Fig. 48 F-35C approaching the ship with 15-degTEFs and 300% IDLC.

Compensation (APC), that resulted in a pilot workload reduction. The combi-

nation of improved handling qualities and reduced pilot workload culminated
in a reduced hook touchdown dispersion, thus increasing the probability of a suc-
cessful hook-to-wire engagement, as seen in Figs. 49, 50, and 51.
The potential benefits that DFP offers are a reduction in training require-
ments (both shore-based and shipboard), an increase in safety margins, and
increased operational capability. As a result of the level one handling qualities,
reduced pilot workload, and small hook point touchdown dispersion, the
initial cadre of fleet aviators (all of whom had previous carrier experience in
other aircraft) to conduct F-35C carrier qualifications used DFP successfully.
During that evolution, the aircraft neither boltered (landing long, missing all
of the cross-deck pendants) nor engaged the one wire (minimum safety
margin tolerated). The consistency with which the aviators were able to safely
land on the deck on the target location with ease leads to the logical question
of reducing training requirements. With respect to safety, the consistency and
repeatability with which the fleet aviators were able to execute approaches to
the deck and precisely place the hook point on or close to the targeted touchdown
point increases safety margins (hook-to-ramp distance consistently greater than
the minimum allowable). The aircraft’s capability to reliably land aboard the ship
on the first attempt without boltering or being waved off serves to improve
boarding rates. Operational capability will be increased as a result of reliable
boarding rates, which affect the time the ship is required to steam into the
wind and allow more time for maintenance personnel to prepare the aircraft
for the next launch.
F-35 CARRIER SUITABILITY TESTING 499

Fig. 49 DT-I hook point touchdown dispersion.

Fig. 50 DT-II hook point touchdown dispersion.

500 T. WILSON

Fig. 51 DT-III hook point touchdown dispersion.

VI. CONCLUSION
Designing an aircraft that will operate from the decks of seagoing vessels is not
easy. The design space is filled with a myriad of constraints and intricacies such
as sizing for deck space considerations, wing fold capabilities to reduce that size
factor, and the proper sizing of control surfaces for the tasks of launching and
recovering. The F-35C couples those baseline constraints with the additional
requirements of low observability and commonality with two other variants,
which affected such things as arresting hook geometry that had second-order
effects on arresting gear dynamics (perturbations of kink waves). These engineer-
ing challenges were faced by the multidiscipline specialization team of carrier
suitability that combines theories from aircraft loads, flying qualities, and per-
formance in a system of systems approach to assess the suitability of an aircraft
to operate from ships and austere sites. In addition to flight sciences, carrier suit-
ability includes many other nuances and facets such as navigation and guidance,
sensor integration, data link interoperability, pilot –vehicle interface, supportabil-
ity, maintainability, and night evaluations, shown in Fig. 52, that were all beyond
the scope of this paper.
Aircraft that are launched from and perform arrested landings to the deck
of a ship require enhanced structure to withstand the loads to which they are
F-35 CARRIER SUITABILITY TESTING 501

subjected. As such, a majority of the carrier suitability process was spent on

conducting specialized loads testing in the form of shore-based catapults and
arrested landings. During the structural survey, the F-35C faced and overcame
many challenges such as the AHS. The original tailhook is proof that although
designs may work on paper, they do not always go right the first time. The
AHS redesign process was validation of the engineering problem-solving
process.
Additionally, strides towards the improvement of the carrier suitability disci-
pline were also made by challenging the current assumptions and philosophies.
The F-35C was approached with a blank canvas mentality that paved the way
for the use of automation in the pursuit of CVS testing that increased efficiency.
The advanced approach modes and IDLC provided the pilot with an unprece-
dented ability to control the aircraft. This technology, specifically DFP, and
FTAs provided tools that allowed challenging test points to be achieved with a
high level of repeatability without relying on a test pilot’s skill to finesse
aircraft performance.
Undeniably, the CLAW, which provides the basis for the advanced approach
modes and FTAs, demonstrated many benefits to both the F-35C and the entire
F-35 program. Delta Flight Path is a game changer. It was an instrumental tool
in allowing the test team to complete the structural survey efficiently, effectively,
and safely. DFP, which has been attainable only with the implementation of
fly-by-wire controlled aircraft and advances in computing power, will be the
basis for changes in how the discipline of CVS flight test is approached in the
future. DFP has not only been beneficial to flight test, but also has provided
the fleet aviator with an unprecedented ability to control the aircraft during
shipboard approaches, which will increase operational capabilities and make
naval aviation safer.

Fig. 52 First night shipboard evaluation.

502 T. WILSON

REFERENCES
[1] Jakab, P., “Eugene Ely and the Birth of Naval Aviation—January 18, 1911,”
Smithsonian National Air and Space Museum, 18 Jan. 2011, https://airandspace.si.
edu/stories/editorial/eugene-ely-and-birth-naval-aviation—january-18-
1911?lipi¼urn%3Ali%3Apage%3Ad_flagship3_feed%
3BB3eSYDI7RcKZ3oim3aV9Kg%3D%3D [retrieved 22 Aug. 2018].
[2] Kandler, E., and Pope, J., F-35C Ship Suitability Third Sea Trials (DT-III) Joint Test
Plan, F-35 Integrated Test Force, Patuxent River, MD, 2016 (unpublished).
[3] Cenciotti, D., “‘F-35 Unable to Land Aboard Aircraft Carriers’ Report Says. U.S. Navy
and Royal Navy Have Something to Be Worried About,” Aviationist, 9 Jan. 2012,
https://theaviationist.com/2012/01/09/f-35c-hook-problems/ [retrieved 22 Aug.
2018].
[4] Thomlinson, J., “A Study of the Aircraft Arresting-Hook Bounce Problem,” Ministry
of Supply, Aeronautical Research Council Reports and Memoranda, London,
England, 1957.
[5] Wedertz, R., Hess, E., Taylor, B., Stockton, K., and Antel, J., The Landing Signal Officer
Reference Manual, U.S. Navy Landing Signal Officer School, Virginia Beach, VA, 2010
(unpublished).
[6] Carrier Suitability Testing Manual, SA FTM-01, Naval Air Warfare Center Aircraft
Division, Patuxent River, MD, 1994.
[7] “NATOPS Landing Signal Officer Manual,” NAVAIR 00-80T-104, Naval Air
Technical Data and Engineering Service Command, San Diego, CA, 2001.
[8] “Aircraft Launching Bulletin No. 33-76,” Steam Catapult Performance Engineering,
Naval Air Technical Data and Engineering Service Command, San Diego, CA, 2016.
 CHAPTER 14

F-35 Aerodynamic Performance

Verification
David G. Parsons , Austin G. Eckstein† and Jeff J. Azevedo‡
Lockheed Martin Aeronautics Company, Fort Worth, TX

The combined efforts of the airframe contractor, propulsion contractor,

and government program office culminated in verifying the primary
conventional performance requirements of the F-35 aircraft. The joint
group collaborated to develop a cost-effective and credible modeling
and simulation-based approach to validate the engineering databases
used to calculate aircraft performance. Instrumental to the group’s
success was a rigorous weight management process, supported by an
incremental burndown of conservative factors applied to preflight test
performance predictions. These approaches ensured that the F-35 met
the key contractual performance requirements. Throughout our
efforts, we maintained attention to detail when performing focused
analyses of test data, from individual sensor measurements to
calculated performance parameters. These efforts allowed a minimal
flight test matrix to be sufficient in resolving even minor adjustments
to preflight test aerodynamics. Ultimately, the flight test results proved
that the design of the F-35 aircraft exceeded requirements. Further, the
results formed the basis of the operational performance capability
provided to the operator throughout the flight envelope.

I. NOMENCLATURE
a angle of attack
aX longitudinal acceleration
aZ normal acceleration
CD drag coefficient
CDf friction drag coefficient
CDmin minimum drag coefficient

 Aeronautical
Engineer, Senior Staff, Flight Sciences.
†
Aeronautical Engineering Manager, Flight Sciences.
‡
Aeronautical Engineer, Senior Staff, Flight Sciences.

503
504 D. G. PARSONS ET AL.

CDT test drag coefficient

Cf skin friction coefficient
CL lift coefficient
CLT test lift coefficient
d ambient pressure ratio
D drag
DE engine drag
FG gross thrust
g acceleration due to gravity
g flight path angle
K equivalent sand grain roughness
L lift
nZ normal load factor
nX axial load factor
q dynamic pressure
S reference wing area
u pitch attitude
W gross weight
V velocity

II. INTRODUCTION
This paper provides a top-level understanding of the approach taken and analyti-
cal techniques used on the F-35 during our performance validation. With these we
verified the conventional performance requirements of the F-35, with particular
emphasis on the Key Performance Parameters (KPPs) of the Joint Contract Spe-
cification (JCS). Our modeling and simulation-based verification process success-
fully validated the aerodynamics and performance databases used to calculate
performance with a minimal flight test matrix. Ultimately, the process indicated
that the mission performance of all three variants of the F-35 exceeded require-
ments. This paper presents the conventional aircraft performance management
approach and highlights some of the challenges, issues, and successes involved
in the verification process.

III. BACKGROUND
The F-35 is a fighter aircraft designed for use by the U.S. Air Force, Marine Corps,
and Navy, as well as the current 12 partner nations. The aircraft comes in three
largely common configurations/variants: F-35A Conventional Takeoff and
Landing (CTOL), F-35B Short Takeoff and Vertical Landing (STOVL), and
F-35C Carrier Variant (CV). Figure 1 provides an overview of the variant plan-
forms and basic configuration details.
F-35 AERODYNAMIC PERFORMANCE VERIFICATION 505

Fig. 1 F-35 variants.

Performance requirements for the F-35 were defined in the JCS. The primary
conventional aircraft performance requirement for each variant was the KPP
design mission radius. The F-35B (STOVL) had additional KPP requirements
for short takeoff distance and vertical landing bring-back to emphasize operations
aboard Navy LH-class amphibious assault ships. The approach to verifying each of
those requirements followed processes similar to those used for the conventional
performance requirement approach presented herein [1]. Vertical landing bring-
back is a measure of the aircraft’s ability to recover to the ship with unexpended
munitions and fuel. The F-35C (CV) had a separate KPP requirement for approach
speed to emphasize operations on Navy aircraft carriers [2].
The F-35 program used a modeling and simulation-based approach to deter-
mine and verify aircraft aerodynamics and performance. The wind tunnel testing
component of the program consisted of approximately 50,000 hours of testing
using 15 models in 23 tunnel facilities in seven states and four countries.
We used two force and moment models: a one-twelfth-scale model to baseline
for all aerodynamic databases, and a one-fifteenth-scale model for aerodynamic
store increments and most stability and control databases. We validated both
models to ensure the accurate representation of the aircraft’s outer mold line.
Figure 2 shows two of the installations.
We incorporated several innovations into the larger one-twelfth-scale model
to improve data quality. The model had two aft ends, which allowed for different
balance/sting installations. The smaller aft end accommodated a 2-in. balance and
sting, allowing it to be completely accurate to the airframe/nozzle interface. The
2-in. balance was also gauged to a lower capacity that resulted in higher-sensitivity
data with greater accuracy. We used the smaller aft end installation to gather data
across the Mach range to 15 deg a. The larger aft end had a 2.5-in. higher-
capacity/lower-sensitivity balance/sting and was used to gather data from 15 to
40 deg a.
506 D. G. PARSONS ET AL.

Fig. 2 F-35 wind tunnel model installations.

The one-twelfth-scale model also incorporated remote actuation of the hori-

zontal tails. This permitted us to change the configuration without opening the
tunnel. That, in turn, enabled us to gather far more data within the budget and
time constraints imposed on any wind tunnel program. The increased amount
of trailing-edge flap/horizontal tail interaction data greatly improved the accuracy
of the final aerodynamic database.
We used Computational Fluid Dynamics (CFD) extensively to complement
the wind tunnel testing. This included evaluating dozens of trade study options
with CFD to determine the most effective options to test in the wind tunnel
[3]. We also used CFD to account for items that could not be tested in the
tunnel, such as nacelle ventilation and the propulsion thermal heat exchanger
inlet. In addition, we used it to account for any minor configuration changes
that occurred after the wind tunnel testing had concluded. This ensured that
the aerodynamic database accurately represented the flight test aircraft before
entering the flight test phase.
F-35 wind tunnel testing also designed a secure network for communications
and data exchange between the wind tunnel test facilities and Fort Worth, Texas,
and between different tunnels. The wind tunnel data were automatically synced
among the source tunnel facility, other specified tunnel facilities, and Lockheed
Martin in Fort Worth. This allowed a limited number of engineers to run and
monitor multiple tests worldwide. At one point in the test program, 11 tests
were being run simultaneously.
Table 1 provides the total wind tunnel testing hours from the F-35 System
Development and Demonstration (SDD) program across all disciplines.
Figure 3 illustrates the F-35 three-degrees-of-freedom (3DOF) force and
moment accounting system. The data obtained during the wind tunnel testing
were supplemented with CFD analyses [3– 5]. The idea behind this approach
was to invest in accurate modeling of the configurations early in the program.
Doing so minimized, to the greatest extent possible, the relatively expensive
flight testing used to validate the databases.
F-35 AERODYNAMIC PERFORMANCE VERIFICATION 507

TABLE 1 F-35 SDD PROGRAM WIND TUNNEL TESTING HOURS

User Occupancy Hour Summary (Actuals thru Mar. 30, 2010)

Actual Hours per Test Discipline

Year Aero Stability & External Store Propulsion Flutter Total

Analysis Control Environment Separation Aero
2002 3975 1435 868 1175 2269 192 9914
2003 3464 3168 677 578 2060 – 9947
2004 1991 2582 810 373 2892 – 8648
2005 4054 3319 556 732 2935 – 11,596
2006 2677 2315 468 342 1253 – 7055
2007 780 – 308 445 250 – 1783
2008 – – – 264 – – 264
2009 – – – – – – –
2010 – – 777 – – – 777
Actual 16,941 12,819 4464 3909 11,659 192 49,984
thru
Mar.
30,
2010
Plan 16,941 12,819 4464 3909 11,659 192 49,984
Complete 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0%

Fig. 3 F-35 3DOF force and moment accounting system.

508 D. G. PARSONS ET AL.

IV. PERFORMANCE MANAGEMENT

At contract award, the F-35 program mandated the use of a conservative factor on
aircraft performance calculations to cover risks in several aspects of the new
design. The factor was implemented as a 5% increase in predicted fuel flow to
cover immaturity in the design, among other things. It also covered the uncer-
tainty in predicted aerodynamics and propulsion databases, as well as the possi-
bility of weight growth during the maturation of the design. Further, it covered
the potential for configuration changes resulting from discoveries during flight
testing that could adversely affect performance. As the aircraft design matured
and flight testing was accomplished, uncertainty in each of these areas was to
be retired. In tandem, the factor was to be progressively reduced until no conser-
vatism was to be applied for the final calculation of specification performance.
Two of the largest areas of uncertainty at contract award were the accuracy of
the predicted aerodynamics database and the process of adjusting subscale-based
testing results to a full-scale aircraft.

Fig. 4 F-35 aircraft performance calculation process.

F-35 AERODYNAMIC PERFORMANCE VERIFICATION 509

Fig. 5 F-35 performance capability tracking against tripwire and requirement.

From a simplified perspective, aircraft performance calculations involve com-

bining inputs from aerodynamics, propulsion, and mass properties databases, as
illustrated in Fig. 4. These are used to calculate mission and maneuver capabilities,
and changes in any of the databases can either improve or degrade calculated
aircraft performance. The F-35 program implemented an aircraft performance
management process early in the development program to ensure that the final
aircraft design would meet the KPP requirements. We calculated the F-35 capa-
bility against the JCS performance requirements and tracked the levels of per-
formance as the databases matured through wind tunnel testing and analyses.
For each requirement we developed tripwires to indicate when the margin to
the requirement had eroded enough to compromise meeting the requirement. If
predicted performance fell below a tripwire, we would initiate a recovery plan to
recover margin and improve performance. Figure 5 provides an illustration of how
the performance for a particular requirement changed over time. Each symbol in it
indicates a formal calculation of the requirement. For the early phases of the
program, aerodynamics, propulsion, and mass properties databases were being
updated frequently. As part of this, performance calculations for the specification
requirements were executed roughly once a month. As the design matured and stabil-
ized, changes that affected performance calculations became less common. Conse-
quently, the frequency of calculations of specification performance was reduced.
For the performance requirement in Fig. 5, calculated performance dipped
below the tripwire level in late 2012. A subset of maneuvers from the dedicated
performance flight test matrix was executed to evaluate the relative accuracy of
510 D. G. PARSONS ET AL.

the preflight test aerodynamics database. Results of this dataset analysis indicated
that the pretest databases provided a good match of measured performance, and
that uncertainty in the final level of aerodynamics was not high. From this analy-
sis, and in agreement with the F-35 Joint Program Office (JPO), we reduced the
fuel flow conservatism factor used for specification performance calculations
from 5% to 4%. The change in conservatism was enough to allow the calculated
performance to exceed the tripwire level. The conservatism in the fuel flow
factor was then incrementally reduced as the remaining dedicated performance
flight test matrix was executed and analyzed. Incremental reductions were also
the product of the flight test program maturing and not indicating the need for
configuration changes that would adversely affect performance.
In Fig. 5, the final point on the plot indicates the official JCS performance ver-
ification calculation where the fuel flow conservatism factor had been completely
removed. This represented the final validated flight test – level aerodynamic data-
bases, validated propulsion installation effects, and actual measured production
aircraft weight and fuel volume.
A key component of managing aircraft performance was to control weight
growth during the program’s design and development phase. Aircraft gross weight
typically grows as the design matures, so F-35 weight was actively managed, and
weight growth was limited to a projected increase with time [6]. The calculations of
aircraft performance against program requirements used the projected weight at

Fig. 6 F-35 weight management.

F-35 AERODYNAMIC PERFORMANCE VERIFICATION 511

the end of the SDD program. This was in contrast to the status weight at that particu-
lar point in time. Figure 6 illustrates F-35 weight growth with time and shows the
plan-to-perform line that the program used as a ceiling on weight increase with
time. We waited to implement design changes that might cause the weight to cross the
plan-to-perform line. Only after we could implement offsetting changes that reduced
weight to keep status weight below the target did we proceed with design changes.

V. FLIGHT TEST APPROACH

The overarching strategy chosen by the F-35 aerodynamics and performance ver-
ification team was one of database validation rather than performance demon-
stration. Aircraft performance predictions were calculated by combining
databases for aerodynamics, propulsion, mass properties, and landing gear (for
basing performance) through the laws of physics. The predictions’ calculations
also used the equations of motion defining the mission or maneuver in question.
Each variant had a unique design mission profile, but the approach to verify each
of those mission capabilities was common among the variants. The team’s strategy
was to acquire the data and complete the analysis necessary to validate the data-
bases used in those calculations. It then used the validated databases to calculate
verification performance.
The team’s approach provided several tangible and cost-effective benefits.
Among these, the amount of dedicated flight testing necessary to validate the data-
bases was considerably less than that for demonstrating performance for all air-
craft configurations, in all ambient operating conditions. Minimizing flight
testing provided significant cost savings. As another benefit, the contributions
of aerodynamics and propulsion to the overall performance of the aircraft were
able to accurately be allocated and validated. As a result, the amount of testing
for future aircraft or propulsion system modifications can be minimized
because testing can be initiated from a well-established baseline. Although this
effect does not generally provide immediate cost benefits at this phase of the
program, it will provide significant cost avoidance for future F-35 moderniz-
ation/enhancement contracts.

A. IN-FLIGHT THRUST AND PERFORMANCE WORKING GROUP

A key contributor to the success of the aerodynamics and performance verifica-
tion team’s approach was the early establishment of the In-Flight Thrust and Per-
formance Working Group (IFTPWG). Shortly after contract award, the IFTPWG
was established with representatives from Lockheed Martin, BAE Systems, Pratt &
Whitney, and Air Force and Navy personnel at the F-35 JPO. The group’s goal was
to collaborate on the testing and analysis necessary to credibly validate the data-
bases and successfully verify aircraft performance specification requirements.
512 D. G. PARSONS ET AL.

The IFTPWG leveraged its combined expertise in aerodynamics, perform-

ance, and propulsion testing to define and agree upon the analyses and testing
necessary to achieve database validation. By agreeing on the approach prior to
initiating testing, scope growth in expensive flight testing could be minimized
or prevented. The IFTPWG served as a forum for vetting and discussing all pro-
blems and results, and was key to fostering a cooperative working relationship
among all parties.
A critical component of the working group’s efforts was the In-Flight Thrust
Calculation Deck (IFTCD). The IFTCD was a Pratt & Whitney – supplied tool for
calculating thrust at any discrete point from flight or ground testing of the aircraft,
based on measured engine and aircraft parameters. Two independent methods of
calculating thrust were included as options for the IFTCD: airflow square root of
temperature and flow area pressure [7]. Each method relied on different par-
ameters to determine propulsive thrust for inflight conditions. One of the benefits
of having two independent methods was obtaining the ability to better identify
any systematic bias within one method.
Thrust and aerodynamics cannot be independently measured in flight. As
such, the IFTCD was crucial for determining aircraft aerodynamics from flight
test maneuvers. In Fig. 7, values for W, aX and aZ, V, and a can be directly
measured; however, the aerodynamics values of L and D and the propulsive
terms of FG and DE cannot be directly measured. To determine the aerodynamics
parameters, the propulsive terms are required to be known.
The IFTCD was calibrated against the results of engine development testing in
altitude chambers with measured load cells to improve the fidelity of the two cal-
culation methods. The IFTPWG continually evaluated options for increasing

Fig. 7 Conventional forces acting on an aircraft in level flight.

F-35 AERODYNAMIC PERFORMANCE VERIFICATION 513

fidelity in the flight test results. It used improved methods for individual par-
ameter measurement, along with specific ground and flight test techniques, to
reduce uncertainty in the end product. Each option was evaluated for its technical
feasibility and to assess its cost/benefit tradeoff. Any additional cost to the
program had to provide realistic, obtainable, and significant benefits to be
included as part of the verification process.

B. TEST MATRIX
The F-35 Aerodynamics and Propulsion teams entered the flight test phase of the
SDD program with a high-fidelity wind tunnel analytical database. Extensive work
was accomplished to ensure that all forces and moments identified in the force
and moment bookkeeping system were well-defined from wind tunnel testing
and CFD. This was particularly important when objectives were difficult to
achieve from testing. The comprehensive nature of the preflight databases was
key to successfully verifying aircraft performance requirements with a minimal
matrix of flight test maneuvers.
The matrix of dedicated aircraft performance flight test maneuvers was
designed to minimize required test flights. At the same time, it continued to
provide the data necessary to verify the KPP performance requirements and vali-
date the performance databases for all phases of flight. The goal was to validate a
credible collection of the databases that formed the basis of the performance pro-
ducts provided to the operator (e.g., flight manual, pilot checklist, performance on
the glass). The IFTPWG balanced fidelity in the final databases with the cost of
testing and analysis. Its members were constantly evaluating and refining the
test matrices due to programmatic pressure to reduce flight test costs and
shorten the schedule. In the end, the number of dedicated test points flown for
any one variant was roughly half the number used for previous fighter aircraft.
Table 2 summarizes the test points for clean configuration.
One element of the dedicated flight test matrix that supported verifying the
KPP design missions was the focus on the critical portions of each mission.
The Aerodynamics and Performance Validation team analyzed each design
mission to determine which segments consumed the most fuel. One or more dedi-
cated flight test point(s) were defined to anchor the validation of the databases.
This was done for each flight condition that represented a significant portion of
the design mission fuel usage (e.g., optimum altitude/optimum Mach cruise).
Figure 8 illustrates the flight envelope coverage of dedicated performance flight
test maneuvers.
The validation team created predictions for each flight test maneuver. It then
spotted the predictions onto the predicted drag polars and lift curves to ensure
sufficient coverage and avoid unnecessary duplication of conditions. It used a
combination of steady-state and quasi-steady-state maneuvers to evaluate the
lift and drag of the F-35 at a full range of sustainable a. Examples of such man-
euvers were cruise points, fixed-power sustained turns, Roller Coaster Maneuvers
514 D. G. PARSONS ET AL.

TABLE 2 CLEAN-AIRCRAFT DEDICATED PERFORMANCE MANEUVERS

Unique Dedicated Performance Test Points for Clean Aircraft Configuration

Maneuver F-35A F-35B F-35C
Speed Powers (cruise) 51 56 54
Powered Approach (gear down) Speed Powers 7 12 28
Roller Coaster Maneuvers 8 9 9
Accelerations 9 9 9
Decelerations 6 6 6
Sustained Turns 14 14 14
Climbs 8 8 8
Descents 4 4 4
Landings 8 8 8
Takeoffs 0 0 0
 Maneuver analyzed from ride-along data for other disciplines.

Fig. 8 Dedicated performance flight test maneuvers.

F-35 AERODYNAMIC PERFORMANCE VERIFICATION 515

Fig. 9 Example lift curve and drag polar.

(RCMs), and fixed-power accelerations. The Flying Qualities and Loads teams’
maneuvers were used to assess higher a aerodynamics and performance, rather
than executing dedicated maneuvers, for cost savings and greater efficiency.
Figure 9 illustrates the relative locations of multiple types of maneuvers on a pre-
dicted lift curve and drag polar. In this scenario, cruise maneuvers would have
been executed at multiple altitudes to provide the spread of CL; higher-altitude
maneuvers require higher CL values to sustain level flight at a given weight as a
result of lower dynamic pressure.
The team also utilized cruise maneuver repeats flown at identical W divided
by d conditions to evaluate the repeatability of aerodynamics results from flight
test. Flying the same Mach condition at a different weight and determining the
test altitude from the W/d equation drives both test points to the same CL. There-
fore, it drives them to the same point on the drag polar as well. For instance, if the
original test point was flown at a W of 45,000 lb and an altitude of 35,000 ft,
the W/d would be approximately 190,500. Then, a repeat test point could be
flown at 42,000 lb, and the test altitude would need to be approximately 36,400
ft. Consistency in drag levels from the two maneuvers, once standardized for alti-
tude, center of gravity, and engine operating conditions, would demonstrate the
repeatability of the flight test data.

C. UNCERTAINTY ANALYSIS
To make a minimal test matrix work, accurate and redundant sources of critical
instrumentation were required. An uncertainty analysis was accomplished early in
the program. With it, we identified the parameters that drive accuracy in the aero-
dynamics and performance data calculated from flight test maneuvers. The results
of the analysis provided insight into how best to utilize limited resources to
improve the final product.
The uncertainty analysis process involved creating influence coefficients for
each sensor measurement that was an input to the aerodynamics and performance
data reduction process. The influence coefficients varied with test condition and
516 D. G. PARSONS ET AL.

maneuver type and needed to be evaluated across the expected test operating
range. These influence coefficients were then combined with values for sensor
accuracy to determine the uncertainty in the calculated aerodynamics and per-
formance parameters. Realistic sensor measurement accuracies are required for
a realistic uncertainty analysis. The results of the analysis allowed the team to
focus its efforts on the parameters that were significant drivers in the uncertainty
of the results. In so doing, the team avoided wasting resources on measurements
that had little or no effect on the results.
The detailed uncertainty analysis also identifies where one method of inflight
thrust calculation would have less uncertainty than another. Just as the influence
coefficients varied with condition and maneuver type, so did the analytical accu-
racy of the two inflight thrust methods.

VI. FLIGHT TEST ANALYSIS RESULTS

The effort to validate aerodynamics and performance databases overcame several
technical challenges during the course of analysis. Aircraft a is a critical measure-
ment that influences the accuracy of the derived aerodynamics data from flight
test. The Aerodynamics and Performance Verification team worked with the
Air Data team to refine the accuracy of the flight test nose boom data beyond
what was required for the other disciplines. The teams used all of the dedicated
performance cruise data acquired for each variant to develop a secondary a cali-
bration. They analyzed the difference between the primary corrected nose boom a
and an inertially derived a determined from the inertial navigation sensors on the
aircraft. The primary nose boom correction mostly adjusts for up wash and pitch
rate. Equation (1) shows the formula used in the teams’ calculation for inertial a.
aINERTIAL ¼ u  g (1)
Equation (1) assumes there are no ambient vertical winds, and this is
generally a reasonable assumption; however, a detailed analysis conducted by
the Validation team revealed that certain maneuvers exhibited higher levels of
vertical wind than others. This applied to maneuvers conducted near the moun-
tains around Edwards Air Force Base, compared to maneuvers conducted away
from the mountains or over the water at Naval Air Station Patuxent River.
These results are consistent with ambient winds traveling up and around the
mountains. The winds can also be seen generating increased levels of turbulence
during the maneuvers. High-rate aircraft acceleration and a data were also filtered
to reduce spurious noise in the signals. All these refinements led to more accurate
lift and drag data with reduced scatter.
Aircraft W is another parameter for which the team looked to reduce uncer-
tainty. The test team employed a process to increase the fidelity of aircraft W at
any point in time. The test aircraft was weighed with a set of portable scales fol-
lowing on-ground fueling and prior to any test flight. The aircraft was weighed
F-35 AERODYNAMIC PERFORMANCE VERIFICATION 517

again following the test flight. The pretest and posttest weighings provided anchor
points used to determine aircraft W at any point during the flight. The amount of
fuel burned from each anchor point was integrated for the entire flight. This deter-
mined the amount of fuel and the weight of the jet at any discrete point during the
flight. This process reduced uncertainty from the fuel gauging system due to fuel
slosh or aircraft attitude.
Fuel flow is a primary driver in aerodynamics and performance calculations. As
mentioned previously, fuel flow is used to help determine aircraft weight at any dis-
crete point in time. It is also a primary input for the inflight thrust calculation
methods and a primary driver in inflight thrust calculation accuracy. The team
used test flow meters for both gas-generator and afterburner fuel flow. This
increased the accuracy in the fuel flow measurement over the production source of
fuel flow for the engine control. The team intended to use the production fuel flow
measurement as a backup in case of sensor failure of the test flow meter; however,
the installation of the test flow meter caused an unintended consequence for the
Initial Service Release (ISR) engine standard used in F-35A and F-35C testing.
The software for the ISR production engine control included several correc-
tions to improve the accuracy of the production fuel flow measurement. With
the test flow meter installed, one of the corrections was no longer applicable,
introducing an error of up to 3% to the production measure of fuel flow. The
team did not discover this problem until the difference was revealed by an
engine stand run to calibrate the test flow meter. Correcting the error would
have required a modification to the production engine control software, which
was deemed unacceptable due to its effects on cost and schedule. As a result,
for the ISR engine standard, the production fuel flow measurement could not
be used as a backup for the test flow meter. This problem was not present for
testing with the first flight release engine standard used for F-35B conventional
performance testing.
After having acquired all test data for a particular variant, we processed the
raw test data through a set of data reduction routines. With these, we calculated
test-day aerodynamics and performance parameters, and then standardized those
parameters to a set of reference conditions to compare them to the pretest pre-
dicted databases. The standardization process involved analytically adjusting
data from test-day conditions to the reference conditions via predicted incre-
ments. Standardization allowed multiple maneuvers to be evaluated against
pretest predictions, rather than evaluating each point against a test-day prediction.
This process allowed us to readily identify trends between/among prediction and
flight test results.
Aerodynamics data for the aircraft in an up-and-away configuration (i.e., gear
up) were standardized to the nearest database breakpoint Mach number, 36,089-ft
pressure altitude, a constant center of gravity, and a set of engine operating par-
ameters generally consistent with cruise conditions. The altitude serves as the
baseline condition for the aerodynamics database. Aerodynamics data for the air-
craft in a powered-approach configuration (i.e., gear down) followed that process
518 D. G. PARSONS ET AL.

but were modified due to the difference in typical operating altitudes between an
up-and-away and a powered-approach configuration; the latter occurs nearer to
sea level so that sea level serves as the baseline for powered-approach configur-
ation aerodynamics. All aerodynamics standardization was completed holding
test a constant because the databases are a function of a, rather than of CL.
Performance data standardization varied slightly by maneuver type. All per-
formance data were standardized to flight design gross weight, center of gravity,
and standard day temperature. Cruise maneuvers were standardized to the
nearest 50-foot pressure altitude and wings level, constant altitude, and unaccel-
erated flight conditions holding test Mach number constant. Acceleration and
deceleration maneuvers were standardized to the chosen power setting (idle, mili-
tary, or maximum afterburner), nearest 5000-ft pressure altitude, and wings-level
attitude. We held the test Mach number constant for each discrete point in the
maneuver. Climb maneuvers were standardized to the chosen power setting
and desired climb schedule, holding the test altitude constant for each discrete
point in the maneuver.
The standardized aerodynamics data were evaluated against predicted lift
curves, drag polars, and trimmed horizontal tail deflections at each database
Mach number. We used this approach, with the a and acceleration filtering dis-
cussed previously implemented to reduce scatter. In our assessment, the flight-
derived aerodynamics data were generally found to reconcile among maneuver
types with varying power settings and test altitudes. As expected, hysteresis was
noted in some maneuvers (mainly RCMs) near the lift curve break. This occurred
where the flow separations differ when pitching up and pitching back down, being
attached to separated when pitching up and separated to attached when pitching
down. The resultant data quality was very good, and the final flight test-basis aero-
dynamics levels were readily defined using standard regression curve fitting
techniques.
By evaluating the data as a whole, the team was able to identify a trend with
altitude that had not been present in the predictions. The difference was attributed
to the predicted skin friction roughness used to adjust the subscale pretest wind
tunnel – based database to full-scale conditions. Figure 10 illustrates the predicted
variation of minimum coefficient of drag with altitude for different levels of rough-
ness. The pretest aerodynamics database estimated a roughness of 250 m in.,
whereas the flight test data indicated the roughness to be closer to 450 m in.
The skin friction roughness was updated to be more representative of the
results of flight test and the comparison process accomplished again. Further aero-
dynamic investigations revealed transonic drag rise differences from the wind
tunnel – predicted aerodynamics data. These were attributed to differences in
shock locations in flight, as well as complex interactions among the close-coupled
wing/empennage, including power influences. Wind tunnel and flight data from
the F-35 Loads and Flying Qualities teams were leveraged for these analyses.
Fairings of the flight test data were created to facilitate the development of
untrimmed flight test increments to the pretest databases. The implementation
F-35 AERODYNAMIC PERFORMANCE VERIFICATION 519

Fig. 10 Effect of roughness on variation of CDmin with altitude.

of the increments provided a flight test – basis aerodynamics database across the
full operating envelope.
In addition to conducting the aerodynamics analysis, we calculated residual
differences for each flight test maneuver from the adjusted database. A thorough
evaluation of these residuals allowed the team to identify adjustments to the pre-
dicted jet effects that further improved the fidelity of the aerodynamics databases.
Figure 11 illustrates a residual drag variation with nozzle exit area present in the
data. After correcting the aerodynamics databases for the observed flight test jet
effects, the resolution among all flight test maneuvers was approximately five
counts of drag (0.0005 CD) for cruise conditions. Samples of a lift curve and
drag polar illustrating the final flight test results against original and flight
test –based predictions are provided in Figs. 12 and 13, respectively.
After the aerodynamics analyses were completed, the adjusted database was
used to create predictions for the standardized performance data. As had been
done with the aerodynamics analysis, the performance data were evaluated as a
whole rather than maneuver by maneuver. The team compared the performance
predictions and the standardized test data. This allowed its members to validate
that the adjustments made to the aerodynamics databases produced performance
data that matched flight test results. Where differences still existed, further inspec-
tion led to identifying differences between the predicted engine model and the
actual test engine. Some of these differences were due to different levels of
engine bleed flow between the predicted model and the test engine. By contrast,
others could be attributed to the number and severity of operating hours on the
test engines. Although the inflight thrust methods typically accounted for these
differences, the predicted steady-state engine model used to create predicted per-
formance did not.
520 D. G. PARSONS ET AL.

Fig. 11 Residual drag difference with nozzle area.

Fig. 12 F-35 flight test lift curve.

F-35 AERODYNAMIC PERFORMANCE VERIFICATION 521

Fig. 13 F-35 flight test drag polar.

Fig. 14 Flight test cruise fuel flow vs predictions at 35,000 ft.

522 D. G. PARSONS ET AL.

Fig. 15 Flight test cruise net thrust required vs predictions at 35,000 ft.

Figures 14 and 15 compare original predictions, flight test – based predictions,

and standardized flight test data for cruise conditions at 35,000 ft. Figure 14
depicts the fuel flow, and Fig. 15 depicts the net thrust required. The original pre-
flight test predictions are represented by the solid blue line, with dashed lines
representing a deviation from that level of 5%. The flight test – based predictions
are represented by the solid green line, and the standardized flight test data are
represented by the symbols, one for each flight test maneuver. These plots are
representative of the comparison of flight test data to predictions for all maneu-
vers in all conditions.
After completing all flight test analysis and obtaining the F-35 JPO’s concur-
rence, we used the validated aerodynamics databases to calculate the KPP mission
performance for each variant.
The results showed that each of
the variants exceeded the JCS
requirement for mission range
by more than 10%, as illustrated
in Fig. 16.

Fig. 16 Mission radius capabilities

vs JCS requirements per F-35 variant.
F-35 AERODYNAMIC PERFORMANCE VERIFICATION 523

VII. CONCLUSION
On the F-35 program, we successfully implemented a modeling and simulation-
based approach to aerodynamic performance verification. Applying conservatism
to performance calculations early in the program protected against potential
uncertainties in configuration, weight, or aerodynamics levels. Our rigorous
process controlled aircraft weight growth and helped to ensure that the perform-
ance of the final F-35 design met the KPP requirements of the program specifica-
tion. The efforts of a government/contractor team culminated in delivering a
credible, flight test – based aerodynamics and performance database that accu-
rately represents the performance of the F-35. This will be applicable for not
only specification verification, but also the operational performance products
used by the fleet.

ACKNOWLEDGMENTS
D. G. Parsons thanks Brian Losos, Scott Rudin, and Jason Westly for their contri-
butions to this paper and the analysis of the F-35 performance flight test data.

REFERENCES
[1] Levin, D. A., Parsons, D. G., Panteny, D. J., Rask, M. R., and Wilson, P., “F-35 STOVL
Performance Requirements Verification,” 2018 (to be published).
[2] Wilson, M. A., “F-35 Carrier Suitability (CVS) Testing,” 2018 (to be published).
[3] Wooden, P. A., and Azevedo, J. J., “Use of CFD in Developing the JSF F-35 Outer
Mold Lines,” AIAA Paper 2006-3663, 2006.
[4] Wooden, P. A., Smith, B. R., and Azevedo, J. J., “CFD Predictions of Wing Pressure
Distributions on the F-35 at Angles-of-Attack for Transonic Maneuvers,” AIAA
Paper 2007-4433, 2007.
[5] Karman, S. L., Jr., and Wooden, P. A., “CFD Modeling of F-35 Using Hybrid
Unstructured Meshes,” AIAA Paper 2009-3662, 2009.
[6] Counts, M. A., Kiger, B. A., Hoffschwelle, J. E., and Houtman, A. M., “F-35 Air
Vehicle Configuration Development,” 2018 (to be published).
[7] Vorwerk, A. V., and Ciszek, R. S., “Use of In-Flight Thrust on JSF Program,”
International Powered Lift Conference, Oct. 2010.
 CHAPTER 15

F-35 High Angle of Attack Flight

Control Development and Flight
Test Results
Daniel G. Canin Jeffrey K. McConnell† and Paul W. James‡
Lockheed Martin Aeronautics Company, Fort Worth, TX

The defining requirements of the F-35 relate to its range, performance,

and 5th Generation combat capabilities, including all-aspect stealth,
sensor fusion and network-enabled operations. Meeting signature and
performance requirements while providing a robust high Angle-of-
Attack (AOA) maneuvering capability presented challenges in configur-
ation and flight control design. The selection of Nonlinear Dynamic
Inversion (NDI) as the flight control methodology, with its dependence
on highly accurate aircraft state information and onboard aerodynamic
modeling, presented additional challenges when applied to the high
AOA regime. This paper describes the F-35 high AOA flight control
system development and testing conducted during the System Develop-
ment and Demonstration (SDD) phase, with an emphasis on control law
development, technical challenges, and key flight test results (Fig. 1).

Fig. 1 Author flying F-35C test aircraft CF-5 through a high AOA maneuver
(8 January 2014).

 Experimental Test Pilot (High AOA Lead, F-35B/C), Lockheed Martin.

†
F-35 Stability and Control Lead Engineer, Lockheed Martin.
‡
F-35 Control Law Design Engineer (High AOA Lead), Lockheed Martin.

525
526 D. G. CANIN ET AL.

I. NOMENCLATURE
a angle of attack
a-dot rate of change of angle of attack
b sideslip angle
b-dot rate of change of sideslip angle
w, u, c Euler roll, pitch and yaw angles
Cl, Cm, Cn body-axis aerodynamic rolling, pitching, and yawing moment
coefficients
Clb, Cnb body-axis aerodynamic rolling and yawing moment due to
sideslip
Cnb, dynamic departure susceptibility parameter
p, q, r body-axis pitch, roll and yaw rates
Pb body-axis roll rate
Pb-dot body-axis roll acceleration
q, qbar dynamic pressure
Nz longitudinal acceleration

II. INTRODUCTION
The unique tactical capabilities of the F-35, including all-aspect stealth,
advanced sensor fusion, and network-enabled operations, support a strong
argument that Within-Visual-Range (WVR), high Angle-of-Attack (AOA) man-
euvering (i.e., dogfighting) may be a thing of the past. Indeed, the advent of
high-off-boresight air-to-air missiles alone makes the close-in, low-speed engage-
ment something to be avoided in any fighter. That said, there are cases in which a
WVR engagement may be unavoidable, such as when rules of engagement require
a visual identification; when onboard missiles have been depleted; or when the
aircraft is required to dwell in a threat area to defend assets on the ground.
Even if these scenarios are rarely encountered in combat, training for them
results in routine exposure to the high AOA regime, with the attendant risk of
Out-of-Control Flight (OCF).
With that understanding, the F-35 was required to provide high AOA maneu-
vering capabilities comparable to that of legacy fighters without compromising
its more critical 5th Generation tactical capabilities. Specifically, the F-35 was
required to:

. Be capable of performing air-to-air tracking tasks up to the stall AOA.§

. Provide positive and predictable pitch, yaw, and roll control response in the
post-stall flight regime.
§
For the purpose of these requirements, the F-35’s stall AOA was defined as the AOA for maximum coefficient of
lift (CLmax).
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 527

. Be departure resistant.}
. Recover from all out-of-control modes using aerodynamic control with
minimum pilot input.
The challenge for the F-35 design team was to extract the required high AOA
maneuvering capability from an aircraft whose configuration was driven almost
exclusively by other requirements.

III. AIRCRAFT DESCRIPTION

The F-35 is a family of single-seat, single-engine, all-weather stealth fighters
designed for ground attack and air superiority missions. The family consists of
the three variants shown in Fig. 2: the F-35A Conventional Takeoff and
Landing (CTOL) variant, the F-35B Short Takeoff and Vertical Landing
(STOVL) variant, and the F-35C Carrier-based Variant (CV).

A. FLIGHT CONTROL SURFACES

The flight control surfaces of the F-35 are conventional, consisting of Trailing
Edge Flaps (TEFs), Horizontal Tails (HTs), dual rudders, and full-span Leading
Edge Flaps (LEFs). Ailerons are incorporated on the CV variant only. The LEFs
are scheduled symmetrically with Mach number and AOA to optimize perform-
ance and to improve lateral/directional stability at elevated AOAs but are not
actively used for control.

Fig. 2 F-35 variants.

}
By departure resistant, the intent was not necessarily to be carefree or departure-proof, if doing so would
unnecessarily restrict capability. The desire was to incorporate appropriate flight control features that would
protect the aircraft from loss of control during reasonably expected maneuvers.
528 D. G. CANIN ET AL.

The design features that affect high AOA aerodynamics differ significantly
among the three variants:
. The STOVL variant has a similar planform to CTOL, but the upper fuselage
contours associated with its powered lift system have an important influence
on vortex behavior at elevated AOAs. More significantly, the lower
maximum speed and Nz requirements of the STOVL aircraft allowed the
area of the HT to be reduced by nine percent relative to CTOL’s. While the
STOVL aircraft has a vectoring nozzle for powered lift operations, it is not
used at high AOA.
. The CV planform is the most obviously different from the other two. The
requirement for shipboard launch and recovery resulted in a wing that is 45
percent larger than that of the other two variants, and the addition of ailerons
to enhance roll and flight path control at low speeds. Additionally, the CV’s
horizontal and vertical tails are much larger due to the aircraft’s lower
approach speed and greater inertia.

B. CONTROL SURFACE MONITORING AND ACTUATION

The F-35 is the first manned fighter to use electric power to drive its primary flight
control surfaces. The TEFs, HTs, and rudders are driven by Electro-Hydrostatic
Actuators (EHAs), each of which is comprised of an integral 270V motor, hydrau-
lic pump, and reservoir [1]. Relative to legacy hydraulic systems, EHAs improve
the maintainability, reliability, and survivability of the flight control system, but
introduce challenges of their own. Since the direction of the pump rotation deter-
mines the direction of surface movement, the motor and pump must reverse
direction with each control surface reversal. The motor is assisted in this reversal
by regenerative braking during the deceleration cycle. At low airspeed and high
AOA, the control surfaces are often operating at their rate limits, with frequent
reversals, resulting in a significant amount of regenerative energy. Dissipating
this energy presented a significant challenge in the design and the cooling of
the EHAs and associated power switching equipment.
Control surface rates and deflections, the limits of which are shown in Table 1,
are monitored continuously by the flight control system. This feedback is used not
only for failure detection and accommodation, but also to improve high AOA
flying qualities and Departure Resistance (DR), as described later.

C. FUEL SYSTEM AND STORES MANAGEMENT SYSTEM

The F-35’s use of Nonlinear Dynamic Inversion (NDI) requires the aircraft’s mass
properties to be known with sufficient accuracy and redundancy to be used as
primary inputs to the Control Laws (CLAW). This dependency places stringent
requirements on the accuracy, reliability, and failure detection capabilities of
the aircraft’s fuel quantity measurement and stores management systems. The
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 529

TABLE 1. CONTROL SURFACE DEFLECTION LIMITS

(IN DEGREES)

Surface F-35A/B F-35C

LEF 3 40 3 36
Flaps 30 30 30 30
HT 30 25 24 27
Rudder +30 +30
Aileron 30 30

payoff for this investment is the ability to optimize handling qualities over a wide
range of loadings and fuel states, without pilot interaction. This capability is par-
ticularly valuable at high AOA, where maneuvering performance and command
limits can be closely tailored to mass properties, obviating the need for overly
conservative (worst case) limits as are typically invoked on legacy fighters.

D. AIR DATA SYSTEM AND TACTICAL NAVIGATION SYSTEM

The F-35 pneumatic Air Data System (ADS) uses two dog-leg probes and two
flush-mounted ports that are located on the forebody below the chine, just aft
of the radome (Fig. 3). These sensors are used to measure AOA, angle of sideslip,
static pressure (Ps), and total pressure (Ptotal). At very low speeds or extreme AOA
or sideslip angles, where the pneumatic measurements become unreliable, the
system transitions to inertially derived AOA, sideslip, and Ptotal. To support the
calculation of these inertial parameters, the velocity of the air mass (wind) is con-
tinuously computed whenever the aircraft is in the pneumatic air data envelope.
When outside the pneumatic air data envelope, the stored wind information is

Fig. 3 Air data system flush port and multifunction probe.

530 D. G. CANIN ET AL.

Fig. 4 Pneumatic and inertial air data envelopes.

used along with the aircraft’s inertial velocity to derive the required air
data parameters.
The AOA thresholds for transitioning between pneumatic and inertial
sources, illustrated in Fig. 4, evolved significantly during the high AOA test
program. Although a flight test nose boom was available, it was not installed
during high AOA testing because of its potential aerodynamic impact. At the
beginning of each flight for which excursions outside the pneumatic envelope
were anticipated, a “wind calibration” climb or descent through the test altitude
band was performed. Wind data collected in this manner combined with
aircraft inertial attitude and velocity provided an independent source of speed,
AOA, and sideslip for real-time monitoring and post-flight analysis.

E. STABILITY AND CONTROL

The following section provides a brief overview of the aerodynamic characteristics
that most impact the F-35’s ability to maneuver at high AOA and to recover
from OCF.

LONGITUDINAL STABILITY AND PITCH CONTROL POWER

The low-speed pitching moment characteristics of the STOVL variant at aft
Center of Gravity (CG) are presented in Fig. 5. The STOVL variant was selected
 Aircraft
attitude and inertial velocity are provided with very high reliability in the F-35 by the Tactical Naviga-
tion System (TNS). This system provides a redundant, six-degree-of-freedom (6-DOF) inertial solution, required
to support hover in STOVL mode, and is an integral part of the flight control system in all variants of the F-35.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 531

for this illustration because it is the most pitch-challenged of the three variants as a
result of its smaller HT. The solid lines represent the pitching moments associated
with maximum nose-up, maximum nose-down, and zero HT positions, while the
dashed line shows the reduction in nose-down capability with the Weapon Bay
Doors (WBDs) open.
All three variants are neutrally or negatively stable at low AOA, and strongly
stable at high positive and negative AOA. The reduced size of the STOVL
aircraft’s horizontal tail makes it the most susceptible to deep stall (also known
as a pitch hang-up) at AOAs beyond the CLAW limiter. The CTOL variant
exhibits the potential for a weak deep stall only at extremely aft CGs achievable
at very low fuel states. The CV variant, which has the largest HT of the three,
does not exhibit deep stalls even in worst-case conditions. Inverted deep stalls
are not a concern in the F-35 since all variants can generate positive pitching
moment for all loadings across the full range of negative AOAs.
Each of the F-35’s two weapon bays are equipped with two doors, a small
door near the aircraft centerline and a larger outer door. The pitch effect of the
inner doors is negligible, but the outer doors generate significant nose-up
moment when open, shown by the dotted line in Fig. 5. Normally, the WBDs
close automatically after internal weapons employment, so a deep stall due to
their effect would occur only if they failed open while at high AOA with an
aft CG. In keeping with the objective of ensuring recovery from the worst-case
departures, recoveries were demonstrated for these conditions despite their
remote probability.

Fig. 5 Longitudinal stability.

532 D. G. CANIN ET AL.

LATERAL/DIRECTIONAL STABILITY
The lateral/directional stability characteristics of the F-35 are typical of twin-
tailed, swept-wing fighters. Fig. 6 provides a representative illustration of the
variation in static lateral/directional stability with AOA for a low-speed, aft-CG
condition. Directional stability (Cnb) is slightly stable at low AOA but becomes
unstable at higher AOA as the vertical tails are increasingly immersed in the
wing/forebody wake. Dihedral effect is stable (Clb negative) at all AOAs, with
a minimum at moderate AOA followed by a steady increase beyond that.
Overall, the bare airframe provides some level of departure resistance (positive
Cnb, dynamic) throughout the positive AOA range, with a notable reduction in
the medium a range.

ROLL/YAW CONTROL POWER

For any fighter, high AOA agility is determined primarily by the amount of
control authority available to generate roll and yaw moments. The effects of
AOA on roll/yaw surface effectiveness in the F-35 are shown in Figs. 7 and 8.
In each figure, the vertical scales are common for the three effectors to show
their relative capability. The red, green, and blue lines represent increasing
deflection of the control surfaces.
Plots of roll control (Fig. 7) show that the HTs and TEFs produce similar levels
of rolling moment across the full AOA range. As AOA increases, the change in
roll moment per degree of deflection reduces notably for the TEF and HT near
the maximum trailing-edge-down deflection limits.

Fig. 6 Lateral/directional stability.

F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 533

Fig. 7 Roll control power of the left rudder, HT, and flap.

In the yaw axis (Fig. 8), the most obvious control challenge is the near-
complete loss of rudder effectiveness as AOA increases in the medium AOA
range. By contrast, differential HT produces little yaw at low AOA but becomes
the primary yaw effector at higher AOAs. Note that at high AOAs, the incremen-
tal increase in yaw control power of differential HT (i.e., the change in yaw effect
with a change in HT deflection) decreases at large deflections. This has important
consequences when the HTs are commanded to large deflections (symmetrically)
to meet pitch axis demands, as their ability to generate yaw is greatly reduced.
To illustrate how limitations in rudder control power impact maneuverability
at high AOA, consider a 1 g roll produced by 15 degrees of differential flap.
Figure 9 compares the amounts of rudder and differential HT required to coor-
dinate such a roll when each is used independently. As AOA increases the
yawing moment required increases, with the rudder being the preferred surface
since it produces relatively pure yaw with minimal opposing roll. With further
increases in AOA, the decrease in effectiveness of the rudder ultimately result
in it becoming saturated, requiring differential HT to coordinate the roll. Differ-
ential HT, however, generates rolling moment opposing that of the flaps, thereby
reducing the roll rate achievable. Additionally, the amount of differential HT
available for yaw may be limited by the amount of symmetric HT needed for

Fig. 8 Yaw control power of the left rudder, HT, and flap.
534 D. G. CANIN ET AL.

Fig. 9 Rudder vs. differential

HT needed to coordinate a
representative roll.

pitch, which has priority.

Proper allocation of the
HT between the pitch and
yaw axes is critical to maxi-
mizing maneuverability at
high AOA while avoiding
departure.

IV. CONTROL LAW DESIGN

The F-35 is the first production fighter to incorporate NDI as the control allo-
cation scheme for its Fly-By-Wire (FBW) control system [2, 3]. As distinguished
from legacy FBW schemes, which use preprogrammed (scheduled) gains, the NDI
approach allows the aircraft to determine a control solution on the fly, based on a
detailed Onboard Model (OBM) of the aircraft’s mass properties and stability and
control characteristics.
In the NDI implementation (Fig. 10), the desired flying qualities (aircraft
responses to stick and pedal inputs) are contained in the front end, the command
and regulator modules. The outputs of these modules are the pitch, roll, and yaw
accelerations needed to achieve those responses. The incremental effector com-
mands needed to provide these accelerations are computed in the back end by
the Effector Blender (EB) based on information received from the OBM.
To support this task, the OBM provides two types of information as a function
of surface positions, configuration, mass properties, and flight condition: The
aircraft’s linear and angular accelerations, and the effectiveness of each control
surface pair. Using this information, the EB determines the additional moments

Fig. 10 Control law structure.

F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 535

required to produce the desired changes in angular acceleration, and the changes
in surface deflections needed to produce those moments. An error in the OBM
(i.e., a difference between the aerodynamic model and the actual aircraft aerody-
namics) for either the accelerations or the control effectiveness estimates will
result in persistent errors in the control solution.
Of the many considerations that led to the selection of NDI for the F-35, a
primary factor was NDI’s ability to manage the complex propulsive and aerody-
namic control allocation challenge in STOVL. The applicability of NDI to the high
AOA flight regime, however, was not seen as an NDI strongpoint. Specific con-
cerns included the ability to accurately model highly nonlinear and asymmetric
stability trends, complex control interactions, and significantly reduced control
surface effectiveness. Prior to initiating detailed design, therefore, it was necessary
to evaluate the potential benefits and disadvantages of using NDI at high AOA.
The conclusion of that analysis was that, although there were areas of concern,
there were no issues that would preclude the use of NDI throughout the high
AOA envelope or would require fundamental changes to the basic NDI approach.
Indeed, the extensive information contained in the OBM had the potential to
provide significant control benefits at high AOA.

A. HIGH AOA CLAW MODES

The Up-and-Away (UA) control laws of the F-35 at low-to-moderate AOA are
typical of most FBW fighters: Longitudinal stick commands pitch rate at low
speed and Nz at high speed; lateral stick commands roll rate around the flight
path; and pedals command sideslip.
As AOA increases, however, the directional response to large pedal inputs
blends from sideslip command to yaw rate command (Fig. 11). In this mode,
the pilot has no direct control over the roll axis. CLAW controls roll rate and
sideslip, exploiting the natural stability of the aircraft (via Cnb and Clb) to drive
roll and yaw. This stra-
tegy significantly enhances
roll and yaw performance
beyond that which could
be achieved using control
surfaces alone.
In the high AOA mode,
the pilot commands yaw
rate primarily with the
pedals. Lateral stick input
increases or decreases the

Fig. 11 Up-and-away
control laws.
536 D. G. CANIN ET AL.

commanded rate, depending on whether the stick is displaced into or opposite

the pedals. The intent of this approach is to allow the pedals to be used for gross
maneuvering, with the stick used for fine tuning of that command.
Figure 12 and Fig. 13 compare bank-to-bank maneuvers using the two lateral-
directional command modes: roll rate command using lateral stick only, and yaw
rate command using stick and pedal.
. For the stick-only case (Fig. 12), the commanded parameter is stability axis
roll rate, which exhibits a well-controlled first-order response. Yaw rate
follows the roll rate, resulting in near-zero sideslip.
. When pedal is used (Fig. 13), the primary command parameter is body axis
yaw rate, which is well-controlled with only minor overshoots. Rather than
attempting to coordinate the roll, CLAW commands proverse sideslip, using
the natural stability of the aircraft (Cnb and Clb) to generate roll and
yaw moments.

Fig. 12 Bank-to-bank roll using roll stick only (roll rate command).
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 537

Fig. 13 Bank-to-bank roll using full pedal and roll stick (yaw rate command)

B. DEPARTURE PREVENTION
One of the most commonly held perceptions among pilots of highly augmented
fighters is that CLAW features incorporated to prevent departures (most
notably AOA and roll/yaw rate command limiters) are unnecessarily intrusive
and prevent pilots from getting the maximum maneuverability from the
airplane. Historically, there is some validity to this perception. Less sophisticated
designs required more conservatism to provide the required levels of departure
protection. The complexity of the original F-16 CLAW, for example, was con-
strained by limits in computing power and aircraft state information available
to the CLAW, resulting in compromises and simplifications in the design of its
DR features. The F-16 pitch axis limiter, for example, was based on a simple
schedule of AOA and G. The consequence of that simplification was that as
AOA increased, the available G decreased (with AOA still below the absolute
AOA limit), leaving some amount of maneuvering capability unavailable. By
contrast, the F-22 was able to exploit its greater computing power and aircraft
538 D. G. CANIN ET AL.

state information to incorporate more complex and less intrusive DR features. The
F-35 continued this trend, using its NDI control structure and significantly
increased state information to support a DR logic that extends well beyond tra-
ditional command limiters. By continuously monitoring measured and projected
aircraft states (Mach, dynamic pressure, angular rates, mass properties, effector
positions, and OBM-predicted accelerations), the F-35 CLAW anticipate and
respond to potential departure conditions in the least intrusive manner possible.
Brief descriptions of the F-35 AOA and command limiters are presented in the
following subsections. Although the limiters are described individually, it is
important to note that they do not act in isolation. Whereas legacy systems set
limits in each axis based on the worst-case inputs in other axes, the F-35 takes
a more adaptive approach, establishing baseline limits for each axis and then
modifying them in real time based on the rates and commands in the other
axes. Designing and tuning these features was one of the more challenging
aspects of the high AOA CLAW development. The result is an aircraft that is
extremely resistant to departure while wringing as much capability as possible
out of its controls.

AOA LIMITERS
The AOA limiter in the F-35 represents a significant leap in complexity and
performance relative to legacy aircraft. Inputs to the limiter computation
include longitudinal CG, lateral asymmetry, Mach, external stores configuration,
WBD position, control surface failures, and failures of the air data, fuel gauging,
electrical, and tactical navigation systems. These various AOA limiter schedules
run in parallel, with the most restrictive being applied.
At low speed, the primary focus of the AOA limiter design was the preser-
vation of nose-down pitch acceleration capability (pitch margin). There is a
rich history of research in this area, and a number of guidelines were referenced
in selecting pitch margin criteria that would be acceptable across all variants,
loadings, and fuel states [4– 6]. The result of this study was the selection, as
the low-speed criteria, of – 0.20 rad/sec2 in UA mode and – 0.15 rad/sec2 in
Powered Approach (PA) mode. Based on these criteria, AOA limiter schedules
were developed as a function of CG, WBD position and other factors affecting
pitching moment. Secondary effects on pitching moment, such as inlet momen-
tum, jet effects, and non-zero sideslip, were also accounted for to ensure that
the aircraft was robust to these variations.
The F-35 fuel system helps to maximize the AOA capability by keeping the
CG as far forward as possible. Figure 14 shows the CG movement as a function
of fuel burn and internal store loading for the CTOL variant.
Regardless of the specific AOA limiter criteria in effect, the goal of the limiter
mechanization was to allow the pilot to attain the limit as rapidly as possible, with
acceptable overshoot, following a full-aft-stick pull. In some cases, extensive logic
and gain scheduling were required to achieve this goal. One particular challenge
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 539

Fig. 14 Variation of the AOA

limiter with fuel (CTOL clean-wing
configuration).

was presented in the case of

a maximum slowdown turn
from high speed, where the
rapidly changing relationship
between Nz and AOA results
in a switch from an Nz
limit at high speed to an
AOA limit at lower speed.
This required the inclusion
of airspeed and airspeed
rate-of-change parameters in the limiter design structure to ensure that both
the AOA and the Nz limits were attained.

ROLL/YAW LIMITERS
Command limits in the lateral and directional axes are scheduled as a function
of AOA, dynamic pressure, Mach, stores loading, and CG, and then adjusted
based on feedback of actual aircraft dynamics. The feedback-based adjustments
are selective and temporary, reducing the limits in appropriate axes based on
observed aircraft rate and acceleration parameters (a-dot, b-dot, Pb-dot, etc.).
These adjustments are essential to maximizing maneuverability in that they
preclude the need for overly restrictive, preemptive command limits. While
important for any fighter design, adaptive limiters of this sort are especially
important for an NDI-based high AOA CLAW, as they provide robustness to
aerodynamic modeling errors that are common in this regime.
At low speed, where control authority is limited, the design of the roll/yaw
command limits was guided by handling qualities considerations rather than
DR concerns. The limits on roll rate, for example, were defined such that a roll
initiated in one direction can be arrested and reversed without unacceptable
bank angle overshoot (roll coast). In the absence of guidance defining acceptable
levels of coast (MIL-STD-1797 is silent on this point [7]), command limits were
based on piloted simulation and flight experience, including lessons learned
from the F-16 and F-22.
At high speed, the lateral/directional control limits were determined by
control surface loads limitations and by departure considerations related to
gyroscopic (inertial) coupling, since the roll rates in this regime can be high.
For illustration, a typical inertia-coupled departure might begin with an aggressive
roll at low AOA combined with an abrupt pitch command. As AOA increases, the
yaw controls may be unable to balance the inertial and aerodynamic moments
540 D. G. CANIN ET AL.

generated in this maneuver, resulting in sideslip increase and potential departure.

A simple way to prevent the coupled roll-pitch departures would be to limit
roll rate based on the yaw control available and the maximum pitch rate achiev-
able at that condition. While effective, this solution would be overly restrictive
for cases in which the pilot has no intention of applying simultaneous roll
and pitch inputs. By contrast, the F-35 limits roll rate based on the pilot’s
actual control inputs. With accurate knowledge of the yaw control available,
and accurate predictions of the yaw control required as a function of mass prop-
erties and roll rate, CLAW will reduce the roll command only by the amount
required to prevent a departure. If the aircraft is already rolling at high rate
when a pitch input is made, CLAW will temporarily reduce the pitch rate
command until the roll rate is reduced, at which point the pitch rate command
is restored to full authority. The combination of command and feedback-based
limits provides enhanced agility and predictability at high AOA, while maximiz-
ing maneuvering capability.

C. AUTO RECOVERY MODES

While the F-35 CLAW is extremely effective at preventing departures from the
normal flight envelope, the aircraft is not completely carefree. CLAW will not,
for example, prevent departures that result from sustained high-pitch-attitude
climbs, where AOA may stay within the nominal range until airspeed drops
below that required for aerodynamic control. The most extreme example of a
departure of this sort is the entry into a tailslide, where the aircraft remains at
low AOA right up to the point where its speed goes to zero. CLAW functions
could be incorporated that intercede and maneuver the aircraft away from such
conditions preemptively, as does the Eurofighter Typhoon’s Automatic Low
Speed Recovery (ALSR) system [8]. The F-35 design philosophy, however, was
to give the pilot the greatest freedom to maneuver possible, as long as controllabil-
ity cliffs – if they exist – are recognizable and avoidable. An enabling element of
that philosophy was the incorporation of automatic recovery modes that recog-
nize a departure from controlled flight and affect a quick recovery.
There are two auto-recovery modes in the F-35: an anti-spin mode to counter
uncommanded yaw rate, and an Automatic Pitch Rocker (APR) to respond to
locked-in deep stalls. Given the criticality of these modes, it was essential that
they function in a robust and predictable manner and not be sensitive to modeling
errors or sensor failures.

AUTOMATIC PITCH ROCKER MODE

APR is designed to recognize and recover from deep stalls without pilot input. The
recovery strategy used by APR is derived from the Manual Pitch Override (MPO)
technique developed for deep stall recovery in the F-16, in which the pilot takes
direct control of the tails using MPO and manually pitch-rocks the aircraft out
of the deep stall. Recovery from deep stall in the F-16 relies on the pilot to
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 541

apply well-timed inputs, the success of which varies with the pilot’s familiarity
with the technique. Lessons learned from the F-16, combined with reliable aircraft
state data in the F-35, allowed the development of a highly effective, robust, and
autonomous F-35 APR.
APR declares a deep stall when AOA is stabilized above CLAW limits and
pitch rate has dropped below a threshold value, indicating that the nose is hung
up. APR then pitch-rocks the aircraft, using pitch rate feedback, until AOA is
returned to within the basic CLAW envelope.
The initial APR design was extremely effective, but refinements were made to
the entry criteria and to the mode itself, several of which are discussed in the Flight
Test Results section. The result of this development was an APR system that con-
sistently and autonomously recovers the aircraft from the worst-case deep stall
conditions, including the most aft CG combined with the WBDs failed open.

ANTI-SPIN MODES
The anti-spin functions include a spin recovery mode and a yaw rate suppression
mode. As their names indicate, the primary functions of these modes are (a)
to arrest the large, self-propelling yaw rates associated with incipient or fully
developed spins and (b) to suppress yaw rate during high AOA conditions
where sustained spins are possible. These two functions, described below, work
in a complementary fashion in conjunction with APR to recover automatically
from any type of departure, regardless of loading or fuel state.
. Spin recovery logic is engaged if body-axis yaw rate exceeds a defined
threshold while at high (positive or negative) AOAs where sustained spins
are possible. The yaw rate threshold is a function of speed and is beyond
that which would be expected during the most aggressive maneuvers. When
spin recovery logic is activated, the relationship (gain) between the HT and
yaw rate is scheduled based on lateral CG and uses yaw rate and yaw accelera-
tion feedbacks to ensure a prompt, smooth recovery for all loadings.
. Yaw rate suppression logic is invoked when the positive AOA limiter has
been exceeded by a specified amount. The AOA schedule is designed so as
not to cause engagement in response to minor AOA limiter overshoots. Yaw
rate suppression is not invoked at negative AOA, as the aircraft has no inverted
deep stall and the rudders and vertical tails are effective at negative AOA.
A key element of the anti-spin CLAW is a change in the prioritization of the HTs.
During normal flight, CLAW prioritizes pitch control (versus yaw control) in
its use of the HT. This priority ensures that AOA or Nz is maintained as precisely
as possible, even at the expense of lateral/directional control. In a high AOA
departure, however, minimizing yaw rate is essential since even a low sustained
yaw rate can significantly delay—or completely prevent—a successful recovery.
Accordingly, when a spin mode is activated, HT priority is shifted to the roll/
yaw axes and pitch control is sacrificed (if necessary), even if APR is active.
542 D. G. CANIN ET AL.

Once yaw rate is reduced to a low value, the magnitude of differential HT needed
to keep it low is fairly small, leaving a substantial proportion of symmetric HT
available for APR or a direct nose-down recovery.
In addition to axis prioritization, another key to the effectiveness of the anti-
spin recovery mode is an advanced system of adaptive filters applied to body-axis
yaw rate. The original F-16 design was overly aggressive in its attempts to control
yaw rate in a deep stall, resulting in ineffective pitch rocking and, in one case,
an extremely delayed recovery. This event led to a filter design that reacts to
steady-state yaw rate but avoids over-responding to high-frequency yaw rate oscil-
lations. The F-35 variable filter design is a direct result of these lessons learned.

AUTO RECOVERY SWITCH

To minimize nuisance engagement, the entry criteria for the auto recovery modes
require a relatively high certainty that a sustained departure has occurred. The
result of this conservatism is that there are cases where the pilot might recognize
the departure before the automatic system engages. An AUTO RECOVERY
switch (Fig. 15) allows the pilot to expedite auto recovery engagement by relaxing
or eliminating some of the entry criteria (AOA, yaw rate, pitch rate, dynamic
pressure, etc.). A beeping aural tone indicates to the pilot that activating the
switch will engage one of the recovery modes.

MANUAL PITCH ROCKER MODE

A Manual Pitch Rocker (MPR) mode is incorporated into the F-35 CLAW,
activated by pressing and holding a button on the throttle. This mode gives the
pilot direct control of the HT in a manner similar to that of the F-16 MPO
mode. MPR was intended primarily as a backup to APR for deep stall recovery

Fig. 15 AUTO RECOVERY switch.

F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 543

during flight test. It was used extensively during initial pitching moment veri-
fication, as it allowed the pilot to set the HT manually at various positions
when outside the normal CLAW AOA envelope.

D. NONLINEAR DYNAMIC INVERSION IMPLEMENTATION AT HIGH AOA

LATERAL CG OFFSETS
One of the first issues encountered during high AOA analysis related to the way
NDI handles lateral CG offsets.
In conventional aircraft, a lateral CG offset results in a rolling moment that
the pilot naturally counters with a roll stick or trim input. The roll input
creates a non-zero sideslip that the pilot either ignores or removes with yaw trim.
By comparison, the F-35’s NDI CLAW is aware of the mass asymmetry and
automatically adjusts control surfaces to zero the sideslip and roll rate. This auto-
trim capability works well as long as control power is abundant. However, at low
speed and high AOA, the reduced control surface effectiveness can result in very
large differential control deflections. While an acceptable static trim may be
achieved, this control solution results in highly asymmetric roll performance,
with high rates being achievable in one direction and very low rates in the other.
Pilots of conventionally controlled swept-wing fighters learn that the solution
to the roll control problem at high AOA is to use pedal (versus lateral stick) to roll
the aircraft; the pedal input creates proverse sideslip, and dihedral effect (Clb) pro-
duces the roll. By contrast, a baseline NDI design counters the roll-off using
rolling surfaces only, while seeking to maintain zero sideslip, resulting in large
asymmetries in trimmed surface deflection and roll response.
In the F-35, the remedy to the control asymmetry involved incorporating fea-
tures that command non-zero sideslip to offload the roll control requirement
when trimmed. The result is an augmentation of the classic NDI control strategy
that provides symmetric roll performance for very large lateral CG offsets across
the high AOA envelope.

ONBOARD MODEL DEVELOPMENT

Providing an accurate aerodynamic OBM, which is fundamental to the success
of NDI-based CLAW, is particularly challenging at high AOA. The aircraft’s
aerodynamics at high AOA are dominated by a complex system of vortices and
transitions between attached and separated flow that result in nonlinear control
effectiveness and strong control interactions. Extensive wind tunnel data were col-
lected in this regime to develop a truth model and the OBM. These data were also
used to establish control surface deflection limits and optimum control surface
trim positions, with the objective of eliminating aerodynamic deadband and
ensuring the effective use of the limited control power.
One particularly challenging modeling issue involved asymmetric flow separ-
ation resulting in uncommanded roll. This behavior, observed early in the flight
544 D. G. CANIN ET AL.

test program at moderate AOAs and transonic speeds (referred to as transonic

roll-off, or TRO), was also observed in the high AOA regime. The conditions
under which the roll-offs occurred followed a trend of increasing AOA with
decreasing Mach. The precise trigger point, however, varied considerably as a
function of surface positions, aircraft rates, AOA, Mach rate, and other par-
ameters. Since precise modeling of these aerodynamic anomalies was not practi-
cal, logic was added to augment the predicted OBM moments using aircraft
feedback. This implementation, discussed in Ref. [2], was similar to that devel-
oped to handle TRO and was used in other areas of the flight envelope where
modeling errors were expected.

EFFECTOR BLENDER LIMITATIONS

A central element of the NDI control scheme is the EB, which uses high-fidelity
OBM data to determine the optimal effector solution as a function of aircraft
state. In the lateral/directional axes, the EB performs the function of a traditional
control surface mixer and, at low AOA, nominally arrives at the same surface
allocation as would a human designer. In the high AOA regime, however, cases
arise (e.g. when control power is limited) in which the standard NDI approach
may not result in the optimal effector solution. In these cases, the human designer
would invoke a variable (adaptive) control surface strategy. Unfortunately, the
basic NDI architecture restricts the designer’s ability to do so directly, requiring
modifications to the EB structure.
An example may help to clarify this point. Rudder effectiveness is extremely
limited at high AOA, leading to rudder saturation during maneuvers such as
full-command rolls and roll reversals. When the rudder saturates, the EB auto-
matically overflows the unmet yaw requirement to the next best source of yaw:
differential HT. If differential HT cannot satisfy the yaw demand (e.g., if the
HTs are symmetrically deflected for pitch), the EB overflows the remaining yaw
requirement to the next best effector: differential flap. Fundamentally, this
solution is no different from what a human designer would implement with a tra-
ditional mixer. The human designer, however, can identify consequences of this
solution that the nominal EB design will not recognize. For example, since the
EB prioritizes yaw control over roll, it will do so using as much differential flap
as needed, even if the flaps are weak yaw effectors, with the undesirable effect
of producing an opposing moment in roll. Knowing this, the human designer
might decide that, where rudder effectiveness is inadequate, the optimal solution
would be to conditionally relax the priority on yaw. The designer could implement
this by setting limits on the amounts of differential HT and differential flap used,
allowing some amount of sideslip to occur. Incorporating functionality of this sort
in a traditional mixer architecture is straightforward. In an OBM/EB structure,
however, where the designer does not have direct control of the effector solution,
incorporating such a function is a significant challenge. For the F-35, the resol-
ution of this type of problem involves actively varying control surface and axis
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 545

weightings using a combination of aircraft feedbacks. The logic evaluates the

current state of the aircraft, including control surface usage, and determines
whether full yaw axis priority will result in the best control surface utilization
or a relaxed yaw priority would be preferred.

V. FLIGHT TEST PROGRAM

This section provides a general overview of the high AOA test program, including
scope, schedule, and a description of prerequisite testing and the major test
phases. This provides a backdrop for the next section, which discusses key
test results.

A. OVERVIEW
The primary objective during the early stages of the System Development and
Demonstration (SDD) program was to certify an initial operational envelope
for the fleet. The AOA envelope for that initial capability, 2108a to þ208a,
became the de facto definition of “low AOA” for the F-35 program, and testing
outside that range was deferred until more than four years after the beginning
of SDD flight test. High AOA testing commenced in late 2012 on CTOL, with
the other two variants starting approximately one year later.
The high AOA test program followed a phased approach that was similar
for all three variants. Figure 16 presents a high-level description and list of the
objectives of each phase, along with a schedule and test-point breakdown.
Approximately 320 flights and 3,500 test points were flown, the majority of
which were dedicated to DR testing.
The rest of this section presents a more detailed description of the scope and
objectives of each phase, and how each fit in the overall high AOA clearance
process. A discussion of flight test results is presented in Section VI.

B. PREPARATION
Several aircraft modifications and system checks were performed as prerequisites
to high AOA testing. In addition to the risks directly associated with high AOA
flight, the potential for dwelling at zero or negative g during tail slides and
other nose-high departures raised concerns related to the engine oil system, the
fuel system, the electrical power system, and the Power and Thermal Management
System (PTMS).
. To mitigate the risk of engine flameout at high AOA, engine airstart testing
was completed in the CTOL and STOVL variants prior to starting high
AOA testing. (The CV variant was cleared by its similarity to CTOL.)
. Waivers were received from the engine manufacturer for a limited number of
excursions beyond the normal negative G (low oil pressure) time limit.
546 D. G. CANIN ET AL.

Fig. 16 High AOA test phases.

. The fuel system in the CTOL and CV variants was modified to allow greater
time at negative g, and fuel-boost-pump pressures were monitored closely
during all zero and negative g operations.
. Hardware-in-the-loop (laboratory) testing was performed to verify that the
electrical system could withstand the increased demand of sustained periods
of large amplitude surface motion during worst case departures.
. A Flight Test Aid (FTA) was used to extend the amount of negative g time
available on the integrated power package (the central component of the
PTMS) before it would automatically perform a self-protective shutdown.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 547

. A supplemental cabin pressurization system, installed in CTOL for air-starts,

was retained for the initial high AOA testing to mitigate the risk of injury
to the pilot in the event of a flameout at high altitude. This system was not
incorporated into the F-35B and F-35C variants, but cabin leak checks were
performed in those aircraft to ensure that adequate cockpit pressure would
be retained in the event of a flameout at high altitude.

Provisions were made for the installation of a Spin Recovery Chute (SRC) on
one aircraft of each variant. The SRC installation consisted of a chute canister
mounted to a four-legged support (shown in white in Fig. 17), a cockpit control
panel, and four brackets to carry the chute loads into the aircraft structure
(shown in orange in Fig. 18). Although this design allowed for the rapid
removal/installation of the support structure and canister, the same cannot be
said for the brackets, which required several days to remove or install. This recon-
figuration time provided a strong incentive to group the SRC-required testing into
blocks in order to minimize the number of on/off cycles.
After careful consideration, the decision was made to forgo an in-air checkout
of the chute. An end-to-end check of the SRC system was conducted during a
65-knot taxi on the CTOL aircraft (Fig. 19). This test went smoothly and was
the only time the chute was ever deployed on the F-35 program. The final
check before testing above 208a was to investigate the effects of the SRC instal-
lation on flying qualities
at low AOAs. A one-
flight checkout was flown
with each variant, which
included standard maneu-
ver blocks (doublets, side-
slips, rolls, and windup
turns) at a range of condi-
tions from 0.95 Mach
down to 1 g/208a in UA
mode, and across the full
PA envelope. Minor aero-
dynamic effects due to
the SRC installation were
noted, but they were con-
sistent with wind tunnel
predictions and had only
minor effects on aircraft
handling.
Pilot-selectable FTAs,
many of which were

Fig. 17 SRC canister.

548 D. G. CANIN ET AL.

Fig. 18 SRC installation.

developed specifically for

high AOA testing, were
used extensively and con-
tributed immensely to the
efficiency of these tests.
The FTAs most commonly
used during high AOA
testing were those that
. modified the fuel burn
sequence (fuel balla-
sting) to set a desired
longitudinal CG;
. set reduced Nz or AOA
limits;
. enabled hands-on-stick-and-throttle control of the WBDs;
. disabled the DANG COND tone for CLAW AOA limit exceedance (to facili-
tate communication),
. disabled automatic APR engagement (to allow MPR testing);
. disabled anti-spin logic and allowed the pilot to command specified spin rates
using pedal inputs; and
. biased the CG reported to CLAW to allow maximum AOA to be achieved over
a wider fuel range.

Fig. 19 SRC ground deployment test.

F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 549

C. EXPANSION TO 50-DEGREE AOA

The objective of this phase was to validate the aerodynamic model to the limits
of the CLAW envelope. Since departure recovery capability had not yet been
established, all testing in this phase was done at the most conservative CG
(forward) with the SRC installed. The expansion was performed in increments
of 58a, building up from 208a to the CLAW limit of 508a. The test cards for a
typical envelope expansion integrated test block are shown in Fig. 20.
The first maneuvers conducted at each new AOA were a trim, a doublet series
and a pushover, illustrated on the left card in Fig. 20. These were performed
at an AOA five degrees beyond that to be used for the lateral directional tests,
illustrated on the right card in Fig. 20. The pushover was performed to ensure
that adequate nose-down pitch acceleration was available prior to proceeding to
the more challenging lateral/directional maneuvers. At each AOA, a stringent
set of flight test continuation criteria was applied before proceeding to the next
increment in AOA.
Maneuvers in this phase were limited to modest bank angles and did
not involve simultaneous pitch and yaw/roll inputs. The AOA limit FTA was
used extensively to assist the pilot in maintaining constant AOA (using full
aft stick) and minimizing cross-axis contamination. Testing was initially done
with the clean aircraft, followed by various air-to-air external stores and
WBD configurations.

Fig. 20 Test card showing envelope expansion integrated test block maneuvers
and buildup.
550 D. G. CANIN ET AL.

D. DEPARTURE RECOVERY
The goal of this phase was to demonstrate that the F-35 recovers from OCF
without pilot input in all air-to-air configurations. The departure recovery
phase consisted of two distinct sub-phases: (1) pitch-axis-only departures
executed using a controlled, wings-level buildup in AOA beyond the 508a
CLAW limit; and (2) dynamic departures using multi-axis inputs to fully check
the auto-recovery capability of the aircraft. The primary objective of these
demonstrations was to gain confidence that we could remove the SRC for the
next phase, DR testing, so the DR tests could be performed in a production-
representative configuration.
The departure recovery demonstrations began with a series of climbs using
progressively higher pitch attitudes to produce low-speed, ballistic recoveries
that resulted in momentary exceedances of the CLAW AOA limit. These
initial tests were conducted at forward CGs, where recovery was guaranteed, in
order to assess any tendencies to yaw or roll at extreme AOAs and to provide a
check of air data and engine behavior. Next, the MPR feature was used to
command various symmetric HT positions to establish stabilized trim points
beyond CLAW limits. These trims were used to verify pitching moment pre-
dictions, to investigate deep stall susceptibility, and (on recovery) to verify
nose-down control power. The initial excursions beyond the CLAW limit were
conducted in the clean configuration at nominal CG. Subsequent tests progressed
to more aft CGs and more critical configurations, including WBDs open. (The
WBDs normally close automatically after stores release, but the outer doors
produce a significant nose-up pitching moment, so it was important to test this
failure mode.)
Once the pitching moment predictions had been verified, a comprehensive
evaluation of APR recovery logic began. Having established adequate control
power for recovery using MPR, the APR testing was done almost exclusively at
the most aft CGs. Initial APR engagements were done in a controlled fashion
by using MPR to establish a stable trim beyond the CLAW AOA limit, then releas-
ing it to allow APR to engage. Since the F-35 does not have a negative-AOA
deep stall, the majority of the MPR-induced APR checks were performed
upright, with a handful initiated inverted to validate the negative-AOA logic.
Preflight analysis indicated that all variants of the F-35 were highly resistant
to spins, an attribute credited both to aerodynamics and the anti-spin logic in
CLAW. The initial concept, therefore, was to validate the spin recovery logic
using simulation only. Based on past experience, however, the Navy insisted
that a spin recovery be demonstrated in flight before proceeding with more
aggressive departures. Since no means of creating a spin could be devised using
the normal CLAW, the demonstration required the development of a dedicated
pro-spin FTA. The FTA disabled the nominal anti-spin logic and allowed a high-
rate, spin-like condition to be created using pedal commands, after which the
FTA would be turned off, allowing the spin recovery mode to engage.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 551

The graduation exercise for the departure recovery phase was the performance
of dynamic departures. With the confidence gained from the more controlled
testing described above, and the successful correlation of those results with ana-
lytic models, a buildup in configuration and CG was not required for this
testing. Dynamic departure testing was therefore done almost exclusively at the
most aft CG and in the worst case aerodynamic configuration, including WBDs
open. Departures were generated by conducting high-pitch attitude climbs,
including tail slides, applying stick and pedal inputs to generate as much rate in
all three axes as possible at the point of departure. The results from the first
round of maneuvers were used to devise even more diabolical combinations of
pitch attitudes and control inputs in an attempt to create sustained departures.
This no-holds-barred approach to demonstrating recoveries from any conceivable
situation was deliberate, in order to build a convincing body of evidence to justify
removing the SRC for DR testing. As a final verification of the recovery capability,
a set of the most critical configurations and maneuvers was repeated with the
SRC removed to ensure that the SRC did not have a notable effect on recovery.

E. DEPARTURE RESISTANCE
The majority of the high AOA test program was dedicated to DR testing, i.e.,
verifying that the aircraft was highly resistant to departures for the full range of
configurations and pilot control inputs. Having fully verified the recovery capa-
bility of the aircraft in the previous phase, almost all DR testing was done in a
production-representative configuration with the SRC off.
Removing the SRC for the majority of this phase was somewhat controversial
from a safety perspective but was crucial to the efficiency of the F-35 high
AOA program. DR testing represented about two-thirds of the high AOA test
program. Performing these tests in a production-representative configuration
eliminated the potential for SRC effects to impact test results, thereby avoiding
the need for costly SRC-off regression. Just as importantly, it allowed the data col-
lected in this phase to be used to refine the aerodynamic OBM and associated
truth model, without concerns about its validity. Had testing a production-
representative configuration been deferred until the end of the program, the
data would have been severely limited and would have been acquired too late
to be integrated during SDD.
DR tests involved the application of aggressive, maximum, single- and multi-
axis commands, reversals, and assaults on both the positive and negative AOA
limiters. To illustrate the types of maneuvers performed, test card examples are
shown in Fig. 21. Because some of the control inputs and timing were relatively
complex, graphic illustrations of stick and pedal inputs reduced pilot error and
added significantly to test efficiency.
A standard buildup was done in altitude (high to low), speed (low to high),
and configuration (clean, then air-to-air loadings, then air-to-ground). The
primary test altitudes for gear-up testing were 35,000 and 20,000 feet, although,
552 D. G. CANIN ET AL.

Fig. 21 Sample departure resistance test cards.

as discussed in the Flight Test Results section, some testing at lower altitudes was
conducted when simulation or flight test results indicated it was necessary.
Longitudinal CG was held at target condition with an FTA, and lateral
CG variations were established using internal and external stores. The majority
of DR testing was performed with air-to-air loadings since CLAW limits are
close to controllability boundaries for these configurations. Critical external
air-to-ground loadings were also tested to their (more restrictive) limits. Since
departure recovery had not been demonstrated (or required) for air-to-ground
loadings, the SRC was installed for the initial DR testing of those configurations.
Gear down DR testing was also performed for critical air-to-air and air-to-
ground loadings.
Offline simulation was used as the primary means of verifying that each
CLAW update performed as intended and was safe to operate over the range of
conditions in the test plan. For each software release, roughly 30,000 simulation
runs were conducted, among them a routine analysis set known as the Matrix
of Death. Automated tools were used to scan the results to identify cases that,
due to flying qualities and/or loads concerns, required further scrutiny. The
simulation results were also used to identify potential trouble spots that required
investigation in flight, and to design critical stressing maneuvers.

LOW-SPEED DEPARTURE RESISTANCE TESTING

The first objective of DR testing was to clear the low-speed envelope. While this
testing covered the same flight conditions explored during the expansion to 508a,
DR tests were done with aggravated, multi-axis inputs over a range of CG and
AOA combinations. An FTA was used to change CG expeditiously during the
flight, greatly enhancing test efficiency.

HIGH-SPEED DEPARTURE RESISTANCE TESTING

The approach to DR testing at higher speeds was similar to that used for
low speeds (aggressive inputs, worst-case loadings, etc.), with a few unique
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 553

considerations. First, since the effect of CG on the AOA and sideslip command
limits blends out at high Mach, testing in this regime was done almost exclusively
at a single, most aft CG. Second, since many of these points would be performed at
high dynamic pressure, there was a potential for exceeding airframe load limits.
Accordingly, while the initial testing at 35,000 feet (and relatively low dynamic
pressure) was done with a non-loads-instrumented jet, the testing at 20,000 feet
and below required close coordination with loads engineers and real-time moni-
toring with the instrumented aircraft.
Prior to testing, the AOA limits at high speed had been established based on
inlet compatibility alone, although it was acknowledged that there were several
issues that had the potential to require more restrictive limits:
. Unlike at low speed, high-speed DR would reach combinations of Mach and
AOA that fell outside the envelope where wind tunnel data were acquired.
This meant relying on OBM data that were based on extrapolated trends.
. Testing below 208a had confirmed predictions that the F-35 was subject to
abrupt flow field changes at transonic speeds and elevated AOAs, resulting
in unsteady rolling moments on the aircraft. The expectation was that these
effects would subside at higher AOA, but only testing would tell.
The approach was to start with the AOA limiter set by inlet compatibility require-
ments and reduce it only if necessary. It was fully understood that departures were
possible—even likely—with this test approach. However, this was considered to be
a fully managed risk based on the confidence gained in our recovery capability.

F. OPERATIONAL ASSESSMENT
The final phase of high AOA testing for each variant was a limited evaluation
of aircraft performance and handling qualities while conducting operational
tasks above 208a. Unlike the previous phases, which consisted of highly scripted,
open-loop maneuvering, this testing allowed the pilot to operate the aircraft to
achieve an operational task. Testing included an evaluation of high AOA task
elements (tracking and reversals, ditches, minimum speed overhead maneuvering,
nose-high recoveries, etc.), as well as engagements with an adversary aircraft start-
ing from offensive, defensive, and neutral setups. The data gathered from this
testing were qualitative in nature and used primarily to demonstrate specification
compliance and to support describing the high AOA characteristics in the
flight manual.

VI. FLIGHT TEST RESULTS

This section provides an overview of the key findings from the SDD high AOA test
program. In general, the aircraft performed as expected and designed. However,
since the positive results are less interesting than the anomalies, this section
554 D. G. CANIN ET AL.

focuses primarily on the latter. Correcting these anomalies involved combinations

of air data refinements, CLAW logic updates, and OBM corrections.

A. AIR DATA SYSTEM

Several issues with the ADS were discovered during the initial expansion to 508a
and during departure recovery testing. These findings resulted in improvements
to the calibration of pneumatic measurements and the logic used to transition
between pneumatic and inertial sources.

INERTIAL SIDESLIP IN CHANGING WIND

At AOAs outside the pneumatic envelope of the ADS, estimated winds and iner-
tial velocities are used to derive AOA and sideslip, referred to as inertial a and b.
This scheme works well as long as the aircraft remains in an air mass of relatively
uniform velocity.
The example in Fig. 22 shows a deceleration to high AOA while in a descent.
During this maneuver, the aircraft descends through a wind gradient, causing the

Fig. 22 Effect of inertial air data (sideslip) error during descent at high AOA.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 555

sideslip being reported to CLAW to grow at about twice the rate of the
actual value.
Starting at around six seconds, a slight uncommanded roll to the left begins.
The CLAW respond with differential flap to create a right rolling moment, which
should stop the roll and drive sideslip back to zero. However, the overestimated
sideslip being reported to the OBM causes it to over-predict the amount of
rolling moment it will get from dihedral effect (Clb). The roll input by CLAW
is therefore less than it needs to be, which the pilot corrects manually using
right stick. The aircraft eventually stabilizes in a left bank with differential flap
at its deflection limit. At this point, all the available roll control power from the
flaps is being used just to hold the current bank angle, which precludes the aircraft
from rolling right.
The solution to this issue involved both improving the air data accuracy
and making the CLAW more robust to air data and aerodynamic modeling
errors. The updates included:
. Raising the AOA for transitioning to inertial sideslip, as pneumatic sideslip was
found to be accurate to higher AOAs than predicted.
. While the AOA transition point was not changed, robustness was improved
by improving the calibration of pneumatic AOA and by using pneumatic AOA
for wind computations even when the CLAW AOA had transitioned to inertial.

PNEUMATIC AOA ERROR AT LARGE NEGATIVE AOAS

Departure recovery testing identified the need for additional refinements to the air
data algorithm. The most significant issue from this phase emerged during
inverted MPR testing on CTOL. At large negative AOAs, separation off the fore-
body chine created a low static pressure condition that impacted all primary
air data outputs. The static pressure error resulted in a positive AOA reading
that, since it was within the pneumatic envelope, caused the system to accept
the erroneous pneumatic air data while the aircraft was still in an inverted
departure. The aircraft was easily recovered using MPR, but post-flight analysis
showed that this AOA sign error would adversely impact spin recovery. Departure
recovery testing was suspended until logic was added to ensure that the aircraft’s
normal load factor (Nz) was positive before allowing the ADS to accept a positive
pneumatic AOA.

B. ENGINE AND SUBSYSTEMS

A critical objective of the envelope expansion phase was to ensure that the engine
and related emergency backup systems functioned properly at elevated AOAs
and low speeds. During the entire high AOA test program, all critical subsystems
(propulsion, fuel, electrical, environmental control, etc.) functioned flawlessly at
all AOAs, at low speed, and during extended low-g conditions.
556 D. G. CANIN ET AL.

C. AERODYNAMIC MODEL VALIDATION

The most valuable flight data for high AOA aerodynamic model validation were
acquired during the envelope expansion phase. In general, the aircraft response
compared well with predictions for all maneuvers and configurations flown,
including asymmetric air – to– air stores, gun pod installed, and WBDs open.
Most importantly, there was very good agreement between flight test results
and predictions for longitudinal stability and HT control power, as exemplified
by the max push from a high AOA trim condition shown in Fig. 23. Although
minor adjustments to the OBM were made as a result of data collected in this
phase, none were consequential enough to require updates before proceeding
to the DR testing. This result, which was pivotal in keeping the high AOA test

Fig. 23 Nose-down pitch maneuver from 5088a.

F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 557

Fig. 24 MPR trim point showing lower-than-expected pitch damping.

program on schedule, was the payoff for the significant upfront investment
made in collecting low speed wind tunnel data and assembling a high-fidelity
aerodynamic model.
Figure 24 shows an example of how MPR was used to collect data for stabilized
trims beyond the CLAW limit. The pilot initially let the aircraft stabilize near 608a
with neutral HT (no stick input), then slowly pulled aft, causing the HT to go full
trailing-edge-up. After the pull, AOA increased to around 72 degrees but exhib-
ited a persistent oscillatory behavior. In this case, the predicted trim conditions
were well-modeled for both HT settings, but the simulation predicted a much
more dynamic response than that seen in flight. This damping mismatch was
observed for most upright trim cases in all three variants. Based on these data,
models were updated to include a sizable increase in pitch damping to match
the more sedate behavior exhibited by the aircraft.
Similar testing was conducted for a range of configurations, both upright and
inverted. Where necessary, models for basic pitch stability, symmetric tail control
558 D. G. CANIN ET AL.

power, and pitch damping were adjusted. A small subset of the MPR test points
were repeated with the SRC off. As expected based on wind tunnel testing, the
impact of the SRC on trim AOAs was negligible.
Dedicated parameter identification testing was not included in the initial
planning, but preparations (FTAs, analysis tools, etc.) were put in place so this
capability could be used on an as-needed basis. Relatively late in the program,
three dedicated high AOA parameter identification flights were performed with
the STOVL and CV aircraft to verify control effectiveness in localized Mach
and AOA regions. These data were used to adjust several terms in the aerody-
namic truth models used for offline analysis, but the changes were not significant
enough to warrant an update to the OBM, which would have required flight
test regression.

D. DEPARTURE RECOVERY
Of the more than 220 dynamic departures performed, the vast majority self-
recovered without engaging any auto-recovery mode. While no significant issues
with the CLAW were encountered during this phase, test results were used to fine
tune the recovery logic as described below. The end result was a system that, in all
demonstrated and simulated cases, recovered without any pilot involvement
whatsoever, clearly meeting the requirement to recover with “minimum pilot
input.” This result provided high confidence in the departure recovery logic
and was critical to allowing DR testing to be conducted without the SRC installed.

PITCH ROCKER
Initial APR engagements were done in a controlled fashion by stabilizing beyond
the CLAW AOA limit using MPR, then releasing the MPR switch to allow APR
logic to engage. In total, 78 APR engagements were performed in this manner,
all of which recovered the aircraft in one or two pitch-rock cycles and less than
4,000 feet of altitude loss. Of those that required two cycles to recover, most
occurred with WBDs open (a failure mode of the doors).
During recoveries from dynamic departures (steep climbs and tailslides),
APR engaged 45 times, all but one of which required one or two pitch-rock
cycles to recover. The single event that required three cycles, while not operation-
ally relevant since the WBDs were open, highlighted areas where further improve-
ments could be made.
The three-cycle APR event, shown in Fig. 25 occurred with the CV variant fol-
lowing a tailslide at an extremely aft CG with all WBDs open. The maneuver
begins with the aircraft in a vertical climb with a full right roll input applied at
120 knots. As speed drops to near zero, AOA peaks at 90 degrees and the HTs
go to nearly full trailing-edge-down deflection, backing off as needed for yaw
rate suppression. After five seconds of low pitch rate, APR engages automatically,
but the initial nose-up kick is only marginally effective due to the extremely low
speed (less than 40 KCAS). From 22 seconds until the recovery, the amplitude of
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 559

Fig. 25 Three cycle APR recovery.

the pitch rocking gradually increases until the deep stall is broken. Note that while
APR was engaged, the roll and yaw rates subsided to near zero as a result of the
spin suppression logic and increasing airspeed.
In this event, the CLAW functioned as designed and recovered the aircraft
without pilot intervention. The delayed recovery, however, led to improvements
to the APR entry criteria and function. Since the improvements were minor,
inflight regression was not warranted. Offline analysis, however, indicates that
all deep stalls will recover with a single pitch-rock cycle with WBDs closed, and
no more than two cycles with WBDs open.

SPIN RECOVERY
There were 25 engagements of the spin recovery logic during dynamic departure
testing, all of which occurred during post-departure gyrations at low dynamic
560 D. G. CANIN ET AL.

Fig. 26 Spin recovery demonstration.

pressure. The yaw rate suppression logic was effective, and the aircraft exhibited
no propensity to spin, regardless of its dynamic state at departure entry.
The only sustained spins encountered were those that were staged using
the spin mode FTA described earlier, a typical example of which is shown in
Fig. 26. MPR was engaged at the start of the maneuver, allowing the pilot to estab-
lish the aircraft beyond the CLAW AOA limit. With the spin FTA engaged, the
pilot applied full right pedal, and yaw rate built up to the high target value.
(Note in the figure that AOA begins at 55 degrees but increases as the yaw rate
grows and the spin flattens out.) The pilot then disengaged MPR and the FTA
simultaneously, allowing immediate engagement of the CLAW spin recovery
logic, which drove the differential tail to arrest the spin. As the yaw rate subsided,
differential HT backed off to around half of its initial value, freeing up symmetric
HT for pitch control. As the yaw rate reduced further, the associated decrease in
nose-up coupling resulted in a steady drop in AOA, with the recovery completed
around four seconds after spin recovery mode had been enabled.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 561

TAILSLIDE EVENT
An interesting occurrence during departure recovery testing was the discovery
of the perfect tailslide. Whereas the recoveries from the first 71 tailslides were
relatively benign, the 72nd – which was performed with a loading and entry con-
ditions not significantly different from its predecessors – provided a valuable
lesson regarding the statistical nature of departure recovery testing. Figure 27
shows the pitch and yaw rates for all the tailslides flown, with event No. 72 high-
lighted in red. Unlike prior events, where the aircraft began pitching nose-down
before achieving a significant negative velocity, airspeed in this case reached
approximately 60 KCAS before the nose-down movement began. The higher
aerodynamic moment resulting from this higher speed drove the initial nose-
down pitch rate to 85 degrees per second, which was much higher than seen in
any prior event, causing the aircraft to swing through the nose-down attitude
back to near 90-degrees nose-high. With the aircraft now flying backward at
more than 110 KCAS, an extremely dynamic recovery occurred, driving both
pitch and yaw rates beyond the engine design limits.
Since sufficient tailslide data had been collected at this point to validate the
recovery logic, tailslide entries were discontinued because of the risk of engine
damage. No CLAW changes were made in response to this event, but a special
aural tone was added to alert the pilot to an impending tailslide while airspeed
was high enough to allow preventative action to be taken.

Fig. 27 Engine yaw and pitch rate limits exceedance in tailslide.

562 D. G. CANIN ET AL.

E. DEPARTURE RESISTANCE
This section provides examples of some of the more interesting findings from DR
testing, along with a description of the actions taken to resolve them. Improve-
ments in the CLAW logic and aero models derived from data acquired during
this phase resulted in an aircraft that clearly met the requirement to be departure
resistant and would more accurately be categorized as “extremely resistant to
departure” based on the guidance in MIL-F-83691B [9].

CTOL DEPARTURE RESISTANCE

Since the CTOL aircraft was the first to start DR testing, much of the early discov-
ery occurred during these tests, and several of the resultant CLAW updates were
applied preemptively to the other two variants.

Low speed
The example in Fig. 28 illustrates a case in which competing pitch and yaw axis
demands on limited HT control power led to a spin recovery engagement. This
case occurred during one of the first bank-to-bank maneuvers done on the
508a limiter using maximum (roll stick and pedal) yaw rate inputs. For this man-
euver, the yaw rate tracked the command flawlessly on the initial bank to the right
and had a small but controlled overshoot after reversing to the left. A similar over-
shoot occurred in response to the next reversal back to the right, but in this case
the pilot released the pitch stick, creating an additional demand on the HTs to
reduce AOA. The CLAW therefore prioritized nose-down pitch, leaving no differ-
ential HT available for yaw control. Since the rudder is ineffective at this AOA and
the HTs were fully committed to pitch, the unchecked yaw rate was enough to
exceed the spin recovery mode threshold. As noted previously, anti-spin engage-
ment causes priority to be switched to yaw until the rate is arrested. Although the
CLAW was successful in preventing a sustained departure, switching from full
pitch to full yaw prioritization led to a momentary loss of pilot control while
yaw rate was arrested. In later CLAW releases, modifications were introduced
to achieve a better balance between pitch and yaw demands at high AOAs,
where both are important to aircraft control.
Examples of other low-speed issues that were found and addressed are shown
in Table 2.

High speed
Early in the CTOL high-speed DR testing, sideslip excursions were observed
during a roll-and-pull maneuver initiated at high transonic Mach and high alti-
tude. While basic control was never in question, the sideslip excursions were a
concern due to the potential to exceed airframe load limits at lower altitude.
Accordingly, an additional series of tests was conducted to focus on combined
roll and pull maneuvers in this regime.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 563

Fig. 28 Nose-down pitch command during bank-to-bank maneuver results in loss of

yaw control.

More than 20 maneuvers were completed in the narrow Mach-AOA region of

interest using a variety of pull-and-roll inputs. Although yaw control saturation
and loose sideslip were observed, no major handling issues or departures were
noted after completing all planned runs. However, with fuel remaining, the test
team elected to repeat a few worst-case runs with minor variations in entry con-
ditions and roll input timing. On the last of these bonus runs, the full-blown
departure shown in Fig. 29 occurred. The maneuver was a full-aft-stick decelerat-
ing turn, with a max roll over the top initiated at 158a. The initial roll response
was in the correct direction, but yaw control was inadequate to coordinate the
roll, resulting in a buildup of five degrees of adverse sideslip. As AOA increased
through 19 degrees, a large un-commanded left roll occurred due to a combi-
nation of adverse sideslip and sudden asymmetric shock movement on the
wings. This uncoordinated left roll, along with the still-uncontrolled positive
 TABLE 2 CTOL LOW-SPEED DEPARTURE ISSUES AND RESOLUTION.
564

Issue: HT ringing during assaults on the Issue: After stabilizing in a maximum yaw Issue: Excessive overshoots of yaw rate commanded trigger
negative AOA limiter using a full-stick push rate maneuver, a max push caused a seven spin recovery during high yaw rate maneuvers. Yaw rate
followed by roll stick and opposite pedal. second saturation of the symmetric HT and grew steadily to a value well past the commanded value,
The large amplitude, high-rate HT motions a transient excursion above the AOA causing spin recovery logic to intervene and drive the yaw
were very lightly damped and resulted in limiter. This issue was predicted by rate back to a controllable range before returning control to
sustained pitch axis oscillations. simulation and was confirmed in flight. the pilot.

Solution (AOA limiter gain): The negative Solution (CLAW and OBM): Desired Solution (CLAW and OBM): CLAW logic and aero models
AOA limiter was softened by allowing it to dynamics command models were adjusted were improved incrementally over several software releases
begin intervening sooner, with lower gains. when large errors between the requested to resolve issues such as this. Solutions were customized by
This fix was regressed successfully. and actual pitching moment occurred variant and involved a fly-fix-fly approach to ensure that
while maneuvering near the AOA limiter. maneuverability limits were maximized while remaining
within control power constraints.
D. G. CANIN ET AL.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 565

Fig. 29 CTOL high-speed departure.

yaw rate, caused sideslip to reach -24 degrees. The large roll moment due to
sideslip led to a snap roll back to the right, which coupled into an AOA excursion
to 608a before the aircraft self-recovered.
This event led to incorporating several significant changes to the CLAW:
. Extensive modifications were made to the aerodynamic models (truth and
OBM) in the transonic elevated-AOA regime. The changes included updates
to roll and yaw stability, damping, and control power.
. DR logic was bolstered using aircraft states, surface positions, and pilot inputs
to preemptively reduce command limits during aggressive maneuvering.
. The AOA limiter, which had previously been established by inlet compatibility
considerations only, was reduced for high transonic and supersonic Mach
numbers.
566 D. G. CANIN ET AL.

Once the updated software became available, the changes were successfully
regressed and the remainder of the CTOL high-speed DR testing was completed
without event.

STOVL DEPARTURE RESISTANCE

With the benefit of experience gained on the CTOL variant, the STOVL DR
testing was largely uneventful. In fact, the similarity of the two planforms
allowed the overall scope of the STOVL DR testing to be reduced. The most
significant savings were realized by eliminating the buildup loadings for
air-to-ground DR testing in both UA and PA modes, based on commonality
assumptions and the success of the CTOL testing. Despite having a later start,
the STOVL variant was cleared for full AOA capability concurrently with CTOL.

CV DEPARTURE RESISTANCE
The CV variant proved the most challenging from a DR perspective. Due to its
distinct planform, lessons learned during DR testing of the other variants were
generally not applicable, and several CV-unique problems were identified that
required customized solutions.

Roll reversal
CV had the distinction of being the only variant to experience a departure prior
to formal high AOA testing. The event, shown in Fig. 30, was flown relatively
early in the test program for loads evaluation. The maneuver was a constant
208a slowdown turn to the right, with a left roll commanded at 0.93 Mach via
an auto-maneuver FTA. At roll initiation, the rudders and differential flaps satu-
rated for nearly two seconds as adverse sideslip grew to 20 degrees, precipitating
a rapid, un-commanded roll reversal.
As noted previously, this Mach/AOA regime is subject to un-commanded
roll and pitch motions resulting from rapid, often asymmetric, shock movement
on the wing. On CV, the resultant flow separations can be further aggravated by
several CV-unique factors, including the influence of aileron deflection, increased
wing span, and close proximity of the wing and HT. Addressing this issue required
CLAW modifications that were similar to but more extensive than those made
following the CTOL transonic departure described earlier. An extensive rework
of the aerodynamic OBM was also required, including updates to roll damping
and the effectiveness of all the lateral/directional control surfaces.

Sideslip control
The CV aircraft exhibited poor sideslip control while maneuvering near stall AOA
around 0.75 Mach. Figure 31 shows one of the more extreme examples that
occurred while reversing a maximum yaw rate command at high altitude. The
initial input was applied at low AOA and generated the expected response.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 567

Fig. 30 CV roll reversal.

When the input was reversed just below the limit AOA, the yaw rate responded
immediately but the roll rate continued to increase, leading to a larger-than-
desired sideslip.
While the roll response and sideslip control were less than ideal, basic control
was maintained. However, the large sideslips had to be addressed, as they had
been predicted to create excessive loads at lower altitudes if left uncorrected.
Several CLAW updates were developed to correct the issue, including modi-
fications to desired dynamics and the addition of logic to improve sideslip
control. To ensure that potential structural concerns were fully mitigated, the
test plan was modified to include a progressive altitude build down to lower
values than originally planned, using the loads-instrumented aircraft.
568 D. G. CANIN ET AL.

Fig. 31 CV excessive sideslip during control reversal.

Anti-spin mode engagements

Another CV-specific issue was the intervention of the spin recovery logic
during stick-only rolls. Figure 32 shows a case that occurred as AOA increased
through 25 degrees during a level pull with a full-stick roll input (pedal-free)
at 15,000 feet and 0.6 Mach. Roll and yaw rates tracked command initially,
but yaw rate continued to grow and eventually exceeded the spin recovery engage-
ment threshold. Once the anti-spin logic engaged, CLAW took control of the
aircraft and reduced yaw rate, leading to an uncommanded increase in the roll
rate and some fluctuation in AOA. This issue was determined to be the result
of small errors in Mach-dependent OBM coefficients that were only observable
at low altitude, as discussed in the following paragraphs.

ALTITUDE EFFECTS
The initial plan for DR testing called for flight testing at 35,000 and 20,000 feet
only, relying on simulation to clear lower altitudes. The rationale for this
approach, used in past high AOA programs, was that the primary aerodynamic
drivers are Mach and AOA, not altitude or dynamic pressure. However, a
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 569

number of anomalies observed during low altitude testing on CV called this

strategy into question.
Poor sideslip control was initially seen at 35,000 feet, but the issue did not pose
a threat to aircraft control and was not considered to be operationally significant
At lower altitude, however, stick-only rolls initiated at around 0.6 Mach and
between 208a and 308a resulted in significant yaw excursions and occasional
spin recovery engagements, as described above.
Analysis showed that the altitude dependence was not the result of changes
in aerodynamic coefficients with altitude but to the increased effect of
errors in Mach-dependent coefficients at low altitude. The increased effect was

Fig. 32 CV anti-spin engagement during maneuvering.

570 D. G. CANIN ET AL.

attributable to (1) the higher dynamic pressure (for a given Mach) generating
greater physical moments, making the effects of the errors more observable,
and (2) the aircraft’s performance being greater at low altitude, causing it to
decelerate more slowly and thus dwell longer in the troublesome Mach-AOA
regions.
Once the impact of these errors was understood, additional testing was
conducted with the CV aircraft at lower altitude (down to 7,000 feet MSL),
which provided the data to make the necessary corrections. Based on the CV
testing, lower-altitude DR testing was then performed on the CTOL and
STOVL variants, although no similar anomalies were discovered. As a result of
this additional unplanned effort, all three variants were ultimately cleared for
unrestricted high AOA maneuvering at all altitudes.

VII. SUMMARY
The F-35 high AOA development and flight test program proceeded with extra-
ordinary efficiency and resulted in a maneuvering capability that met or exceeded
all high AOA requirements. This success can be attributed to a number of factors,
among them:
. Highly accurate modeling and simulation. Fundamental to NDI is the
requirement for a highly accurate aerodynamic model, a requirement that is
inherently difficult to achieve in the high AOA regime. Numerous refinements
were made to the truth model and OBM as a result of flight test discoveries,
several of which are described in this paper. In general, the accuracy of these
models provided extremely accurate predictions prior to flight, allowing
the high AOA test program to proceed quickly with minimal downtime
for analysis.
. CLAW design. In addition to accurate models, NDI requires extensive real-
time information to be provided to CLAW regarding the aerodynamic state
and mass properties of the aircraft. Exploiting this state information in the
high AOA regime allowed limiters to be precisely tailored to the aircraft
state, the result being a design that is highly departure-resistant while extract-
ing the maximum capability from the aircraft.
. Robust OCF recovery capability. Implementing a proven (gain-scheduled)
approach for the automatic recovery modes leveraged lessons learned from
past programs and eliminated dependence on NDI for these critical modes.
This design decision, and the demonstration of consistent recovery from
worst case conditions, gave the team and leadership confidence to remove
the SRC for the vast majority of the DR phase. Doing so allowed the DR
phase, which comprised nearly two-thirds of the high AOA flight test
program, to be conducted in a production-representative configuration,
precluding the need for costly regression testing.
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 571

. Flight test aids. FTAs (particularly the ability to set AOA limiters and to
modify fuel burn to control CG) were critical to the efficiency and precision
with which data were collected throughout the high AOA program. Specialized
FTAs, such as those used to support the spin recovery demonstration, enabled
testing that would not have been possible otherwise.
. Flight test preparation and training. Although not discussed in this paper,
extensive planning and rehearsal preceded every test flight, resulting in an
extremely low (less than 2 percent) test point re-fly rate.

As can be expected of any test program of this scope and complexity, many lessons
can be learned that, even if impractical for implementation on the F-35 program,
could potentially be used in future programs. These are:
. Start high AOA testing earlier. The programmatic need to provide an initial
fleet operational capability (below 20 degrees AOA) delayed the start of high
AOA testing until nearly four years after the SDD flight test program began.
The delayed start, and the requirement to clear all variants to 50 degrees
AOA by mid-2015, compressed the calendar time and reduced the number
of Operational Flight Program (OFP) release cycles available to resolve high
AOA issues. Additionally, the schedule compression forced testing to be
done on all three variants concurrently, limiting the extent to which lessons
learned on the lead aircraft (CTOL) could be exploited by the others. An
earlier look would have provided greater efficiency and more fly-fix-fly
opportunities (OFP cycles) for refining the OBM and high AOA CLAW.
. Streamline the OFP release cycle process. Other factors limiting the number
of OFP releases possible during the SDD program were the personnel and
laboratory time required to develop, test, and verify each software update.
Each OFP was a comprehensive air vehicle release that combined updates
for all disciplines. This software release process, appropriate for more deter-
ministic disciplines and systems, was a pacing item in the development of
the high AOA CLAW, which is intrinsically a fly-fix-fly process. Since the
CLAW (in particular, the OBM) functions independently of all other vehicle
functions, the use of an agile process for turning CLAW-unique software
updates during high AOA testing is worthy of consideration in order to
maximize the number of iterations available for optimization.
. If implementing NDI at high AOA, expect to need innovative solutions. The
unsteady, nonlinear nature of aerodynamics at high AOA makes accurate
modeling, which is intrinsic to NDI, a challenge. Errors in the OBM in this
regime are inevitable and cannot always be remedied with model updates
alone. Adaptive, customized solutions that fall outside the normal NDI struc-
ture may be required.
. Allow freeform testing within the cleared envelope. Perhaps to a greater
extent here than in any other area of flight test, modeling uncertainty gives
572 D. G. CANIN ET AL.

rise to the potential to discover unknown unknowns (i.e., issues that the
designers had not thought of or were not evident in simulation) during high
AOA testing. Accordingly, once the basic envelope has been cleared and a
robust recovery capability has been established, some amount of time should
be allowed for unscripted exploration. Allowing the test pilot to explore the
handling characteristics of the airplane and to fly in an operationally
representative manner would allow issues to be discovered that may not
have presented themselves during structured test point execution.

VIII. CONCLUSION
Although high AOA was not a primary consideration in the aircraft’s design,
the F-35’s advanced flight control system provides a high AOA maneuvering
capability that meets or exceeds all program requirements, including post-stall
maneuvering, extreme DR, and automatic recovery from OCF. In providing
this capability, the F-35 not only pioneered the use of NDI on a production
fighter but also demonstrated its application in the inherently challenging
environment of high AOA.

ACRONYMS
AOA Angle of Attack
APR Automatic Pitch Rocker
CG Center of Gravity
CLAW Control Laws
CTOL Conventional Takeoff and Landing
CV Carrier Variant
6-DOF Six Degree of Freedom
DR Departure Resistance
EB Effector Blender
EHA Electro-hydrostatic Actuators
FBW Fly-By-Wire
FTA Flight Test Aid
HT Horizontal Tail
LEF Leading Edge Flap
MPR Manual Pitch Rocker
NDI Nonlinear Dynamic Inversion
OBM Onboard Model
OCF Out-of-Control Flight
OFP Operational Flight Program
PA Powered Approach
PTMS Power and Thermal Management System
SDD System Development and Demonstration
F-35 HIGH ANGLE OF ATTACK FLIGHT CONTROL DEVELOPMENT AND FLIGHT TEST RESULTS 573

SRC Spin Recovery Chute

STOVL Short Takeoff and Vertical Landing
TEF Trailing Edge Flap
TNS Tactical Navigation System
TRO Transonic Roll-Off
WBD Weapons Bay Door
UA Up-and-Away

REFERENCES

[1] Robbins, A., Bobalik, J., De Stena, D., Plag, K., Rail, K., and Wall, K., “F-35 Subsystems
Design, Development, and Verification,” AIAA Aviation Forum, Atlanta, GA, June
2018 (unpublished).
[2] Harris, J., and Stanford, J., “F-35 Flight Control System Design, Development and
Verification,” AIAA Aviation Forum, Atlanta, GA, June 2018 (unpublished).
[3] Bordignon, K., and Bessolo, J., “Control Allocation for the X-35B,” 2002 Biennial
International Powered Lift Conference and Exhibit, AIAA 2002-6020,
November 2002.
[4] Ogburn, M. E., et al., “High Angle of Attack Nose Down Pitch Control Requirements
for Relaxed Static Stability Combat Aircraft,” NASA High Angle of Attack
Technology Conference, NASA Langley Research Center, Hampton, Virginia,
30 October to 1 November 1990, NASA CP-3149, Part 2, Paper No. 24, 1992.
[5] McNamara, W. G., et al., “Navy High Angle of Attack Pitch Control Margin
Requirements for Class IV Aircraft,” NAVAIRWARCENACDIV Technical
Memorandum 91-167 SA, 25 June 1992.
[6] Nguyen, L. T., and Foster, J. V., “Development of a Preliminary High Angle of
Attack Nose Down Pitch Control Requirement for High Performance Aircraft,”
NASA Technical Memorandum 101684, February 1990.
[7] MIL-STD-1797B, Flying Qualities of Piloted Aircraft, 2006.
[8] Bowman, M., and Bemridge, A., “The Automatic Low Speed Recovery Function of
the Eurofighter Typhoon Aircraft and how it was Flight Tested,” Society of
Experimental Test Pilots, 2004.
[9] MIL-F-83691B, Military Specification: Flight Test Demonstration Requirements
for Departure Resistance and Post-Departure Characteristics of Piloted Airplanes,
March 1991.
 CHAPTER 16

F-35 Weapons Separation Test

and Verification
Christopher F. Hetreed , Matthew D. Carroll† and
Joe E. Collard‡
Lockheed Martin Aeronautics Company, Fort Worth, TX
Richard C. Snyder§
Naval Air Warfare Center, Naval Air Station Patuxent River, MD

Weapons separation flight testing is the most visible phase (Fig. 1) of

a military aircraft program’s store certification process. Store
separation flight tests, commonly video-recorded from a chase aircraft,
are generally regarded as both verification that the weapon release
requirements have been satisfied and validation that the system will
perform as expected during operational use by the customer. During
the System Development and Demonstration (SDD) phase of the F-35
Lightning II program, Lockheed Martin has provided the F-35 Joint
Program Office (JPO) with a “certification recommendation” to load,
carry, and release 11 weapons in 20 different loadings across the
F-35A, F-35B, and F35C variants. The basis for both Lockheed
Martin’s certification recommendation and verification that the
weapon release requirements have been satisfied is flight test –validated
Modeling and Simulation (M&S) that has been adjusted based on data
gathered during the store separation flight test events, so that
simulated store separation events replicate the actual store separation
events. During the test program, the flight test team rapidly and
successfully performed 183 safe store separation events, without the
need for any reflown points.

I. NOMENCLATURE
6DOF six degrees of freedom.
F-35A/B/C the 3 F-35 Joint Strike Fighter (JSF) variants.
IMU inertial measurement unit; measures store rates and accelerations.

 AeronauticalEngineer Senior Staff, F-35 Weapons.

†
Aeronautical Engineer Staff, F-35 Weapons.
‡
Aeronautical Engineer Manager, F-35 Weapons.
§
F-35 Store Separation Lead and Task Area Expert, F-35 Lightning II Joint Program Office/NAVAIR-4321.

575
576 C. F. HETREED ET AL.

Fig. 1 F-35 Lightning II weapon separation flight test.

S&RE/AME aircraft devices used for carriage, suspension, employment, and

jettison; Suspension and Release Equipment (S&RE) includes
missile and bomb carriage/ejector racks, and pneumatic power
systems; Alternate Mission Equipment (AME) includes adapters,
missile rail launchers, pylons, etc. In this paper, S&RE includes
AME, for convenience.
store any device intended for internal or external carriage and
mounted on aircraft suspension and release equipment, whether
or not the item is intended to be separated in flight from the air-
craft. Stores include missiles, rockets, bombs, nuclear weapons,
mines, torpedoes, pyrotechnic devices, detachable fuel and spray
tanks, line-source disseminators, dispensers, pods (refueling,
thrust augmentation, gun, electronic-counter measures, etc.),
targets, cargo-drop containers, and drones.
TM kit telemetry kit; for transmitting 6DOF/IMU measured store rates
and accelerations.

II. INTRODUCTION
The F-35 Lightning II Joint Strike Fighter (JSF) is a 5th-Generation fighter attack
aircraft designed for survivability, lethality, supportability, and affordability, a cul-
mination of the JSF program [1]. In order to meet the needs of multiple military
services and international customers, while still maintaining affordability, there
are three unique F-35 variants [2] with many common aircraft and systems com-
ponents; for example, the F-35A and F-35C share a common weapon bay, capable
of carrying larger air-to-ground weapons than the vertically landing F-35B. But
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 577

unlike the F-35C variant, which requires larger wings to enable slower approach
speeds for landing on naval aircraft carriers, the F-35A and F-35B variants share a
weight-saving, similarly shaped smaller wing. During the SDD phase of the JSF
program, 11 weapons were certified for carriage and release in 20 different load-
ings across these three aircraft variants (Fig. 2)—a challenging process that
includes design, modeling, simulation, and testing across multiple engineering
disciplines.
Many engineering disciplines follow the same general path (Fig. 3), starting
from system requirements and proceeding to design, then to Modeling and Simu-
lation (M&S), and then to testing for the purposes of M&S adjustment/validation
and verification that requirements have been met. In the end, requirements ver-
ification is accomplished using both testing and test-validated M&S.
The left-to-right process for the weapons (store) separation discipline is
similar, with additional detail depicted in Fig. 4 for the store separation modeling,
simulation, and flight test path to store certification, and summarized within this
paper. Maintaining affordability is critical in both the development and sustain-
ment of the F-35 Lightning II, and it is shown in Ref. [3] that several techniques,
such as Computational Fluid Dynamics (CFD) studies, were used to minimize
preflight wind tunnel aerodynamics testing costs and reduce risk, prior to flight
testing. But the most significant cost in the assessment of safe store separation,
on the path to store certification, is usually flight testing. So with attention to

Fig. 2 F-35 Lightning II SDD weapons, certified for carriage and release.
578 C. F. HETREED ET AL.

Fig. 3 Requirements to verification.

affordability, the philosophy on the F-35 Lightning II program was to extract as

much accurate information during the fewest number of weapon separation
flight test events, to confidently validate modeling and simulation in order to
support this high-confidence safe store separation assessment and certification
recommendation to the JPO.
Due to challenging requirements, 183 store separation test points were
required for M&S validation over the weapons release flight envelope. In order
to maintain an aggressive schedule, rapid progression through these test points
was necessary, with no tolerance for delay—the Edwards Air Force Base (EDW)
and Patuxent River Naval Air Station (PAX) flight test teams needed the ability
to execute store separation events on back-to-back days. Within hours of
receipt of flight test data, it was necessary for Lockheed Martin to understand

Fig. 4 Store separation path to certification.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 579

and communicate not only what happened during a separation event, but also
how it happened, especially if there was an unexpected store trajectory.
Lessons learned and recommendations from previous store separation pro-
grams were combined with inputs and detailed planning from data processing,
instrumentation, survey/metrology, mass properties, store separation, flight
test, and weapons experts within Lockheed Martin, EDW, PAX, U.S. Naval Air
Systems Command (NAVAIR), U.S Air Force SEEK EAGLE Office (AFSEO),
U.K. Ministry of Defense (UK MoD), Arnold Engineering Development Center
(AEDC), Boeing, Raytheon Missile Systems, MBDA Missile Systems, and
Raytheon Systems Limited. This planning, analysis, cooperation, and execution
resulted in the F-35 SDD weapon separation test program, with the following
highlights:
1. 183 safe store separation flight test events, across three aircraft variants, at
several ranges at two test sites
2. 121 in-bay weapon releases:
a. 58 large and/or bay-defining air-to-ground stores
b. 41 air-to-air missiles
c. 22 air-to-ground stores from four-place ejector racks
3. 62 external weapon releases:
a. 32 air-to-air rail-launched missiles
b. 30 air-to-ground pylon-mounted stores
4. One- to two-day typical approval/clearance to next test event:
a. On-board high-speed camera and 6DOF store telemetry data analysis,
computation of flight test-based aerodynamics increments, and adjusted
predictions for the next store separation test event were usually accom-
plished within one day.
5. 17 flights with multiple store releases; “in-control-room” clear-to-next-point:
a. Three flights with four store releases
6. One weapon fin-activation lanyard redesign/reconfiguration
7. Zero changes in ejector rack pitch settings
8. Zero unsafe store trajectories; zero store separation– related test program
delay/pauses
9. Zero reflown test points

III. F-35 LIGHTNING II AND VARIOUS SDD WEAPONS

In the SDD phase of the F-35 program, there was a requirement for 11 weapons to
be loaded, carried, and employed from three aircraft variants, from internal
weapon bays and underwing; detailed descriptions of design considerations and
hardware are presented in F-35 Weapons Design Integration [4]. In summary,
580 C. F. HETREED ET AL.

Fig. 5 F-35A SDD weapons carriage.

the F-35 Lightning II has 11 weapon stations (Fig. 5) and carries and releases 250-,
500-, 1000-, and 2000-lb-class air-to-ground bombs from selectable-performance
pneumatic ejector racks mounted within space-constraining weapons bays and
underwing pylons. Air-to-air missiles are carried and released from pneumatic
ejector racks within the bay, and from underwing rail launchers.
The pneumatically actuated, two-piston BRU-68 ejector rack (Fig. 6) is
capable of releasing 500-, 1000-, and 2000-lb-class bombs from the F-35A/C
internal weapons bay, or from the F-35A/B/C external weapons stations on the
wing when installed within an air-to-ground pylon. Similar to the BRU-68, the

Fig. 6 BRU-67/BRU-68 ejector rack (top), in bay (bottom left), and in pylon (bottom right).
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 581

smaller pneumatically actuated BRU-67 ejector rack is capable of releasing 500-

and 1000-lb-class bombs but is designed for use within the F-35B internal
weapons bay. Both the BRU-68 and BRU-67 have multiple pitch valve settings
that allow for selectable pitch performance based on which weapon is installed.
Inside the weapons bay, the Pneumatic Power System (PPS) pressurizes the
bomb rack to its operating pressure. Air-to-ground pylons each contain a separate
PPS for pressurizing bomb racks at external/wing weapons stations.
The LAU-147 (Fig. 7) is a pneumatically actuated, two-piston ejection
launcher that can be installed on the F-35A/B/C at either internal weapons bay
stations. LAU-147s are capable of launching medium range air-to-air missiles
such as the AIM-120 AMRAAM. Along with pressurizing any BRU-68 or
BRU-67 in the weapons bay, the same in-bay PPS pressurizes any LAU-147 to
its operating pressure.
The BRU-61 (Fig. 8) is a four-station pneumatically actuated ejector rack that
is capable of releasing up to four 250-lb-class bombs including the GBU-39 Small
Diameter Bomb (SDB) and the GBU-53 SDB-II. The BRU-61 can be installed in
the F-35 weapons bay on a fixed adapter or on a BRU-68 connected to an
air-to-ground pylon on the aircraft wing. The BRU-61 has a self-contained com-
pressor, independent of the PPS in both the weapons bay and the air-to-ground
pylons. The four stations on the rack include an aft-inboard, forward-inboard,
aft-outboard, and forward-outboard rack station, each with selectable-
performance pitch valves and two pistons. The release order on the BRU-61
bomb rack is customizable.
During the internal weapon bay design phase, weapons were rigged in
Computer-Aided Design (CAD) software, such that the stores would all be

Fig. 7 LAU-147 missile ejector rack.

582 C. F. HETREED ET AL.

Fig. 8 BRU-61 with four-pack of GBU-39 small diameter bombs.

sufficiently far away from in-bay hardware during carriage and release from the
F-35 weapon bays. Reference [4] contains detailed descriptions of internal
weapon bay design considerations, along with descriptions of on-aircraft
weapons-mounting hardware, such as the Ordnance Quick Latch System
(OQLS). Stores have different sizes, shapes, and carriage lug locations, so it was
not feasible to have a single ejector rack location that would be sufficient for car-
riage and employment of all stores. Some stores required a more-forward ejector
rack location, relative to the aircraft OQLS, whereas others needed to be carried
higher within the weapon bay, for instance. Therefore, various adapters are
used (Fig. 9) to ensure adequate aircraft hardware miss distances during store car-
riage and release from the F-35. And because some of these adapters are attached
to the in-bay OQLS differently, carriage and ejection loads are transferred differ-
ently to the aircraft structure.
The LAU-148 and LAU-151 (Fig. 10) are rail launchers that are installed on
air-to-air pylons at the most outboard external F-3A/B/C weapons stations,
and are capable of launching rail-launched missiles, such as the AIM-9X Sidewin-
der and the AIM-132 ASRAAM. The LAU-148/151 restricts the missile’s vertical
and lateral motion, while the missile’s thrust propels the missile forward on
the rail.
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 583

Fig. 9 BRU-67/68 within two different adapters (blue), each with different mounting to
in-bay OQLS.

In addition to the differences in aircraft variants, weapon stations, and pre-

viously described S&RE from which stores are released from the F-35, each
store has different features that result in unique release behavior for each store
type, with respect to miss distance to aircraft hardware, aerodynamics, trajectory
motion, and stability and control while in the near-field vicinity of the aircraft. For
instance, the GBU-31 JDAM and AGM-154 JSOW stores are large, fill a majority
of the internal weapon bay, and are in close proximity to aircraft hardware, such as
weapon bay doors, when released from the internal bay. The Paveway II (GBU-12)
and Paveway IV stores include forward canards and aft tailfins, the latter of which
are deployed with the use of a lanyard device. The GBU-39 SDB is a lightweight
store and also features deployment of stabilizing fins during the store separation

Fig. 10 AIM-132 and AIM-9X air-to-air rail-launched missiles from LAU-148/151.

584 C. F. HETREED ET AL.

event. The rail-launched AIM-9X Sidewinder missile includes forward canards,

whereas the AIM-132 ASRAAM is absent of such forward aerodynamic surfaces.
Several stores have relatively low mass moments of inertia, which can result in fast
response to aircraft flow field aerodynamic forces, compared to stores with higher
mass and inertia. Some stores feature active controlled flight, using moving fin
control surfaces, when in close proximity to the F-35, whereas other stores
delay such control or propulsion activation—the ejected AIM-120 AMRAAM,
for instance, ignites its rocket motor at a delay sufficient to ensure that forward
motion starts when the store is at a safe distance below the aircraft.
Therefore, modeling, simulation, and testing must all be conducted in con-
sideration of these unique store features, behaviors, and manners in which they
are released or separated from the F-35. The store separation event is a
dynamic event, influenced by its own aerodynamic, propulsion, and inertia
forces, and by interaction with the S&RE and aircraft.

IV. MODELING
Modeling and simulation are both on the store separation path to certification, but
before considering models of the aircraft, S&RE, and stores previously described,
it is useful to understand details of typical real store separation events, followed by
consideration of how and whether such details could be modeled and simulated.
Store separation problem descriptions, hardware requirements, compatibility
guides, and weapons flight testing information are found in Ref. [5– 10, 23].
Although some of the techniques have been replaced by modern equipment,
tools, and methods, the store separation problem and general testing techniques
and goals still apply. These particular references are provided because they
served as the foundation documentation for understanding store separation fun-
damentals, aiding the implementation of newer modeling techniques, and reinfor-
cing the importance of adequate flight test instrumentation for properly validated
models and simulations. Reference [11] includes “lessons learned,” providing rec-
ommendations for focusing resources in modeling, test, and analysis, which were
generally adhered to during the F-35 store separation preflight, flight test, and
postflight activities. But this was somewhat unintentional and coincidental, yet
reassuring—the F-35 store separation team’s grassroots approach was to 1) try
to understand actual typical events; 2) consider available general-purpose model-
ing tools and techniques used in the aerospace and automotive industries; 3) try to
reasonably model detailed mechanical, structural, and aerodynamic behavior; and
4) consider how to objectively judge and adjust the simulated model’s behavior vs
the real hardware’s behavior, based on available instrumentation and typical
testing procedures. As with other disciplines such as loads, dynamics, flutter,
aero/performance, and stability and control, modeling and simulation adjust-
ments were expected as part of the validation process. Store separation component
models were built, simulated, and validated during and after component ground
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 585

testing. Store separation full-aircraft models were built, simulated, and validated
during and after full-aircraft ground and flight testing. And like those other dis-
ciplines, validated models and simulations are the basis for verification that
store separation requirements have been met.

A. THE ACTUAL EVENT

The actual event is described in a manner to help think about construction of a
virtual functional prototype—a simulation model that will behave like the real
system. During captive carriage and prior to a store separation event, the store
is constrained or restrained to the aircraft by S&RE. The store separation event
begins at the moment the store carriage state changes, and the store is released,
partially restrained, and/or pushed away from the carriage position. During
this event, the aircraft and store each respond in a rigid and flexible manner
due to mechanical, structural, inertial, gravitational, aerodynamic, and/or propul-
sive dynamic forces. Consider two examples.

1. RELEASED/EJECTED STORE EXAMPLE

An ejector S&RE system retracts its store restraint mechanism (hooks or clamps)
and the store becomes free to fall. Almost immediately at this point, the ejector
S&RE system extends pressurized pistons against the store, pushing the store
away (Fig. 11). The interactive response of the aircraft, S&RE mechanism, and
store may be complex.
The aircraft response depends on the aircraft mass and inertia, S&RE location
relative to the aircraft CG, mass of the released store, S&RE ejection force, aircraft
structural flexibility near the S&RE location, pilot aircraft control inputs, and
flight control system response, for instance. The aircraft response may be impor-
tant during the store separation event, or it may be negligible.
The S&RE behavior is influenced by the attachment constraint configuration
to the aircraft or wing pylons, local aircraft and/or multiple-store carriage system

Fig. 11 Ejected store.

586 C. F. HETREED ET AL.

structural flexibility and response, contact constraint with the ejecting store, store
mass and inertia, store CG, behavior of the extending ejection pistons from the
pneumatic or pyrotechnic charging inputs, and aerodynamic forces on the store
during the ejection process.
The store response is influenced by its structural flexibility, mass properties,
contact constraint and ejection forces from the S&RE ejector system, store
control laws inputs, and aerodynamic forces acting on the store.

2. RAIL-LAUNCHED STORE EXAMPLE

During captive carriage, two or more of a missile’s hanger brackets are constrained
within a long inner or outer rail channel on a rail-launcher S&RE system. When
the missile’s rocket motor ignites, the missile starts to move forward, and its
hanger brackets push against and retract the launcher’s forward-motion
restraint/detent mechanism. The rocket motor then continues to push the
missile forward, down the rail (Fig. 12), while the missile is partially restrained
during this launch event by its hanger brackets within the rail channel. After
tip-off, or the moment the last missile hanger bracket exits the rail channel, the
missile is no longer restrained by the launcher on the aircraft and continues to
be propelled forward. And like the ejected store example, the response of the air-
craft, launcher S&RE, and missile may be complex.
The aircraft response during a normal rail-launched missile separation event
is usually much less pronounced than during an ejected store separation event and
is generally observed in the localized structural response near the S&RE location.
The launcher S&RE behavior is influenced by the attachment constraint con-
figuration to the aircraft or wing pylons, launcher structural flexibility and
response, rail constraint contact with the missile’s hanger brackets, store mass
and inertia, aerodynamic forces on the missile and launcher, and missile plume
impingement forces, for instance.
The store response may be influenced by its own structural flexibility, mass
properties, launcher flexibility, rail channel constraint contact, propulsion force,

Fig. 12 Rail-launched store.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 587

aerodynamic forces, and fin or canard surface movements due to missile flight

control law system inputs, for example.

B. THE MODEL AND SIMULATION EVENT

The prior two actual event examples illustrate the physical system (aircraft,
S&RE, and separating store) behaving dynamically, in an interrelated
manner—structurally, mechanically, and aerodynamically. In M&S of the
dynamic behavior of such systems, the simulation portion can be regarded as
the time-dependent solution of the set of mathematical dynamics equations
representing the model of the actual physical system. Such models can range
from simple to complex, depending on the desired fidelity and intent of the simu-
lation. For example, in some M&S instances, it may be desirable to consider
mechanical S&RE behavior during the separation event, beginning at the
moment of store first motion, whereas in other instances it may be beneficial
to exclude these interactions from the model and simply start simulations after
End-of-Stroke (EOS) or after End-of-Rail (EOR), after the store would normally
lose contact with the S&RE.
Typical store separation software involves the solution of mathematical
equations of motion representing a rigid aircraft and a 6DOF store system, and
includes store aerodynamics, propulsion, controls, S&RE-to-store interactions,
and other provisions [12, 13]. In such software environments, the model, or math-
ematical equations of motion, are customized to include specific behaviors of
specific systems; when additional components, constraints, or other mechanical
effects are to be added, the equations of motion may need to be modified or cus-
tomized. Manually customizing mathematical equations of motion, to capture
modeling changes, is a task that requires tremendous skill and experience, argu-
ably shifting valuable resources away from the important analysis task.

C. LOCKHEED MARTIN’S DYNAMIC SYSTEM MODEL

Lockheed Martin’s store separation M&S software (Fig. 13), ASEP, is a plug-in to
the commercially available VI-Aircraft and ADAMS (Automatic Dynamic Analy-
sis of Mechanical Systems) mechanical system simulation software. In the model-
ing phase, virtual models of multibody rigid and flexible mechanical systems are
built up, assembled, and archived in a CAD-like environment. So, when consider-
ing physical details of an actual event, such as friction, flexible components,
surface contact, complicated restraint mechanisms, and so forth, these physical
effects are easily “built” into the virtual model. The model can include detailed
structural and mechanical behavior, as necessary. Similarly, “plugged-in” store
aerodynamics, controls, and propulsion forces can be simple or complex, and
may be derived from various sources, such as wind tunnel tests, CFD, or proprie-
tary “black box” modules furnished from ejector rack and store suppliers.
588 C. F. HETREED ET AL.

Fig. 13 Dynamic system model.

During ASEP’s simulation phase, each unique assembled system model’s

mathematical equations of motion are automatically generated and solved. So,
the inclusion of additional hardware components and physical behavior
becomes an act of simply adding them to the model (Fig. 14), without the need
to manually reformulate mathematical equations of motion.
Lockheed Martin’s ASEP M&S software features two modeling approaches—a
simpler system dynamics model to focus mainly on the store, and a more complex
system dynamics model to include complex interaction among the aircraft, S&RE,
and store. The modeling approaches are summarized as follows.

Fig. 14 ASEP full model.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 589

1. SIMPLE DYNAMICS SYSTEM: INITIAL CONDITIONS MOTION M&S

This system modeling approach, involving two or three 6DOF bodies, includes a
maneuvering rigid aircraft model, a rigid store model, and an optional 6DOF/
IMU TM Kit rigid body. This approach is used when intentionally focusing
only on the less-complicated unrestrained store free-flight region, excluding the
direct S&RE mechanical connection between the store and aircraft. The
primary focus is usually on a store’s trajectory after EOS or tip-off. The rigid
bodies are as follows.
The rigid aircraft body is “flown” in space, according to either a prescribed
steady maneuver or an exact flight-test aircraft maneuver response. In both
types of aircraft maneuvering, the aircraft is flown to match the desired aircraft
translational accelerations and rotational rates.
The rigid store body motion can be simulated in two ways. Store trajectory
motion begins from prescribed initial store displacement and velocity conditions.
Alternatively, store is constrained to follow a prescribed store motion trajectory
time-history, relative to the moving aircraft, and then optionally becomes
detached from the aircraft at any point in the store trajectory time-history at
the instantaneous displacement and velocity conditions. In this latter scenario,
the source of the prescribed store trajectory time-history may be either from a
store separation wind tunnel trajectory simulation, from a flight test event’s
photogrammetric solution store trajectory, or from a store trajectory solution
derived from flight test measured rates and accelerations.
In both of these store model motion scenarios, only after the store rigid part
becomes completely detached from constrained motion with the aircraft part and
becomes its own independent 6DOF body (Fig. 15) can its dynamic trajectory be
influenced by the forces acting on it: aerodynamics, propulsion, gravitational, and
so forth. So, such force models are also included in the system model, usually
through software links to aerodynamics databases or black box modules.
Like the rigid aircraft body and rigid store body, the optional 6DOF/IMU TM
Kit rigid body is a moving part and is used to: 1) automatically replicate a flight
test store’s actual separation trajectory motion, and 2) automatically compute the
applied aerodynamic forces that influenced this store trajectory motion. During a
posttest simulation, this 6DOF/IMU TM Kit part is initially attached to the rigid
store. At a selected point in the rigid store’s trajectory (whether at carriage or
when moving according to a prescribed photogrammetry-based motion relative
to the aircraft), the 6DOF/IMU TM Kit part becomes completely detached
from the store rigid body and “flies” according to the measured rates and
accelerations.

2. FULL DYNAMICS M&S

This system modeling approach includes a maneuvering flexible or rigid aircraft
model, a model of the S&RE components, and a flexible or rigid store model
590 C. F. HETREED ET AL.

Fig. 15 Simple dynamics.

(Fig. 16). This approach is used when simulating structural, mechanical, and aero-
dynamic behaviors of the combined aircraft/S&RE/store system—when attempt-
ing to achieve a detailed virtual model of the physical system. The simulation
starts at store carriage; continues through the release, launch, or ejection phase;
and includes the system mechanical dynamic response during the time the
store and S&RE are in contact during the separation event. The simulation con-
tinues after the store loses contact with the S&RE, after ejector piston EOS, after
launcher rail/missile tip-off, or after release of a pylon or fuel tank’s hook/ball/
socket restraint mechanism.
The aircraft model may be flexible or rigid (Fig. 17). It may be attached to a
landing gear mechanical model resting on the ground, to simulate aircraft
dynamic response during a store static ejection pit test. Or it may be “flown,”
or moved in space during simulations, according to either a prescribed steady
maneuver or an exact flight-test aircraft maneuver response.
The flexible or rigid S&RE system model is attached to the aircraft model, in a
rigid or flexible manner similar to that of the actual system, to ensure proper inter-
dependent motion and intended structural load paths between the S&RE and the
aircraft. The S&RE model (Fig. 18) could be a mechanically actuated model of an
ejector rack, including rack supplier – provided gas or pyrotechnic piston pressure
modules. Or the S&RE model could be of a flexible rail missile launcher, obtained
from Finite Element Modeling (FEM) software. The model may also include flex-
ible or rigid adapter and pylon structural models. And the S&RE is built to model
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 591

Fig. 16 Full dynamics.

Fig. 17 Flexible aircraft.

592 C. F. HETREED ET AL.

Fig. 18 S&RE system models in ASEP.

the physical interaction with the store. The S&RE subsystem model may have as
much or as little detail as desired.
The flexible or rigid store model is fully or partially constrained or restrained
by the S&RE mechanical model during the store separation simulation. In order to
simulate the interaction forces between the store and S&RE, contact models
between the store and S&RE are usually included, and may or may not include
friction. This contact could be between a store body and S&RE ejector pistons
and/or piston sway brace pads. Or this contact could be between store hangers
and a launcher’s rail channel. This contact could even be a complicated hook/
ball/socket constraint typically seen in a rear-pivot mechanism connecting a
fuel tank to a pylon, or a pylon to an aircraft wing. And these contact models
may be as simple or as complex as necessary to replicate the physical interaction
between the store and S&RE. In the simulated separation event, during and after
constraint by the S&RE, other forces also act on the store, such as aerodynamics,
propulsion, gravitational, and so forth. So, such force models are also included,
usually through software links to aerodynamics databases or black box modules.

D. AERODYNAMICS MODELS
The store aerodynamics models (Fig. 19) are usually derived from wind tunnel
testing [14], but can be replaced by, or combined with, CFD-generated store aero-
dynamics [15]. In building store separation aerodynamics models for released
stores from all SDD loadings on the three F-35 aircraft variants’ internal and
external weapon stations, 11 total wind tunnel tests were conducted, with over
3900 User-Occupancy Hours (UOH) and over 1700 air-on hours. One-
fifteenth-scale aircraft and store models were used in order to minimize wind
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 593

Fig. 19 Store aerodynamics model.

tunnel test costs by using the smaller 4-ft transonic/supersonic wind tunnel at
AEDC, instead of AEDC’s 16-ft transonic/supersonic wind tunnel. Details of
store separation wind tunnel testing are best-described in wind tunnel literature
and test reports, so only a summary is provided herein. During these tests,
store aerodynamic forces were obtained from a load-measuring balance, to
which the store was attached, during up to four test scenarios.

1. ATTACHED LOADS
The store and balance are attached to a strut or other mounting hardware directly
on the aircraft model, usually within the weapon bay (Fig. 20). In this scenario,
aerodynamics forces are obtained for the store, usually in a “carriage” configur-
ation for two purposes: computing store-applied airloads onto the aircraft and
obtaining aerodynamic forces in regions, such as a weapon bay cavity, where a
sting-mounted store model cannot reach.

Fig. 20 Attached loads.

594 C. F. HETREED ET AL.

2. STORE-ONLY FREESTREAM
The store and balance are attached to a sting and swept through various angles
relative to the freestream airflow direction (Fig. 21). Comparisons are usually
made to larger-scale (full-scale, 1/2-scale, 1/4-scale, for instance) freestream aero-
dynamics models, obtained from weapon suppliers, to quantify uncertainty associ-
ated with the 1/15-scale model’s “small-scale freestream” aerodynamics model.

3. FLOWFIELD GRID
The store and balance are attached to a sting, identically to that in the freestream
store-and-balance sting arrangement, but the aircraft model is also present in the
wind tunnel test section, mounted on an independent sting. In this arrangement,
the aircraft and store can be moved independently, enabling the store to be posi-
tioned relative to the aircraft, while the aircraft is positioned relative to the free-
stream air flow to simulate various aircraft angles of attack and angles of
sideslip (Fig. 22). This type of testing is known as the Captive Trajectory
System (CTS) at AEDC, or the Two-Sting-Rig (TSR) system at Aircraft Research
Association (ARA). The main goal in this arrangement is to measure store aero-
dynamics forces at various positions in the vicinity of the aircraft that populate the
store flowfield aerodynamics database.

4. FLOWFIELD TRAJECTORY SIMULATION

In the same arrangement as in the flowfield grid scenario, the aircraft is positioned
at the desired angle of attack and angle of sideslip, relative to the wind tunnel free-
stream air flow, and the store is positioned, relative to the aircraft, as close to

Fig. 21 Freestream.
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 595

Fig. 22 Flowfield CTS.

carriage as practical. Measured store aerodynamics forces are input into the wind
tunnel facility’s trajectory simulation software, where 6DOF equations of dynamic
rigid-body motion are solved in time, for the aircraft flight condition of interest,
and the resulting “next time step” store position is output to the store sting con-
troller, resulting in a store movement to this next position. Store ejection forces
and constraints are usually included in the online trajectory simulation model,
as well as weapon supplier– provided black-box autopilot control and propulsion
forces. This mutual-feedback measure/simulate process results in a simulated
store trajectory, which is used in several ways. The collection of trajectories
results in a range of combined store positions and attitudes, which can influence
the test’s aerodynamic flowfield database grid density and ranges, while at the
wind tunnel. Online trajectory simulations at the wind tunnel are usually a
quick source for comparing trajectories of different aircraft weapons loadouts.
During and after the wind tunnel test, comparisons of 1) simulations involving
the same exact trajectory aerodynamics, and 2) simulations involving the grid-
based aerodynamics database are conducted—“wind tunnel emulation” simu-
lations can provide additional insight into grid-based aerodynamic model contri-
butions to uncertainties in aerodynamics forces. More importantly, these wind
tunnel emulations are used to ensure that offline aerodynamic grid-based simu-
lations adequately replicate wind tunnel online trajectory simulations.
596 C. F. HETREED ET AL.

5. COMPUTATIONAL FLUID DYNAMICS

Although the industry-standard source for store aerodynamics models has usually
been wind tunnel tests, CFD is gradually gaining acceptance. During the F-35
SDD phase, CFD has not been a primary source for store separation aerodynamics
database models, primarily because a rigorous comparison with flight test –
derived aerodynamics has not yet been completed. Many comparisons between
CFD and wind tunnel have been made, with more-favorable comparisons
taking place for underwing positions vs in-bay positions. Ultimately, however,
the more important CFD comparisons will take place between CFD-based and
flight test –derived “truth” aerodynamics, instead of between CFD-based and
wind tunnel –based aerodynamics. The derivation of flight test “truth,” or flight
test –derived aerodynamics, is presented in the flight testing section of this paper.
During F-35 SDD, store separation CFD analysis has been used as a low-cost
alternative to wind tunnel testing [16]. For instance, CFD was used to identify
similar flowfield regions among the three different F-35 aircraft variants, in
order to more efficiently perform wind tunnel testing. During such CFD
studies, it was discovered that underwing store separation testing could effectively
be performed with either the F-35A or F-35B variant (Fig. 23), and the wind
tunnel store separation aerodynamics were applicable for both variants. Further-
more, it was also discovered that transonic in-bay store separation wind tunnel
testing on either the F-35A or F-35C was equally applicable to both variants
(Fig. 24). It was not necessary to perform duplicative store separation wind
tunnel testing using all three F-35 aircraft variant models.
Store configuration effects on the aircraft flowfield were also analyzed. In the
early stages of GBU-39 Small Diameter Bomb (SDB) integration on the F-35, the
GBU-39 fins were set to deploy at a time after first motion such that the store was
out of the region of influence of the aircraft flowfield. F-35/GBU-39 wind tunnel
testing and preliminary store separation analysis were therefore conducted in a
fins-stowed store configuration. Following this F-35/GBU-39 wind tunnel
testing, the default fin deployment timeline was shortened during GBU-39

Fig. 23 Flowfield comparison.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 597

Fig. 24 Variant comparison.

development testing on non-F-35 store separation programs. As a result of this

store change, when released from the F-35, the GBU-39 fins would then deploy
while the store would still be in the region of influence of the F-35 aircraft flow-
field. There were concerns that the underbay flowfield might be different for the
GBU-39 with fins stowed and the GBU-39 with fins deployed. With F-35/GBU-39
flight testing on the near-term schedule, CFD analysis was chosen as an efficient,
cost-effective method of quantifying the flowfield differences with GBU-39 in the
fins-deployed configuration, possibly without the need for additional wind
tunnel testing.
Simulated trajectories with a purely fins-stowed configuration (Fig. 25) still
applied for the GBU-39 jettison case (cyan store). For the GBU-39 employment
scenario with fins deploying early (green, yellow, and red stores), simulated
store trajectories that used the supplemental CFD aircraft flowfield coefficients
from the GBU-39 in the fins-deployed configuration were similar to simulated tra-
jectories using wind tunnel aircraft flowfield coefficients derived from the GBU-39
in the fins-stowed configuration. Because of the availability of CFD analysis and
the aforementioned trajectory similarities, store separation analysts were able to

Fig. 25 GBU-39 CFD with fins deployed (left); simulated trajectories (right).
598 C. F. HETREED ET AL.

Fig. 26 AIM-9X surface pressure comparison LAU-148 vs LAU-151.

conclude that the fins-deployed effect on the aircraft flowfield was negligible.
Unfortunately, aerodynamics modeling schemes, such as combining store free-
stream and flowfield aerodynamics to obtain total aerodynamics, are beyond
the scope of this paper. However, it was concluded that for simulations with
the fins-deployed configuration, complete GBU-39 store aerodynamics, as a com-
bination of fins-deployed GBU-39 freestream aerodynamics plus fins-stowed
GBU-39/F-35 flowfield aerodynamics from the original F-35/GBU-39 wind
tunnel test, were applicable; the adverse budget and schedule effects of an
additional wind tunnel test were avoided.
CFD has also been used to quickly assess changes in the air-to-air launcher
surface geometry, with respect to quantifying differences in aerodynamic forces
on the launching AIM-9X missile (Fig. 26). In performing such a comparison,
the goal was to demonstrate that it was reasonable to conclude that the
AIM-9X aerodynamics model and store separation flight test trajectory results
applied equally for two different launchers.
Postflight CFD studies have also been used, for example, to confirm that
differences between the GBU-32 JDAM wind tunnel – based aerodynamics
model and flight test – derived aerodynamic forces were likely due to store rear-
ward movement into a high pressure gradient region underneath the aircraft aft
end (Fig. 27). This particular study highlighted the importance of obtaining
store aerodynamics grid data, beyond the limits of the fore – aft grid that were
included in the relevant wind tunnel test.

Fig. 27 CFD to investigate flight test store aero differences.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 599

V. GROUND TESTS: VALIDATION OF THE DYNAMIC SYSTEM MODEL

The usefulness and value of both detailed dynamic systems models and aero-
dynamics models diminish if such models are not judged and validated, or appro-
priately adjusted, to be representative of the real dynamic system and real
aerodynamics. There are a series of ground-based validation tests (Fig. 28) in
which information is gathered and used for preflight system model validation,
whereas store aerodynamics model validation can only be performed using infor-
mation gathered in flight test.
Recall that the dynamic system includes the store, S&RE, and aircraft. Some-
times it is impractical or difficult to successfully validate the full system that
includes the aircraft. Even though individual component behavior may change
when it becomes part of a larger system of components, it remains good practice
to validate isolated component models before attempting to validate the entire
system, especially because it can sometimes be difficult to isolate areas of
concern, in the event that full-system validation becomes unsatisfactory.

A. STATIC S&RE EJECTION TEST

Hundreds of BRU-67, BRU-68, and LAU-147 static rack ejection tests were con-
ducted by the rack supplier, EDO LLC, in support of required performance,
environmental, durability, and qualification testing. These ejections were con-
ducted at room temperatures, at hot and cold temperatures, at various starting
piston accumulator pressures, and with various ejector pitch performance settings

Fig. 28 Dynamic system model validation.

600 C. F. HETREED ET AL.

(if applicable). Because the BRU-67 and BRU-68 racks were designed to carry and
release different store weight classes, 500-lb, 1000-lb, and 2000-lb stores were
ejected. In addition, different racks with different production serial numbers
were used. Furthermore, acceptance testing requirements are such that every
ejector rack must be ejection-tested at least once, prior to customer delivery. In
all of these tests for the F-35’s BRU-67, BRU-68, and LAU-147 ejector racks,
store EOS vertical velocity and pitch rate were measured or computed. This
wealth of data was useful for quantifying rack-to-rack performance variation
for uncertainty simulations/analysis.
Lockheed Martin’s BRU-67, BRU-68, and LAU-147 S&RE ejector system
models are based on EDO’s mechanism model and include EDO’s proprietary
black box pneumatic piston force gas model, so the ejector rack S&RE model vali-
dation consisted of comparing Lockheed Martin’s ASEP ejection simulations of
the BRU-67 and BRU-68, using 500-lb, 1000-lb, and 2000-lb store models with
mass properties identical to those in EDO’s tests (Fig. 29). The same model vali-
dation was performed for the LAU-147 S&RE ejector system model, using a store
model with the same mass properties as EDO’s AIM-120-like test store. Lockheed
Martin’s test-validated BRU-67, BRU-68, and LAU-147 S&RE system models
were the basis for store separation preflight predictions.

Fig. 29 Rack static ejection test.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 601

Unfortunately, in the alternate dynamics modeling scenario of using a simple

S&RE model, where only EOS displacement and velocity initial conditions are
specified for specific store types and mass properties, these EDO test data are
useful only for establishing nominal simulation store initial condition vertical vel-
ocity and pitch rates if the store separation simulations are performed with store
mass properties identical to those tested by EDO.

B. STORE VIBRATION TEST

When the dynamic response of flexible dynamic stores was important, such as
during AIM-9X and AIM-132 air-to-air missile launches, the store and pylon
system simulation models included a flexible representation, usually from the
store supplier’s finite element model. The store supplier typically performs its
own finite element model validation using store vibration tests. Lockheed
Martin’s flexible missile store models used test-validated AIM-9X and AIM-132
missile FEM models from Raytheon Missile Systems and MBDA Missile
Systems, respectively.

C. GROUND VIBRATION TESTS (GVT)

The Lockheed Martin dynamic loads integrated product team conducted
on-aircraft ground vibration tests to validate finite element models, which
included flexible aircraft, pylon, launcher, and missile models, in addition to
contact stiffness or freeplay in various degrees of freedom at every wing-to-pylon
attachment bushing, every pylon-to-launcher attachment point, and every
launcher rail channel-to-missile hanger/lug contact point. After adjusting their
own flexible dynamics systems models, the dynamic loads team also similarly
adjusted the flexible ASEP models shared with the store separation team. These
GVT test-validated flexible AMR/S&RE system models were used in store separ-
ation preflight predictions for rail-launched flight test events.

D. ON-AIRCRAFT STATIC EJECTION PIT TEST

Some 158 on-aircraft SDD store ejection events were conducted from three F-35
test aircraft that were used for a majority of store separation flight testing, with
each of these aircraft variants parked (and resting on the landing gear) over a
pit filled with foam (Fig. 30). The main purposes were to 1) validate the aircraft-
installed S&RE ejector rack model performance, 2) validate S&RE-to-aircraft
loads models, 3) verify on-aircraft stores management system communication
and function prior to flight testing, and 4) verify on-aircraft instrumentation func-
tion prior to flight testing. The on-aircraft static ejection pit test was the only
opportunity to perform store þ S&RE þ aircraft dynamic system model validation,
prior to flight testing, using the same aircraft and weapon stations, store types, and
instrumentation that would ultimately be used during subsequent store separation
in-flight testing.
602 C. F. HETREED ET AL.

Fig. 30 On-aircraft static ejection pit test.

During system model validation, aircraft landing gear models were also
included in the system model to replicate the as-tested system, where during
the downward ejection event, the aircraft always reacted with an upward motion
with some aircraft roll. The general effect was a slight reduction in ejector forces
and performance, relative to a nearly rigid S&RE-only pit ejection test, while also
imparting some store EOS roll motion during ejections taking place at external
wing stations. At that time, there was an expectation that the in-flight aircraft
would also exhibit an upward and roll response during store ejection, which
later proved to be true. After on-aircraft static ejection pit tests, the validated
store þ S&RE þ aircraft system models were used in final preflight test store sep-
aration trajectory predictions for all of the planned flight test events, during which
validation of the store aerodynamics model was expected to occur.

VI. TYPICAL STORE SEPARATION FLIGHT TEST INSTRUMENTATION

A brief description of typical store separation flight test instrumentation analysis
is included in this section, in recognition of its existence prior to the F-35 flight
test program. These same types of instrumentation would eventually be used
during the F-35 store separation test program, so an understanding of the basic
instrumentation was necessary. F-35-specific planning, considerations, and
additional uses are described in the next section.
The act of judging the extent to which store separation simulation matches
flight test has typically been focused on “what happened”—comparing simulated
and actual store displacement trajectories and rotational rates. The determination
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 603

of the actual store displacement trajectory and store rotational rates has evolved
from estimation to high-confidence calculations and direct measurement, using
one or both of the following: photogrammetric analysis of on-aircraft weapon sep-
aration camera video, and numerical integration of measured translational accel-
erations and rotational rates from either an in-store IMU or a 6DOF telemetry kit,
collectively called “6DOF/IMU TM Kit” herein [17 – 19].

A. PHOTOGRAMMETRIC ANALYSIS
Video-dependent high-quality photogrammetric analysis results in an accurate
computation of store translational and rotational displacement, relative to the air-
craft. Such accurate computations are usually completed within 8 –16 hours, and
careful numerical differentiation of this store position solution results in a store
velocity solution, and even a store acceleration solution. Photogrammetric analy-
sis includes using high-capture-rate video imagery (Fig. 31) of the store and air-
craft, and uses reference locations, such as high-contrast photogrammetry
(photoG) targets and store geometric features.

1. PHOTOGRAMMETRY PROS
The direct product of the photogrammetry solution is position, and is therefore
considered the position “truth,” which is especially useful in computing store
miss distance to aircraft hardware.
The photogrammetric position solution is usually a better initial condition
starting point for complementary 6DOF/IMU TM Kit– based trajectory sol-
utions, especially when the store, S&RE, and/or airframe maneuver-dependent
flexibility and free-play results in mismatch between ideal CAD position and
actual inflight carriage position.
Even if targets become obscured, a successful position solution is still possible
by overlaying CAD wireframe onto video imagery, if the store outline is still
visible.

Fig. 31 Store separation video cameras [17] and typical overlaid solution.
604 C. F. HETREED ET AL.

The cameras required for photogrammetric analysis are also useful for captur-
ing video evidence of other events, such as store, S&RE, or lanyard function.

2. PHOTOGRAMMETRY CONS
Too much or too little light may render imagery unusable for photogrammetric
analysis. In addition, vapor condensation may obscure store and/or reference
geometry and targets. Also, unless properly accounted for, vibrating cameras
may decrease solution accuracy. Further, unless properly smoothed, numerically
differentiated positions may produce noisy velocity solutions, and especially noisy
acceleration solutions. Compounding this, typical video frame rates of 200 –400
frames per second may not provide enough resolution for computing store aero-
dynamics from acceleration solutions.

B. 6DOF/IMU TM KIT ANALYSIS

The main strength of the in-store 6DOF/IMU TM Kit (Fig. 32) is that it provides
a direct measurement of translational acceleration and rotational rate relative to
the ground; when combined with numerical integration of the translational accel-
eration, the result is regarded as the best source of store 6DOF “velocity truth.” An
added strength is that 6DOF/IMU TM Kit measurements are usually available
immediately to engineers in the flight test control room, as the store separation
flight test event takes place. Careful numerical integration of the store velocity

Fig. 32 6DOF TM Kit with store nose installation [19].

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 605

results in a store position solution, relative to the aircraft, as long as the aircraft
motion is accounted for.

1. 6DOF/IMU TM KIT PROS

The output is a telemetered fast and direct measurement of store rates and accel-
erations, so little to no data manipulation is needed to interpret basic store behav-
ior, thereby enabling in-control-room “continue to next point” decisions for
missions with multiple separation events per flight.
High sample rate (. 4500 samples per second) enables many filtering strat-
egies, which is useful when differentiating rotational rates to obtain rotational
accelerations.
Modern 6DOF TM Kits have 1) delayed turn-on, to save battery during long
missions; and 2) delayed transmission, which increase the opportunity to resolve
regions of data loss by “stitching” recorded real-time and delayed transmission
data streams.

2. 6DOF/IMU TM KIT CONS

Measured data are telemetered, sometimes resulting in data dropout and/or noise.
Also, computing accurate store positions, relative to the maneuvering aircraft,
requires accurate initial conditions and knowledge of aircraft maneuver response.
In addition, data sometimes require alignment/nulling with aircraft rates and
accelerations. Further, sometimes data from all 6DOF are not properly time-
synchronized.

3. FURTHER ANALYSIS
It may initially appear as if redundancy exists when both of these flight test data
sources are available, because they can indeed be used to compute similar infor-
mation; however, these sources should instead be considered as complementary
[20]. For instance, during several F-35 in-bay store separation events, the
6DOF/IMU TM Kit data stream had excessive or unresolved data noise while
the store was still in contact with the S&RE, which resulted in a “TM Kit-solution”
drift in store position and velocity, resulting in inaccurate late-trajectory solutions;
however, when the photogrammetry solution was substituted for the TM Kit sol-
ution in the early separation phase, the late-trajectory solution accuracy was dras-
tically increased. Furthermore, during approximately 10% of the missions during
F-35 SDD store separation flight testing, one data source failed or became unusa-
ble, but the availability of the other data source resulted in data sufficient for postt-
est analysis, without the need to refly the test point. Many times, the more accurate
photogrammetry position “truth” was needed in the earlier part of the separation
trajectory where store miss distance to aircraft hardware was critical, whereas the
more accurate (and higher sample-rate) 6DOF TM Kit store velocity and
606 C. F. HETREED ET AL.

acceleration was needed after the store became free from the S&RE, so that the
critical store aerodynamics computations could be performed.

C. F-35: USE THE SAME INSTRUMENTATION FOR AERODYNAMICS MODEL VALIDATION

Prior to the F-35 program, typical store aerodynamics models were adjusted
during a trial and error process where postflight simulations were iteratively per-
formed with simple modifications, such as scaling or shifting/biasing, to aerody-
namic coefficients in one or more degrees of freedom. The criteria for successful
aerodynamics model validation was subjective, and usually based on comparing
successive postflight simulated trajectories or rates with those computed from
the available instrumentation, such as on-aircraft high-speed video or store
6DOF/IMU TM Kits. Favorable comparisons were not always achieved in all
six degrees of freedom.
During the F-35 store separation program, Lockheed Martin implemented a
new strategy of also including a focus on where and how it happened—comparing
simulated and actual store applied aerodynamics, in which the quality of the simu-
lation aerodynamics model is directly judged in all six degrees of freedom, as a
more-detailed function of distance from carriage. Using the same type of instru-
mentation that was used during prior store separation test programs, Lockheed
Martin has created a rapid and repeatable process for computing accurate store
6DOF aerodynamic forces, and adjusting the aerodynamics model, such that
the validated postflight simulations reproduce the exact flight test store trajectory
in all six degrees of freedom. So, iterative trial and error adjustments and subjec-
tive interpretation have been replaced with a standard repeatable process.
However, successful store aerodynamics model validation requires extra attention
to the quality and detail of the flight test instrumentation, which is the subject of
the following section.

VII. F-35 STORE SEPARATION FLIGHT TESTING: VALIDATING STORE

AERODYNAMICS MODELS
Recall that the basis for both Lockheed Martin’s certification recommendation
and verification that the weapon release requirements have been satisfied is
flight test – validated M&S—modeling and simulation that has been adjusted
based on data gathered during the store separation flight test events, so that simu-
lated store separation events replicate the actual store separation events. It is gen-
erally reasoned that if the store trajectory and rotational rates are similar between
simulation and flight test, then the total forces and moments acting on the store
during the simulation are representative of those acting on the store during the
actual event, and the simulation is, therefore, validated. Ideally, the store separ-
ation preflight predictive simulation should match that of the flight test event.
In reality, a nominal prediction doesn’t always match its corresponding actual
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 607

flight test trajectory. When such discoveries occur during flight testing, it is impor-
tant to understand and explain such differences as quickly as possible, in order to
adjust the model and prediction simulation for the next store separation flight test
event, without adversely affecting the flight testing schedule.
In both the actual and simulated store separation events, the following occur:
1. Six degrees of freedom total forces and moments are applied to the specific
store, which results in
2. Store translational and rotational acceleration, which then results in
3. Changes in store translational and rotational velocity, which then results in
4. Changes in store translational and rotational displacements.
The actual aircraft maneuver response, which may be different than that during a
nominal simulation, must also be considered. So, the simulated store separation
event is only identical to the real event when all of the following are true:
. Simulated and real store mass properties are identical.
. Simulated and real aircraft response are identical.
. Simulated and real store 6DOF applied forces and moments are identical.
Measurements (or calculations from measurements) of these real items are the
basis for postflight model adjustments. The simulation and model, which includes
the aerodynamics model, become validated (Fig. 33) only after verification that
simulated and real store 6DOF response (displacement trajectory, velocity, and
acceleration) are identical. Measurement sources were identified as follows:

Fig. 33 Flight test and aerodynamics model validation.

608 C. F. HETREED ET AL.

. Store mass properties: Preflight measurements

. Aircraft response: On-aircraft instrumentation
. Store displacement, velocity, and acceleration: 6DOF/IMU TM Kit and
on-aircraft camera photogrammetry
. Store aerodynamics: 6DOF/IMU TM Kit
The requirement of performing mass properties measurements for each store
test asset prior to flight testing ensured the availability of store mass properties.
Availability of aircraft maneuver response measurements was reasonably assured,
because this information was telemetered to the control room and also recorded
by the on-aircraft data acquisition system; however, to ensure high-confidence
photogrammetric and 6DOF/IMU TM Kit postflight analyses, which were cri-
tically important to successful aerodynamics model validation, Lockheed
Martin store separation engineers coordinated with data processing, instrumen-
tation, survey/metrology, mass properties, flight test, and weapons experts on
the F-35 test team to plan and enact standard processes in preflight measure-
ment, in-flight data acquisition, and postflight computation and comparison/
validation.
With M&S validation as the primary reason for store separation flight
testing, and because of the importance of the quality of these measurements,
an understanding of planning and processes are relevant and are described
herein.

A. PLANNING PLAYBACK OF THE REAL FLIGHT TEST EVENT

In consideration of the moving parts and associated instrumentation during store
separation flight test events, reviewing the store separation event would involve
the playback of three items: the maneuvering aircraft, the store trajectory, and
the 6DOF/IMU TM Kit response. All three would be treated as unique rigid
bodies in a single-simulation playback of the event within the Lockheed Martin
store separation modeling and simulation environment, ASEP. Using ASEP’s
simple dynamicsm model, described earlier, these three moving rigid bodies
would be forced or constrained to move:
. The aircraft body would be “flown” according to on-aircraft response
measurements, using customized artificial applied forces that yield the
same aircraft rates and accelerations from the on-aircraft instrumentation
measurements.
. The store would be constrained to move according to the trajectory displace-
ments solution (provided by photogrammetry engineers).
. The 6DOF/IMU TM Kit body would be “flown” according to the 6DOF/IMU
TM Kit response measurements, using customized artificial applied forces that
yield the same 6DOF/IMU TM Kit rates and accelerations.
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 609

The common ingredient in the movement of each of these bodies within the play-
back simulation is the set of simplified rigid body equations of motion, Eqs. (1 – 6):

SFx ¼ max (1)

SFy ¼ may (2)
SFz ¼ maz (3)
SMx ¼ Ixx v̇x  (Iyy  Izz )vy vz (4)
SMy ¼ Iyy v̇y  (Izz  Ixx )vz vx (5)
SMz ¼ Izz v̇z  (Ixx  Iyy )vx vy (6)
. This form of the common F ¼ ma equations of motion was used, simply as a
result of the available measurements.
. Because products of inertia were unavailable, the rigid store body axes were
conveniently assumed to also be the principal axes.
. All three parts (aircraft, store, and 6DOF/IMU TM Kit) were treated as rigid,
without rotors, liquid sloshing, and so forth.
The ASEP-specific application of these simplified equations of motion to each of
the three bodies is described in relevant sections that follow.

B. STORE MASS PROPERTIES: PREFLIGHT MEASUREMENTS

As observed in Eqs. (1 –6), the store mass and inertia must be known, in order to
eventually compute the total forces and moments. During 6DOF/IMU TM Kit
posttest processing, the relative positions of the actual store’s center of gravity
and the actual 6DOF/IMU TM Kit must be known. Furthermore, store features,
such as photogrammetry reference targets, fin and strake corners, S&RE attach-
ment locations (such as bomb lugs or missile hangers), nose, tail, and other body
references must be known well before the store disappears into the flight test
range’s desert floor or ocean bottom at the end of the store separation test event.
Nominal mass properties are usually provided by the store supplier, but the
nominal store center of gravity sometimes differs from tested stores, which typi-
cally results in simulation vs test EOS pitch rate performance differences. So, when
performing validation exercises, or comparisons between test results and pretest/
posttest simulations, it was necessary to be confident that the as-tested store mass
properties were also the ones used in simulations. In order to provide measure-
ment confidence and to quantify measurement uncertainty, several types of
stores were mass-propertied, repeatedly, by different personnel at the same and
different test sites. Part of the store separation preflight process was to ensure
that all stores used during store separation on-aircraft pit ejection testing and
flight testing were mass-propertied by the ordnance team at PAX, EDW, or
610 C. F. HETREED ET AL.

Fig. 34 GBU-12 mass properties and model.

AFSEO, where the store mass, center of gravity, and mass moments of inertia were
quantified from a combination of measurements and calculations. The store was
required to be in its as-tested configuration, complete with fins, strakes, and even
installed instrumentation, such as the 6DOF/IMU TM Kit. As a result, every
unique store separation test asset was represented by its equivalent ASEP model
(Fig. 34).

C. AIRCRAFT RESPONSE: ON-AIRCRAFT INSTRUMENTATION

During store separation events with an air-to-ground store, the S&RE ejector for-
cibly pushes the store away from the aircraft; however, because the S&RE, store,
and aircraft are all part of the dynamic system, it was reasonable to expect that
the aircraft would also have a reaction to this S&RE ejector impulse. Furthermore,
with the reality of occasional atmospheric turbulence, and a potential for aircraft
transient reaction to opening weapon bay doors, there was sufficient reason to
believe F-35 store separation flight test aircraft would inevitably experience tran-
sient responses during store separation events.
Considering reference frames, the 6DOF/IMU TM Kit provides measure-
ments of store rates and accelerations relative to an inertial (Earth/ground) refer-
ence frame, independent from the separately moving aircraft; however,
photogrammetric store trajectory solutions are relative to the moving aircraft,
because the store separation cameras are attached to the aircraft. So, 1) transform-
ing 6DOF/IMU TM Kit solutions into the moving aircraft’s reference frame, and
2) transforming photogrammetric store trajectory solutions into an inertial refer-
ence frame both require consideration of the aircraft motion.
Furthermore, basic aircraft flight condition information is necessary to trans-
form the computed aerodynamic forces and moments into nondimensional aero-
dynamic coefficients familiar to aerodynamicists, because the main goal in
aerodynamics model validation is to compare these flight test – derived store aero-
dynamic coefficients to those used in the simulation.
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 611

So, as a standard process for every store separation flight test event, when post-
processing store 6DOF/IMU TM Kit data for the purposes of computing a store’s
trajectory relative to the aircraft (and ground), Lockheed Martin has also included
the aircraft response. The final motivating reason for this approach is simply to
include all relevant data and avoid shortcuts based on assumptions, especially if
these assumptions have the potential for adding confusion when troubleshooting
potentially problematic postflight solutions, which could cause unnecessary delays
in continuing to the next test point.
During store separation event playback simulations within ASEP, the aircraft
rigid body was “flown” by applying customized artificial forces at the aircraft
body’s center of gravity, according to Eqs. (1 –6), with the following considerations:
. The applied forces onto the aircraft body are dependent on measured aircraft
response, where ax, ay, and az are the measured/transferred translational air-
craft accelerations, and vx, vy, and vz are the measured aircraft rotational rates
(Fig. 35). Simple differentiation of these aircraft rotational rates yields aircraft
rotational accelerations: v̇x , v̇y , and v̇z .
. These applied forces are also dependent on the aircraft model’s mass m and
moments of inertia Ixx, Iyy, and Izz; however, in order to ensure that the
ASEP playback aircraft model response is not influenced by other forces
(which are beyond the scope of this paper), the aircraft body’s mass and
moments of inertia were artificially increased by several orders of magnitude,
in both the equations of motion and the aircraft rigid body ASEP model to
which the aircraft “flying” forces are applied.
. The initial condition is coincident with store first motion, or the moment the
store separation event begins. When examining aircraft measurement par-
ameters, this moment is identified by certain behavior in the aircraft response,
such as a sudden change in the aircraft roll rate and vertical acceleration, in

Fig. 35 Aircraft rate and acceleration response.

612 C. F. HETREED ET AL.

response to strong store ejection forces, for instance. The red vertical line in
Fig. 35 marks this point in a typical flight test aircraft response during an
air-to-ground store separation event.
. The aircraft body’s initial playback flying conditions are set by other aircraft
measurements, such as Mach, angle of attack, angle of sideslip, altitude,
bank angle, pitch angle, yaw rate, pitch rate, and roll rate.
The PAX and EDW aircraft maneuver response data were usually available to
the store separation engineers in Fort Worth within four hours after the flight
test event.

D. STORE DISPLACEMENT: ON-AIRCRAFT VIDEO PHOTOGRAMMETRY

When validating the store aerodynamics model database, the flight test – derived
aerodynamic forces are ultimately compared to the store aerodynamics model.
The main input into the store’s aerodynamics model is the store position and
orientation. Furthermore, the photogrammetry-based trajectory solution is the
best store position “truth” with which to sanity-check 6DOF/IMU TM Kit –
based trajectory solutions. In addition, the primary consideration in safe store
separation is store clearance to aircraft hardware and adjacent stores. For these
aforementioned reasons, it is critically important to confidently know the real
store trajectory positions and orientations.
Photogrammetric analysis includes using video imagery of the store and air-
craft, and is dependent upon reference locations, such as high-contrast photoG
targets and store geometric features; the key requirement for such video
imagery is that these store and aircraft reference locations are visible and identifi-
able. So, prior to F-35 flight testing, several tasks were undertaken to ensure that
reliable and accurate photogrammetric analyses could be performed for every
SDD store, at aircraft station, on each of the three F-35 variants.

1. CAMERAS: LOCATION, ORIENTATION, AND FIELD OF VIEW

Well before flight testing, the store separation test aircraft were not available, so
Lockheed Martin’s Flight Test Instrumentation and Store Separation groups
worked together to solidify high-speed digital Airborne Separation Video
System (ASVS) camera locations and orientations on the three F-35 variants.
This process involved assessing simulated CAD image views, with identical
fields of view as the real ASVS cameras (Fig. 36).
For determining store position relative to the aircraft, Lockheed Martin’s
photogrammetric software, Digital Store Separation Analysis System (DSSAS)
relies on visibility of at least three or four store location references from each
camera, and store position computation accuracy is increased with additional
location references and multiple camera views. Even though DSSAS is capable
of producing reasonably accurate photoG solutions from as few as one camera’s
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 613

Fig. 36 Virtual camera views and flight test camera layout.

imagery, there is a need for achieving more than reasonably accurate photoG sol-
utions, due to expected low store clearance distance from in-bay hardware.
Although multiple cameras are used to capture imagery during a store separation
flight test event, it is sometimes not known whether one or more cameras will mal-
function or provide store imagery that is washed out due to bright sunlight, dar-
kened due to poor ambient lighting, or obscured by water vapor condensation
(Fig. 37). Although there was consideration of attempting prediction of likelihood
of flowfield-induced water vapor condensation, based on forecasted atmospheric
conditions [21], it was believed that flight test missions would not be cancelled due
to the possibility of store-obscuring water vapor condensation, if other cameras
could capture unobstructed store views.
Therefore, a collection of 15 ASVS camera locations were selected for F-35
store separation flight testing, in order to capture store imagery during separation
events from the left-side and right-side weapons bays and external wing stations.

2. STORE TARGET LAYOUT

Because it was not practical to select specific store test assets and assign them to
separation testing at specific aircraft weapon stations, each store would require
sufficient coverage of high-contrast photoG targets so that a sufficient number

Fig. 37 Water vapor condensation temporarily obscuring separating store.

614 C. F. HETREED ET AL.

Fig. 38 Virtual store target layout.

of these targets could be identified (seen) within any of the 15 ASVS camera fields
of view. It was also impractical to use the actual flight test aircraft to load weapons
in order to view ASVS imagery of all SDD store types at all weapons stations and
interactively assess photoG target schemes. So, with recognition of the balance
between 1) too many photoG targets, which would result in time-consuming
store metrology survey exercises for each store; and 2) too few photoG targets,
which could result in inaccurate photoG solutions, Lockheed Martin relied on
CAD and AEDC’s store trajectory visualization software, TVIS, to place virtual
photoG targets on each store (Fig. 38) and aircraft (Fig. 40), and to assess the
resulting target schemes as seen by each camera’s field of view.
For every store that was to be used in store separation flight testing, the PAX
and EDW metrology experts applied photoG target stickers onto the store, per the
virtual photoG target scheme. Finally, metrology surveys (Fig. 39) were performed
twice, for confidence, to measure 1) all of these target locations, and 2) key features
of the store, such as lug locations, nose, tail, fin/strake root and corner points,
and even installed 6DOF/IMU TM Kit location and orientation. Because the
6DOF/IMU TM Kit’s measured rate and acceleration data would be used in post-
test computations at other store reference locations, it was necessary to confi-
dently know the position and orientation of the 6DOF/IMU TM Kit (and its
sensors) in the store reference frame.

Fig. 39 Store metrology survey.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 615

Fig. 40 Virtual and real aircraft photogrammetry targets.

3. AIRCRAFT TARGET LAYOUT

Aircraft-mounted ASVS cameras can move or vibrate in flight, which means that
the ASVS camera focal point’s location and field-of-view orientation can move
from its nominal installed location, relative to the aircraft reference frame;
however, with visibility of known references, it is possible to know where the
camera moved. DSSAS relies on seeing known and stationary location targets
on the aircraft, within the weapon bay and on the exterior surface. This enables
the instantaneous computation, within DSSAS, of the instantaneous location
and orientation of each ASVS camera, so that the final DSSAS photoG solution
for the store position is relative to the fixed aircraft reference system, instead of
the moving camera system’s reference frame.
Using the virtual aircraft photoG target scheme as a guide, PAX and EDW
metrology experts applied high-durability painted photoG targets to the store sep-
aration aircraft external surfaces. During this time, photoG target stickers were
also applied within the left and right weapon bays, interactively with the real
ASVS camera views. Finally, metrology surveys were performed twice (for confi-
dence), to measure all of these target locations, in what would become the “known
locations” of all the aircraft photoG targets in the aircraft reference frame.
616 C. F. HETREED ET AL.

4. ONE-TIME ACCURACY ASSESSMENT

The photogrammetry-based store trajectory solution has been regarded as the store
position “truth,” but this confidence was only realized after a measurement exercise
at Northrup Grumman, well before flight testing commenced. Actual ASVS
cameras were set up within an F-35 mockup, with similar fields of view as those
that were being considered for F-35 store separation flight testing. PhotoG
targets were attached to the aircraft mockup, an air-to-ground store mockup,
and an air-to-air store mockup (Fig. 41); these photoG target locations were sur-
veyed, just as they would be on real flight test aircraft and stores, to establish the
known target locations in the aircraft’s and stores’ coordinate systems, respectively.
The following steps were then accomplished:
1. ASVS image capture: The store was placed at the first location, and the image
was captured on each ASVS camera, simultaneously. This image (along with
separate camera and lens calibration images) was to be used in the normal
DSSAS photogrammetry position solution.
2. Real store position: With the metrology laser tracker system established within
the coordinate system of the aircraft mockup, several store photoG targets
were surveyed, and the metrology software subsequently revealed the store
mockup’s true location and orientation in the aircraft axis system. This real
store position was locked up and hidden from the photogrammetry engineers
until after the DSSAS store position solutions, from the ASVS camera images,
were eventually computed.
3. Repeat: These two previous steps were repeated for five more real store pos-
itions below carriage, and then repeated again with the other store mockup.
4. DSSAS photogrammetry store position: Using the ASVS imagery, along with
the known store and aircraft photoG target positions in their respective refer-
ence systems (as would identically be done during real flight test photogram-
metry), the DSSAS displacement solution was computed for each of the six
store mockup positions below carriage, for both stores.
The resultant DSSAS photogrammetry store positions were finally compared
to the previously “locked-up and hidden” real store positions, revealing photo-
grammetry-based store displacement solutions 1) within 0.2 in. of the real store

Fig. 41 DSSAS photogrammetry accuracy assessment.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 617

three-dimensional translation position, 2) within 0.2 deg of the real store’s yaw
and pitch rotation angle, and 3) within 1 deg of the real store roll position.
These differences were within the resolution needed for confidently determining
store clearance (or miss distance) to aircraft hardware during real store separation
flight test events. Because high accuracy was demonstrated for these solutions, the
DSSAS photogrammetry solutions have routinely been considered the store dis-
placement and rotation position truth.

5. FLIGHT TEST TRAJECTORY SOLUTIONS

Following all store separation events where there was little clearance margin to
aircraft hardware, or where the photogrammetry solutions were needed to
confirm, enhance, or substitute for 6DOF/IMU TM Kit solutions, DSSAS photo-
grammetry store position solutions were computed using the ASVS video imagery
and identifiable store photoG targets. This digitization can be seen from the green
points in Fig. 42. These store position solutions were ultimately provided to the
store separation engineers within one to two days of the flight test event.
Occasionally, wireframe overlay representations of the final trajectory were
produced (Fig. 43), but in all instances, the solution was in the form of tabular
store translational and rotational displacements, expressed in the aircraft refer-
ence frame. These tabular photogrammetry solutions were used by store separ-
ation engineers during postflight simulations in ASEP, as the time-dependent
store body motion constraint to the aircraft, resulting in the store trajectory play-
back in ASEP.

E. STORE DISPLACEMENT, VELOCITY, AND ACCELERATION: 6DOF/IMU TM KIT

As observed in Eqs. (1 –6), the rates and accelerations must be known in order
to compute the total forces and moments. The store rotational rate and transla-
tional acceleration “truth” is best-obtained from direct measurement, using an
IMU and/or 6DOF TM Kit. Fundamental to the computation of the total

Fig. 42 PhotoG target digitization during DSSAS photogrammetry analysis.

618 C. F. HETREED ET AL.

Fig. 43 DSSAS photogrammetry solution, wireframe overlay on ASVS video.

applied moment are the rotational accelerations, v̇x , v̇y , and v̇z , which are
obtained simply by numerical differentiation of the measured rotational rates,
vx, vy, and vz. Although it is possible to alternatively obtain store rate and accel-
eration information from photogrammetry solutions, the comparatively low
sample size and numerical differentiation of positions to obtain store velocities
and accelerations require extra time and skill. Store aerodynamics are more
quickly and reliably obtained using directly measured store rate and acceleration
“truth.” Nevertheless, overall confidence is further increased when accurate
photogrammetry and 6DOF/IMU TM Kit solutions are available, with one
source’s solution confirming or even improving the accuracy of the other.

1. STORE IMU AND 6DOF TM KIT

Certified weapons in the F-35 SDD phase are all actively controlled, guided stores.
Such guided stores usually have an IMU to measure store rotational rates and
translational accelerations, which are passed to the store’s guidance control
system. The store separation flight test versions of these guided stores also trans-
mit the IMU’s measured data to nearby antennas.
There are also unguided versions of these stores, which are used for various
captive-carriage tests and for store separation jettison testing. A majority of
F-35 SDD store separation testing involved these unguided jettison stores 1)
because of their lower cost, and 2) because at similar in-flight release conditions,
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 619

jettisoned stores’ trajectories are generally not as safe and well-behaved as guided
stores’ trajectories. So, when unguided stores are used during separation testing, a
6DOF TM Kit (Fig. 44) is usually installed within these stores, enabling the
measurement and telemetered transmittal of similar rotational rate and trans-
lation acceleration parameters as a guided store’s IMU.
When these telemetered data are used only to compare rotational rates and
translational acceleration between simulation and flight test, it is sometimes poss-
ible to use telemetered data that have localized loss, or dropouts, provided that the
available rate time history at least reveals peak values, ejector end-of-stroke rates,
or rail-launched end-of-rail tip-off rates, depending on the region of interest in
the trajectory.
However, when these telemetered data are used primarily for computing
the time-dependent store position and orientation trajectory and especially
the store aerodynamic forces, data loss (dropout) becomes much less tolerable.
Dropouts and inadequately filtered rotational rates will result in inaccurate
flight test – derived aerodynamics forces and moments, and therefore, an incom-
plete or inaccurate aerodynamics model validation.

2. PREPARING FOR THE INEVITABILITY OF DATA LOSS

For the F-35 SDD store separation flight test program, several steps were taken to
minimize the chances for untimely data loss.

Fig. 44 Store 6DOF/IMU TM Kit acceleration and rate data.

620 C. F. HETREED ET AL.

For unguided/jettison stores, most of the installed 6DOF/IMU TM Kits

included a TM-delay feature, enabling a delayed retransmittal of the sensed rates
and accelerations. A delay duration of approximately 0.5 seconds was chosen, to
ensure that the retransmittal occurred when the store was sufficiently far from the
aircraft, without obstruction to the ground or support aircraft antennas.
For in-bay store separation events, when the store hasn’t yet fallen completely
out of the bay and past the weapon bay doors, ground-based antennas may not be
able to receive the 6DOF/IMU TM Kit telemetered signal. Therefore, for F-35
in-bay store separation tests, Lockheed Martin used in-bay telemetry reception/
retransmittal equipment, or Re-Rad Kit, to independently receive the close proxi-
mity store 6DOF/IMU TM Kit’s telemetered signal and retransmit it without bay
and door obstruction.
With both the F-35 Re-Rad Kit hardware and 6DOF/IMU TM Kit TM-delay
feature, up to four separate signal transmission sources would be available for
postflight use. In many cases, the best of the four sources was selected for postpro-
cessing, without need of enhancement; however, in some cases where the best
source still had dropouts in critical areas of the trajectory, one or more of the
remaining sources were used to fill in the dropouts with better data.
For external underwing store separation events involving a guided weapon,
only one data stream from the store’s IMU was telemetered from the store,
without the possibility of using the F-35 in-bay Re-Rad Kit. For such events, in
order to ensure successful IMU data acquisition, multiple ground-based reception
antennas were used, along with occasional use of support-aircraft data acquisition
antennas.

3. REPLAYING THE 6DOF/IMU TRAJECTORY

The PAX and EDW 6DOF TM Kit response data were usually available to the
store separation engineers in Fort Worth within four hours after the flight test
event. During store separation event playback simulations within ASEP, the
6DOF/IMU TM Kit rigid body was “flown,” separately and independently
from the store model, by applying customized artificial forces at the TM Kit’s
center of gravity, according to Eqs. (1 –6), with the following considerations.
Applied forces onto the TM Kit part are dependent upon TM Kit measure-
ment, where ax, ay, and az are the measured translational accelerations, and vx,
vy, and vz are the measured rotational rates. Simple differentiation of these
6DOF TM Kit rotational rates yields TM Kit rotational accelerations: v̇x , v̇y ,
and v̇z .
In ASEP, because the TM Kit part is a moving rigid body, this part’s response
to the applied forces is also dependent upon the TM Kit part’s model’s mass m and
moments of inertia Ixx, Iyy, and Izz. The TM Kit part is simply a convenient rigid
body to move, from which model-based playback measurements of displacement,
velocity, and acceleration measurements are requested. Conveniently, massless
store geometry is attached to this rigid body for playback visualization. So, the
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 621

Fig. 45 Store 6DOF/IMU TM Kit and DSSAS photogrammetry trajectory, visualized.

TM Kit part’s mass and moments of inertia were artificially set to a conveniently
low value of one unit, in both the equations of motion and in the ASEP TM Kit
part rigid body model to which the TM Kit part “flying” forces are applied.
The ASEP TM Kit part’s initial condition is coincident with store at the
CAD carriage position, or the initial condition could be at any point in a
photogrammetry-produced trajectory, for instance. Because most store separation
test events yielded high-quality 6DOF/IMU TM Kit data, almost all trajectory
computations were begun at first motion, corresponding to the moment immedi-
ately before the rapid increase in vertical acceleration (Fig. 44), and also included
the timespan when the actual store was still in contact with the ejector.
As often as was necessary or practical, the photogrammetry-based and 6DOF/
IMU TM Kit – based trajectory solutions were both plotted or visualized (Fig. 45),
to ensure favorable comparisons.

F. STORE AERODYNAMICS
Lockheed Martin’s additional focus on actual store aerodynamics involves no
additional measurement sources, relying on the measured 6DOF/IMU TM Kit
rates and accelerations, measured store mass properties, measured aircraft flight
conditions, and a standardized posttest process, usually within one day. The
store total applied forces and moments in all six degrees of freedom are computed
from the simplified rigid body equations of motion, Eqs. (1 – 6). This form of the
common F ¼ ma equations of motion was used, simply as a result of the available
measurements.
From the existing store mass property information, m is the measured actual
store mass, and Ixx, Iyy, and Izz are the measured store 3D rigid body axis moments
622 C. F. HETREED ET AL.

of inertia. Because actual store mass properties measurements did not include
products of inertia, the rigid store body axes were conveniently assumed to also
be the principal axes.
After the in-flight 6DOF/IMU TM Kit measurements are transferred to the
store’s center of gravity, ax, ay, and az are the measured/transferred translational
store accelerations, and vx, vy, and vz are the measured store rotational rates.
Simple differentiation of these store rotational rates yields store rotational
accelerations: v̇x , v̇y , and v̇z .
Other simplifying assumptions were made, for convenience. The store was pre-
sumed to be rigid, without rotors, liquid sloshing, and so forth. The 6DOF/IMU
TM Kits were never located near stores’ centers of gravity, and measurements fre-
quently revealed flexible store response, for instance, which was filtered in order to
enable rigid body – like store aerodynamics computation. “Sanity checks” were fre-
quently performed during postprocessing: 6DOF/IMU TM Kit– based trajectory
solutions were compared with photogrammetry-based trajectory solutions. For
the air-to-air and air-to-ground stores in the F-35 SDD store separation test
program, these frequently matching solutions provided confidence that the simpli-
fied F ¼ ma equations were sufficient for replicating photogrammetry-based actual
store trajectories and for extracting valid aerodynamics.
Although the idea is simple, and is certainly not new, several considerations
must be addressed in order to successfully implement this simple idea to under-
stand where and how it happened—to achieve confident aerodynamics model
validation. Although it might be convenient to further simplify Eqs. (4 –6) as
was described in Ref. [22], such a simplification did not yield flight test –computed
store aerodynamics forces and moments sufficient for posttest simulation replica-
tion of the flight test trajectories in all six degrees of freedom.
During the store separation phase between carriage and EOR or EOS, the total
forces and moments applied to the store include the interaction between the store
and S&RE. Distinguishing between store aerodynamics and S&RE-applied forces
relies on knowledge of the S&RE performance. Although this behavior can gener-
ally be characterized from other testing and roughly subtracted from the total
forces and moments, the Lockheed Martin approach, when strictly performing
aerodynamics (and propulsion) model validation, is to remove the S&RE
portion from consideration and focus only on the phase beyond EOR/EOS
(Fig. 46), where the store is no longer influenced by the mechanical and structural
interaction with the aircraft and S&RE.
Additional forces may act on the store beyond EOR/EOS, but it should be
noted that the stores involved in F-35 SDD store separation flight testing were suf-
ficiently large, such that the forces associated with sudden mechanical deployment
or rapid actuation of moving control surfaces, for instance, were extremely small,
compared to applied aerodynamic and propulsion forces.
Considering an example of computing flight test – based aerodynamic pitching
moment, and recalling Eq. (5), the applied pitching moment forces depend mainly
on the store pitch acceleration v̇y . However, measured 6DOF TM Kit pitch rate
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 623

Fig. 46 Post-EOS store pitching moment flight test aero increments, derived from filtered
6DOF TM Kit pitch rates.

vy is not usually smooth (Fig. 46, green curve in top plot), and differentiation
typically results in noisy pitch acceleration v̇y producing an equally noisy pitch
moment computation, observed in the green curve in the lower plot of Fig. 46.
However, filtering the 6DOF TM Kit’s pitch rate helps produce smoother pitch
accelerations, yielding smoother 6DOF TM Kit –based aero pitch moment com-
putations, observed in the successively filtered black, magenta, and blue curves
in the lower plot of Fig. 46.
The 6DOF TM Kit –based aero pitch moment is the flight test “truth” against
which the original store aerodynamics model will be compared and adjusted (vali-
dated). For this exact flight test trajectory, the photogrammetry-produced or
6DOF TM Kit– produced store trajectories are useful for obtaining the aerody-
namics database model’s table look-up pitching moment, observed in the red
curve in the lower plot of Fig. 46. For this flight test example, the difference
between the 6DOF TM Kit – derived aerodynamics pitching moment and the
aerodynamics model’s pitching moment is the aero flight test increment.
For several disciplines such as Loads, Dynamics, Flutter, Aero/Performance,
and Stability and Control, modeling and simulation adjustments were expected
624 C. F. HETREED ET AL.

Fig. 47 Flight test – derived and aero database model pitch moment coefficient (CLM):
a) two separation events at different Mach numbers, and b) three separation events at same
Mach numbers.

as part of the flight test – based validation process, especially because wind tunnel
models frequently exhibit transonic aerodynamics differences from full-scale air-
craft in flight. The same is true for the Store Separation discipline. One of the goals
during store aerodynamics model validation was to strive for adjustments as a
function of Mach and/or aircraft angle of attack, for instance, instead of a
flight test event-specific adjustment. Although various store separation events,
even at nearly identical flight conditions, can sometimes produce slightly different
store trajectories, it was reassuring to group similar events (when available) and
observe similar store aerodynamics. In Fig. 47, the left plot’s blue and red
curves reflect the store aerodynamics pitch moment for test events at two different
Mach numbers, whereas all of the right plot’s curves reflect the store pitch
moment during three different separation events at the same Mach number.
The Flight Test aero increment (FTinc), or the difference between the flight
test –derived store aerodynamics and store aerodynamics models, is a function
of store distance below the aircraft. Such increments are obtained for the forces
and moments in all six degrees of freedom, producing the complete 6DOF aero
adjustment, contributing to a flight test – validated aerodynamics model for the
specific flight test trajectory. Finally, grouping 6DOF flight test aero increments
from different flight test events produces a more general set of flight test aero
increments, perhaps as a function or aircraft Mach or angle of attack, for instance.
The lower middle plot in Fig. 48 represents the store aerodynamics pitch moment
FTinc for the same events in Fig. 47. The three red-tinted FTinc curves in Fig. 48,
associated with events with the same Mach numbers, reflect nearly identical aero
model adjustments and are conveniently averaged (white dotted central curves in
Fig. 48) to represent the nominal FTinc for that particular Mach number. The blue
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 625

FTinc curve, associated with a store separation event with a different Mach
number, clearly reflects a different (and nearly zero) adjustment to the store
aerodynamics model.
Explanation of reasons for differences between scale model wind tunnel–
based aerodynamics and actual in-flight aerodynamics are beyond the scope of
this paper, but it is clearly observed in Fig. 48 that such aerodynamics differences
can be dependent upon store distance and aircraft (and wind tunnel) Mach, at
the very least. So, separate studies (if deemed necessary) to investigate reasons
for such differences can be focused directly on store aerodynamics, instead of
indirectly on the trajectories and rates produced by the store aerodynamics.
But more importantly, it becomes possible to focus on aerodynamics model
validation, separately from the mechanical, dynamic system model. Furthermore,
analysis of such FTinc summaries enables a flight test – based estimation of aero-
dynamics coefficient uncertainties, represented by the white dashed curves/
boundaries on either side from the white dotted central nominal FTinc curves.
Such uncertainties are later used in Monte Carlo uncertainty simulations
during analyses supporting certification recommendation.
The final verification, that both the aerodynamics and system models are
validated, is conducted following posttest simulations, comparing the photogram-
metry-based and 6DOF TM Kit– based playbacks (solutions) trajectories of
the flight test event, the preflight (prior to validation) simulation trajectory, and
the postflight validated simulation trajectory. Well-validated aerodynamics
models produce favorable trajectory comparisons in all six degrees of freedom,
as shown in a sample air-to-ground store separation flight test event in Fig. 49,
where the validated M&S trajectory (gold curve) identically matches the trajectory

Fig. 48 Flight test– derived aerodynamics increments in five degrees of freedom (axial not
shown).
626 C. F. HETREED ET AL.

Fig. 49 Store trajectories compared: DSSAS photogrammetry – based (green), 6DOF/IMU TM

Kit– based (red), preflight prediction (blue), postflight validated-aero simulation (gold).

playback from the DSSAS photogrammetry (green curve) and 6DOF TM Kit
solution (red curve).

G. TIME CONSIDERATIONS
When planning for the rapid posttest analysis required for “next-day” clearance/
approval to continue to the next test point, it was presumed that the information
from the following data sources was never going to be time-synchronized when
the Lockheed Martin store separation team received them:
1. 6DOF/IMU TM Kit data streams:
a. Real-time data stream, directly from store
b. Delayed data stream, directly from store
c. Real-time data stream, retransmitted by test or supporting aircraft
d. Delayed data stream, retransmitted by test or supporting aircraft
2. Store separation video and associated photogrammetry trajectory solution
3. Aircraft maneuver response data
Furthermore, posttest data streams that were telemetered to the control room
were not time-synchronized with the same data streams that were recorded on
the test aircraft. Yet all of these data were to be merged and used together.
Because of aircraft system delays associated with electronic triggers, such as the
cockpit pickle switch, or insufficient sample-rate resolution in S&RE-associated
electronic messages, such as “hooks open,” the search for reliable store
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 627

first-motion evidence was biased toward evidence of physical response. The F-35
store separation posttest analysis process included an exercise of identifying the
time at store first motion, independently in each data source, and then synchro-
nizing the independent data sources with this common start time:

1. 6DOF/IMU TM Kit data: Vivid store vertical or longitudinal acceleration

reflected the store response to the S&RE impulse or store rocket motor
ignition.
2. Store separation video: Video includes time-stamped video frames, but
sometimes the identification of store first motion was subjective, with not-so-
obvious small first movements; however, during review of processed photo-
grammetry trajectory solutions, the start of store translational displacement
changes was easily identified.
3. Aircraft maneuver response data: The aircraft response to the S&RE store ejec-
tions was usually observed in sudden aircraft vertical acceleration and roll rate
response. Selecting the start time within aircraft maneuver response data from
rail-launched missiles events was slightly more difficult, requiring a less-
accurate time selection from low-sample-rate Stores Management System
(SMS) and flight controls information, such as a drop in total aircraft
weight due to “store release” electronic messages, followed by a more-refined
examination of aircraft maneuver response, such as yaw rate and longitudinal
acceleration.

VIII. F-35 STORE SEPARATION FLIGHT TEST VIGNETTES

Even with commonality among aircraft variants and similarity amongst stor,
there still exist differences unique to aircraft variants as well as the various
stores, resulting in a significant amount of unique cases to be considered for
store separation modeling, analysis, ground testing, and flight testing. Store sep-
aration analysis and flight testing has been conducted over the last several years
by the industry partner/government integrated test forces at EDW and PAX,
where 183 stores were released. This number of test points would have been
higher if aircraft variant commonality wasn’t proven in wind tunnel testing and
CFD analyses and leveraged during flight test planning. For example, simulation
model validation for GBU-31 JDAM Mk-84 weapon releases from the internal
bays of both the F-35A and F-35C variants only required seven store separation
test events.
Because F-35 store separation flight testing was conducted from aircraft that
also performed testing required by other disciplines, such as loads, flutter,
environment, and flying qualities, the store separation test flights would fre-
quently be scheduled to occur in small blocks of time requiring expeditious clear-
ance to subsequent test points. However, there was also a safety constraint on
proceeding to the next test point: the flight clearance authority and flight
628 C. F. HETREED ET AL.

test teams at PAX and EDW required that the previous test point was proven to
be safe and that there was sufficient confidence that the next test point would also
be safe. Several store separation flight vignettes are presented, and include
instances of failures, discoveries, confirmations, and a process to increase test
efficiency.

A. AERODYNAMICS EFFECTS ON S&RE PERFORMANCE AND INITIAL CONDITIONS

Actual and simulated store trajectories are highly dependent on the end-of-stroke
conditions of the store as it leaves the ejector rack or launcher. Predicting
end-of-stroke conditions for any given flight condition is a complex problem
because the store end-of-stroke conditions are dependent on the flight condition,
especially for stores released from a pylon on the wing or from any other station
that is exposed to external flow when the store is at the carriage position.
End-of-stroke conditions from static ejection testing can provide some insight
into ejector rack performance variability as well as indicate expected end-of-stroke
conditions at low-dynamic-pressure flight conditions. However, the end-of-stroke
conditions observed in static ejection testing do not adequately predict the
end-of-stroke conditions of stores subject to aerodynamic influence during the
ejector stroke.
An enhanced simulation prediction can be accomplished by modeling the
forces the store experiences on the ejector as soon as the store begins moving
from the carriage position. Modeling the sway braces at the end of the
pistons ensures that the modeled store is constrained laterally in the same
manner as the actual store. Modeling the friction of the sway brace pads
ensures that the store can be held in place with the same frictional force as
an actual store on the ejector. Modeling the ejector piston forces using the EDO-
provided black box gas model ensures that the S&RE ejector model accounts for
the variable inertial force of the store under flight-condition-dependent aerody-
namic loading. Simulation trajectories conducted with the aforementioned mod-
eling considerations, together with aerodynamics from wind tunnel testing,
match flight test trajectories better than simulation trajectories initialized using
end-of-stroke conditions from static ejection testing and wind tunnel
aerodynamics.
As previously discussed, validation of the S&RE-only models and the complete
dynamic system models (including the aircraft, S&RE, and store) were accom-
plished following both store/S&RE ejection testing and on-aircraft ejection pit
testing. Also, from previous discussion, store aerodynamics models were
created primarily from grid aerodynamics data from store separation wind
tunnel testing. In such tests, store aerodynamic forces and moments were
obtained with the store positioned at many store locations and angles relative
to the aircraft. For stores that are ejected, one of these important store positions
is at or near carriage, because the store aerodynamics at this location are represen-
tative of those that the store will see during the mechanical ejection phase. So, per
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 629

the normal preflight prediction process, the validated dynamic system model was
combined with the store aerodynamics model to perform simulations of all antici-
pated flight test events. Certain simulation cases revealed different store/S&RE
performance than that observed during the ground-based ejection testing,
which was fully expected.

1. AIM-120 AMRAAM MISSILE

Numerous AIM-120 (Fig. 50) ejections from the LAU-147 during on-aircraft pit
ejection tests were performed, with EOS pitch rates nearly identical to those
observed during AIM-120/LAU-147 S&RE-only ejection stand tests. Lockheed
Martin’s fully dynamic (aircraft þ S&RE þ store) “no airload” AIM-120 store sep-
aration simulations yielded identical EOS pitch rates; however, flight test simu-
lations performed with the same dynamic system model yielded substantially
more EOS nose-down pitch rate, when including wind tunnel –based aerody-
namic applied store loads during the ejection phase.
Store separation flight testing included AIM-120 separation events from the
in-bay near-door position, where the AIM-120 was exposed to the air flow,
instead of being hidden deep within the weapons bay. During these test events,
EOS store pitch rate performance was proven to be nearly identical to that
from the preflight prediction simulations. The significance of this favorable com-
parison with flight test behavior was that the dynamic system model was validated,
not only in the no store aero state, but more importantly in the aero-loaded
in-flight state—store separation simulations using the fully dynamic aircraft þ
S&RE þ store þ aerodynamic model could now be performed more confidently
for many different aircraft flight conditions and maneuvers in the AIM-120
employment flight envelope.

Fig. 50 AIM-120 store separation: static ejection pit test and flight test.
630 C. F. HETREED ET AL.

2. EXTERNAL GBU-12
Store separation simulations were performed using ASEP’s fully-dynamic
(aircraft þ S&RE þ store) model for underwing GBU-12 (Fig. 51) ejections
from the in-pylon BRU-68 ejector rack, matching on-aircraft pit ejection test
results for the same configurations. Similar to the aforementioned AIM-120
case, GBU-12 store separation flight test simulations performed with the full
dynamic system model yielded different EOS nose-down pitch rate, when includ-
ing external GBU-12 wind tunnel –based applied store aerodynamic loads during
the ejection phase. Such “with airload” simulations matched flight test ejector per-
formance very well.
But pitch wasn’t the only relevant degree of freedom. Lockheed Martin’s fully
dynamic system model included a frictional contact model for BRU-68
swaybrace-to-store interaction. So, like the actual GBU-12 interaction with the
BRU-68 ejector rack, the simulated GBU-12 store was restrained (not con-
strained), such that the store could yaw relative to the BRU-68 and pylon
during the ejection phase. At each swaybrace point, the store surface was able
to frictionally slide fore – aft and sideways, while in contact with the four sway-
braces during ejection. Simulated external GBU-12 store separation EOS yaw
rates were consistent with those from actual GBU-12 external store separation
flight test events. The fully dynamic system model was validated in the aero-
loaded in-flight state, so like the AIM-120 case, GBU-12 store separation simu-
lations using the fully dynamic aircraft þ S&RE þ store þ aerodynamic model
could now be performed confidently for many different aircraft flight conditions
and maneuvers in the GBU-12 external release flight envelope.

B. MULTIPLE STORE SEPARATION EVENTS PER FLIGHT

Before F-35A GBU-39 small-diameter bomb store separation flight testing began
at EDW, F-35 store separation flight test events were limited to one separation

Fig. 51 Flight test: GBU-12 store separation from BRU-68 rack in underwing pylon.
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 631

Fig. 52 GBU-39 store separation multistep criteria: chase pilot and 6DOF TM Kit assessment.

event per flight, partly in order to allow store separation engineers enough time to
download, process, and analyze the data before making a clear-to-next-point call
that would allow the test site Integrated Test Force to continue to the next store
separation flight test point. For in-bay store separation events, this approach was
in line with the strategy of primarily performing separations from the left-hand
bay, because telemetry retransmission equipment, required to increase likelihood
of successful 6DOF/IMU TM Kit data acquisition, was installed where stores
would normally be carried in the right-hand bay. Furthermore, in-bay air-to-
ground store separation events typically required a different weapons test range
than the in-bay air-to-air AIM-120 AMRAAM store, further contributing to
the necessity for single-release-per-flight missions.
With the four-place GBU-39 carriage arrangement from the BRU-61 within
the weapon bay, however, there was an opportunity to plan and execute store sep-
aration missions involving the release of up to four of these stores from the
weapons bay, thereby increasing test efficiency by reducing the total number of
required store separation test flights. Fully realizing this opportunity required a
clearly understood process, with unambiguous criteria for in-control-room
approval for continuation to the next GBU-39 separation event. After gaining
acceptance by the Integrated Test Force team, a three-step process (Fig. 52) was
implemented, which led to a continuation decision within five minutes.

1. CHASE PILOT ASSESSMENT

The first rapid decision was based on the visual observation from the chase pilot.
Before the mission, the predicted trajectories for all planned store separation
632 C. F. HETREED ET AL.

events were briefed to the test team, including the test and chase pilots. This brief-
ing includes animated movies from the prediction simulations, along with a
description of important behaviors to consider and communicate back to the
control room, such as pitch attitude (down vs flat or nose-up) near the aircraft
and further away from the aircraft. The chase pilot’s description of the visually
observed event, relative to the preflight briefed store trajectory behavior, was
audible to all engineers in the control room.

2. ON-EJECTOR PERFORMANCE
The second rapid communication was based on an in-control-room assessment of
whether the ejector performed as expected. Store 6DOF/IMU TM Kit telemetry
data were examined, with primary focus on store vertical acceleration, pitch
rate, and roll rate. On-ejector parameter plots from prior static pit ejection and
flight test events served as a guide for the in-control-room store separation engin-
eer. For each of the four GBU-39 positions, there was an expectation of specific
peak vertical acceleration during the ejection, in addition to specific store pitch
and roll rates at the completion of the ejection phase, EOS. A rigid, documented
tolerance band was allowed; if these tolerances were exceeded, continuation to the
next GBU-39 was not allowed. On the other hand, verification of expected
on-ejector performance allowed the store separation engineer to move to the
next criterion.

3. OFF-EJECTOR EARLY TRAJECTORY

The final rapid decision was based on the off-ejector store trajectory between EOS
and the time that the GBU-39 moved below the open weapon bay doors. Preflight
predictions included Monte Carlo uncertainty analyses, accounting for variations
in store mass properties, aircraft flight conditions, ejector rack performance, and
store aerodynamic forces. Safe separation trajectories were gathered from these
simulations and formed the basis for documenting acceptable ranges, or vari-
ations, in store linear acceleration and rotational rates during this brief phase in
the store separation event. The criteria for acceptable early trajectory store behav-
ior, as observed from the store 6DOF/IMU TM Kit data in the control room, were
that these accelerations and rates must remain within the documented acceptable
ranges. Exceedance of the allowed variation, with any one of the 6DOF par-
ameters, automatically resulted in ending the mission—continuation to the
flight’s next release event was not allowed.

4. EVENTS
To minimize the in-control-room store separation engineer’s comparison assess-
ment subjectivity and decision time, the referenced criteria parameter plots were
presented identically to those plotted in the control room playback of the teleme-
tered 6DOF/IMU parameters—reference plots included the same time and
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 633

parameter plot scales, and parameter variation tolerances were clearly marked.
Complete confidence in all three of these assessments was required. With atten-
tion to safety, continuation to the next separation event in the flight mission
was allowed only after a firm, clearly communicated favorable assessment in all
three steps in the process. As a result, the EDW F-35 Integrated Test Force
(ITF) team eliminated 12 GBU-39 store separation flights by efficiently conduct-
ing the following:
. Two flight missions, each including four store separation events
. Two flight missions, each including three store separation events
. Two flight missions, each including two store separation events
This same in-control-room process of analyzing 6DOF/IMU telemetered par-
ameters within the control room was utilized during flights with multiple separ-
ation events with other stores:
. The PAX F-35 ITF team eliminated seven store separation flights by conduct-
ing multiple separations per flight for the GBU-12, GBU-32, AIM-132
ASRAAM, and Paveway IV weapons, including a flight during which all
four external GBU-12 stores were released.
. The EDW F-35 ITF team eliminated four more store separation flights by con-
ducting multiple GBU-12 separations in three flights.

C. GBU-12 TAILKIT LANYARD RECONFIGURATION

The GBU-12 store is an air-to-ground weapon that uses deploying tail fins for
increased aerodynamic stability; the opening of these tail fins is mechanically acti-
vated by a wire lanyard arrangement. With rapid and unsteady air flow around the
separating store, intricate lanyard arrangements are sometimes susceptible to
improper function, especially when arrangements differ from typical externally
carried, underwing arrangements on many prior aircraft programs. Because
weapon bay and door geometry constraints are present, the GBU-12 tail fins
could not be opened until the store was past the doors during an ejection event,
which necessitated a longer and more complex lanyard arrangement.
Based on prior AFSEO experience with lanyards and GBU-12-like stores
(Fig. 53), it was recommended that a failed lanyard condition, where fins do
not open, should not be considered a statistically insignificant event—store separ-
ation simulation events should include both a normally functioning tail fin
arrangement and a nonopening tail fin arrangement.
The unfavorable store stability associated with the latter case, with completely
stowed fins, led the store separation team to mandate a BRU-67/BRU-68 rack
ejector split setting with nose-down pitch rate bias in the GBU-12 store loading
procedures, which ensured that the stowed-fins GBU-12 would not point up
and fly back to the aircraft during the store separation event.
634 C. F. HETREED ET AL.

Fig. 53 F-15 aircraft with flyback of GBU-12-like store with stowed tail fins [23].

During one of the first four GBU-12 store separation events from the F-35, the
tail fins did not deploy, remaining stowed. As a result, the store pitched down safely
and as expected, due to the mandated nose-down pitch rate bias on the BRU-67/68
rack. In addition to reaffirming the prudence of including the lanyard failure
scenario when selecting the appropriate ejector split setting, this store trajectory
associated with this fins-stowed event proved to be beneficial because the store
was in one of the exact store model arrangements during preflight wind tunnel
testing. Preflight store separation wind tunnel testing for the GBU-12 included
the model in various states of fins and canards, and aerodynamic forces and
moments were measured for these states, including one state with fully stowed
fins (Fig. 54). As it turned out, the flight test trajectory for this event identically
matched the preflight simulation’s trajectory (with fins stowed), effectively validat-
ing the GBU-12 store aerodynamics model in this fins-stowed configuration.
However, the demonstrated failure of the lanyard to properly function
(Fig. 55) resulted in a brief root cause investigation. Because air flow between

Fig. 54 F-35 GBU-12 fins stowed and fins opening: flight test events and wind
tunnel models.
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 635

Fig. 55 Proper and improper (circled) GBU-12 lanyard function.

and near the weapon bay doors is both rapid and unsteady, the initial lanyard
routing arrangement resulted in portions of the lanyard pulling prematurely
under various conditions. A simpler yet more reliable arrangement was
implemented after this event, resulting in increased consistency in lanyard behav-
ior within the unsteady underbay aircraft region.
Several noteworthy items were associated with this event:
. The importance of ASVS camera video evidence was emphasized during the
root cause investigation.
. The GBU-12 lanyard arrangement redesign, for in-bay releases, happened
early in the test program, with enough remaining in-bay separation events
to provide statistical lanyard function reliability metrics.
. Valuable GBU-12 aerodynamics model validation data were still obtained.
. Flight testing was not delayed; because of the intentionally biased nose-down
fins stowed store trajectory behavior, flight testing could continue safely
without the prospect of future unsafe trajectories.

D. IMPORTANCE OF INCLUDING AIRCRAFT MANEUVER IN POSTTEST ANALYSIS

During a GBU-31 store separation flight test event, test engineers in the control
room noticed that a replay of the ASVS in-bay camera video revealed store
636 C. F. HETREED ET AL.

lateral movement that was different than that typically observed in prior GBU-31
store separation events. Postprocessed 6DOF/IMU TM Kit – based and
photogrammetry-based store trajectory solutions confirmed this additional
lateral store movement relative to the aircraft. An ASVS in-bay camera image
of this event, superimposed with an image from this same camera, but for a differ-
ent event, visually reveals this difference (Fig. 56).
However, the raw, unprocessed store 6DOF/IMU TM Kit data for these
two different events were nearly identical, even in the lateral acceleration measure-
ment parameter. Air-to-ground store separation simulations typically reveal
similar store trajectories, relative to the ground, or Earth, following release from
the aircraft, even if the aircraft has different acceleration response, provided

Fig. 56 GBU-31: Two separation events, superimposed.

F-35 WEAPONS SEPARATION TEST AND VERIFICATION 637

that the aircraft initial forward velocities are the same. So, the focus of the lateral
store movement needed to include consideration of another culprit.
Because the Lockheed Martin posttest analysis process includes the maneu-
vering aircraft response data for each flight test condition, the combined store
separation posttest simulation replays of the store and aircraft trajectories, for
the two different flight test events, revealed the reason for the lateral store move-
ment, relative to the aircraft: the two aircraft were moving differently. In fact, the
aircraft flight control system, during the store separation event with the appear-
ance of lateral movement, had responded to sudden pilot control stick inputs at
the time of the separation event, briefly taking the aircraft out of the specified
release acceleration limits. Although the image appears to show the store
moving laterally during the separation event, it was instead the aircraft that was
moving laterally.
Based on visual observation of the replay of the store separation event, the
team thought that this might be considered an unexpected event, requiring a
pause in the GBU-31 store separation test program; however, because of the
rapid posttest analysis process, which includes combined store and aircraft
response, continuation to the next flight test point was not delayed, with the
full explanation provided to the test team within two days. So, the main lesson
from this example is that when the aircraft response is not included in posttest
replay simulations, incomplete or incorrect explanations of events similar to
this one might have had cascading consequences. These may have included poss-
ible delays in the test program or attributing the perceived unexpected store
motion to increased uncertainty or general factors of nonsimulation—unknowns
that weren’t directly included in the nominal or Monte Carlo uncertainty
simulations.

IX. CONCLUSION
Many variables can prevent a store separation flight test event from taking
place: bad weather, excessive clouds, chase aircraft issues, refueling aircraft
issues, range-clearing aircraft or boat issues, instrumentation issues, test aircraft
issues, telemetry issues, relatively short test range times, and even boats or whales
in an off-shore test range. Difficult maneuvers required additional preparation to
increase the probability of success—stressing high-g and high-angle-of-attack
maneuvers were rehearsed many times within flight simulators, to minimize
time-consuming (and fuel-consuming) in-flight cold-pass rehearsals. Store
separation flight test events are expensive and challenging to achieve, requiring
no barriers to stand in the way of a successfully executed store separation
mission.
The basis for both Lockheed Martin’s certification recommendation and ver-
ification that the weapon release requirements had been satisfied is flight test-
validated M&S—modeling and simulation that has been adjusted based on data
638 C. F. HETREED ET AL.

gathered during the store separation flight test events, so that simulated store sep-
aration events replicate the actual store separation events. Failure to obtain the
necessary instrumentation measurement data for verification of safe store separ-
ation and proper M&S validation would mean that the test point would need to be
reflown—an expensive, schedule-stressing situation, considering the aforemen-
tioned difficulty associated with achieving successful store separation flight
test missions.
Overall confidence in the final M&S and certification is directly dependent on
the quality of the dynamic system models, the quality of the aerodynamics models,
and the quality of the adjustment and validation of these models. Increasing the
quality of baseline dynamic system models was accomplished by using functional
mechanical models of the aircraft, S&RE, and store in the ASEP modeling and
simulation environment. The quality of the aerodynamics models, based on
industry-standard scale-model wind tunnel testing, was sufficient for preflight
simulations required to obtain flight clearance, and was adjusted and validated
during the progression of flight testing. The high quality of the aerodynamics
model validation and the rapid flight test pace was a direct result of extensive plan-
ning, with attention to ensuring that the instrumentation critical to confident vali-
dation was obtained reliably.
During F-35 weapons testing, the pieces were in place to support high-quality
photogrammetry-based trajectory solutions, 6DOF/IMU TM Kit – based trajec-
tory solutions, and store aerodynamics model validation. The body of knowledge
gained in these ground-based and in-flight tests, combined with adjusted and vali-
dated model simulations, support the confident safe store separation assessments
required for a certification recommendation to release weapons from the F-35.

ACKNOWLEDGMENTS
The authors would like to recognize several individuals for their contributions,
specifically in regards to understanding or increasing confidence. Rob Crandall,
from the PAX F-35 Integrated Test Force, is recognized for his persistence and
insistence in quantifying 1) accuracy in DSSAS photogrammetry – based store
position solutions, and 2) uncertainties in store mass property measurements.
Paul Sasenbury, from the EDW F-35 Integrated Test Force, is recognized for
his diligence and commitment to ensuring that the store separation team
thoroughly understand detailed, real-hardware store-to-S&RE mechanical inter-
action, so that models reflect actual hardware behavior. Greg Harding, from the
Air Force SEEK EAGLE Office, is recognized for his continual reinforcement of
the importance of objectively assessing predictive confidence. The mere existence
of numerous vivid flight test videos contributes to initial impressions that store
separation flight test events were successful; however, the success of the store sep-
aration flight test program was dependent primarily upon behind-the-scenes
efforts by the PAX, EDW, and Fort Worth F-35 Integrated Test Force teams,
assuring high-quality instrumentation and measurements at the heart of
F-35 WEAPONS SEPARATION TEST AND VERIFICATION 639

high-confidence modeling and simulation validation leading to the rapid flight

testing pace and certification of 11 stores across the F-35A, F-35B, and F35C
JSF variants.

REFERENCES
[1] Sheridan, A. E., Rapp, D. C., and Burnes, R., “F-35 Program History—From JAST
to IOC,” June 2018 (to be published).
[2] Counts, M. A., Kiger, B. A., Hoffschwelle, J. E., Houtman, A. M., and Henderson, G.,
“F-35 Air Vehicle Configuration Development,” June 2018 (to be published).
[3] Purdon, M. L., Hetreed, C. F., and Hudson, M. L., “F-35 Pre-Flight Store Separation
Analyses: Innovative Techniques for Affordability,” AIAA Paper 2009-102, Jan.
2009.
[4] Hayward, D. M., Duff, A. K., and Wagner, C., “F-35 Weapons Design Integration,”
June 2018 (to be published).
[5] Schindel, L. H., “Store Separation,” AGARD-AG-202, June 1975.
[6] Arnold, R. J., Epstein, C. S., and Bogue, R. K., “Store Separation Flight Testing,”
AGARD-AG-300-Vol-5, 1986.
[7] Nadar, O., “Aircraft/Stores Compatibility, Integration and Separation Testing,”
AGARD-AG-300-Vol-29, Sept. 2014.
[8] MIL-HDBK-1763, “Aircraft/Stores Compatibility: Systems Engineering Data
Requirements and Test Procedures,” 1998.
[9] MIL-STD-2088A, “Bomb Rack Unit (BRU), Aircraft, General Design Criteria for,”
1997.
[10] MIL-HDBK-244A, “Guide to Aircraft/Stores Compatibility,” 1990.
[11] Keen, K. S., Morgret, C. H., Langham, T. F., and Baker, W. B., Jr., “Trajectory
Simulations Should Match Flight Tests and Other Lessons Learned in 30 Years of
Store-Separation Analysis,” AIAA Paper 2009-99, Jan. 2009.
[12] Morgret, C. H., Moore, D. A., and Smith, M. E., “The FLIP 4 Store-Separation
Trajectory Simulation Code,” AIAA Paper 2009-100, Jan. 2009.
[13] Keen, K. S., “Equations for Store Separation Motion Simulations and Instrumented
Model Data Reduction,” AEDC TR-95-12, Aug. 1996.
[14] Veazey, D. T., “Current AEDC Weapons Separation Testing and Analysis to Support
Flight Testing,” AIAA Paper 2004-6847, Nov. 2004.
[15] Davis, M. B., Yagle, P., Smith, B. R., Chankaya, K. M., and Johnson, R. A., “Store
Trajectory Response to Unsteady Weapons Bay Flowfields,” AIAA Paper 2009-547,
Jan. 2009.
[16] Hudson, M. L., and Charlton, E. F., “Many Uses of CFD in JSF Store Separation,”
International Aircraft-Stores Compatibility Symposium XIV, Fort Walton Beach,
FL, 11 – 13 April 2006.
[17] Crandall, R., “Airborne Separation Video System,” presented at Aircraft-Store
Compatibility Symposium, Destin, FL, March 2001.
[18] Forsman, E., Getson, E. S., Schug, D., and Urtz, G., “Improved Analysis Techniques
for More Efficient Weapon Separation Testing,” Proceedings of the 19th Annual
640 C. F. HETREED ET AL.

Symposium of the Society of Flight Test Engineers European Chapter, Manching,

Germany, September 2008.
[19] Getson, E. S., “Telemetry Solutions for Weapons Separation Testing,” presented at
ITEA Aircraft-Stores Compatibility Symposium and Workshop, Destin, FL,
Feb. 2003.
[20] Forsman, E., and Schug, D., “Estimating Store 6DOF Trajectories Using Sensor
Fusion between Photogrammetry and 6DOF Telemetry,” presented at ITEA Test
Instrumentation Workshop, 2011.
[21] Harding, G. C., and Barton, K. M.., “Predicting Aircraft Flowfield Induced Water
Vapor Condensation for Store Separation Flight Tests,” AIAA Paper 2007-4169,
June 2007.
[22] Tutty, M., Akroyd, G., Cenko, A., and Piranian, A., “Stores Separation from
Weapons Bays,” presented at 30th Congress of the International Council of the
Aeronautical Sciences, Daejeon, Korea, Sep. 2016.
[23] Roberts, E., “Lessons Learned: Limitations of Modern Tools and Applications for
Store Separation Prediction,” presented at Aircraft-Store Compatibility Symposium,
Destin, FL, March 2001.
 CHAPTER 17

F-35 STOVL Performance

Requirements Verification
David G. Parsons and Daniel E. Levin†
Lockheed Martin Aeronautics Company, Fort Worth, TX
David J. Panteny‡ and Peter N. Wilson§
BAE Systems, Warton, United Kingdom

Michael R. Rask}
Cummings Aerospace, Niceville, FL

Brad L. Morris
Naval Air Systems Command, Naval Air Station Patuxent River, MD

The F-35B is the Short Takeoff and Vertical Landing (STOVL) variant of
the F-35, with two challenging and unique requirements. First, the
aircraft must take off from both LHA/LHD-class flat-deck ships and
ski-jump – equipped Queen Elizabeth –class ships while carrying the
contractually specified fuel and payload. The second requirement is to
return and perform vertical landings on the same ships or at austere
sites while carrying unexpended ordnance. Flat-deck, short-takeoff
distance and vertical landing bring-back capability requirements are
two STOVL aircraft Key Performance Parameters (KPPs) that must be
met for the aircraft to be successful. The verification of achieving these
KPPs spanned more than a decade. It involved sub- and full-scale
ground-based testing and model development, and culminated in
flight testing in the System Development and Demonstration phase of
the F-35 program. This paper focuses on the flight test element of the
F-35 program’s STOVL performance requirements verification effort.
Flight test verification methods and results, as well as operational
observations, are presented for both shore- and ship-based testing.

 Aeronautical Engineer, Senior Staff, Flight Sciences.

†
Lockheed Martin Site Lead Test Pilot, F-35 Integrated Test Force, NAS Patuxent River.
‡
BAE Lead STOVL Propulsion Engineer, F-35 Integrated Test Force, NAS Patuxent River.
§
STOVL Lead Test Pilot, F-35 Integrated Test Force, NAS Patuxent River.
}
Principal Aerospace Engineer.
 Tactical Fixed Wing Performance Lead Engineer, NAWCAD, Patuxent River.

641
642 D. G. PARSONS ET AL.

I. INTRODUCTION
The F-35 Joint Program Office required the Short Takeoff and Vertical Landing
(STOVL) F-35 variant to have substantially improved range, payload, and bring-
back capabilities over the legacy Harrier. With 5th-Generation stealth capabilities
and supersonic speeds, the STOVL variant provides the operator with enhanced
lethality, survivability, maintainability, and affordability.
The Joint Contract Specification (JCS) identifies two challenging Key Per-
formance Parameters (KPPs) for STOVL operations. The first specifies that the
aircraft must be able to launch from a flat-deck LHD/LHA-class ship with a
full complement of internal stores and the fuel required to support a designated
combat radius. The second specification establishes the aircraft’s Vertical
Landing Bring-Back (VLBB) requirement, which is based on recovery with
those same internal stores. VLBB also requires an appropriate amount of fuel
on board to conduct operations in the pattern. Figure 1 shows an F-35B hovering
abeam an LHD prior to a Vertical Landing (VL).

Fig. 1 F-35B hovering abeam an LHD.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 643

These requirements provide a metric for comparing competing designs and

design changes. Although generally representative of the expected aircraft capa-
bilities, the requirements are specified for ideal operating conditions that are
not necessarily consistent with real-world operations. The actual capabilities of
the aircraft that are provided to the operator must be validated across the expected
range of operating conditions and configurations.
This paper describes the process used to verify the achievement of the two
KPPs. It details how additional testing and analysis led to the operational perform-
ance capabilities of the F-35B. The discussion is limited to flat-deck Short Takeoff
(STO) and VL. As such, it does not cover the additional launch and recovery capa-
bilities that were developed and validated, such as ski jump STOs, vertical takeoffs,
and slow landings. In addition, it discusses several operationally significant issues
and lessons learned that were discovered during flight test.

II. BACKGROUND
The STOVL variant of the F-35 is intended for use by the U.S. Marine Corps and
the U.K. and Italian ministries of defense as a replacement for the AV-8B Harrier.
The F-35B flight regime is divided into conventional and STOVL modes. Conven-
tional mode is used for nominal, fixed-wing-type flight operations. STOVL mode
provides the operator with capabilities essential to the F-35B mission. This
includes the abilities to take off from naval vessels or austere runways without
using a catapult, and to return to land on the same surfaces.
In STOVL mode the propulsion system generates thrust that is exhausted
through four nozzles: the main or core nozzle, the LiftFanw nozzle, and the left
and right roll-post nozzles [1]. The core nozzle can pitch downward and yaw
left and right through the three-bearing swivel module. The LiftFan is powered
by diverting horsepower through a driveshaft from the main engine’s low-
pressure turbine via a clutch and gearbox. The LiftFan exhaust can be vectored
forward and aft via a Variable Area Vane Box Nozzle (VAVBN). The roll-post
nozzles are located under the wings just outboard of the main landing gear.
The STOVL configuration is shown in Fig. 2.
To meet other requirements of the JCS, including a supersonic capability and
low observability, the high-drag, nonstealth components of STOVL mode are
hidden behind doors. STOVL operation is incorporated into the design as a
unique mode that is used for only a small portion of a mission and is enabled
through a process called conversion. STO is a submode of STOVL in which the
Control Laws (CLAW) automatically manage thrust, Thrust Split (TS), and
control effector positioning for optimal STO performance. TS is the ratio of
core nozzle thrust to LiftFan nozzle thrust. For a STO, the pilot selects STOVL
mode prior to takeoff (Fig. 3) and converts back to the Conventional Takeoff
and Landing (CTOL) mode once airborne. In CTOL mode, the aircraft is prepared
to use its supersonic, stealthy, 5th-Generation capability. At the end of the
644 D. G. PARSONS ET AL.

Fig. 2 STOVL features.

Fig. 3 F-35B STO from an LHA.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 645

mission, the pilot converts back to STOVL mode in preparation for a slow
landing or VL.

III. F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION

The verification of the two KPPs was the culmination of years of development, as
Fig. 4 shows. All performance verification work on the F-35 program followed the
same approach. Aerodynamics models were developed from subscale wind tunnel
and full-scale force and moment evaluations, and uninstalled propulsion perform-
ance was measured in indoor and outdoor test facilities. Flight testing was accom-
plished to validate the models used to calculate performance. Once validated,
those models were then used to calculate performance against the specification
requirements. Flight testing to verify the JCS requirements discussed in this
paper is a subset of the larger performance test matrix. It was designed to sub-
stantiate the performance of the aircraft throughout the flight envelope to validate
operational performance mission-planning products.
Through years of development testing on the F-35 program, the test team
completed numerous maneuvers. Table 1 indicates the experience and knowledge
the team developed for STOVL-related tasks, along with the robust capability of
the F-35B. The data in the Unique column under Vertical Landings and Short
Takeoffs indicate events completed to satisfy the requirements of specific
test points.

Fig. 4 F-35 STOVL performance verification timeline.

646 D. G. PARSONS ET AL.

TABLE 1 TOTAL VLS AND STOS AS OF 14 FEBRUARY 2018

Vertical Landings Short Takeoffs

Aircraft Unique Total Aircraft Unique Total
BF-1 360 782 BF-1 271 1407
BF-2 187 286 BF-2 92 358
BF-3 18 25 BF-3 20 29
BF-4 89 181 BF-4 158 458
BF-5 104 228 BF-5 105 470
Grand Total 758 1502 Grand Total 646 2722

IV. SHORT TAKEOFF PERFORMANCE VERIFICATION

The JCS flat-deck STO requirement is to take off within a specified distance from
an LHD/LHA-class ship. The aircraft taking off must have internal weapons and
only the fuel necessary to complete a specified mission profile. The total aircraft
weight, including the fuel required to complete the mission and stores, affects
the spotting distance required to launch the aircraft. Verifying the requirement
entails validating both the conventional performance databases (for the fuel
required to complete the mission) and STOVL-mode databases (for the distance
required to launch the aircraft). The verification of the fuel required to complete
the specified mission was part of the conventional performance requirements ver-
ification process [2]. Validating the STOVL-mode performance databases for STO
calculations required both shore- and ship-based testing.

A. SHORE-BASED SHORT TAKEOFF TESTING

Shore- and ship-based STO performance validation was accomplished in a comp-
lementary manner. Shore-based testing was conducted as a prerequisite for ship-
based testing, but it provided data for use in validating areas of the databases
common to both. These databases were updated with and validated through
these results, and the validated databases provided the foundation for ship-based
STO predictions and testing.
Pretest analysis and modeling provided recommended rotation speeds, which
are a critical element of the STO, for the beginning of shore-based testing. The
STOVL propulsion effectors (core nozzle and LiftFan nozzle thrust vector
angles, and TS) provide considerable rotation power. As such, these alone are
capable of rotating the aircraft at a speed below which the aircraft can fly away.
The recommended rotation speed needed to satisfy several criteria:
. The aircraft must reach 45 knots Calibrated Airspeed (KCAS) before the
control system will allow rotation.
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 647

. The rotation and liftoff must provide sufficient clearance to avoid a nozzle or
STOVL door strike.
. The aircraft must exhibit a positive climb rate and flight path acceleration
throughout the climb out.

The STO predictions surveyed rotation speeds for each mass properties/power
setting/ambient atmospheric condition. These determined the optimum speed
to achieve a height of 50 ft above ground level in the shortest ground distance
while meeting the three criteria above.
From a performance analysis standpoint, the STO maneuver can be broken
into three major elements [3]. These are the ground acceleration to rotation
speed, the rotation to flyaway attitude, and the climb out to a height consistent
with Out-of-Ground Effects (OGE) conditions.
The ground acceleration is straightforward and dominated by thrust from the
propulsion system. Aircraft control surface positions are relatively constant
throughout the acceleration. The contributions from unpowered aerodynamics
(mainly drag) are minimized due to the slow rotation speeds (low dynamic
pressure generates low drag force). Test data validated the predicted levels of
ground acceleration performance, and no adjustments were necessary for the
models to match flight test. This was a crucial aspect of model validation that
was required to be completed prior to moving to ship-based STOs.
STO rotation, shown in Fig. 5, is a dynamic event. Multiple control effectors
(conventional control surface positions and STOVL propulsion system effectors)
move in unison to increase pitch attitude on the aircraft. This, in turn, increases
conventional aerodynamic lift while simultaneously providing thrust in the

Fig. 5 STO rotation.

648 D. G. PARSONS ET AL.

vertical and longitudinal axes. Use of the effectors is managed by the flight control
system and is a function of many variables [4].
Owing to the dynamic nature of the rotation event, model comparisons to
flight test data are generated by feeding actual flight test control effector usage
into the model. This is done as a function of groundspeed, and the model
ground track, pitch path, and flight path are integrated. Ideally, true airspeed
would be used instead of groundspeed to ensure a match of dynamic pressure;
however, the true airspeed signal is occasionally noisy and subject to wind gusts
that drive the true airspeed trace to be nonmonotonic with either time or distance.
Because groundspeed is single-valued with both time and distance, it is used to
determine the correct effector positions from flight test data. Time is not used
to determine control effector usage from test because the entire maneuver is ana-
lyzed at once. If there were any deviation from the test airspeed and model air-
speed at rotation, the causes of differences between test and model flyaway
characteristics would be more difficult to ascertain. (Such a deviation would be
due to differences in ground acceleration.) This approach forces rotation model-
ing to be done at airspeeds consistent with flight testing to achieve representative
conventional lift and drag forces.
The modeling of the rotation control effectors is conducted in an open-loop
fashion, for performance analysis purposes, until the model achieves nose gear
liftoff. After model nose gear liftoff is achieved, the remaining rotation is con-
ducted in a closed-loop fashion using horizontal tail deflection to close the loop
on the flight test pitch attitude. For this portion of the prediction, the model
uses whatever predicted tail deflection is necessary to match the flight test pitch
vs the groundspeed characteristic. Modeling in this manner effectively uses the
horizontal tail to account for any pitching moment errors in the predicted
databases.
Analyzing the rotations of multiple STO maneuvers across a range of mass
properties and ambient conditions showed that the model was a good match
for the test aircraft. Minor adjustments to the model pitching moment and lift
coefficient in ground effect at compressed gear height were included as well.
These enabled us to better match test aircraft rotation characteristics. No other
adjustments were required.
Flight test STO climb out to a height of 50 ft was conducted with the pilot out
of the loop (as much as possible) for performance purposes. As a result, the air-
craft control surfaces remained in nearly constant positions once flyaway attitude
was attained. The F-35 flight control system schedules the pitch attitude of the air-
craft for STO flyaway as a function of calibrated airspeed, with attitude gradually
decreasing as airspeed increases. The modeling of the flyaway was conducted in
the same manner as the last portion of the rotation. A model horizontal tail pos-
ition was used to close the loop on aircraft pitch attitude as a function of ground-
speed. Initial comparisons of STO climb out showed the predicted model to be
conservative in terms of climb rate; however, conventional Powered Approach
(PA) configuration performance flight test data indicated a lift benefit for the
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 649

test aircraft versus predictions. Once this increment for conventional lift was
included in the STO modeling, predictions for climb out to a height of 50 ft
matched test data within acceptable tolerances. Some residual pitching moment
differences existed during climb out, evidenced by differences in the horizontal
tail deflections between the model and test; however, the differences in tail deflec-
tion had only minor effects on the climb out characteristic and were considered
acceptable for performance modeling purposes.
Figure 6 provides an example comparison of the predicted model to the results
of a complete shore-based STO. Climb out and ground acceleration matched the
test data well. In the figure, the lower two plots illustrate how the comparisons
were developed. Prior to liftoff, the prediction was calculated in an open-loop
fashion, with control effectors in the model forced into the flight test positions
as a function of groundspeed. This is demonstrated in the figure by the tail pos-
ition in the Tail Behavior plot and was done prior to separating prediction and test
data. Once the nose wheel lifted off, the pitch attitude in the model was forced to
match the pitch attitude from flight test as a function of groundspeed. In the
figure, this is depicted in the Pitch Attitude plot after approximately 83 kt. The
residual pitching moment difference in the model is evidenced in the Tail Behav-
ior plot as a difference between the flight test and model predicted tail positions.
Figure 7 summarizes the results of the shore-based STO analysis, obtained
after all adjustments noted above had been included in the model. Data are pre-
sented as flight test distance to a height of 50 ft vs the predicted distance to a

Fig. 6 STO performance characteristics.

650 D. G. PARSONS ET AL.

Fig. 7 Shore-based STO model comparison to flight test.

height of 50 ft. In general, all predicted data are within 5% of the test values for all
conditions, configurations, and power settings tested.

B. SHIP-BASED SHORT TAKEOFF TESTING

Successful shore-based testing resulted in the development of a high-quality STO/
STOVL model that permitted the team to enter the ship-based testing phase with
confidence. It also allowed our team to focus mainly on the characteristics unique
to ship-based operations. For ship STOs, the pilot does not initiate rotation at a
target airspeed, as is done for shore-based STOs. On the ship, a reference rotation
line is placed 225 ft from the ship’s bow, and all rotations are initiated at this point
on the deck. Predictions are generated to determine the proper starting distance
(spot) from the bow of the ship from which to initiate the STO. The STO is
initiated in a manner such that the aircraft will reach the rotation line at the
correct speed. Spots are markings on the carrier deck in 50-ft increments from
the bow, with the largest usable distance being 750 ft. The successful correlation
of predictions to the shore-based ground acceleration data provided a solid foun-
dation for this portion of the ship-based STO predictions.
An element of ship operations that could not be evaluated during shore-based
testing was the variation of ambient wind over the deck of the ship. The amount of
headwind and crosswind the aircraft experiences during a STO is a function of the
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 651

magnitude and direction of the ambient wind and the speed, pitch, and roll of the
ship. It is also affected by the position of the aircraft along the STO launch tram
line and its location relative to the ship’s superstructure. The environment can be
very unsteady and unpredictable.
The aircraft’s ability to successfully fly away from a ship STO launch is a func-
tion of airspeed and aircraft attitude at the deck edge. Figure 8 shows the aircraft
approaching the deck edge as the flyaway attitude is established. A deck spot is
used to target the airspeed and attitude for a given combination of aircraft
weight and ambient conditions. The airspeed can be generated through a combi-
nation of ambient wind, ship speed, and aircraft longitudinal acceleration poten-
tial along the deck.
Similar to the shore-based STO, the preferred launch spot for a given ship-
based STO is a balancing act of multiple criteria:
. The aircraft must achieve 45 KCAS prior to the rotation line for the CLAW to
enable STO mode. Without STO mode, the aircraft cannot rotate to the atti-
tude necessary to achieve a successful launch.
. The aircraft must have sufficient clearance to prevent a nozzle or STOVL door
strike during rotation prior to leaving the deck.
. The aircraft must have margin to the launch conditions that yields 10 ft of air-
craft Center of Gravity (CG) sink once the aircraft leaves the deck.
The 10-ft-of-sink criterion is the only one of the three that is mandated by the
contract specification. The other criteria have been developed for use only with
the F-35B.

Fig. 8 STO from the USS Wasp.

652 D. G. PARSONS ET AL.

The JCS-based sink criterion is the focus of dedicated performance STO

testing on the ship. To ensure that the aircraft has 15 kt of margin to 10 ft of
sink off bow, the 10-ft-of-sink condition must be validated. Obviously, these cri-
teria cannot be evaluated during shore-based testing. The complex environment
experienced by the aircraft during operations at the ship is difficult to predict.
That drives the focus in this area of testing and is mostly responsible for the
requirement for operations to have 15 kt of margin to this criterion.
Predictions were generated prior to the ship trials to understand the variation
of aircraft sink with wind over deck. At the ship, testing progressed in a controlled
fashion by reducing the amount of wind over deck in measured steps. In this way,
we quantified the wind over deck vs aircraft sink characteristic. For an F-35B flat-
deck STO, the wind over deck must decrease from the operationally planned value
by as much as 8 to 10 kt before the aircraft CG will experience any appreciable sink
(less than 2 ft) relative to the ship deck. As wind over deck is reduced beyond the
8-to-10-kt level, the amount of sink off bow increases fairly rapidly. At the
10-ft-of-sink point, the slope of the airspeed vs the sink curve is very steep,
which makes minimum-end airspeed testing very challenging.
Due to the sensitivity of aircraft sink to the maneuver technique, all perform-
ance ship-based STOs were conducted as auto STOs. An auto STO is a technique
with which the pilot enters the deck spot through the cockpit displays. The flight
control system determines the distance traveled from throttle up and brake release
to the rotation line. Then, it initiates and executes rotation to capture flyaway
attitude automatically. This STO type was designed to eliminate possible pilot
rotation errors and variability, and it is the primary STO technique used at
the ship.
The ship-based testing approach included mitigation procedures that could
be implemented if the actual sink was expected to be excessive. This was based
on the real-time monitoring of aircraft conditions in the control room by per-
formance engineers. Shore-based training leading up to the ship trials was used
to help the control room engineers recognize the characteristics of a poor
launch. They did this by monitoring the aircraft height rate trace and the absolute
level of sink below deck level. The mitigation procedure for a poor launch
included advancing the throttle to request the maximum thrust available and
increasing the pitch attitude to arrest the sink rate. The procedure proved effective
because the F-35B flight characteristics provided sufficient time to recognize and
respond to excessive sink.
Ship-based performance STO testing involved two series of build-downs to 10
ft of sink. The first series was conducted with no external stores on the aircraft
during the second development test ship trial (DT-2). The flight test team estab-
lished an approach to incrementally reduce wind over deck margin to gradually
increase sink toward 10 ft. For DT-2, the build-down steps involved reducing
spot distance and/or wind over deck to generate sink. The STOVL propulsion
system performed strongly, so the team planned for tailwind conditions to gener-
ate any amount of sink at the maximum gross weight possible without external
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 653

stores carriage. The other launch criteria, such as airspeed at rotation line and
nozzle clearance, were still respected in the process. As a result, the aircraft
could not simply be spotted as far forward as possible to generate sink.
The tailwind requirement proved to be problematic for successful testing.
With ambient wind coming from behind the ship, the aircraft entered a somewhat
unknown wind condition as it left the deck and sank below deck level. The
ambient wind was not as constant in this airspace as it had been along the
deck. This was likely due to the turbulence effects of the wind swirling off the
front of the boat. The final test point conducted for performance during DT-2
exhibited more sink than desired, and the mitigation procedures were executed.
As expected, once the pilot had increased the pitch attitude, the sink rate was
arrested and the aircraft flew away without incident.
Figure 9 shows the aircraft height rate response to the pilot recovery action
during the high-sink event. The blue line shows the raw Inertial Navigation
System (INS) vertical velocity. The orange line shows the vertical velocity of the
aircraft’s CG, which is the reference parameter for aircraft sink off the bow. As
the pilot increased the pitch attitude of the aircraft, the inertial velocity showed
an immediate positive increase due to its location forward on the aircraft.
Approximately one second from pilot input, the vertical velocity of the CG
became positive. This indicated that the sink rate had been arrested and that
the aircraft had begun to climb.

Fig. 9 Height rate change with pilot response action.

654 D. G. PARSONS ET AL.

Figure 10 illustrates the characteristic of aircraft sink off bow with aircraft
speed at the rotation line for DT-2. The solid line represents the predicted
minimum height of the aircraft’s CG. This is relative to deck edge vs speed at
the rotation line for the target test weight and a nominal wind over deck con-
dition. The symbols represent the flight test sink off bow adjusted to the reference
conditions, such as weight, CG, pressure altitude, ambient temperature, and wind
over deck. The sink is consistent with the predicted line. The final test point is
circled, indicating that the maximum sink was achieved after pilot action.
The testing that was accomplished during DT-2 indicated that the predicted
model was generally conservative; however, there was uncertainty in measuring
the tailwind magnitude and direction in front of the ship. Owing to this, the vali-
dation team was hesitant to adjust models to reflect the results. Another element
of the testing that proved problematic to understanding the results was varying
both the lineup spot and wind over deck to progress from test point to test
point. By changing more than one variable at a time, the team struggled to under-
stand the complex combination of variables that determines aircraft sink off bow.
A posttest detailed analysis of each performance launch provided evidence that
the predicted model was conservative, relative to the flight test results (i.e., had
less real sink than predicted). Further, it demonstrated that the tailwind con-
ditions tested had not been particularly operationally representative. As such,

Fig. 10 Aircraft sink vs deck edge airspeed for DT-2.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 655

Fig. 11 Sink vs wind over deck for DT-3.

the team chose to leave the model unchanged prior to testing for the third devel-
opment test ship trial (DT-3).
Testing during DT-3 utilized external stores to increase the weight of the jet so
that minimum-end airspeed testing could be accomplished in positive headwind
conditions. For this series of tests, aircraft weight and spot were held constant,
leaving wind over deck as the only means to affect wind over deck margin
changes. By holding weight and spot constant, the aircraft would be at nominally
consistent groundspeed (relative to the deck) and attitude at deck edge for each
test point. Again, the team took a measured approach to gradually increase sink
off bow toward 10 ft. The increased weight, positive headwind conditions, and
controlled spot provided higher-quality data than those from DT-2. DT-3 test
data confirmed that the predicted sink characteristic was conservative (i.e., had
less real sink than predicted sink) for a given wind over deck. The data also
showed that the sink vs wind over deck curve characteristic was steeper than pre-
dicted. Figure 11 shows the results from DT-3. As with DT-2 data, the flight test
data have all been adjusted to the same reference conditions: weight, CG, altitude,
and ambient temperature; however, they have been held at a constant deck spot
rather than wind over deck to illustrate the test results against a single predicted
line. The final test point in Fig. 11 shows that the predicted wind over deck charac-
teristic had approximately 2 kt of conservatism at the 10-ft-of-sink condition.
Deviations in ship pitch attitude and heave can impart initial vertical velocity
to the aircraft as it exits the deck. This is difficult to account for accurately, and
656 D. G. PARSONS ET AL.

minor changes in initial vertical velocity can make measurable differences in the
achieved sink at these minimal performance conditions. The test team made every
effort to gather performance minimums test data in sea conditions that were as
calm as possible; however, some deviations always exist and must be accounted
for when assessing results.
Prior to DT-3, the verification team had not updated the models for the
results of conventional aircraft testing in the PA or gear-down configurations.
They also had not made any adjustments for the conservatisms observed
during DT-2 in tailwind conditions. The conventional PA configuration testing
showed increased flight test lift, relative to pretest predictions. After adjusting
models for the results of this testing, each of the minimum-end airspeed perform-
ance test points was compared to the revised model predictions. When detailed
model comparisons were made to the as-flown conditions, the predicted flight
path matched the flight test results within acceptable tolerances. Figure 12 com-
pares the flight test results for the largest sink on DT-3 to the detailed model. The
model is an excellent representation of the flight test results for a very dynamic
maneuver.

C. SHORT TAKEOFF PERFORMANCE VERIFICATION RESULTS

Shore-based and ship-based STO testing provided the performance verification
team with the data necessary to update and validate the models used to predict

Fig. 12 Comparison of flight test and detailed modeling for sink.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 657

Fig. 13 Flat-deck STO verification results.

STO performance. There were only a few changes necessary to have the pretest-
predicted models match the flight test results. These consisted of minor adjust-
ments to ground effect pitching moment and lift, and an update to the conven-
tional PA flight test results.
The STO flight testing showed the performance of the F-35B to be better than
originally predicted. The results of the F-35B conventional performance validation
testing also reduced the amount of fuel required to complete the mission associ-
ated with the flat-deck STO requirement [2]. The combination of improvements
in both conventional mission performance and STO performance allowed the
F-35B to meet the JCS requirement for flat-deck STO. Moreover, it did so with
a margin greater than 20%, which will greatly benefit future growth initiatives
for the weapons system platform. Figure 13 illustrates this achievement.

D. OPERATIONAL CONSIDERATIONS
Performance testing focused on the capability of the aircraft in controlled and
repeatable conditions; however, the aircraft must be operationally robust in a
far wider range of operating conditions and techniques. The problem of sink,
and the pilot’s reaction to it, was discussed in the preceding subsection. Robust-
ness testing also highlighted a potential problem that will eventually affect the
operational capability of the aircraft.
Winds experienced by the aircraft during a STO deck run were anything but
constant. Wind can vary in magnitude and direction throughout the STO event.
With strong ambient winds from starboard during ship STO, the aircraft was in
the lee of the island for the majority of the ground roll; however, it experienced a
large change in crosswind as it cleared the island on the way to the bow. Figure 14
illustrates the difference in wind magnitude and direction as measured at the bow
of the ship and farther aft near landing Spot 7. This was assessed during a wind
calibration event conducted during DT-2.
An interesting result occurred from a wind gust during rotation and unstick
during the first ship trial. The integrated flight and propulsion controls prioritized
sideslip control over performance and used differential horizontal tail for that
purpose. Planned performance was based on the tails being in their maximum
lift position. A deviation resulted from the expected performance and dynamic
nature of the rotation in the presence of a crosswind gust. This, in turn, resulted
in a low nozzle clearance event that exceeded established flight test continuation
658 D. G. PARSONS ET AL.

Fig. 14 Comparison of wind magnitude and direction for bow and Spot 7.

criteria. This event is illustrated in Fig. 15. In it, a nominal STO appears shortly
before unstick on the top-left, and the CLAW reaction to a starboard crosswind
is shown on the bottom-left. The nozzle clearance during the actual event is

Fig. 15 Nozzle clearance comparison.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 659

shown on the right. This was barely noticed in the cockpit. However, the magni-
tude of the exceedance and the risk of a nozzle strike in gusty crosswind con-
ditions ultimately led to adjustments in the aircraft CLAW to modify the
prioritization scheme.
The net result of ship trials testing in crosswind led to the inclusion of
additional margin for STOs in the presence of higher values of crosswind. This
margin is carried in the form of additional speed at the rotation line over that
which would be determined for nominal wind conditions down the deck. It is
implemented as a farther-aft deck spot for these conditions over nominal winds
down the deck.

V. VERTICAL LANDING BRING-BACK PERFORMANCE VERIFICATION

In the air, STOVL-mode operation has two major flight regimes: Semi-Jetborne
(SJ) and Jetborne (JB). SJ covers speeds at which the lift required for flight is pro-
vided by both the aerodynamic surfaces and propulsion system thrust (approxi-
mately 45 KCAS and higher). JB is optimized for flight below 45 KCAS, for
which nearly all lift is generated by propulsion system thrust. The CLAW auto-
matically configures the propulsion system and effectors to optimize performance
while ensuring sufficient control power. For a VL, the jet is decelerated into JB
flight as it is positioned for landing.
The VLBB KPP requirement defines the combination of reserve fuel and unex-
pended stores that must be recovered at the end of a mission. The VLBB task is
described in Fig. 16, and its analysis consists of elements in both the JB and SJ

Fig. 16 VLBB pattern profile.

660 D. G. PARSONS ET AL.

regimes. The maximum hover capability of the aircraft must be determined (JB),
and the weight required for the specification hover must be determined. The latter
is determined from a pattern flown around the ship during a recovery (SJ). In
addition, the ability to land vertically at the required weight must also be validated
(JB). The performance team validated each of these regimes separately.
Forces affecting aircraft performance in STOVL mode flight are:
. Propulsive forces, including thrusts from the four nozzles and inlet and
nozzle drags
. Unpowered aerodynamics, or classical aircraft lift and drag
. STOVL Jet Effects (SJEs), or the effects of entrained and reflected flow, the
component of SJEs in the vertical axis commonly referred to as suck-down
because the result is an effective reduction in propulsive lift
The preflight test SJE database was created based from the results of extensive
wind tunnel testing [5]. Figure 17 shows the relative magnitudes of the three

Fig. 17 Relative magnitude of forces on the STOVL aircraft.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 661

categories of forces as a function of airspeed. During a stabilized hover at OGE

heights, SJE effectively offset propulsive total vertical thrust by approximately
2.5%. Otherwise put, the actual usable vertical thrust is 97.5% of the net vertical
thrust.
Prior to flight test, Pratt & Whitney and Lockheed Martin concentrated con-
siderable resources on validating propulsion system performance characteristics
in the JB flight regime [6]. As a result, it was one of the most accurately
modeled regions of the entire flight envelope in conventional or STOVL mode.
For the inflight thrust methodology [7], data from an instrumented propulsion
system fed a propulsion cycle model to provide an accurate estimate of thrust.
This methodology was a critical element in validating aircraft SJE for low-speed,
JB flight. Pratt & Whitney developed the In-Flight Thrust Calculation Deck
(IFTCD) for use by the performance validation team. The IFTCD was calibrated
for propulsion system performance at hover conditions based on extensive thrust
stand testing at Pratt & Whitney facilities in West Palm Beach, Florida.
Reconciliation of aircraft performance in hover was also improved by using
data from the force and moment hover pit testing [8, 9]. The force and
moment testing on the hover pit in Fort Worth, Texas, provided the first oppor-
tunity to evaluate multiple factors. These included the consistency of IFTCD cal-
culations, wind tunnel– based SJE databases, and measured forces and moments
from load cells installed in the pit restraint system for simulated OGE conditions.
The results of the pit testing confirmed that the combination of IFTCD-calculated
thrust and the preflight test SJE database matched the measured forces within
defined goals. These results allowed the validation team to enter the flight test
phase with confidence in the hover performance prediction model.
Having an aircraft gross weight capability in hover in OGE conditions does not
necessarily directly indicate having a VL capability. Hot Gas Ingestion (HGI) is
a significant factor affecting performance and operability during VL. It reduces
total propulsion system thrust, potentially influencing aircraft descent rate and
handling qualities and reducing engine and LiftFan stall margins. When executing
a VL, the aircraft must not experience excessive uncommanded acceleration or
adverse handling qualities during the landing as the jet enters ground effects.
HGI has two main mechanisms of action: ground sheet roll-up ingestion and
fountain ingestion. These two mechanisms are visually depicted in Fig. 18.
The aircraft exhaust plumes impinging on the ground form a ground sheet
that flows away from the aircraft in all directions. Ground sheet roll-up occurs
when a strong environmental wind picks up this buoyant layer of gas and turns
it back toward the aircraft. Due to the well-mixed nature of the flow returning
to the aircraft, the HGI experienced from this mechanism is of a low magnitude.
Usually, it is only seen near the ground in the last stages of a VL. The
flight-test-measured levels of HGI from this mechanism have not been significant
and do not require mitigation through performance limitations.
There are certain heights above ground where the propulsive jets do not merge
before impacting the ground. There, the ground sheets from the individual
662 D. G. PARSONS ET AL.

Fig. 18 HGI mechanisms.

exhaust plumes of the LiftFan, core nozzle, and roll-posts flow inward toward one
another under the aircraft. An upward fountain is formed where they collide, and
it then impacts the undersurface of the aircraft. Under certain conditions, exhaust
gas from the fountain can enter the main engine inlets and LiftFan inlet. HGI in
this scenario can result in a high temperature rise and high-temperature inlet dis-
tortion. This leads to concerns for both aircraft performance and engine/LiftFan
operability that must then be mitigated.
For the F-35B, the amount of HGI is a function of height above ground, air-
craft relative wind, aircraft Fuselage Station Center of Gravity (FSCG), and
ground slope. Fountain HGI is typically only seen with tailwind or crosswind con-
ditions relative to the aircraft and not with headwind conditions relative to the air-
craft. For the aircraft to remain stationary in the hover in the environmental wind,
the exhaust plumes must be directed aft for a headwind and forward for a tailwind.
Additionally, the environmental wind acts on both exhaust plumes as they travel
toward the ground and on the fountain as it returns. Therefore, a headwind tends
to angle the fountain aft away from the inlets, reducing HGI levels. By contrast, a
tailwind tends to angle the fountain forward toward the inlets, resulting in poten-
tially higher levels of HGI.
An aft aircraft FSCG increases the trim TS required. Higher TSs result in
greater exhaust flow from the engine nozzle vs the LiftFan nozzle. This reduces
the strength of the ground sheet from the LiftFan exhaust, compared to the
ground sheet from the engine nozzle exhaust that forms the fountain. With an
aft FSCG, the point at which the fountain forms underneath the aircraft moves
farther forward. Consequently, the fountain flow will angle forward toward the
inlets, increasing the likelihood of encountering higher HGI levels for a given con-
dition. With the ground sloping away from the nose of the aircraft (downhill), the
fountain will also be angled farther forward toward the inlets. This increases the
likelihood of encountering high HGI levels for a given condition.
Subscale wind tunnel testing [10] was used to create an HGI database to
use in six-degrees-of-freedom flight simulation tools. With these, we performed
sensitivity analyses under piloted conditions and developed flight test HGI limit-
ations for VL envelope expansion, allowing VL flight testing to commence.
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 663

A. SHORE-BASED TESTING
Accurately determining the ambient wind conditions in which an aircraft operates
is critical to comparing actual and predicted flight test data. In the JB regime,
determining the wind conditions in which the aircraft operates is challenging.
At JB speeds, the aircraft air data system is ineffective for airspeed determination.
Other sources of environmental winds (e.g., control towers, portable wind anem-
ometers) are usually located significantly far from the testing location; however, if
these other sources were near the aircraft, they could be influenced by the environ-
ment the aircraft propulsion system creates. That would lead to erroneous indi-
cations of winds acting upon the aircraft. Therefore, an alternative method for
determining wind magnitude and direction was developed.
For the aircraft to remain stationary in JB flight, relative to the ground in a
headwind condition, the propulsive jets must be vectored aft to counter the pre-
vailing wind. Similarly, for the aircraft to remain stationary in a crosswind con-
dition, the aircraft must be banked into the wind. This enables the propulsive
jets to counter the prevailing wind. Using pretest models, correlations were
derived between the headwind component (relative to the aircraft) and the
average longitudinal angle of the exhaust plumes. A similar correlation was devel-
oped for the crosswind component and the acceleration in the aircraft lateral axis.
The correlations were empirically adjusted following the hover pit test and as
flight testing progressed. During flight testing in the JB regime, the winds
derived from these correlations were regarded as the truth source of winds
acting on the aircraft. These allowed for more accurate comparisons to be made
between flight test data and prediction.

1. HOVER CAPABILITY
Flight test maneuvers used for validating hover performance typically have tighter
tolerances and longer dwell times than those for other disciplines. Each hover
required the pilot to establish hover conditions as close to target fuel states as
possible, then remain effectively hands-free for at least 30 seconds. The dwell
time provided for any transients or feedback loops in either the aircraft or propul-
sion control systems to dampen out. This provided data with lower levels of
scatter (uncertainty). In some cases, the ambient wind was sufficient for a
target condition; for others, the pilot had to create a wind relative to the aircraft.
The pilot did this by flying the aircraft along a particular ground track into or
away from the ambient wind to achieve the desired test conditions. Any gustiness
in ambient winds increased the scatter in hover performance results due to aircraft
and control system responses to those gusts.
Aircraft gross-weight fidelity in hover was improved by weighing the aircraft
postflight. Following a performance hover test point, the engine was shut down as
soon as practical after completing the test point. The aircraft was then weighed,
and those data were combined with the integration of measured fuel flow. The
664 D. G. PARSONS ET AL.

data set determined the amount of fuel consumed for engine run-time, which we
then used to back-calculate the weight of the aircraft at the beginning of the
test point.
Dedicated performance flight test hover data were standardized to a nominal
set of reference conditions to provide a meaningful comparison with pretest pre-
dictions. Because hover maneuvers are conducted at various aircraft mass proper-
ties and atmospheric conditions, the data must be adjusted to a set of common
reference conditions. Only in this way can we understand the results in a more
holistic manner than single point by single point. Analytical adjustments are
applied to each measured test point to account for the predicted differences
between the test operating conditions and the desired reference conditions.
Figure 19 illustrates the results of this effort for hover performance testing. Test
aircraft weight is plotted vs propulsion system thrust, with the predicted
thrust-vs-weight characteristic shown as the green line. Combining a robust
pretest SJE database, accurate inflight thrust, and rigorous determinations of
test aircraft weight, we resolved aircraft hover capability to within approximately

Fig. 19 Thrust-vs-weight prediction accuracy.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 665

200 lb. This provided credence to the validity of the JB SJE database. With the
hover performance database validated, maximum hover capability for the VLBB
calculation was possible.

2. VERTICAL LANDING
VL flight test began by performing landings in the portion of the flight envelope
where HGI was not predicted to be a factor. Once a basic HGI-free understanding
of the performance of the aircraft during VL had been established, landings were
performed in the region of the flight envelope where HGI was predicted.
During the flight test of VLs, engine and LiftFan inlet temperature rake instru-
mentation was used to measure temperature rises and temperature distortion
levels due to HGI. Performance margin limitations were modified as flight
testing progressed based on results observed. Temperature rake instrumentation
data from VL testing were used to validate the HGI database. These provided
insight into the aircraft control system’s response to HGI.
Flight test – measured levels of HGI generally matched predictions from sub-
scale tests. Low levels of HGI were measured in headwind conditions, relative to
the aircraft, that were relevant to the VLBB KPP criteria. Therefore, mitigating
HGI with performance margin was not pertinent to the VLBB requirement.
However, as predicted by subscale test, crosswind and tailwind conditions relative
to the aircraft produced moderate to high levels of HGI. As such, they required
limitations on aircraft weight for the VL to protect against high-sink-rate landings
and/or undesirable handling qualities. These margins are discussed in Section V.D.
A VL can have a preceding In-Ground Effect (IGE) hover, which is a stabilized
hover at a low height above ground that is influenced by ground proximity. The
VL with this is not, however, a standard maneuver for shore-based operations.
Shore-based VLs are typically initiated from heights above ground that are signifi-
cantly above the IGE/OGE boundary. An IGE hover occurring prior to a VL can
happen during shipboard operations and, potentially, during shore-based
workups ahead of aircraft carrier operations (field carrier landing practice). The
typical hover height over an aircraft carrier deck is low enough such that the air-
craft is considered IGE. Therefore, several shore-based IGE hovers were per-
formed prior to ship testing at various wind conditions to ascertain any HGI
impacts from the dwell IGE. Results from this testing, as well as nominal VL
testing, were used to define the initial limitations for ship testing.

3. PATTERN FUEL
Validating the fuel required to complete the VLBB pattern (and set the hover
weight requirement for VLBB) involved performing trim maneuvers. These
were executed in the SJ regime at speeds consistent with pattern speeds. The
team approached this regime with a plan similar to that for the JB regime. We exe-
cuted constant-altitude, constant-speed trim maneuvers and exercised inflight
666 D. G. PARSONS ET AL.

thrust methods to validate the contributions of conventional aerodynamics, SJE,

and propulsive forces.
Unfortunately, the SJ regime proved to be far more difficult in terms of isolat-
ing the individual forces. The higher forward speeds of SJ flight, relative to JB
flight, moved the operating conditions of the propulsion system farther from
the cases where the inflight thrust tool was calibrated. As a result, the team was
unable to utilize inflight thrust in SJ for the primary validation approach.
Even though contributions from individual forces could not be isolated, the
team successfully validated the end performance of the model. Engineering judg-
ment was used to adjust various elements of the SJ databases to drive predictions
from the model to be consistent with measured flight test results. Figure 20 shows
fuel flow vs calibrated airspeed for SJ trim maneuvers. The flight test data have
been adjusted to a nominal weight and tropical day conditions. In this way, we
evaluated multiple maneuvers at once to identify any potential trends for devi-
ation from predictions. The flight-test-adjusted predictions are shown as the
blue line, and the flight test data are the orange circles. The model matches the
test data well across the SJ speed range. Most of the SJ pattern for the requirement
was conducted in the 150 to 180 kt True Airspeed (KTAS) range, where fuel flow
is minimized.
All validation work for the pattern fuel was accomplished with shore-based
testing. There is nothing unique about operations at the ship that required
further validation during ship-based testing. The fuel required to fly the VLBB
specification pattern was calculated using the SJ flight-test-validated databases.

B. SHIP-BASED TESTING
As with the STO, confidence gained from shore-based testing led to successful
initial recoveries at the ship.
Figure 21 shows an early VL on
the USS Wasp during DT-1.
Three ship trial evolutions were
completed to confirm the deliv-
ered capability in the more chal-
lenging shipborne environment.

1. HOVER CAPABILITY
For hover alongside the ship
prior to crossing the deck for
the VL, as seen in Fig. 22, the
data from shore-based testing in

Fig. 20 SJ fuel flow predictions.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 667

Fig. 21 F-35B landing aboard the USS Wasp (LHD 1).

OGE conditions apply directly. However, once the aircraft crosses over the deck of
the ship, the aircraft immediately enters a near-IGE condition. There, the hover
can be more challenging than it is at OGE heights. SJE forces (suck-down)
increase as the height above ground decreases. Also, the winds the aircraft experi-
ences above the deck of the ship can be vastly different from the winds experi-
enced in hover over the water next to the ship. This is because the ambient
winds are affected by the presence of the ship and the ship superstructure. The
winds experienced by the aircraft above the deck are offset from the ambient
wind magnitude and direction, and can become turbulent and inconsistent.
This phenomenon may also alter the HGI characteristics from those seen
during shore-based testing. In such an instance, it would affect the maximum
gross weight from which a VL can be initiated. Figure 23 provides two cases of

Fig. 22 OGE hover abeam the USS Wasp (LHD 1).

668 D. G. PARSONS ET AL.

Fig. 23 CFD visualization of the wind flow field around the ship.

Computational Fluid Dynamics (CFD) analyses completed to visualize the flow

field around the carrier. The figure also illustrates the potentially turbulent
nature of winds around the ship.
The variations in hover capability with wind magnitude, direction, and deck
landing spot were evaluated during the three F-35B ship trials. Although not
directly applicable to the VLBB specification calculation, these data are necessary
to determine the capability provided to the operator via shipboard operating bul-
letins and mission planning products.

2. VERTICAL LANDING
Testing for VL at the ship began with nominal winds down the deck of the ship,
where the expectation for variances in wind from ambient were minimized. The
wind envelope was then progressively expanded in both magnitude and direction.
VL testing at the ship included landing at different designated landing spot
locations. It also had high hovers over the landing spots with winds from the
direction of the island superstructure to assess Ship-exhaust Gas Ingestion
(SGI) levels. In addition, VL testing involved landing with another aircraft oper-
ating on the deck in front of the landing spot to assess the HGI levels created by
the on-deck aircraft. Supplementing this, VL testing examined landing with the
aircraft facing aft down the deck, and landing with the aircraft facing cross
deck, perpendicular to the ship.
From the results of the ship trials testing, no significant effects of the island
superstructure or other ship effects exist for wind notionally down the deck. Per-
formance results from shore-based testing are directly applicable to operations at
the ship within this envelope. Outside this envelope, the performance margin is
required for mitigating HGI, SGI, and other effects resulting from the ship
environment. Section V.D describes this in greater detail.

C. VERTICAL LANDING BRING-BACK PERFORMANCE VERIFICATION RESULTS

The validation of the models used to calculate VLBB analytically was completed
successfully. The use of inflight thrust methodology was critical for validating
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 669

aircraft SJE for JB flight. The effort culminated in the resolution of aircraft hover
capability to within approximately 200 lb. VL flight testing verified that HGI is not
relevant to the VLBB KPP wind envelope and is generally considered only for
operational limitations outside the KPP envelope. Test results showed that the
pretest models for hover and VL were accurate and did not need adjustment
for the VLBB KPP calculation. The fuel required to execute the specification
pattern around the ship in SJ conditions was also validated successfully. This
was done in a top-level manner rather than isolating the contributions of individ-
ual forces. Adjustments to the model for SJ conditions had a very minor effect on
the fuel required to perform the pattern. Ultimately, validated VLBB performance
exceeded the requirement by more than 7%, as illustrated in Fig. 24.

D. OPERATIONAL CONSIDERATIONS
Outside the VLBB KPP envelope, the operator still has significant performance
capability, even though it can be less than the VLBB due to multiple factors.
Testing during the System Development and Demonstration program validated
the performance capability and highlighted some problems that can affect per-
formance levels provided to the operator.

1. HOVER CAPABILITY
The positioning task in the hover prior to a VL requires the aircraft to be capable
of translating forward, sideways, and rearward as the pilot maneuvers the aircraft.
The maximum groundspeed for rearward translations is protected by the CLAW,
preventing rearward acceleration at the groundspeed limit. The maximum tail-
wind relative to the aircraft is a pilot-observed limitation. During flying qualities
testing to the maximum tailwind limit, ETR consistently increased with increasing
tailwind, often to 100%. It was frequently accompanied by uncommanded sink on
the order of 10 to 20 ft. The observed behavior was driven by TS moving forward
significantly while accelerating into a rearward translation. The predominant
cause is intake momentum drag creating a nose-down pitching moment with tail-
wind conditions. This prompts the flight control system to respond to the nose-
down moment by adjusting TS forward. Conversely, intake momentum drag
causes a nose-up pitching moment with a headwind. The tailwind condition is
depicted in Fig. 25. The individual intake forces (the LiftFan and the auxiliary
and main engines) from the intake momentum drag act forward of the aircraft

Fig. 24 VLBB verification results.

670 D. G. PARSONS ET AL.

Fig. 25 Intake momentum drag-induced pitching moment.

FSCG. There, they are close to the direction of the freestream creating the nose-
down pitching moment.
The forward TS reduces total available system thrust, and the LiftFan nozzle
angle will point well forward of the vertical if necessary to achieve rearward accel-
eration. When combined, these effects reduce the available vertical thrust such
that ETR saturates and the aircraft starts to sink.
An aircraft CLAW change mitigated the problem by constraining the LiftFan
to less forward nozzle angles when uncommanded sink was detected. This change
reduced the rate at which the pilot could accelerate rearward but did not affect the
maximum rearward speed that could be achieved. The net result was an improve-
ment in the handling qualities for this area of the flight envelope.

2. VERTICAL LANDING DESCENT RATE

Areas of the operationally cleared wind envelope can experience high levels of
HGI, predominantly in crosswind and tailwind conditions relative to the aircraft.
Figure 26 shows an example grouping of flight test landings in tailwind conditions
relative to the aircraft. Each flight test VL is indicated by a different colored
symbol on the wind rosette on the left in Fig. 26. The same symbols are used in
the plot on the right in the figure to show temperature rise as a function of
height above ground level. The flight test average is indicated by the solid black
line, aligning well with the average from subscale testing (shown by the blue line).
Flight test experience confirmed, as expected, that the aircraft descent rate
increases over the commanded descent rate during VLs in the presence of signifi-
cant HGI on warmer days. In the JB regime, the propulsion control system
is designed to produce a maximum specified level of thrust. This level is indepen-
dent of ambient temperature up to a high-temperature theta-break threshold.
Once the sensed inlet temperature rises above the theta-break threshold,
maximum thrust is reduced. Operations with a sensed inlet temperature above
the theta break require an ETR rise to maintain the same level of thrust, as
shown in the upper-left plot of Fig. 27.
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 671

Fig. 26 Example flight test results for winds and associated temperature rises.

The aircraft also experiences a pitch derotation during high-HGI events, and
TS is commanded forward to counter the pitch-down moment. Total available
thrust is reduced as TS moves forward due to propulsion system tent curve

Fig. 27 Theta break and tent curve effects on aircraft in the presence of HGI.
672 D. G. PARSONS ET AL.

effects. This also leads to a rise in ETR for the same total thrust, as depicted in the
upper-right plot of Fig. 27. Once ETR saturates at 100%, any further rise in inlet
temperature or decrease in TS will result in a reduction in thrust and an uncom-
manded increase in descent rate.
Depictions of the theta break and TS-vs-thrust schedules are provided in
Fig. 27. There is also an example of their effect on overall system thrust and the
aircraft descent rate during a VL in a tailwind condition on a warmer ambient
temperature day.
The lower plot in Fig. 27 shows an example VL time history illustrating the
effect of HGI on the theta break and TS-vs-thrust schedules. It also depicts the
ultimate impact on the aircraft descent rate. The time history begins just before
encountering HGI (point A in the in plot) and ends at VL touchdown (point
D). Between points A and B in the plot, the sensed inlet temperature (solid
green line) is below or near the theta-break threshold (dashed green line). TS
(orange line) is relatively constant at a nominal value, and the resultant system
total thrust (black line) and ETR (red line) remain relatively constant. The aircraft
descent rate (solid blue line) meets the descent rate command (dashed blue line)
as well. Between points B and C in the plot, the sensed inlet temperature starts to
rise significantly above the theta-break threshold, resulting in an increased ETR
for the same total thrust. Between points C and D in the plot, ETR has saturated
at 100% and the inlet temperature continues to rise. TS is also commanded
forward, resulting in a reduction in total system thrust and an increase in aircraft
descent rate above command.
There was an uncommanded decrease in the descent rate during VLs experi-
encing HGI at colder ambient temperatures. Although this is an unpredicted
phenomenon for VL, it is commonly seen during flight test. The mechanism
for this is similar to that for the derotation characteristic. On a colder day, a
VL with HGI results in sensed inlet temperatures remaining below the theta
break. In such instances, the propulsion system attempts to compensate for the
HGI to produce the same level of thrust as was present before the HGI event.
However, the engine control generally over-reads the inlet temperature during
an HGI event. Consequently, the propulsion system momentarily produces
more thrust than necessary for the condition, resulting in a decrease in descent
rate. Inevitably, the HGI environment is difficult to characterize for all possible
scenarios with any sensor mechanism short of a full 40-probe inlet rake. As
such, some conservatism in control methodology is prudent. The CLAW responds
to the reduction in descent rate by decreasing commanded ETR. However, the air-
craft experiences HGI at a low enough height that it generally lands before the
descent rate increases back to command.
The aircraft descent rate sets the operational performance margin required to
mitigate HGI during VL. If the desire were to ensure that HGI would have no
noticeable impact on the aircraft descent rate, the resultant limitations would
be unnecessarily restrictive. Rather than trying to eliminate the effect of HGI,
the effect can be controlled via performance limitations to ensure that the
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 673

Fig. 28 Example HGI-region limitation.

uncommanded increase in descent rate by touchdown is not excessive. Uncom-

manded descent rate thresholds are set by handling qualities assessments and
landing-gear or aircraft load limits.
Within the wind envelope where HGI is expected, the performance margin for
the mitigation of HGI is a function of both aircraft weight and ambient tempera-
ture. Additional margin is required, however, for warmer ambient temperatures,
as depicted in Fig. 28. This plot also shows the effect of the theta-break thrust
schedule within the no-HGI region, as previously discussed.

3. VERTICAL LANDING DEROTATION

VLs with high HGI levels exhibit a pitch-down characteristic just prior to touch-
down. HGI levels measured in flight testing for these conditions match those in
the subscale prediction; however, the magnitude of the aircraft pitch response
was not predicted by models. Posttest analysis indicated that engine inlet tempera-
ture sensor lag and HGI compensation logic in the control system were the
primary factors contributing to this previously unmodelled characteristic.
During VL in the presence of HGI, the exhaust fountain is circulated into
the main engine inlets. It is then directed primarily to the lower portion of the
engine inlet face. The engine control receives inlet temperature feedback from
two sensors located in the lower portion of the engine inlet face. Therefore, it
senses a higher temperature than the average over the engine face during a VL
in the presence of HGI. An example is shown in Fig. 29 for a flight test VL
with a tailwind relative to the aircraft. In the figure, the T2 sensors measure
inlet temperature.
In Fig. 29, the plot on the right shows engine inlet temperature rise vs height
above ground for the average face value measured by the inlet rake (blue line). It
also indicates the average T2 location value measured by the inlet temperature
674 D. G. PARSONS ET AL.

rake (red line) and the value used by the control system (green line) based on the
inlet temperature sensors. The contour plot on the left shows the inlet rake
measured temperature over the engine inlet face at the height of the peak
average engine face temperature rise. The engine inlet temperature sensor
locations are represented by the black circles in the lower half of the engine
face. HGI compensation logic within the control system is designed to account
for sensor-placement-induced errors resulting from the temperature distortion.
It also factors in the attempts to correct the inlet temperature values used in
the control for this discrepancy. However, the corrected temperatures are still
warmer than the average over the engine face, as measured by the flight test
inlet rake instrumentation.
The temperature difference between the average engine inlet face and that
sensed by the control system means that core engine thrust is momentarily
higher than the control system calculates. This results in a brief nose-down pitch-
ing moment. The aircraft CLAW responds to the pitch attitude change by
commanding a nose-up TS to counter the pitching moment. However, the dero-
tation pitch rate is low enough that the control loop does not act immediately. An
example derotation time history during a flight test VL with a tailwind relative to
the aircraft is shown in Fig. 30, which is the same VL shown in Fig. 29.
In Fig. 30, the time history starts just prior to HGI being encountered and ends
at landing touchdown. The aircraft pitch attitude (purple line) is at a nominal
value until the control system inlet temperature (green line) starts to diverge
from the rake average inlet face temperature (blue line). At this point, the aircraft
starts to pitch down and TS (orange line) is commanded forward.
This derotation characteristic was consistently observed during VLs and
appears to be proportionate to the level of HGI encountered. The pitch attitude
can decrease by as much as 2 deg before aircraft touchdown. These changes are

Fig. 29 Engine face distortion impact on sensed temperature rise.

F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 675

Fig. 30 Derotation illustration.

perceptible to the pilot, but no pilot response is possible because the pilot does not
have direct control over pitch attitude during a VL.
A CLAW change was implemented to increase the aircraft pitch attitude
during VL descent by 2 deg. The updated VL pitch attitude was intended to
prevent nose landing gear touchdown prior to the main landing gear. It also
aimed to avoid any load concern accompanying such a maneuver.

4. SHIP OPERATIONS
Ship VL operations with wind down the deck (similar to the VLBB relevant wind
envelope) were generally unremarkable. They showed no requirement for
additional performance margin over shore-based VLs performed with similar
wind conditions. Outside this benign wind envelope, multiple features required
additional performance margin for mitigation.
Figure 31 shows the USS America and its superstructure. With starboard
winds from the direction of the ship superstructure, increased thrust is required
to mitigate the effects of SGI. It also counters amplified HGI due to the reflection
of the aircraft exhaust ground sheet from the ship’s island. Further, it mitigates
the downwash effect of the relative winds from the presence of the island
superstructure.
The ship’s engines release exhaust from stacks on the island superstructure.
With an ambient wind relative to the ship that directs the exhaust over the
intended landing spot, this exhaust can be ingested by the aircraft. To quantify
a worst-case scenario for SGI, we performed high hovers over the intended
landing spot. We used winds that placed the aircraft within the ship’s exhaust
plume and measured the temperature rise at the LiftFan and engine inlet faces
using inlet rake instrumentation. Figure 32 shows an F-35B hovering at a
nominal height above deck prior to VL, with an illustration (in red) of the
676 D. G. PARSONS ET AL.

Fig. 31 USS America (LHA 6).

ship’s exhaust. High hovers were performed at approximately 20 ft higher than the
nominal height to gather data in a worst-case scenario for SGI.
HGI and SGI only require mitigation at warmer ambient temperatures where
their effect reduces available system thrust above the temperature of the theta
break. This can result in a reduced hover-over-deck capability and/or increased
descent rate during VL. However, the effect of downwash on the aircraft results

Fig. 32 F-35B and stack gas ingestion for the USS America (LHA 6).
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 677

Fig. 33 Overhead view of the layout of the USS America (LHA 6).

in a greater required thrust to hover over the deck and, hence, additional perform-
ance margin. This effect is independent of ambient temperature. The downwash
effects from the island superstructure were clearly seen when operating at the
landing spot adjacent to the island superstructure, labelled in Fig. 33.
An F-35B at this landing spot adjacent to the island superstructure is shown in
Fig. 34. During one recovery to this landing spot and immediately after the deck
cross, the aircraft experienced a significant increase in ETR, as depicted in Fig. 35.
The saturation of ETR resulted in an uncommanded descent prior to the VL
initiation. The increase in ETR was attributed to an aerodynamic effect from
landing close to the island superstructure, and no significant HGI/SGI was

Fig. 34 F-35B at landing spot adjacent to the ship superstructure on the USS America (LHA 6).
678 D. G. PARSONS ET AL.

Fig. 35 ETR time history for VL to spot adjacent ship’s superstructure.

experienced at the time. The VL was subsequently completed without issue, but
the landing highlighted the need to account for downwash on the aircraft.
With port winds, SGI is no longer a concern, and at most landing spots the
island superstructure no longer influences the winds that the aircraft experiences.
However, influences of the ship operating environment on both the level of HGI
experienced and downwash on the aircraft are still seen and require performance
margin to mitigate. With winds from port, the proximity of the aircraft to deck
edge and the near vertical side of the ship are expected to contribute to a more
adverse flow field. This applies to both HGI and aerodynamic effects, and is eval-
uated through a comparison with similar wind conditions during shore-based
operations.

E. PILOT PERSPECTIVE
STOVL JB and VL testing included performance trims and VLs across the func-
tional range of altitudes and wind conditions at various FSCGs and hover-weight
ratios. (The ratios are calculated by dividing aircraft gross weight by the maximum
hover weight capability.) The trims and VLs provided unexpected challenges from
the pilot’s perspective. These were a function of nuances in the integration of flight
and propulsion controls and engine cutbacks due to the wear of engine com-
ponents. The effects were also the product of the integration of our testing into
a busy airfield at extremely low fuel states.
In the first case, the system is designed to enter JB mode with a 3% thrust
margin. In this way, the CLAW will suspend the deceleration from SJ to JB if
that margin cannot be maintained. The CLAW feature, a performance deficit pro-
tection, is a welcome capability that will prevent entry into JB without sufficient
performance. However, it added challenges to performance testing that required
JB entry at the highest hover-weight ratio possible. Tolerances were tight and
translated to approximately one minute to get from SJ flight to a stabilized
hover, providing little room for error or unexpected events.
F-35 STOVL PERFORMANCE REQUIREMENTS VERIFICATION 679

The importance of working closely with airfield management, Air Traffic

Control (ATC), and tower controllers cannot be overstated. JB performance
testing frequently put the aircraft in a very low fuel state (approximately two
minutes of flight time) with one landing option. Orchestrating the arrival of the
aircraft into the complex traffic pattern at Naval Air Station Patuxent River
took detailed coordination and communication. Face-to-face discussions, joint
simulator sessions with the ATC team, and pre- and postmission briefs were
essential elements to the success of this process.
Although JB performance testing presented significant operational challenges,
performance VLs were generally unremarkable. The performance VL technique
did not differ from that of normal VLs. HGI was encountered on numerous
occasions, usually noticeable in the cockpit as a sudden increase in height rate
in the last 10 to 20 ft of the descent. Notwithstanding, no pilot response was
required or desired.

VI. CONCLUSION
The F-35B STOVL flight test program built upon a solid foundation of preflight
modeling and simulation evidence that enabled efficient envelope expansion and
posttest results verification. Testing was accomplished over more than seven years,
demonstrating that the aircraft provides repeatable performance with an intuitive
pilot interface. The interface allowed testing at both minimum fuel states and
maximum gross weight capability to feel routine.
The verification of the KPPs of flat-deck STO and VLBB was relatively
straightforward. Analyses of data acquired from both shore- and ship-based
maneuvers proved that the F-35B exceeded the flat-deck STO requirement.
Moreover, it had considerable margin that may be used to accommodate
future weapons system growth potential. JB maneuvers (hovers and VL) and
SJ maneuvers were utilized to validate the databases used to calculate VLBB
performance. Ultimately, these verified that the F-35B design exceeded the
requirement.
Verification testing for the flat-deck STO and VLBB requirements was suc-
cessful. Accordingly, the F-35B flight test program focused on validating the
STOVL capability provided to the operator throughout the operating envelope.
The unexpected problems the team overcame increased the value of the flight
test effort and contributed to making the F-35B more robust. Changes in aircraft
CLAW to accommodate unpredictable wind conditions during a STO will help
the fleet to avoid nozzle strikes. It will also enhance the fleet’s understanding of
the effects of HGI on VL performance, which will protect the fleet against the
possibility of potentially dangerous touchdown sink rates.
The team planned extensively to conduct safe, efficient, and effective testing to
achieve program objectives. Expected problems regarding sink off bow character-
ization in support of ship-based STOs and HGI concerns with respect to VL
680 D. G. PARSONS ET AL.

performance validation were quantified. They were then effectively mitigated

through detailed test hazard analyses and the implementation of a cautious
buildup approach. Particularly important was the development and implemen-
tation of a recovery strategy in the event of excessive sink off bow.
The end product of the flight test process discussed herein was the capability
provided to the operator of the F-35B weapons system. Innovations in flight test
and analysis methods contributed to successful STOVL requirements verification
and increased pilot confidence in the aircraft’s robust performance in this flight
regime. The positive performance test results demonstrate that the F-35B provides
a considerable upgrade in range, payload, and bring-back capability over the
legacy Harrier. Along with this, it retains supersonic speed and 5th-Generation
stealth characteristics.

REFERENCES
[1] Wurth, S. P., Smith, M. S., and Celeberti, L., “F-35 Propulsion System Integration,
Development & Verification,” AIAA Aviation Forum, June 2018 (to be published).
[2] Parsons, D. G., Eckstein, A. G., and Azevedo, J. J., “F-35 Aerodynamic Performance
Verification,” AIAA Aviation Forum, June 2018 (to be published).
[3] Mason, J. R., “F-35 STOVL Aircraft Performance Model Validation of Short
Take-off Flight Test Data,” AIAA Aviation Forum, Aug. 2013.
[4] Wurth, S. P., Walker, G. W., and Fuller, J., “F-35B IFPC Development,” AIAA Paper
2013-4243, Aug. 2013.
[5] Mange, R. L., and Hoggarth, R., “Highlights of the Lockheed Martin F-35 STOVL Jet
Effects Program,” Royal Aeronautical Society, International Powered Lift
Conference, 2008.
[6] Hoggarth, R., and Stovall, M. B., “F-35 STOVL Mode Flight Test Analysis
Techniques and Initial Results,” International Powered Lift Conference, Oct. 2010.
[7] Vorwerk, A. V., and Ciszek, R. S., “Use of In-Flight Thrust on JSF Program,”
American Helicopter Society International, International Powered Lift Conference,
Oct. 2010.
[8] Hoggarth, R., and Stovall, M. B., “The Lockheed Martin F-35 STOVL Force and
Moment Ground Test Analysis Techniques and Results,” International Powered Lift
Conference, Oct. 2010.
[9] Pinero, E., “Utilization of the STOVL Operations Test Facility During the F-35
Program,” IPLC Paper IPLC2010_30, Sept. 2010.
[10] Cook, R., Curtis, P., and Fenton, P., “State of the Art in Sub-scale STOVL Hot Gas
Ingestion Wind Tunnel Test Techniques,” SAE International 2005-01-3158.
 CHAPTER 18

F-35 Climatic Chamber Testing

and System Verification
Victorio J. Rodriguez and Billie Flynn†
Lockheed Martin Aeronautics Company, Fort Worth, TX
Marc G. Thompson‡
BAE Systems, Warton, United Kingdom

Steven Brelage§
Naval Air Warfare Center Aircraft Division, Naval Air Station
Patuxent River, MD

The F-35 is expected to withstand the most formidable of weather

extremes over its service lifetime. Validating this capability with three
unique F-35 variants, most notably the Short Takeoff and Vertical
Landing (STOVL) version, was expected to be extremely challenging.
The F-35 program was required to create a test environment where a
pilot could essentially hover an F-35 indoors, operating at maximum
power through extreme weather conditions. To do this, the test team
had to stretch ingenuity and imagination to develop never before seen
structures and test procedures. The result was a one-of-a-kind test
program, housed inside the McKinley Climatic Laboratory (MCL) at
Eglin Air Force Base, Niceville, Florida, to prove the F-35 could operate
in those climates.

I. INTRODUCTION
The F-35 System Development and Demonstration (SDD) phase included the
development and maturation of the aircraft design and configuration as well as
the execution of a large and dedicated ground and flight test evaluation
program. A major subportion of the SDD test program—and the focus of this
paper—was the F-35 Climatic Chamber Testing (CCT), executed in a six-month
period from October 2014 through March 2015. The CCT was an on-aircraft test
executed at a wide range of environmental conditions to gather data on aircraft

 Manager, F-35 Propulsion and Thermal Analysis.

†
Senior Experimental Test Pilot, F-35 Flight Test Program.
‡
Principal Engineering Specialist, F-35 Propulsion Flight Test.
§
Branch Head, Propulsion and Subsystems Flight Test.

681
682 V. J. RODRIGUEZ ET AL.

operation. Historically, this phase of testing is conducted early enough in the life
of an airplane to ensure that any issues that are discovered can be resolved before
the airplane is fielded. This paper will discuss the unique attributes of the F-35
program that challenged traditional testing methods and the difficult decisions
required to define, develop, and schedule the testing. The paper will also
discuss unique and key capabilities and enablers that allowed successful execution.
Finally, the paper will conclude with a high-level review of lessons learned and the
relevance of performing complex and expensive full-scale, air-system-level,
environmental testing in a world of increased simulation and modeling capabili-
ties and pressure to keep cost and schedule threats to a minimum.

II. F-35 VS PREVIOUS AIRCRAFT CLIMATIC TESTING

The program-level, functional requirements for the F-35 air system are defined
in the Joint Strike Fighter (JSF) Contract Specification (JCS). The JCS defines all
aspects of the aircraft-level requirements, including the capability to be parked
outside of a hangar and operation in a multitude of extreme environments that
naturally occur throughout the world. The environmental requirements are
based on the guidelines for design established in the MIL-HNBK-310,
“Global Climatic Data for Developing Military Products.” The requirements
were tailored early in the program development for the anticipated F-35 aircraft
customers and expected operational usage. These requirements served as the
basis for the type and extent of testing required for the F-35 SDD program
and included a range of ambient temperatures, humidity levels, rainfall, icing,
and other environmental conditions. Nearly all new major military aircraft pro-
grams include some sort of climatic environment test and evaluation require-
ment. The testing to satisfy this requirement is traditionally performed at a
dedicated facility, performed in naturally occurring environmental conditions,
or a combination of both approaches. In legacy aircraft programs, test
methods (including natural weather exposure) and facilities often yielded
unpredictable, unrepeatable, and unreliable test conditions and inconclusive
results that would not have satisfied the rigorous performance specifications
demanded for the F-35. Figure 1 shows examples of the lines of testing per-
formed during the F-35 CCT.
For most aircraft programs there is one design configuration to address and
evaluate. The F-35 program, however, consists of an aircraft built in three variants:
Conventional Takeoff and Landing (CTOL, F-35A), Short Takeoff and Vertical
Landing (STOVL, F-35B), and a Carrier Variant (CV, F-35C). The three variants
have significant commonality in mission systems and vehicle systems, and simi-
larities in airframe design despite unique service mission requirements and
concept of operations. Development of a “one size fits all” environmental test
approach was required to plan a test and evaluation effort that maximized effec-
tivity across all three variants and eliminated the need for multiple variant testing.
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 683

Fig. 1 F-35 Climatic Chamber Testing examples.

This led to a challenge to achieve a focused balance of requirements, priorities,

objectives, and tough programmatic decisions prior to proceeding to climatic
test execution. The unique aspects of the F-35B variant were the major driving
factor in the F-35-specific testing that included essentially hovering the aircraft
inside the facility.
Advances in computational methods available to the F-35 program, combined
with capability improvements to the climatic chamber facility, enabled elimin-
ation of the necessity to test in natural environments, which can be both costly
and inefficient, often waiting for appropriate weather conditions to be available
or located around the world. By slightly expanding the scope of targeted
regions of the climatic chamber testing, it was possible to eliminate some orig-
inally planned in-air flight testing, resulting in risk reductions to the flight test
program and significant program cost and schedule savings. However, the same
advances that enabled the usage of the chamber for testing also led to many dis-
cussions regarding the possible elimination of the entire line of testing based on
perceived fidelity of modeling and computational methods. Ultimately, the critical
importance of system-level demonstration of capabilities and data required for the
formal validation of models allowed the testing to proceed.
684 V. J. RODRIGUEZ ET AL.

III. MCKINLEY CLIMATIC LABORATORY TEST FACILITY

The F-35 program elected to use a dedicated and controlled environmental facility
to execute the primary system-level environmental testing. The McKinley Cli-
matic Laboratory was contracted via the JPO organization to perform the
testing. Owned by and operated for the U.S. Air Force (USAF), the MCL is argu-
ably the foremost climatic environmental testing venue in the world and is able to
simulate virtually any required environmental condition. Built in 1947 to provide
a refrigerated hangar to produce environmental extremes under controlled con-
ditions, the facility was renamed the McKinley Climatic Laboratory in 1972 in
honor of Lt. Col. Ashley C. McKinley, one of the initiators of the facility and a
major leader behind the requirement to have such a facility available to the U.S.
armed forces. Renovated and remodeled between 1993 and 1997, the facility
now supports both military and commercial product testing, ranging from low
to high ambient temperatures, icing cloud and ice buildup, rainfall, fog, humidity,
wind, sand, solar radiation and diurnal cycles, and other environments. It is a full-
service facility able to support almost any line of testing or product. The MCL
facility has an Air Make-Up Unit (AMU) system that conditions the air supply
to specified requirements and allows for extended engine operation within a
closed environment. The AMU is capable of supporting up to 1000 lb/second
airflow makeup at –658F in the main test chamber. A total of six chambers are
available for specialized evaluation of various systems. The facility, whose
motto is “Creating the weather you can’t afford to wait for,” was designated as
a National Historic Mechanical Engineering Landmark in 1987.
The bulk of the F-35 aircraft CCT program was executed in the main chamber,
which covers an area measuring 250  262  70 ft. Additional off-aircraft Logis-
tics Test and Evaluation (LT&E) and Support Equipment (SE) evaluation testing
was executed in the smaller equipment test chamber (ECT) (130  30  25 ft)
in parallel with the aircraft – level testing. Figure 2 shows an aerial photo of the
MCL facility taken upon the arrival of the F-35 test aircraft to Eglin AFB.
Figure 3 shows the test aircraft in front of the main test chamber of the MCL
facility.

IV. CHALLENGES AND DECISIONS IN DEFINING AND SCHEDULING

CLIMATIC TEST ACTIVITIES
The decision to execute an air vehicle system – level CCT was established early in
the SDD test planning effort with the testing scheduled to occur several years after
the formal start of the flight test program once system maturity and capability
were established. A program-level decision was made to use only one variant as
the means to capture and evaluate all three variant requirements (based on
intended commonality), with the F-35B variant selected because the addition of
the lift fan, roll post offtakes and nozzles, and three-bearing swivel module
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 685

Fig. 2 Arrival of F-35 test aircraft BF-5 at Eglin AFB over MCL.

Fig. 3 F-35 BF-5 positioned in front of the MCL main chamber.

686 V. J. RODRIGUEZ ET AL.

along with the main engine and the STOVL capabilities provided the most com-
plexity and incorporated the greatest number of operating doors and surfaces. As
a result, the F-35A and C variants would be cleared via similarity, and any unique
systems cleared by a combination of dedicated component-level testing and
system-level analyses. In 2013, the decision was made to execute the testing
using an interim Mission Systems (MS) and Vehicle Systems (VS) hardware
and software configuration that supported both the U.S. Marine Corps (USMC)
F-35B Initial Operating Capability (IOC) milestone for 2015 and remained appli-
cable to the planned final fleet configuration. The high-level overall test objective
shown as part of the pretest planning briefings and the CCT Test Readiness
Review (TRR) meeting held in August 2014 were:

. Evaluate system-level operational capabilities at a range of climatic

environments.
. Provide sufficient evaluation of air vehicle effectiveness ahead of major fleet
operations.
. Gather required supporting data for
W Specification criteria compliance
W Certification criteria
W Removal or deduction of initial operating limitations
. Update Joint Technical Data (JTD) and Flight Series Database (FSD)
information.

Full operation of the aircraft systems (to the extent possible) under each environ-
mental condition was a stated requirement from the outset. This included the
operation of the installed propulsion system from low through high power set-
tings. The program selection of the F-35B as the test variant was in no small
part to evaluate operation in powered lift mode (wherein the propulsion system
effectors are the prime lift and control mechanism) as well as conventional (tra-
ditional) mode of operations. As such, a key factor in the test planning was to
show the capability for the system operation in powered lift mode at all environ-
mental conditions, including hover and slow speed flight regime. Operations at all
power settings within an enclosed setting required sufficient air makeup required
to feed the propulsion system and to be able to expel the aircraft exhaust out of the
main chamber. This was essential for safety of personnel in the chamber as well as
to maintain the specific environmental conditions for extended periods of time.
As a direct result of this stated objective, the operation and restraint of the aircraft
in both operating modes at all power settings (through full afterburner in conven-
tional mode and hover in powered lift mode operations) was a major consider-
ation in the development of the test objectives, test sequences, and facility
requirements and modifications. In the powered lift mode of operations, the lift
fan and engine nozzles are vectored downward and forward to provide increasing
lift forces required to support the aircraft at slow airspeeds. This characteristic and
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 687

objective mandated a unique requirement to elevate the aircraft above the main
chamber floor to be able to accommodate the necessary exhaust ducting and air-
craft support structure to support the aircraft in all modes of operation.
During planning efforts, cost and schedule constraints forced the program to
review the lines of testing planned and traditionally performed in climatic evalu-
ation testing vs program objectives and priorities. The formal lines of testing that
were accepted and planned for the CCT included:
1. A standard day baseline assessment
2. Incremental testing to elevated ambient temperatures
3. A repeat of the standard day test sequence
4. Incremental testing to colder ambient temperatures
5. Ground and flight icing cloud conditions (including ground vortex icing)
6. Static and blowing rainfall conditions
7. Freezing rain conditions (nonoperating)
8. High relative and absolute humidity
9. A final repeat of standard day test sequence
Several originally planned lines of testing were ultimately eliminated because
either sufficient information was achievable via analysis or other means of data
gathering and evaluation, or they were constrained by other design and test
requirements. One example of this approach was the deletion of the snow
loading testing, where analysis could be used to determine structural loads
under a given snow accumulation. Additionally, by choosing to use and modify
an existing force-and-moment restraint system to secure the aircraft via the
landing gear (described in a later section), a planned sequence to swing the
landing gear at each environmental condition was also removed from the plans.
Though it would have been possible to design a system to restrain the aircraft
and allow full movement of the landing gear, this would have involved significant
engineering design and analysis efforts and extensive structural modifications to
the aircraft, and may have compromised other testing requirements. In the end,
this was a good example of the trade-offs that must be made when balancing
cost and schedule against testing requirements.

V. AIRCRAFT MODIFICATION, FACILITY, AND TEST PLAN PREPARATION

Initial and subsequent periodic planning was maintained early in the SDD
program to develop the high-level structure of the testing objectives and require-
ments; however, as the SDD phase progressed, design maturation along
with normal flight test discoveries combined to delay the scheduled entry of
the aircraft into the CCT. For example, improvements to the aircraft thermal
management system were identified early in the F-35 design program, and
688 V. J. RODRIGUEZ ET AL.

the changes were incorporated into the baseline aircraft design and were ulti-
mately part of the aircraft configuration that was tested during the CCT. As the
program schedule and design matured, a more solid test entry date was estab-
lished, and more frequent and focused test planning efforts began. Between
12 and 18 months ahead of planned test entry, a joint team of contractors—
Lockheed Martin, Northrop Grumman, BAE Systems, Pratt & Whitney, and
Rolls-Royce, and including JPO, F-35 Integrated Test Force, and MCL
personnel—met regularly to discuss and mature the test aircraft configuration,
planned objectives, detailed test sequences, detailed schedules, and required
documentation and reviews.
Each line of flight testing executed under the F-35 program is governed by a
Test Execution Package (TEP). The development of the formal TEP for the CCT
was initiated and managed by the F-35 ITF at Naval Air Station (NAS) Patuxent
River, Maryland, with support from the ITF at Edwards Air Force Base (AFB),
California. The ITF is a combined government and contractor team responsible
for the planning, execution, data analysis, and reporting of ground and flight
testing for the F-35 program.
The TEP consisted of the Joint Test Plan (JTP), which details the system under
test and the required sequences to be executed, and the Test Safety Supplement
(TSS), which details system maturity, risk areas, hazards, mitigating procedures,
and the overall risk assessment for the planned testing. It should be noted that
prior to and during this process, there were numerous concerns about the severity
of the proposed testing environments, and the potential for long-term damage to
the test aircraft asset as a result of the planned testing.
Attention to detail during this phase of planning, combined with the F-35’s
unparalleled insight into aircraft subsystem health during testing, allowed the suc-
cessful reintroduction of aircraft BF-5 into the test fleet post-CCT, including its
role as one of the prime assets during F-35B Development Test (DT)-III testing
aboard USS America (LHA-6) in 2016.
Once the TEP was matured, the package proceeded through two main board
review and approval cycles—the Test Readiness Review Board (TRRB) and Execu-
tive Review Board (ERB)—prior to final authorization to execute testing. In total,
the TEP was developed gradually over approximately nine months from a draft
document outline through final ERB approval. The Patuxent River F-35 ITF
Test Operations Team also led the development and planning of the joint detach-
ment team that would travel from NAS Patuxent River, supported by other sites
and organizations, to Eglin AFB to support the test activities. Initial test schedules
planned for a four-to-five-month test period between the aircraft ferry to the MCL
and the scheduled aircraft return to NAS Patuxent River.
Testing at each major environmental condition was structured to provide con-
sistent test flow and set expectations for the test team and MCL personnel. Each
test event consisted of specific periods defined as: aircraft soak and stabilization
period, prestart (maintenance activities), aircraft startup, simulated CTOL oper-
ations, simulated STOVL operations, and an aircraft shutdown period. Figure 4
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 689

Fig. 4 F-35 CCT test event sequence timeline (notional).

illustrates a generalized planned execution timeline for the testing at each

environmental condition.
A Standard Test Sequence (STS) was developed by the joint team for the test
execution. This sequence allowed for a routine flow of the testing at each con-
dition, helped better simulate standard ground operations procedures for the
fleet aircraft, reduced the overall number of test points, and provided an easier
means to compare and evaluate aircraft system performance across the environ-
mental spectrum.
The final STS profile included a normal aircraft startup procedure, a Vehicle
System Built-In Test (VSBIT), the execution of two primary Maintenance Built-In
Tests (MBITs)—designed to be used in the fleet in place of high power restrained
runs after propulsion system maintenance—and simulated conventional mode
takeoff maneuvers (high power). Simulated powered lift mode takeoff and land-
ings (that is, short takeoff [STO] and hover/vertical landing [H/VL]) were also
scheduled to be performed at each major condition to ensure proper aircraft
system operations. This STS was used for most of the testing sequences. For
some unique test objectives (e.g., icing evaluation, alert start maneuvers, rain
intrusion and corrosion inspections, and dedicated maintenance evaluations),
separate specific test sequences were added to the execution plan. Figure 5 illus-
trates the STS profile that was executed vs time.
Aircraft BF-5, one of the SDD test aircraft, was selected for the CCT. Although
BF-5 was a dedicated flight test aircraft, it incorporated many of the production
standard system configurations combined with extensive Flight Test Instrumenta-
tion (FTI) capabilities. The aircraft underwent a significant pretest modification
period at the F-35 Patuxent River ITF facility from June through August 2014
to ensure that the test article had the latest hardware and software packages avail-
able. Care was taken to incorporate a sufficient pedigree of components to
690 V. J. RODRIGUEZ ET AL.

Fig. 5 F-35 CCT STS profile.

adequately evaluate both the Block 2B configuration for the USMC IOC milestone
and the Block 3 configuration planned as the final standard for aircraft deliveries.
Additionally, dedicated and specific FTI measurement devices were added to the
aircraft to serve as primary data collection for validation, reference of environ-
mental conditions, and Safety of Test (SOT) parameters. Most of the modifi-
cations were accomplished prior to the aircraft ferry; however, several
modifications, including preproduction hardware, nonflight-worthy FTI, and
unique test asset installations were completed after aircraft arrival at the MCL.
In preparation for the planned testing at the MCL, the F-35 team coordinated
extensively with the MCL facility team to understand and define what setup,
modifications, and special equipment would be required to accommodate the
test sequences and objectives. A fixture was specifically designed by the MCL
team to accommodate the unique requirements for operation of the F-35B aircraft
in conventional mode and powered lift mode operations. To allow for full power
and articulation planned in both primary modes of operation, the aircraft had to
be elevated in the chamber (approximately 13 ft above the floor). The unique main
chamber configuration for the F-35 testing consisted of the aircraft restraint
system, the exhaust ducting system, the elevated maintenance platform, and air-
craft support equipment garages.
The restraint system was a modified version of a system previously designed,
built, and used for the aircraft force-and-moment system hover pit testing in Fort
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 691

Worth, Texas, in 2007. Reuse of this existing system eliminated the need to phys-
ically modify the aircraft for restraint during testing, and the load paths and forces
for operation imparted on the aircraft had already been analyzed and established
from the previous testing, which significantly reduced the associated cost of design
and acquisition. The backbone of this system was made up of six stanchions
designed to restrain the aircraft in all axes and for all power settings in both con-
ventional and powered lift mode operations. The six stanchions and attachment
hardware were transferred to the MCL team, and the equipment was modified
in length to incorporate the MCL-specific load cells and to attach to an I-beam
frame anchored to the main chamber floor. The asymmetric nature of the restraint
design was such that the system was statically determinant. Real-time monitoring
of the load cell values was performed for each run to assess measured load against
the designated continuation criteria for testing. Figure 6 illustrates a Computer-
Aided Drawing (CAD) model rendering of the aircraft on the modified restraint
system for use in the main chamber.
A requirement was also levied to allow sufficient work area and access to the
aircraft for pre- and posttest maintenance activities because the aircraft would be
significantly above the main chamber floor. This manifested itself in the form of
an elevated platform area around the aircraft that was made up of a grated surface,
supported from below via columns attached to the main chamber floor. Several
iterations were reviewed by all parties before a final agreement was reached
regarding the minimum required area. The platform was designed to be freestand-
ing and did not attach to the aircraft or the restraint system in any way and was
built from medium grade grating, sufficient for most aircraft maintenance activity.
The elevated platform was positioned at a height to closely resemble the normal

Fig. 6 F-35/MCL aircraft restraint system CAD model. (Illustration courtesy of Dwayne Bell,
McKinley Climatic Laboratory.)
692 V. J. RODRIGUEZ ET AL.

ground plane to the aircraft to accommodate the usage of standard F-35 aircraft
support equipment and to perform normal pre- and postrun activities. The area
under the platform was then used to route the facility-provided exhaust ducting
(see the following paragraph) as well as to provide two environmentally controlled
garage areas where common and large support equipment could be housed and
stationed during testing.
The operation of the aircraft and propulsion system inside the facility required
the aircraft exhaust to be routed out of the facility. This was critical to avoid any
issues with exhaust and combustion products buildup and to ensure that the oper-
ation of the aircraft propulsion system did not negatively impact the targeted
environmental test conditions. To that end, the MCL team designed and manu-
factured exhaust ducting configuration unique and specific for this F-35 test
installation. The design of the exhaust collection system was done in coordination
with the Lockheed Martin and Pratt & Whitney Integrated Product Teams (IPTs)
and was based in part upon the current systems in use at Pratt & Whitney test
facilities and from previous test experience by the MCL team. Several iterations
were studied before a final agreement and design was reached. Figure 7 illustrates
several pictures of the as-built exhaust ducting in place after aircraft installation in
the test fixture.
The ducting was designed to channel the exhaust away from the aircraft and
out of the chamber. It was a passive system where the flow moved along via
momentum of the exhaust itself and the pressure differential between the front
and rear openings in the ducts. All the major exhaust products from the aircraft
system were ducted out of the chamber, specifically the main engine, lift fan, roll
post, and Power and Thermal Management System (PTMS). There are no combus-
tion products and the temperature of the air is relatively cool compared to all the
other exhaust airflows, so the exhaust flow from the lift fan was handled slightly
differently than the other exhaust sources. The system was designed to allow
partial recirculation of the lift fan exhaust flow directly back into the main
chamber during operation. This feature relieved pressure on the main chamber
air makeup system, which would have been particularly stressed to support and
maintain the required conditioned airflow when simulating the F-35 hover
where both the main engine and lift fan are operating at maximum airflows. The
implementation scheme allowed improvements in testing efficiency because the
team was able to operate for longer durations before the AMU system needed
recharging. All exhaust ducting was designed to fit as close to the nozzles and
exit planes of the aircraft as possible to minimize induction of the conditioned
chamber air into the exhaust ducts, yet still allow for full aircraft operations in
both operating modes up to full power setting while remaining clear of all struc-
tural movement, operating doors, and major aircraft panel removal paths required
for routine maintenance. Figure 8 illustrates a CAD model rendering of the plat-
form surrounding the aircraft while installed on the restraint system and with
the exhaust ducting in place. Figure 9 illustrates the position of the aircraft and
platform with respect to the entire main chamber facility.
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 693

Fig. 7 F-35/MCL exhaust ducting (lift fan, roll nozzle, PTMS, and main engine) overview.

Fig. 8 Aircraft platform and aircraft installation CAD model illustration. (Illustration
courtesy of Dwayne Bell, McKinley Climatic Laboratory.)
694 V. J. RODRIGUEZ ET AL.

Fig. 9 CAD Model of F-35 aircraft within the MCL main chamber. (Illustration courtesy of
Dwayne Bell, McKinley Climatic Laboratory.)

VI. ENTRY INTO TEST FACILITY AND TEST EXECUTION

After the ferry and arrival of the aircraft to Eglin AFB, a brief aircraft preparation
period was completed to finish required aircraft modifications. On 29 September
2014, the aircraft was moved to the MCL test facility. The aircraft was hoisted and
installed into the test fixture (Figs. 10 and 11) and a subsequent checkout and
several buildup runs were completed at ambient environmental conditions to
gather and assess acoustic data, required to ensure chamber and aircraft accept-
ability to execute the planned testing. This characterization of the acoustic
environment was important because aircraft design requirements do not typically
include full power operations inside an enclosed building, and there was a poten-
tial risk for structural damage to the aircraft as a result. Based on data gathered in
the acoustic surveys, several constraints were established regarding aircraft power
settings and STOVL propulsion effector angle combinations that were sub-
sequently monitored during test execution. Additionally, the early testing for
the acoustic data collection uncovered some unexpected issues with facility
exhaust ducting system performance for the main engine and lift fan that required
modifications prior to the start of formal climatic testing.
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 695

Fig. 10 F-35 Aircraft installation into MCL main chamber test fixture: part 1.

Fig. 11 F-35 Aircraft installation into MCL main chamber test fixture: part 2.
696 V. J. RODRIGUEZ ET AL.

The formal test events that were executed and completed were the standard
day 598F baseline condition (executed a total of three times in the series of
testing), elevated ambient temperatures (including diurnal cycle at 1208F, low
ambient temperatures (down to – 408F), icing cloud conditions (ground and
simulated flight), ground vortex icing, static and blowing rainfall sequences, freez-
ing rain (nonoperating) and de-icing sequences, and high relative and absolute
humidity diurnal (24-hour period) cycle testing. Upon completion of the required
testing, the aircraft was removed from the test fixture and exited the main
chamber on 1 April 2015. This was followed by a period of aircraft restoration
and preparation for ferry back to NAS Patuxent River. The completed time
between entry to and exit from the MCL spanned six months. Some initially
planned lines of dedicated maintenance and LT&E testing were not able to be per-
formed due to priorities and time constraints.

VII. F-35 ICING CLOUD TESTING

A primary objective of and major driver for F-35 testing at the MCL was the icing
environment evaluation for both conventional and powered lift operations. Pre-
vious icing conditions testing in the MCL for a full-scale aircraft had been gener-
ally limited to conventional ground fog and static operations. The challenge for
the F-35 program was to enable sufficient validation that the aircraft and its
installed propulsion systems could operate in icing conditions in both convention-
al and powered lift modes. These two modes of operation required vastly different
airflows, icing cloud size and conditions, and location relative to the aircraft. The
addition of the objective to evaluate F-35 operating conditions with forward air-
speed operations (vs simply ground static conditions) and to gather good quanti-
tative data to compare to icing models further challenged both MCL and the F-35
program engineering teams. Testing required a myriad of new ideas and capabili-
ties that pushed the testing sequences and MCL facilities teams beyond previous
experiences. Recent upgrades to the icing spray bar system at the McKinley facility
allow simulation of icing conditions far in excess of the relatively benign ground
fog conditions previously tested. This new capability, coupled with advances in
icing simulation codes, allowed the F-35 program to favor the decision to elimin-
ate planned in-flight icing tanker trials from the follow-on flight test program. The
high-level icing test objectives for the icing portion of the CCT were:
. Establish and validate total system-level capabilities at the aircraft level.
. Select conditions to represent JCS/Federal Air Regulation (FAR) Part 25
specifications.
. Perform combination of ground-based and simulated low speed flight con-
ditions in both conventional and powered lift modes of operations.
. Use flight test aids to position aircraft doors, propulsion effectors, and aircraft
surfaces to powered lift representative positions.
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 697

The MCL was the only facility capable of testing the F-35 aircraft at full scale
under closely controlled icing cloud conditions. The facility had the capability
to vary ambient temperature, Liquid Water Content (LWC), and water droplet
sizes while generating a limited wind speed, all critical to providing a large
range of required icing cloud conditions. Close coordination was maintained
among Lockheed Martin, Pratt & Whitney, Rolls – Royce, the Propulsion JPO,
and the MCL personnel well ahead of the entry into the chamber to define and
establish the facility capability vs the specific aircraft requirements. The goal
was to develop a test arrangement that provided the greatest combination of
desired conditions to assess and evaluate the F-35 capabilities. The icing
portion of the chamber testing focused primarily on the functionality and per-
formance of the aircraft inlet-mounted ice detector and the Engine Ice Protection
System (EIPS) and Lift Fan Ice Protection System (LFIPS). The secondary objec-
tive was the assessment and evaluation of ice accretion on the airframe structure
areas that influence the airflow to the propulsion system (e.g., inlet lips, inlet duct,
and lift fan inlet bell mouth). A qualitative assessment was also planned of the
impact of icing cloud exposure to the canopy visibility. The data gathered from
this phase of testing were used for icing model validation and assessment of the
installed system capability. The results from the testing were also used by the
Lockheed Martin Propulsion, Pratt & Whitney, and Rolls-Royce teams for certi-
fication documentation after posttest analysis.
The F-35-specific icing cloud testing required a significant change to the MCL
main chamber configuration with the installation of specific test equipment
required to generate the icing conditions. The aircraft position, restraint system,
and exhaust ducting configurations remained as used for the previous test
sequences. The planned testing consisted of both ground (static/low flow
speed) and simulated flight (low/medium flow speed) operations in both the con-
ventional and powered lift modes. MCL facility limitations and constraints on the
achievable conditions defined the allowable configuration. An open jet icing
tunnel fixture was placed centered in front of the aircraft. The tunnel flow was
driven by a total of nine fans (three parallel “legs” each with three fans in
series) pushing conditioned air from within the main chamber through a duct
arrangement fitted with a spray bar system to produce the icing cloud. A CAD
illustration of the overall tunnel design configuration and placement relative to
the test aircraft is shown in Fig. 12. A photo of the as-built configuration is
shown in Fig. 13.
When the icing test was initially conceived, it was intended that there would be
only a single position for the open jet icing tunnel that would cover all test con-
ditions for both conventional and powered lift modes of operation. Once the
design of the icing tunnel was matured and fixed, Computational Fluid Dynamics
(CFD) analysis was used extensively to determine if sufficient coverage of the icing
cloud was available to support all the planned test conditions. It quickly became
apparent that during the powered lift mode of operation, especially at low forward
airspeeds to simulate hover-type conditions, the engine and lift fan combined
698 V. J. RODRIGUEZ ET AL.

Fig. 12 CAD model of F-35 McKinley chamber icing test setup (overview). (Illustration
courtesy of Dwayne Bell, McKinley Climatic Laboratory.)

Fig. 13 Photo of F-35 icing test chamber configuration (postsetup).

F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 699

would pull maximum airflow and the size of the available icing cloud was insuffi-
cient to fully ice both the lift fan and engine at the same time. The speed range
for the powered lift, low speed testing was increased from 30 kt to 45 kt in an
attempt to reduce the required cross-sectional area of the inlet stream tubes but
was still not sufficient to provide the necessary icing cloud coverage. The only
solution easily achievable was to change the approach to employ two distinct
positions for the icing tunnel: one for conventional mode operations and one
for powered lift operations. The new positions that were established relative to
the original location are shown in Fig. 14. For conventional mode operations,
the lower position was utilized, which presented the icing cloud focused on the
main engine inlets. For powered lift mode testing, the higher location was used,
which delivered the icing cloud to the lift fan inlet and the auxiliary air inlet
but did not fully envelope the main engine inlets. This was considered acceptable,
because the limiting icing case for the main engine and inlet was determined to be
during conventional mode flight operations. The two configurations could not
encompass all of the inlet stream tubes for all planned operating cases, but the
joint team agreed they were of adequate size and capability to achieve the majority
of test requirements.
The installation and checkout of the icing tunnel equipment and the cali-
bration of the tunnel configuration for the targeted icing test conditions was
extensive and time consuming, but critical to the establishment of a valid test
sequence. The icing cloud calibration equipment provided by the MCL (shown
in Fig. 15) consisted of a pitot-static system used for flow speed calculation, a
total temperature probe used to establish flow temperature, a multielement
probe used to determine LWC, and a Malvern and/or phase Doppler interferom-
eter device used to determine the Mean Volumetric Diameter (MVD) of the water
droplets. The initial MVD measuring device was unable to operate consistently at
the lowest operating temperatures required for the planned testing, hence the
decision was made to switch probe types.

Fig. 14 F-35 McKinley chamber icing test setup—low and high icing tunnel locations.
700 V. J. RODRIGUEZ ET AL.

Fig. 15 F-35 McKinley chamber icing test setup (spray bar and calibration equipment
shown).

The conventional mode cloud conditions calibration effort consisted of an

iterative process of tweaking facility air and water supply pressure and tempera-
tures to establish the required settings to generate each targeted cloud combi-
nation in the test plan. The flow was allowed to stabilize at each condition and
flow properties measured and assessed. Adjustments were made to the air and
water supply settings if conditions were not within acceptable tolerances based
on the MCL icing expert experience and knowledge. The corresponding facility
air and water supply conditions required to generate each cloud condition were
documented to allow the re-creation of the conditions during the actual test
period. Particular care was taken when calibrating for the test conditions targeting
ice detector performance and included the additional step of having an engineer
standing next to the aircraft with the icing tunnel running, looking into the tunnel
to identify cloud nonuniformity and spray nozzle instability from the spray bar
close to the location expected to influence the aircraft ice detection system. This
inconsistency and instability was more apparent at low LWC conditions, where
water pressure in the spray bars tended to be low and could be overpowered by
the air pressure required for the target droplet size.
One important test at this stage was the icing calibration grid, where it
was possible to measure the thickness of ice that accretes on a predetermined
grid throughout the extent of the icing cloud. Ideally this test would take measure-
ments either upstream of or without the test asset in position. Given the sheer size
and complexity of the test setup for the F-35, this was not possible, so an icing
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 701

calibration grid was required that would conform around the aircraft just ahead of
the main inlet plane (shown in in Fig. 16). Traditionally an icing calibration grid is
constructed by welding together rectangular sections of metal bars; however, these
are costly and time consuming to build. Significant savings were achieved by fore-
going the traditional construction method and using a “chain link fence” approach.
Though not as tightly controlled as the rectangular section bar style of grid, this new
approach was deemed adequate. With this type of setup for an icing calibration
grid, the measured thickness in the center of each link will not be the same for
every link, even for a uniform cloud leaving the icing tunnel exit. This is because
there are local accelerations of the air and droplets around the aircraft, and areas
where the cloud droplets will not necessarily follow the aerodynamic streamlines.
(This is equally true for an aircraft flying through a cloud in the air as it is for this
type of testing on the ground.) Therefore, to make an assessment to determine the
acceptability of the simulated icing cloud, the team was required to compare the
icing grid accretions to the expected local variations in velocity and LWC. CFD
and icing analysis cases were used to aid in this comparison.
The calibration process continued until all the achievable cloud conditions
had been established. The engine and aircraft systems remained off during the
calibration period. The extensive time required to determine the proper flow set-
tings caused a significant (although not realistic normal operating conditions) ice
accretion on the aircraft structure. Example photos of the calibration process and

Fig. 16 F-35 conformal icing calibration grid (surrounding aircraft).

702 V. J. RODRIGUEZ ET AL.

Fig. 17 F-35 MCL icing test calibration process—conventional mode.

the associated ice accretion are shown in Fig. 17. The total calibration period was
divided into several periods due to a combination of aircraft, facility, and test
schedule factors.
The conventional mode icing cloud evaluation test series focused on the
aircraft-mounted ice detection system, the installed EIPS integration and per-
formance, and airframe ice accretion characteristics. Two distinct test objectives
drove the test sequences: the validation of the ice detection system and the assess-
ment of the integrated aircraft and propulsion systems after exposure to represen-
tative ground and flight icing conditions.
The location of the ice detector within the main inlet duct varies slightly
among the three aircraft variants. As a result, a dedicated ice detector was
added to the BF-5 aircraft to provide data at locations that were representative
of all aircraft variants for the verification of ice detector performance across all
three variants. The ice detector validation effort was designed to provide the
final data required to determine ice detector performance and establish a final
ground and flight capability and clearance for the fleet. For performance vali-
dation and verification, an icing cloud temperature of 148F was selected as the
test condition, serving as a midpoint within the icing specification envelope. The
ice detector validation sequence was performed for both small (15-m MVD) and
large (40-m MVD) droplet size. Target Engine Thrust Requests (ETRs) for each
cloud condition were maintained until sufficient data were captured at the ice
detector of focus. This resulted in multiple runs being required to ensure an accu-
rate response and performance characteristics was established.
Evaluation of the integrated operation of ice detection and the installed engine
and inlet focused on three representative ground fog icing conditions and targeted
specific duration at both ground idle and elevated power settings. Test objectives
included operation of the aircraft at each targeted condition for the full engine
specification time duration for ground idle while gathering associated airframe
accretion information to validate models. The goal was the confirmation and
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 703

validation of the integrated propulsion system operation for fleet operations.

Focus then turned to the evaluation of the integrated operation at three represen-
tative flight icing conditions. A tunnel exhaust flow of 100 kt was selected based on
the combination of facility capability and cloud size. The flight regime portion of
testing concentrated on three engine power settings—flight idle, part power, and
military power (full, nonafterburning power)—and targeted maximum continu-
ous and maximum intermittent icing envelope specifications. All testing was exe-
cuted in a buildup approach to manage risk, starting with the lowest power setting
and with the icing conditions considered to be least challenging to the aircraft and
propulsion system.
Similar operations in powered lift mode required the reconfiguration of the
main chamber to the higher tunnel position, approximately 5 ft vertically. A func-
tional checkout was performed followed by calibration runs to establish the
achievable conditions and create a map of the associated facility air and water
pressures. The powered lift mode calibration effort of the calibration runs was per-
formed without the aircraft and propulsion system operating. First phase of
testing focused on powered lift mode and the evaluation of the integrated lift
fan and aircraft inlet at three representative ground fog icing conditions at
ground idle and during a powered lift mode MBIT, a preprogrammed sequence
used during postmaintenance check of the F-35B’s propulsion system. The pre-
scribed icing cloud for each targeted test conditions required a flow of 45 kt
from the icing tunnel (based on facility capability and required cloud size). The
second phase of powered lift mode testing focused on flight operations and con-
sisted of two primary, simulated flight maneuvers: a part-power flight pattern
maneuver and a full-power hover maneuver. Test sequences were performed at
maximum continuous and maximum intermittent icing envelope specifications.
These testing sequences made use of several predefined Flight Test Aids (FTA),
which execute specially programmed aircraft and propulsion system settings to
simulate various phases of powered lift mode flight operations and scheduling.
The final powered lift mode icing test featured a “graduation exercise” in which
a simulated Vertical Landing Bring Back (VLBB) profile was executed in
various icing conditions. The VLBB is a defined aircraft flight profile that rep-
resents the profile and scenario associated with the return of the aircraft to the
ship for analysis and assessment; it represented a key standard for capability
assessment for the F-35B.
The icing test was successfully completed in incremental but effective blocks
of testing. The ice detection system efforts and results were critical to the vali-
dation and calibration of icing code models and analysis for the installed
system. On-aircraft test results were key to providing the evidence that analytical
predictions of system performance were accurate and therefore able to be used for
analysis for all areas of the operating envelope. The testing performed for the
assessment of the installed ice protection system was grueling and time consum-
ing, requiring short periods of operation at key conditions, followed by painstak-
ingly detailed documentation of accretion characteristics and measurements at
704 V. J. RODRIGUEZ ET AL.

key aircraft and propulsion system areas to validate modeling efforts, analysis
tools, and suitability of installed system performance to maximize allowable
fleet operations, establish required limitations and guidance, and certify system-
level capabilities.

VIII. THERMAL MANAGEMENT SYSTEM TESTING

The Thermal Management System (TMS) is composed of the aircraft subsystems
that generate, transport, or dissipate heat from the air vehicle. These systems
include the fuel system, hydraulic system, electrical system, propulsion
system—including engine and lift fan—and the PTMS. The aircraft fuel system
architecture delivers fuel from the feed tank(s) to the main engine. In that
process, heat is rejected from the various aircraft system heat exchangers into
the fuel supply, which then is also used to cool the engine oil before being supplied
to the engine gas generator to be combusted. In exceptionally hot environments,
the summation of the heat sources has the potential to exceed the available fuel
sink supply, resulting in an aircraft fluid (e.g., fuel or oil) temperature reaching
a prescribed limit value. With the addition of the lift fan and subsequent reduction
in fuel volume on board, the F-35B is the most challenging configuration. The
F-35 contract specification requirements for the aircraft TMS performance were
difficult to verify in a normal test environment because of the complex and mul-
tiple systems involved coupled with the rare nature of this exceptionally hot
environment. As such, the requirements were verified using modeling and simu-
lation methods. Calibration and validation of those models were key to providing
final proof of system design and performance.
The hot day testing that was conducted at McKinley provided significant evi-
dence towards the verification of the TMS requirement to operate to the required
time and environmental conditions. During the CCT elevated ambient tempera-
ture testing sequence, the main chamber was set up to simulate the maximum
diurnal cycle (24-hour period, with ambient temperature and solar radiation
effects included) as specified in the requirements. By the time the engine was
started, fuel feed tank temperatures were 108F above the TMS specification
ground rules and assumptions. Despite this elevated system temperature, the air-
craft was still able to perform and complete the required time at engine ground
idle, then successfully performed a conversion to powered lift mode and executed
a simulated short takeoff maneuver before reaching the maximum allowable pre-
flight fuel system temperature. Figure 18 illustrates the feed tank temperature
traces during the executed preflight test run profile.
The main chamber was maintained at the environmental condition present at
the end of the preflight segment previously described, and the aircraft was pre-
pared to execute the postflight segment of the requirement. Defueling was per-
formed to remove a majority of fuel load from the takeoff portion of testing to
reconfigure the aircraft fuel state at conditions associated with return from a
mission. The main chamber was programmed as close to the corresponding
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 705

Fig. 18 F-35B TMS evaluation test—preflight profile.

diurnal cycle setting as possible prior to the start of formal testing. The final test
sequence involving a simulated hover maneuver, vertical landing, and subsequent
postflight ground operation (including shutdown) was performed to simulate the
return from the mission to evaluate postflight performance of the TMS. All the
required sequences and operating time intervals were successfully completed
prior to reaching thermal limit values. Upon engine shutdown, there was less
than 700 lb of fuel remaining in the feed tanks, and the fuel feed tank bulk temp-
erature reached the allowable postflight limit. Figure 19 illustrates the feed tank
temperature traces during the executed postflight test run profile.
The elevated temperature testing that was conducted at McKinley provided
both a demonstration of aircraft system capabilities and significant evidence
needed for the verification of the TMS specification requirement to operate on
a 1% hot day. For both the preflight and postflight test sequences, system compli-
ance to the timeline requirements was demonstrated, removing any concerns
regarding analytical methods, extrapolation of results, and doubt to actual
system capabilities. The data served as key input required for the calibration
and validation of the model inputs and outputs, enabling the models to be used
for validation of the overall system at all required conditions.

IX. F-35 CLIMATIC TESTING RESULTS SUMMARY AND LESSONS LEARNED

All major test objectives were fulfilled, and the aircraft was demonstrated as able to
operate at all environmental conditions tested. Simulations and modeling were
706 V. J. RODRIGUEZ ET AL.

Fig. 19 F-35B TMS evaluation test—postflight profile.

shown to be accurate and predictions very close to actual results for the majority of
the systems. The aircraft Thermal Management System (TMS) performance was a
major success story coming out of the testing. Testing at elevated ambient temp-
erature was a key factor in the ability to provide the data to correlate and validate
the predictions of the TMS to handle heat loads generated by the aircraft and pro-
pulsion systems and remain within required guidelines for allowable fuel system
temperatures. Sufficient data were gathered in icing conditions to validate model-
ing and verify the detection of icing conditions and the protection of the propulsion
system, and establish aircraft accretion characteristics to help guide aircraft limit-
ations and aircrew guidance. The complex chorography of aircraft doors and
control surfaces to move and articulate as scheduled in cold conditions was veri-
fied. Ground operating capabilities were demonstrated in all operating conditions,
and Mission System (MS) and Vehicle Systems (VS) functionality were character-
ized and verified. Airframe and structural characteristics for water tightness, intru-
sion, and drainage were assessed to predictions and documented. Aircrew
evaluation of environmental characteristics in the cockpit was noted for all the
environmental conditions tested. Joint Technical Data (JTD) and defined mainten-
ance procedures were evaluated, verified, and, where required, updated.
Accurately estimating the time required to conduct both ground and flight
testing has always been notoriously difficult to accomplish with any degree of
accuracy. This particular CCT required extended time to complete the primary
objectives despite a rigorous planning effort. The initial order of testing and
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 707

notional schedule was established based on legacy fighter aircraft programs

and similar climatic test objectives. The schedule was further matured as the
F-35-specific operating characteristics, requirements, and unique risks were ident-
ified and studied within the planning phases. Considerations included the order of
test sequences and environmental conditions, the F-35B’s unique configuration
and operation, and the complexity and scope of the planned test objectives vs pre-
vious experiences (e.g., icing evaluation). Care was taken to account for “unknown
unknowns” and discoveries that naturally occur when operating and exploring the
extremes of the environmental conditions.
A total of 151 days was allotted to cover the expected test time in the MCL
facility, which included the initial entry into the facility, installation of the aircraft
into the test fixture, execution of the targeted test sequences, removal from the test
fixture, exit from the facility, and aircraft restoration to full flight status. A 20%
factor for contingencies and an additional 15% allowance to follow the F-35
SDD flight test program planning factor for retest and discoveries were built
into the schedule as well. The actual test effort spanned a total of 184 days
from initial entry to the facility to the final removal of the aircraft from the test
fixture and departure from the main chamber. The final test execution required
a one-month extension of time in the chamber facility despite best efforts to ade-
quately account for nominal contingencies and establish appropriate margins.
This one-month schedule extension required a program management assessment
of priorities of the planned test objectives that resulted in several of the originally
planned test objectives to be removed from the plan (primarily LT&E sequences).
The completion of subsequent aircraft restoration and reconfiguration of the air-
craft resulted in the ferry flight back to NAS Patuxent River coming two months
later than was originally planned.
A second entry was scheduled and performed one year later to complete some
specific LT&E objectives that had not been able to be performed in the original test
period. Testing was performed over a two-week period in February 2016 by the
joint F-35 ITF teams from Edwards AFB and NAS Patuxent River and allowed
a more focused evaluation, documentation, analysis, and reporting on the F-35
air system’s maintainability and supportability in an extreme climatic environ-
ment. The major objective of all-weather testing was to determine how well the
weapon system, including its essential support equipment and attendant mainten-
ance crews, could accomplish the aircraft support in the required climatic
extremes, using Joint Service Technical Data procedures. A test fleet aircraft
asset (BF-07) was used, and testing was performed with the aircraft nonoperating
and on the floor of the main chamber (not elevated or restrained).
The CCT is a punishing final exam for complex aircraft systems. As more and
more complex and increasingly integrated systems are developed, it becomes
crucial to put the system to the test in the extremes of operational environments
at an integrated system level. High-level program requirements are often decom-
posed and flowed from an air system level to lower system and component levels
(often featuring many suppliers and subcontractors), whereby the validation of
708 V. J. RODRIGUEZ ET AL.

capabilities is left to lower subsystem levels, sometimes performed in isolation of

the component itself or assessed under the assumption of perfect operation by
other adjoining or integrated systems. Along with the increasing fidelity of com-
puters and modeling codes to perform complex algorithms and to simulate func-
tionalities and change multiple variables, it becomes increasingly easy and
common for program management teams to consider actual “total system level”
testing to no longer be necessary or value added, or to be expendable in the
name of cost and time savings.
The F-35 CCT testing experience and results adds to the existing record of air-
craft programs that have found that the actual, full-scale, system-level testing at
extreme environmental conditions remains an invaluable means for the under-
standing and identification of where unperceived or unexpected issues can arise
as a result of the complexity of evolving aircraft designs. The MCL facility at
Eglin AFB and the associated testing it allows continues to show its capability
and value as a proving ground for evaluating design, assembly, and operability/
suitability of increasingly complex and costly military systems in real world scen-
arios ahead of deployment to the end users.

X. SUMMARY
The end result of all the testing performed was a very successful demonstration
and confirmation of the F– 35 capability to operate at all the required environ-
mental conditions. All the required test objectives were completed, and the aircraft
was safely returned to the SDD flight test program at NAS Patuxent River with no
major mishaps or damage. A total of 72 test events were conducted corresponding
to 101 hours of aircraft operation to complete the scheduled test program. Though
no dedicated installed performance data were captured as a primary objective, the
thrust values reported by the propulsion system along with the load cell data from
the facility restraint system indicated operation and scheduling to be in line
with predictions and expectations at each condition. The test period contained
many successes in the highly integrated and complex aircraft interactions. This
overall success, however, does not mean that there were no unplanned “discov-
eries.” Results identified several areas requiring additional action via a combi-
nation of design changes, scheduling, timing, and logic changes, and some
procedural changes that will make the F-35 a better product for the users. The
testing was performed early enough in the design phase to allow for the learning
and changes to be incorporated into the final aircraft configuration release.
“The proof is in the pudding” is an expression often used to state that the final
results will provide the support and evidence that proper technique and due dili-
gence were taken on the path to aircraft fielding. The increasing number of F-35
aircraft being fielded worldwide to the various services and customers will soon
enough provide that proof and, in the end, the authors of this paper are confident
that the time spent in the planning and the execution as well as the results and the
F-35 CLIMATIC CHAMBER TESTING AND SYSTEM VERIFICATION 709

actions taken from the F-35 CCT will in no small part be a key factor in the F-35
program’s success for decades to come.

ACKNOWLEDGMENTS
The authors would like to acknowledge and thank all the people who helped in the
planning, development, execution, analysis, and completion of the F-35 Climatic
Chamber Testing. The number of people involved in the process was extensive
and precludes individual naming; however, of specific note, sincere thanks and
acknowledgement are required to the many organizations and companies that
made this testing and this paper a successful reality. Of note, the Joint Strike
Fighter Program Office, the McKinley Climatic Laboratory at Eglin AFB, Lock-
heed Martin Aeronautics Company, Northrop Grumman, the BAE Systems Plc,
Pratt & Whitney, Rolls – Royce, and the F-35 Integrated Test Force organizations
at NAS Patuxent River and Edwards AFB.
About the Authors

CHAPTER 1 – F-35 PROGRAM HISTORY – FROM

JAST TO IOC, ARTHUR E. SHERIDAN
Art Sheridan has served in numerous roles on the
F-35 Joint Strike Fighter program, beginning as
the STOVL Chief Engineer and, at various times,
leading Aerodynamic Development, Systems Engin-
eering and Integration, Weight Reduction, Air-
Vehicle Improvements, and Affordability. Art was
Lockheed Martin’s Program Manager for teammate,
BAE Systems, responsible for the F-35 aft-fuselage
and empennage airframe components, as well as the fuel, crew-escape and life-
support systems. Recently, Art also managed the F-35 Drag-Chute System
program. In his earlier career, Art served as an aerodynamicist, Flight-Sciences
lead, and Chief Engineer for several advanced programs. Art is an AIAA Associate
Fellow and holds Bachelor and Master of Science degrees in Aeronautical and
Astronautical Engineering from the University of Illinois and Stanford.

CHAPTER 2 – F-35 AIR VEHICLE CONFIGURATION

DEVELOPMENT, MARK A. COUNTS
Mark Counts led F-35 Configuration Design and
Air Vehicle Integration activities from conceptual
design, preliminary and detailed design through
flight test. Mark currently serves as a Senior
Manager in the F-35 Engineering Project Office
where he leads a team of project engineers managing
proposal and execution activities across the F-35
Line of Business. He has been with Lockheed Martin
for 32 years and has conceptual design, preliminary

711
712 ABOUT THE AUTHORS

design, and detailed design experience on numerous programs including F-16,

YF-22, and F-22. Mark holds Bachelor and Master of Science degrees in Mechan-
ical Engineering from the University of Arkansas.

CHAPTER 3 – F-35 AIR VEHICLE TECHNOLOGY

OVERVIEW, CHRIS WIEGAND
Chris Wiegand has over 33 years of experience in air-
craft subsystem design development, integration and
test, include development and integration of environ-
mental control, thermal and fluids systems, auxiliary
power and emergency power systems for the F-16,
F-22, F-35 and other aircraft. Chris also has served
as Project Engineer responsible for integration of
reconnaissance systems in special mission aircraft,
and as a Design Engineer responsible for hydrazine
and cryogenic fluids design for Atlas/Centaur and other propulsion systems.
Chris holds a Bachelor of Science degree in Mechanical Engineering from Univer-
sity of Texas at Austin.

CHAPTER 4 – F-35 DIGITAL THREAD AND

ADVANCED MANUFACTURING,
DR. DON A. KINARD
Dr. Don Kinard is Lockheed Martin’s Senior Techni-
cal Fellow for Production Operations, and was
responsible for establishing the F-35 Fighter Pro-
duction System to manage production growth from
1 aircraft per month to an eventual rate of 20 aircraft
per month. Don previously served as Director of F-35
Production Engineering and was responsible for JSF
Tooling, Planning, Manufacturing Engineering, and
Aircraft Systems Testing. Prior to F-35, Don served as Deputy Director of F-22
Production and also held various other positions on F-22 in both Engineering
and Manufacturing. Don also leads the Lockheed Martin Corporate Fellow’s Man-
ufacturing Team, which is chartered with sharing lessons learned and manufac-
turing technologies throughout the corporation. Don’s academic background is
in Chemistry, with a PhD in Polymer Chemistry from Texas A&M University.
ABOUT THE AUTHORS 713

CHAPTER 5 – F-35 WEAPONS DESIGN

INTEGRATION, DOUGLAS M. HAYWARD
Doug Hayward led the F-35 Weapons Integration
Integrated Product Team from the conceptual
definition through the detailed design phases. He cur-
rently serves as Director of F-35 Systems Engineering
and is responsible for leading air system design, inte-
gration and verification across the air vehicle and
ground system products. Doug previously served as
F-35 Air Vehicle Deputy Director, Vehicle Systems
Deputy Director, and Specialty Engineering lead.
Other prior experiences include weapons and subsystems design and integration
on numerous programs including F-16 derivatives, A-12, F-22, and multiple
advanced aircraft studies. He is a graduate of Texas Tech University where he
earned a Bachelor of Science degree in Mechanical
Engineering.

CHAPTER 6 – F-35 SYSTEM DEVELOPMENT AND

DEMONSTRATION FLIGHT TESTING AT EDWARDS
AIR FORCE BASE AND NAVAL AIR STATION
PATUXENT RIVER, DR. MARY L. HUDSON
Dr. Mary Hudson served as the Lockheed Martin
Senior Manager of F-35 Flight Test Operations at
Edwards Air Force Base, and is currently a Chief
Engineer at the Lockheed Martin Skunk Works.
Mary has been with Lockheed Martin for over 30
years with the last 16 years on F-35, including roles
in F-35 aerodynamics, external environment, store
separation, store certification, and as the F-35 Chief Engineer at Edwards. Mary
has a Bachelor of Science degree in Mechanical Engineering from New Mexico
State, a Master of Science in Mechanical Engineering from Stanford and PhD
in Aerospace Engineering from North Carolina State University.
714 ABOUT THE AUTHORS

CHAPTER 7 – F-35 STRUCTURAL DESIGN,

DEVELOPMENT, AND VERIFICATION,
ROBERT M. ELLIS
Bob Ellis led the F-35 Systems Stress Analysis team
through the preliminary and detailed design, first
flight, and flight test phases of the program. Bob cur-
rently serves as the Senior Manager of the F-35 Struc-
tures Development Integration team. Bob has been
with Lockheed Martin for over 33 years and pre-
viously worked in preliminary and detailed structural
design on multiple programs including F-16, A-12,
and T-50. Bob holds a Bachelor of Science in Mechanical Engineering from
Texas A&M University and an MBA in Engineering Management from the Uni-
versity of Dallas.

CHAPTER 8 – F-35 FLIGHT CONTROL LAW DESIGN,

DEVELOPMENT AND VERIFICATION, JEFFREY J.
HARRIS
Jeff Harris is Sr. Manager of the F-35 Control Law
and Air Data Design team. In this role, he led devel-
opment of the flight control laws for all three versions
of the F-35. Jeff has been with Lockheed Martin for
thirty years and previously served as lead of the
F-22 Control Law Design and Analysis Group. Prior
to joining Lockheed Martin, Jeff was a flight test
engineer assigned to the Carrier Suitability Branch
at the Naval Air Test Center in Patuxent River, Maryland. Mr. Harris holds a
Bachelor of Science degree in Mechanical Engineering from Drexel University
and a Master of Science in Mechanical Engineering from Stanford University.
ABOUT THE AUTHORS 715

CHAPTER 9 – F-35 PROPULSION SYSTEM

INTEGRATION, DEVELOPMENT, AND VERIFICATION,
STEVEN P. WURTH
Steve Wurth is the Lockheed Martin Fellow for Inte-
grated Flight Propulsion Control, or IFPC. He cur-
rently serves as the Propulsion Integration lead on
several programs for the Skunk Works. Steve has
also served as the F-35 IFPC Deputy and Software
lead, the Vehicle System’s IFPC Integrator, and
IFPC Deputy Chief Engineer, helping to develop the
STOVL IFPC systems for the F-35B propulsion
system. During the X-35 competition, Steve led the integrated airframe/propul-
sion contractor team of Lockheed Martin, Pratt & Whitney and Rolls Royce in
development and flight demonstration of the X-35B propulsion system. Steve
began his career at the Skunk Works in Burbank, California where he performed
propulsion installations, designed specialized subsystems, and conducted engine
cycle analysis on various programs including the U-2, F-117, SR-71, X-33 and
several other programs. Steve holds a Bachelor of Science in Mechanical Engineer-
ing from The University of Texas at Austin.

CHAPTER 10 – F-35 SUBSYSTEMS DESIGN,

DEVELOPMENT & VERIFICATION, DREW ROBBINS
Drew Robbins is the Deputy Director of F-35 Engin-
eering & Technology. Drew previously served as
Senior Manager of F-35 Utilities and Subsystems, in
which he led the F-35 Power and Control, Fuel, Fire
Protection, Power and Thermal Management, Life
Support, and Crew Escape Systems teams. He has
also served as the F-35 Power and Control Systems
Manager, F-35 Fuel System Integrator, and worked
15 years in the F-35 Control Law Team developing
the control strategy and algorithms for all aspects of
the flight envelope. Drew received his Bachelor and
Master of Science degrees in Aerospace Engineering
from the University of Texas and Virginia Polytechnic Institute.
716 ABOUT THE AUTHORS

CHAPTER 11 – F-35 MISSION SYSTEMS DESIGN,

DEVELOPMENT, AND VERIFICATION, GREG LEMONS
Greg Lemons is Senior Manager of F-35 Mission
Systems Requirements and Design, where he leads
system level design, architecture and software require-
ments specification for the F-35 avionics system. Greg
has held numerous positions covering a broad area of
disciplines on multiple platforms and advanced pro-
grams spanning conceptual design to manufacturing.
Within the Mission Systems area, he has integrated a
wide range of capabilities to include new sensors,
automated flight modes and weapons onto fighter platforms. Greg graduated
with a Bachelor of Science degree in Aerospace Engineering from Texas A&M
University and received a Master’s in Engineering Management from Southern
Methodist University.

CHAPTER 12 – F-35 INFORMATION FUSION,

THOMAS L. FREY JR.
Dr. Tom Frey is the Lockheed Martin Senior Fellow
for Information Fusion. He has provided key contri-
butions to the design, development, and testing of
the F-35, F-22, and F-16. He holds patents in the
areas of adaptive thresholds, digital beamforming,
and advanced antenna design. He has been an
adjunct professor at Texas Christian University and
Columbia College, and currently serves as a member
of the Texas A&M External Advisory & Development
Council. Tom’s has a Bachelor degree from Texas
A&M University, and a Master’s and PhD from the University of Texas at
Arlington.
ABOUT THE AUTHORS 717

CHAPTER 13 – F-35 CARRIER SUITABILITY

TESTING, TONY WILSON
An experimental test pilot for Lockheed Martin,
Tony Wilson – call sign “Brick” - holds the dis-
tinct honor of being the first aviator to execute
an F-35C shipboard arrestment when he suc-
cessfully landed aboard the USS Nimitz. A
combat veteran, Brick has flown operational
missions in multiple variants of the F-18
Hornet. Tony’s test pilot tours included
serving at VX-23 in Carrier Suitability and
F-18 test. Brick has participated in multiple
flight test disciplines for a variety of tactical aircraft programs, and has flown
over 20 different types of aircraft including the Harrier, the U-2, and the
MiG-15. Brick enlisted in the U.S. Navy in 1991 and was later selected to
attend Old Dominion University, where he earned a Bachelor of Science in Mech-
anical Engineering. He also attended the Air Force Institute of Technology and the
United States Naval Test Pilot School, earning a Master of Science in Aeronautical
Engineering.

CHAPTER 14 – F-35 AERODYNAMIC PERFORMANCE

VERIFICATION, DAVID G. PARSONS
David Parsons is the Lockheed Martin Aircraft Per-
formance Flight Test Lead for the F-35 Program,
responsible for defining, executing, and analyzing
the flight testing required to validate the Aerody-
namics and Performance databases across all three
F-35 variants. He has more than 30 years of experi-
ence in Aerodynamics and Performance flight
testing of high performance fighter aircraft, most
notably the F-16, the X-35 Concept Demonstrator, and the F-35. David holds a
Bachelor of Science degree in Aerospace and Ocean Engineering from Virginia
Polytechnic Institute and is a member of the SAE In-Flight Propulsion
Measurement committee.
718 ABOUT THE AUTHORS

CHAPTER 15 – F-35 HIGH ANGLE OF ATTACK

FLIGHT CONTROL DEVELOPMENT AND FLIGHT
TEST RESULTS, DANIEL G. CANIN
An experimental test pilot for Lockheed Martin
for the past 21 years, Dan Canin – call sign
“Dog” – was the lead high angle of attack test
pilot for the F-35B and C variants. Dan has
over 6,000 flight hours in more than 80 types
of aircraft, including over 650 hours in all
three variants of the F-35. As a Navy pilot, he completed 369 carrier landings,
was a distinguished graduate of the US Naval Test Pilot School, and held flight
test leadership positions on the A-12 and T-45A jet trainer programs. In 1997,
Dan joined Lockheed Martin Skunk Works, where he was project pilot on the
Argentine A-4AR, and performed extensive simulation in support of control
law development on the X-35. He’s also logged over 700 hours of testing in the
F-16 E/F, including loads, flutter, weapons integration, store separation, avionics
systems and advanced flight controls development. Dan is a graduate of the
California Institute of Technology.

CHAPTER 16 – F-35 WEAPONS SEPARATION TEST

AND VERIFICATION, CHRISTOPHER F. HETREED
Chris Hetreed is the Lockheed Martin technical lead
for F-35 Store Separation. Chris has been involved
in all aspects of store separation on the F-35
program, including software development, modeling
and simulation, wind tunnel testing, ground ejection
testing, flight testing, and store certification. He has
also performed dynamics modeling and simulation
of landing gear, arresting gear, pilot seat ejection,
and canopy jettison on the F-35 and F-22 programs.
Previously, while with MSC Software as an engineering consultant, Mr. Hetreed
co-developed the ADAMS/Aircraft modeling and simulation software for rapid
and detailed landing gear loads and dynamics analysis. Chris received his Bache-
lor’s and Master of Science degrees in Aerospace Engineering from Virginia Tech.
ABOUT THE AUTHORS 719

CHAPTER 17 – F-35 STOVL PERFORMANCE

REQUIREMENTS VERIFICATION,
DANIEL E. LEVIN
An experimental test pilot for Lockheed
Martin, Dan Levin is the lead F-35 pilot at
Patuxent River. He has flown development
tests in all models and for all disciplines, in par-
ticular, STOVL envelope expansion. Dan was
previously the lead pilot during development
of the F-16 Block 60 aircraft, where he
focused on integration of the F110-GE-132
engine and high angle of attack testing. In
prior assignments, Dan served as the project
pilot for the LANTIRN navigation and target-
ing system program, served as Flight Dynamics Branch Commander and Oper-
ations Officer with the 416th Flight Test Squadron, and performed as the
Deputy Operations Group Commander conducting classified flight test. Dan
has presented papers on STOVL development for AIAA, the American Helicopter
Society, and Society of Experimental Test Pilots. He is a graduate of the US Naval
Test Pilot School and a veteran of Desert Storm, with operational tours in the
F-16, A-10 and OV-10.

CHAPTER 18 – F-35 CLIMATIC CHAMBER TESTING &

SYSTEM VERIFICATION, VICTORIO J. RODRIGUEZ
Vic Rodriguez led the coordination, planning, and
execution of the F-35 Climatic Chamber Test
Program. He is currently manager of the F-35 Propul-
sion and Thermodynamics team, with responsibility
for propulsion system physical and functional inte-
gration, inlet aerodynamics, and the Thermal Manage-
ment System. In a prior assignment, Vic served as the
F-35 Propulsion Flight Test Planning Lead, in which
he led the definition and coordination of flight test requirements between Lock-
heed Martin, the propulsion system contractors, government propulsion organiz-
ation, and the F-35 Integrated Test Force. Vic previously worked for the Boeing
Company as a Test Requirements and Analysis Engineer on the F/A-18 E/F
Super Hornet and T-45A programs, both in St. Louis and at Naval Air Station
Patuxent River. Vic graduated from the University of Michigan with a Bachelor
of Science in Aerospace Engineering.
INDEX

Note: Page numbers followed by “t” indicate table, and “f ” indicate figure.

AAID. See Auxiliary air inlet door (AAID) CFD, 596–598, 596f–598f
A-10 aircrafts, 53 flowfield grid, 594, 595f
A/A tactical situation model (AATSM), flowfield trajectory simulation,
413 594–595, 595f
AATSM. See A/A tactical situation model store-only freestream, 594, 594f
(AATSM) store validation, 606–627, 607f
Abrupt wing stall (AWS), 102 Aeronautical engineering instructions
ACC. See Air combat command (ACC) (AEIs), 226
Accelerated mission test (AMT), 329–330 Aeronautical systems command (ASC), 3
Active electronically scanned array Affordability improvement curves
(AESA) radar, 82 (AICs), 18
Active inceptor system (AIS), 301, 301f AFTI/F-16 J/IST aircraft MEA
ADAMS. See Automatic dynamic analysis technologies, 128f
of mechanical systems (ADAMS) AGM-88 HARM, 184
Aden 25-mm, 203 AGM-84 Harpoon anti-ship missile, 184
ADS. See Air data system (ADS) AHS. See Ammunition handling system
Advanced medium-range air-to-air missile (AHS); Arresting hook system
(AMRAAM), 184, 197, 224 (AHS)
Advanced multi-variable control AICs. See Affordability improvement
(AMVC), 345 curves (AICs)
Advanced STOVL (ASTOVL), 2–3, 6, 313 AIM-132 ASRAAM, 584
AEIs. See Aeronautical engineering AIM-120/B AMRAAM, 184, 584,
instructions (AEIs) 629, 629f
Aerodynamic database development, 259f AIM-120C AMRAAM, 186, 584, 629, 629f
Aerodynamic performance verification, AIM-9X Sidewinder air-to-air
F-35, 503 heat-seeking missile, 224
background, 504–506, 505f–507f, 507t AIM-9X Sidewinder missile, 584
flight test analysis results, 516–519, Airborne separation video system (ASVS),
519f–522f, 522 612–613, 613f
flight test approach for, 511 Air combat command (ACC), 67
in-flight thrust and performance Aircraft launch bulletins (ALBs),
working group, 511–513, 512f 494, 494t
test matrix, 513, 514f, 514t, 515, 515f Aircraft structural integrity program
uncertainty analysis, 515–516 (ASIP), 253–256
nomenclature, 503–504 five pillars of, 254–256, 255t
overview of, 504 certification and force management
performance management, 508–511, development, 282–283
508f–510f design analysis and development
Aerodynamics models, 592–593, 593f testing, 262–267
attached loads, 593, 593f design information, 256–262

721
722 INDEX

Aircraft structural integrity program overview, 139–140, 139f

(ASIP) (Continued ) STOVL lift system, 151–157
force management execution, 284, AIS. See Active inceptor system (AIS)
284f ALBs. See Aircraft launch bulletins (ALBs)
full-scale testing, 267–282 ALIS. See Autonomic logistics information
F-35 multiservice structures design, system (ALIS)
254, 254f Alternate mission equipment (AME), 190
Air data system (ADS), 529–530, 529f, AMAD. See Airframe-mounted accessory
530f, 554–555, 554f drive (AMAD)
Airframe-mounted accessory drive AME. See Alternate mission equipment
(AMAD), 326 (AME)
Air target management (ATM), 413 Ammunition handling system (AHS), 205
Air-to-air weapon, 210f AMRAAM. See Advanced medium-range
Air vehicle configuration development, air-to-air missile (AMRAAM)
F-35, 77–78 AMT. See Accelerated mission test (AMT)
CDA, 87–93, 87f AMVC. See Advanced multi-variable
CDP for, 83–87 control (AMVC)
design approach, 98f AN/AAQ-37 electro-optical distributed
PWSC, 93–97 aperture system, 410–411, 411f
SDD of, 97–102, 98f, 101f AN/AAQ-40 electro-optical targeting
SWAT, 103–106, 106f system, 409–410, 409f, 410f
airframe, 106–107 AN/APG-81 radar, 404–405, 405f, 406f
air vehicle, 107–109, 108f, 109f AN/ASQ-242 communications,
ground rules and capabilities, 110 navigation, and identification
mission systems, 107 system, 412–413, 412f
propulsion, 109–110 AN/ASQ-239 electronic warfare/
results, 110–111 countermeasure system, 406–408,
vehicle systems, 107 407f, 408f
technologies for, 78–82 Angle of Attack (AoA) maneuvering, 7,
test and evaluation, 114–117, 115f, 196, 216, 290, 304, 445, 448–449,
116f, 117f, 117t–118t 464, 465, 474, 477–478, 482–485,
variants, 78f 493, 496, 525, 525f
weight management, 111–114, 112f, aircraft description, 527
113f air data system, 529–530, 529f, 530f
Air vehicle technology, F-35, 121–124, control surface monitoring and
123f actuation, 528
integrated air vehicle subsystems, design features, 528
124–139 flight control surfaces, 527–528, 527f
early design studies, 124–125 fuel system, 528–529
JSF integrated systems technology lateral/directional stability, 532, 532f
demonstration program, 125–136 longitudinal stability and pitch
transition to F-35 program, 136–139 control power, 530–531, 531f
predecessor programs and, 121–122 roll/yaw control power, 532–534,
propulsion technologies, 139–157 533f, 534f
diverterless supersonic inlet, 140–146 stores management system, 528–529
LOAN, 146–151 tactical navigation system, 529–530
INDEX 723

capabilities, 526–527 ASC. See Aeronautical systems command

control law design, 534–535, 534f (ASC)
anti-spin modes, 541–542 ASEP, 587
automatic pitch rocker mode, full model, 588f
540–541 M&S software, 588
auto recovery modes, 540–543 simulation phase, 588
auto recovery switch, 542, 542f S&RE system models in, 592f
departure prevention, 537–540 ASIP. See Aircraft structural integrity
manual pitch rocker mode, 542–543 program (ASIP)
up-and-away (UA), 535–536, A/S tactical situation model (ASTSM), 413
535f–537f ASTOVL. See Advanced STOVL
effector blender limitations, 544–545 (ASTOVL)
flight test program ASTSM. See A/S tactical situation model
departure recovery, 550–551 (ASTSM)
departure resistance, 551–553, 552f ASVS. See Airborne separation video
expansion to 50-degree AOA, system (ASVS)
549, 549f ATM. See Air target management (ATM)
operational assessment, 553 Automatic dynamic analysis of mechanical
overview, 545 systems (ADAMS), 587
phases, 546f Automatic low speed recovery (ALSR)
preparation, 545–548, 547f, 548f system, 540
flight test results, 553–554, 570–572 Automatic pitch rocker (APR), 540–541
aerodynamic model validation, Autonomic logistics information system
556–558, 556f–557f (ALIS), 65, 225, 226–228, 227f
air data system, 554–555, 554f Auto recovery modes, 540–543
departure recovery, 558–561, 559f Auto recovery switch, 542, 542f
departure resistance, 562–570, Auxiliary air inlet door (AAID), 43, 360
563f–565f, 567f–569f Auxiliary power unit (APU), 326, 370
engine and subsystems, 555 AWS. See Abrupt wing stall (AWS)
lateral CG offsets, 543
limiters, 538–539, 539f BAE Systems, 62, 62f
nomenclature, 526 Basic probability assignment (BPA), 432
onboard model development, Berra, Yogi, 424
543–544 BFA. See Blueprint for affordability (BFA)
overview, 526–527 BK-27 27-mm, 203, 204
roll/yaw limiters, 539–540 BLC. See Boundary layer control (BLC)
Anti-spin functions, 541 Block 2B, 65
AoA. See Angle of Attack (AoA) Blueprint for affordability (BFA), 64
APC. See Approach power Blue ribbon action team (BRAT), 27
compensation (APC) Boeing X-32A aircraft, 12, 12f
Approach power compensation (APC), Bottom-up weights (BUWs), 26
444–445, 449 Boundary layer control (BLC), 142
APR. See Automatic pitch rocker (APR) BPA. See Basic probability assignment
APU. See Auxiliary power unit (APU) (BPA)
Arresting hook system (AHS), 45, 245, BRAT. See Blue ribbon action team
450–454, 450f–455f (BRAT)
724 INDEX

BRU-61, 581, 582f shore-based testing of, 443–444,

BRU-67/A, 198 468–470, 469f
BRU-68/A, 198 arrestments, 472–485, 475f
BRU-67 ejector rack, 580–581, 580f catapults, 470–472, 470f, 471f
BRU-68 ejector rack, 580–581, 580f visual landing aids, 457–459, 458f–459f
3BSD. See Three-bearing swivel nozzle Catapults
(3BSD) carrier suitability, 455–457, 456f–458f
3BSM. See Three-bearing swivel module end speed, 490
(3BSM) energy, 493
BTPs. See Build-to packages (BTPs) shipboard launches, 490–495, 490f
Building block test approach, 262f aircraft launch bulletins, 494, 494t
Build-to packages (BTPs), 23 crosswind, 493–494
Burble, 489 high excess end airspeeds, 492, 493f
BUWs. See Bottom-up weights (BUWs) low catapult energy, 493
minimum end airspeed, 444,
CAIG. See Cost Analysis Improvement 491–492, 492f
Group (CAIG) shore-based testing of, 470–472,
CAIV. See Cost-as-independent-variable 470f, 471f
(CAIV) test, 396f
CALF programs. See Common affordable CATB. See Cooperative avionics test bed
lightweight fighter (CALF) (CATB)
programs CCT. See Climatic chamber testing (CCT)
Candidate follow-on weapons, 211f CDA. See Concept demonstrator aircraft
CAPE. See Cost analysis and program (CDA)
evaluation (CAPE) CDDR. See Concept definition and design
Captive trajectory system (CTS), 594 research (CDDR)
Carrier suitability (CVS), F-35, 441 CDP. See Concept demonstration phase
arresting hook system, 450–454, (CDP)
450f–455f CDRs. See Critical design reviews (CDRs)
catapults, 455–457, 456f–458f Certification, F-35, 282–283
control laws, 446, 447t, 448–449 CFD. See Computational fluid dynamics
F-35C Lightning II, 445–446, 446f (CFD)
glideslope geometry, 464–468, Chase pilot assessment, 631–632
465f, 467f Climatic chamber testing (CCT)
landing signal officer, 460, 462, activities, defining and scheduling, 684,
462f, 463 685f, 686–687
Mk 7 arresting engine, 459–460, aircraft modification, facility, and test
459t–460t, 461f, 462f plan preparation, 687–694,
Nimitz-class aircraft carriers, 689f–691f, 693f, 694f
454–455, 456f facility and test execution, 694, 695f, 696
nomenclature, 441–442 icing cloud testing, 696–697, 698f–702f,
overview of, 442–445, 443f 699–704
ship-based testing of, 485–486, 486f McKinley Climatic Laboratory test
arrested landings, 495–499, 496f–500f facility, 684
catapult launches, 490–495, 490f overview, 681–682
need for, 486–488, 487f, 488f vs. previous aircraft, 682–683, 683f
INDEX 725

results, 705–708, 706f software development paradigm,

thermal management system testing, 307–309, 308f, 309f
704–705, 705f structural coupling challenges, 299–301,
Closed loop fusion, 425 300f
Closed-loop sensor data fusion, 414f verification of, 309–310, 310f
Combat identification, 430–433, 431f, 432f Conventional takeoff and landing (CTOL),
Common affordable lightweight fighter 4, 5f, 6–7
(CALF) programs, 6, 77 Cooperative avionics test bed (CATB), 404
Communications, 231 COPTs. See Cost and operational
Communications, navigation, and performance trades (COPTs)
identification (CNI) suite, 82 Cost analysis and program evaluation
Computational fluid dynamics (CFD), 95, (CAPE), 40
142, 506, 596–598, 596f–598f Cost Analysis Improvement Group
Concept definition and design research (CAIG), 38–40
(CDDR), 79, 184 Cost and operational performance trades
Concept demonstration phase (CDP) (COPTs), 8–9, 9f, 94–95
program, 7, 184 Cost-as-independent-variable (CAIV),
competitors of, 7–8 8–9, 9f
of F-35 air vehicle configuration CPITS. See Critical point in the sky
development, 83–87 (CPITS)
preferred weapons system concept, Critical design reviews (CDRs), 18
10–12 Critical point in the sky (CPITS),
requirements for, 8–9, 9f 258–259, 260f
X-35 aircraft, 12–15, 12f, 14f, 15f CT-525 25-mm, 203
Concept demonstrator aircraft (CDA), 7, CTOL. See Conventional takeoff and
12–15, 12f, 14f, 15f landing (CTOL)
Boeing’s X-32, 87f CTS. See Captive trajectory system (CTS)
F-35 air vehicle configuration CVS. See Carrier suitability (CVS)
development, 87–93, 87f
X-35, 289–290 DADT. See Durability and damage
Consilience: The Unity of Knowledge, 422 tolerance (DADT)
Control law design methodology, F-35, 290 DADT Control Plan, 261
active stick and throttles, 301–302, DARPA. See Defense Advanced Research
301f, 302f Projects Agency (DARPA)
aerodynamic modeling challenges, DART. See Data acquisition, recording and
302–307, 303f, 304f, 305f, 306f, telemetry (DART)
307f DAS. See Distributed aperture system
aircraft loads, 296–299, 296f, 297f, (DAS)
298f, 299f Data acquisition, recording and telemetry
electro-hydrostatic actuation system, (DART), 233, 234f
296–299, 296f, 297f, 298f, 299f Data fusion, 422, 424
implementation challenges, 291, Data smog, 422
292f–294f, 294–295 Defense Advanced Research Projects
nonlinear dynamic inversion, 290–291 Agency (DARPA), 79
numerical challenges, 295–296, Delta flight path (DFP), 445, 449, 474–485,
295f, 296f 496–500
726 INDEX

Dempster-Shafer inferencing algorithm, 6DOF/IMU TM Kit, 617–621

430–431 data loss and, 619–620
DESI. See Digital end speed indicator store, 618–619, 619f
(DESI) trajectory, 620–621, 621f
Design and analysis procedures, F-35, Drop test program, 278–279, 279f
262–267, 263f, 264f, 265f, 266f DSI. See Diverterless supersonic inlet (DSI)
Design information task, F-35, 256–262 DSSAS. See Digital Store Separation
Development testing, F-35, 262–267, 263f, Analysis System (DSSAS)
264f, 265f, 266f Durability and damage tolerance (DADT),
DFLCC. See Digital flight control 256, 260
computer (DFLCC) Dynamic system model, 587–588, 588f
DFP. See Delta flight path (DFP) full dynamics M&S, 589–592, 591f, 592f
Digital end speed indicator (DESI), 490 initial conditions motion M&S,
Digital flight control computer 589, 590f
(DFLCC), 131 validation of, 599–602
Digital store separation analysis system ground vibration tests, 601
(DSSAS), 612–613, 617f, 618f on-aircraft static ejection pit test,
Digital thread, F-35, 22–23, 168–176, 601–602, 602f
169f, 170f static S&RE ejection test, 599–601,
automated drilling by, 171f 600f
automation, phase 2, 171–172, 172f store vibration test, 601
3D models for tools, phase 1, 169–171
phase 4, 174–176 EB. See Effector blender (EB)
validation of engineering, phase 4, EDD. See Environmental description
174–176, 175f, 176f document (EDD)
work instruction graphics, phase 3, EDW. See Edwards Air Force Base (EDW)
172–174 Edwards Air Force Base (EDW), 213–216,
Diminishing manufacturing source 217–219
(DMS), 402 aircraft configuration, 225–226
Distributed aperture system (DAS), 82 aircraft maintenance procedures and
Diverterless supersonic inlet (DSI), inspections, 226
140–146, 141f ALIS, 226–228
background, 140–142 control room, 230f
CFD simulation of, 142–143, 143f flight sciences engineering, 219–223
conceptual development, 142–143, 143f F-35 testing, 217f
F-16 flight demonstration, 145f instrumentation/data processing, 233
flight demonstration program, 145, 145f challenges and lessons learned,
integrated aircraft design and validation, 233–234, 234f
143–145, 144f multiple aircraft telemetry
JSF program, transition to, 146 synchronization, 234–235
small-scale testing of, 143, 143f ITF team, 219, 219f
supersonic boundary layer diversion, mission systems engineering, 223
142–143, 143f mission effectiveness, 224, 225f
wind tunnel testing of, 143–144, 144f software releases, 223
DMS. See Diminishing manufacturing weapons surge, 223–224, 224f
source (DMS) sustainment/logistics, 228
INDEX 727

tool control, 228–229 Electro-hydrostatic actuation systems

wheel and tire mitigation plan, 228 (EHASs), 126–131, 162, 296–299,
test operations, 229 296f, 297f, 298f, 299f, 375
communication, 231 installations, 368f
planning, scheduling, and execution, integration qualification effort/
232–233 challenges of, 386–388, 387f
requirements, 231–232 loading fixtures, 387f
roles and responsibilities, 231 overview of, 367–368, 367f, 368f
teamwork and training, 229–230, requirements of, 375
230f technology development of, 378–379
test pilots’ perspective, 235–237, 236f, Electro-hydrostatic actuators (EHAs),
237f 298–299, 298f–299f, 528
Effector blender (EB), 291, 294f, 534–535, Electro-optical targeting system (EOTS),
544 82, 409–410, 409f, 410f
EHAs. See Electro-hydrostatic actuators Engineering inspection requirements
(EHAs) (EIRs), 226
EHASs. See Electro-hydrostatic actuation Engine starter/generator (ESG), 349
systems (EHASs) Environmental description document
EIRs. See Engineering inspection (EDD), 261
requirements (EIRs) EOTS. See Electro-optical targeting system
Ejection seat, F-35, 374, 374f (EOTS)
Ejection system, F-35 EPAD. See Electrically powered actuation
integration qualification effort/ development (EPAD)
challenges of, 395–397, EPGS. See Electrical power generating
396f, 397f system (EPGS)
overview of, 374–375, 374f EPMS. See Electrical power management
requirements of, 378 system (EPMS)
technology development of, 383–384, EPS. See Electrical power system (EPS)
384f, 385f ESG. See Engine starter/generator (ESG)
testing, 397f EW receivers (EWRs), 408
Electrically powered actuation EWRs. See EW receivers (EWRs)
development (EPAD), 130 External weapon bay design, 195–197,
Electrical power generating system 195f, 196f, 198f
(EPGS), 371–372
Electrical power management system F-16 aircrafts, 3, 4, 36, 52, 53–54
(EPMS), 371–372 F-22 aircraft, 6, 7
Electrical power system (EPS), F-35 aircrafts
376–377 aerodynamic performance verification,
architecture, 372f 503–523
installations, 373f carrier suitability testing, 441–501
integration qualification effort/ climatic chamber testing and system
challenges of, 390–394, 391f, verification, 681–709
392f configuration development, 77–118
overview of, 371–372, 372f, 373f demonstration flight testing, 213–250
requirements of, 376–377 digital thread and advanced
technology development of, 381–382 manufacturing, 161–181
728 INDEX

F-35 aircrafts (Continued) force management execution, 284, 284f

flight control law, 287–310 full-scale testing, 267–282, 268f
high angle of attack flight control, drop test program, 278–279, 279f
525–572 durability structural test program,
information fusion, 421–437 273–278
mission systems, 399–420 flight testing, 280–282
program history, 1–74 static test program, 269–273, 270f,
propulsion system, 313–363 271f, 272f, 273f
STOVL performance requirements FLASH. See Fly by light advanced systems
verification, 641–680 hardware (FLASH)
structural design, development, and Fleet replacement squadron (FRS), 65
verification, 253–285 Flight deck compatibility, 243–244
subsystems, 365–398 F-35 Lightning II, 121–122
system development, 213–250 Flight path, 465
technology development, 121–158 Flight sciences engineering, 219–223, 220f,
weapons design integration, 183–212 221f, 222f, 239–240
weapons separation test and verification, Flight test aid (FTA), 480, 546
575–638 Flight test control engineers (FTCEs), 225
F-35A aircrafts, 67–69, 68f, 80f, 214, Flight testing, 49–50, 50f, 214–217, 214f,
288–290, 289f, 314 215f, 217f
F-35B aircrafts, 65–67, 66f, 79, 79f, 80f, FLOLS. See Fresnel lens optical landing
101f, 214, 288–290, 289f, 314 system (FLOLS)
F-35C aircrafts, 69–70, 80f, 214, 288–290, Flow-to-takt assembly line, 165f
289f, 314 Fly by light advanced systems hardware
drop test program, 278–279, 279f (FLASH), 130
Lightning II, 445–446, 446f Fly-by-wire (FBW) flight control standard,
F-117 aircrafts, 53 323, 334
Fan duct heat exchanger (FDHX), FOD. See Foreign object damage (FOD)
368–369, 380 Force management development, F-35,
FATMOPS. See Fusion analysis 282–283
tool—measures of performance Force management execution, F-35,
(FATMOPS) 284, 284f
FDHX. See Fan duct heat exchanger Force structural maintenance plan
(FDHX) (FSMP), 282, 283
Feedback linearization. See Nonlinear Foreign object damage (FOD), 228–229
dynamic inversion (NDI) Fresnel lens optical landing system
Finite element modeling (FEM) (FLOLS), 457–458, 458f
software, 590 Fresnel lens optical landing system
Fire Control and Stores (FC&S) software (IFLOLS), 457–459, 459f
controls, 208 FRIU. See Fuselage remote interface unit
Five pillars of ASIP, 254–256, 255t (FRIU)
certification and force management FSMP. See Force structural maintenance
development, 282–283 plan (FSMP)
design analysis and development testing, FTA. See Flight test aid (FTA)
262–267 FTCEs. See Flight test control engineers
design information, 256–262 (FTCEs)
INDEX 729

Full-scale ground tests, 50–51, 51f, HARM. See High-speed anti-radiation

267–269 missile (HARM)
Full-scale testing, F-35, 267–269, 268f Head-up display (HUD), 82
drop test program, 278–279, 279f Heat exchanger (HX), 369, 380
durability structural test program, Helmet-mounted display (HMD), 47–49,
273–278, 277f 48f, 236, 236f, 416–417, 416f
flight testing, 280–282 Helmet-mounted display system
static test program, 269–273, 270f, (HMDS), 82
271f, 272f, 273f HGI. See Hot gas ingestion (HGI)
Fuselage remote interface unit (FRIU), High excess end airspeeds, 492
207–208 High-speed anti-radiation missile
Fusion, information (F-35), 421, 425–427, (HARM), 184
427–428, 427f HMD. See Helmet-mounted display
autonomous sensor management of, (HMD)
433–435 HMDS. See Helmet-mounted display
cooperative sensing of, 435–437, system (HMDS)
436f Hot gas ingestion (HGI), 661–662,
evidence-based combat identification of, 673–675
430–433, 431f, 432f Hover weight ratio (HWR), 110
information tiers, 428–430, 429f HUA. See Hydraulic and utility actuation
nomenclature for, 421–422 (HUA)
overview of, 422, 423f, 424 HUD. See Head-up display (HUD)
technology, evolution of, 424–425, HWR. See Hover weight ratio (HWR)
425f, 426f HX. See Heat exchanger (HX)
Fusion algorithms, 422 Hydraulic and utility actuation (HUA)
Fusion analysis tool—measures of block diagram, 373f
performance (FATMOPS), 233 integration qualification effort/
challenges of, 394, 395f
GAM. See General actuator model (GAM) overview of, 372–374, 373f
GAU-22/A gun, 205, 205f requirements of, 377–378, 377t
GAU-12 25-mm gun, 203, 204, 204f technology development of,
GBU. See Guided bomb unit (GBU) 382–383, 383f
GBU-12, 630, 633–635, 634f, 635f
GBU-39, 581, 582f, 631f IAT. See Individual aircraft tracking (IAT)
GBU-49 gun, 184 ICP. See Integrated core processor (ICP)
GD-425 gun, 204, 205 IDLC. See Integrated direct lift control
General actuator model (GAM), 325 (IDLC)
Glideslope angle, 465 IFLOLS. See Fresnel lens optical landing
Glideslope geometry, 464–468, 465f, system (IFLOLS)
467f IFTCD. See In-flight thrust calculation
GPS/laser-guided munition, 224 deck (IFTCD)
Ground vibration tests (GVT), 601 IFTPWG. See In-Flight Thrust and
Guided bomb unit (GBU), 184 Performance Working Group
Gun system, 203–206, 204f, (IFTPWG)
205f, 206f IIP. See International industrial
GVT. See Ground vibration tests (GVT) participation (IIP)
730 INDEX

IMRT. See Independent manufacturing power and thermal management

review team (IMRT) system, 138–139
INCONELw, 380 Integrated core processor (ICP), 207
Independent manufacturing review team Integrated direct lift control (IDLC),
(IMRT), 38 445, 449
Individual aircraft tracking (IAT), 282, 283 Integrated product team (IPT), 324
Inflight engagement, 483 Integrated servo-actuators (ISAs), 130
In-Flight Thrust and Performance Integrated test blocks (ITBs), 239–240
Working Group (IFTPWG), Integrated test force (ITF), 214–215, 215f,
511–513 219, 219f
In-flight thrust calculation deck (IFTCD), Internal weapon bay design, 187–195, 189f,
512, 661 190f, 192f, 193f, 194f
Information fusion, F-35, 421, 422 International industrial participation
algorithms, 422, 423f (IIP), 56
approach of, 427–428 International suppliers, 57f
architecture, 425–427, 427f IOC. See Initial operational capabilities
autonomous sensor management of, (IOC)
433–435 IPT. See Integrated product team (IPT)
cooperative sensing of, 435–437, 436f ISAs. See Integrated servo-actuators (ISAs)
evidence-based combat identification of, Isochrone, 437
430–433, 431f, 432f ITBs. See Integrated test blocks (ITBs)
information tiers, 428–430, 429f ITF. See Integrated test force (ITF)
nomenclature for, 421–422
overview of, 422, 423f, 424 JAST. See Joint advanced strike technology
sensor suite, 426, 426f (JAST)
technology, evolution of, 424–425, JBD. See Jet blast deflector (JBD)
425f, 426f JCS. See Joint contract specification (JCS)
Information tiers, 428–430, 429f JDAM. See Joint direct attack munition
Initial operational capabilities (IOC), (JDAM)
65–70, 215 JET. See Joint estimate team (JET)
Integrated air vehicle subsystems, 124–139, Jet blast deflector (JBD), 243–244
126f, 127f, 128f Joint advanced strike technology (JAST)
benefits of, 127f program, 3, 78–79
early design studies, 124–125 concept definition and design
J/IST demonstration program, 125–136 research, 4
EHAS, 126–131 demonstrations of, 3–4
electrically powered flight control industry competitors of, 4–5
system, 130–131, 130f Lockheed Martin air vehicle concept,
electrical power system, 126–129, 5, 5f
129f continuous transition, 6
thermal/energy management hover balance/trim, 6
module, 131–136, 132f, 134f, 135f induced external environment, 6–7
transition to F-35 program, 136–139 vertical thrust-weight margin, 6
electrical power system, 138 Joint contract specification (JCS), 19, 21,
flight control actuation system, 29, 105, 184, 191, 255–259, 468,
137–138 504, 505, 522, 642, 643, 646, 682
INDEX 731

Joint direct attack munition (JDAM), 224 SDD, 15–16, 161

Joint estimate team (JET), 39 alternate engine program, 52–53
Joint operational requirements document CV arresting hook redesign,
(JORD), 4 challenge, 45
Joint precision approach and landing flight tests, 49–50
system (JPALS), 241 fuel tank inerting redesign, challenge,
Joint Program Office (JPO), 161, 46–47
180, 184, 209, 215–216, 254–258, full-scale ground tests, 50–51
510, 511, 575, 688 helmet and seat redesigns, challenge,
Joint Services Specification Guidance 47–49
(JSSG-2006), 257 over target baseline, challenge, 33–41
Joint strike fighter (JSF) program, 1, 4 program, 16–19, 17f, 18f
CDP program, 7 STOVL probation, challenge, 42–45
competitors of, 7–8 strategies, 20–25
preferred weapons system concept, system requirements deployment, 19
10–12 weight growth, challenge, 25–33
requirements for, 8–9, 9f upcoming plans, 70
X-35 aircraft, 12–15, 12f, 14f, 15f Joint technical data (JTD), 226
COPT vs. CAIV trade study, Joint test plans (JTPs), 219–220
8–9, 9f JORD. See Joint operational requirements
integrated subsystems technology, 3 document (JORD)
international participation in, 53 JPALS. See Joint precision approach and
background, 53–54 landing system (JPALS)
foreign military sales, 58 JPO. See Joint Program Office (JPO)
impact and challenge of, 58–59 JSF. See Joint strike fighter (JSF)
international coalition, 54–57 JSF integrated systems technology (J/IST)
supply chain, 57–58 demonstration program, 125–136
JAST program, 3 EHAS, 126–131
concept definition and design electrically powered flight control
research, 4 system, 130–131, 130f
demonstrations of, 3–4 electrical power system, 126–129, 129f
industry competitors of, 4–5 thermal/energy management module,
Lockheed Martin air vehicle concept, 131–136, 132f, 134f, 135f
5–7, 5f requirements, 133
operations, 65–70 system demonstrations, 134–136,
USAF F-35A IOC, 67–69, 68f 134f, 135f
USMC F-35B IOC, 65–67, 66f turbomachine and component
USN F-35C IOC, 69–70 development, 133–134
origins of, 1–3, 2f JSFPO. See JSF Program Office (JSFPO)
production system JSF Program Office (JSFPO), 3
affordability, 62–64, 63f JTD. See Joint technical data (JTD)
background, 59 JTPs. See Joint test plans (JTPs)
facilities, 60–62, 61f, 62f
growing fleet, 64–65, 64f Kalman filter, 429–430
quantity profile, evolution of, KCs. See Key characteristics (KCs)
59–60, 60f Key characteristics (KCs), 170–171
732 INDEX

Key performance parameters (KPPs), 8, 9, PWSC configurations, 10–12

10, 19, 25–26, 27–33, 35, 504, 513 SDD program plan, 18, 18f
Key system developments and integrations vertical thrust-weight margin, 6
(KSDIs), 24–25 X-35 aircraft, 12–15, 12f, 14f, 15f, 83–93,
Kinematic estimation, 414 83f, 89f, 90f, 92f
KPPs. See Key performance parameters LoIs. See Letters of Intent (LoIs)
(KPPs) Low observable axisymmetric nozzle
KSDIs. See Key system developments and (LOAN), 81, 146–151
integrations (KSDIs) aircraft aerodynamic integration,
148–149
Landing signal officer (LSO), 460, 462, background, 146–147, 146f
462f, 463 ground and flight testing, 149–150,
Large scale powered model (LSPM), 82, 149f
82f, 313, 316–319 JSF program, transition to, 150–151
flow field investigation, 319, 319f radio frequency testing, 150, 150f
IGE testing, 318, 318f rapid prototype evaluation,
NFAC wind tunnel and, 319, 319f 147–148, 148f
LAU-147, 581, 581f Low-rate initial production (LRIP), 17,
LAU-147/A, 198, 199 161–162
LAU-148, 197, 582, 583f LRIP. See Low-rate initial production
LAU-149, 197 (LRIP)
LAU-151, 197, 582, 583f LSO. See Landing signal officer (LSO)
LAU-152, 197 LSPM. See Large scale powered model
Launching airspeed, 491 (LSPM)
Left-hand inceptor (LHI), 347
Length-extender table adapter MADL. See Multifunction advanced data
(LETA), 195 link (MADL)
LETA. See Length-extender table M61A2 20-mm, 203
adapter (LETA) Manual pitch override (MPO), 540
Letters of Intent (LoIs), 56 Manual pitch rocker (MPR), 542–543,
LHI. See Left-hand inceptor (LHI) 557–558
LiftFanw, 81–82, 140, 153–154, 162, Manufacturing, F-35, 161–165, 163f
314–335, 339, 341–361 facility layout, 163f
Loads wind tunnel testing, 259f flow-to-takt assembly line, 165f
LOAN. See Low observable axisymmetric full-rate production, 164f
nozzle (LOAN) future of, 176–181
LO Axi configuration, 139, 140, 149–150, global supply chain, 163f
149f, 150, 150f Lean, 164–166
Lockheed Martin air vehicle, 5, 5f, 260 materials and structures, 166–168,
configurations, series of, 6–7 167f, 168f
continuous transition, 6 multivariant final assembly line, 162f
contract, 17 Marine Fighter Attack Squadron 121
CTOL vs. STOVL propulsion systems, (VMFA-121), 65–67, 66f
5f, 6–7 Marine Fighter Attack Training Squadron
hover balance/trim, 6 501 (VMFAT-501), 65
induced external environment, 6–7 Matrix of Death, 552
INDEX 733

MATS. See Multiple aircraft telemetry vision

synchronization (MATS) aircraft concept, 400–401, 401f, 402f
Mauser BK-27 27-mm system, 203 architecture, 402–403, 403f
MEAS. See Minimum end airspeed Mission systems engineering, 223
(MEAS) mission effectiveness, 224, 225f
Megatrends, 422 software releases, 223
MFDS. See Multifunction display suite weapons surge, 223–224, 224f
(MFDS) MITS-12. See Mobile Instrumentation and
MGS. See Missionized gun system (MGS) Telemetry System (MITS-12)
MIL-A-8591, 199 Mk 7 arresting engine, 459–460,
MIL-I-8671, 190–191 459t–460t, 461f, 462f
MIL-STD-1289, 190–191 Mobile Instrumentation and Telemetry
MIL-STD-1530, 253–254 System (MITS-12), 247–248
MIL-STD-1553, 208 MPO. See Manual pitch override (MPO)
MIL-STD-1760, 206–207, 208, 209, 210 MPR. See Manual pitch rocker (MPR)
MIL-STD-2088, 199 MRIU. See Missile remote interface unit
MIL-STD-9490, 299 (MRIU)
Minimum end airspeed (MEAS), Multifunction advanced data link
491–492, 492f (MADL), 82, 400, 426–427,
Missile remote interface unit (MRIU), 207 428–429, 429f, 435
Missionized gun system (MGS), 206, 206f Multifunction display suite (MFDS), 417
Mission systems, F-35, 399–400 Multiple aircraft telemetry
data into information, fusing of, synchronization (MATS),
413–415, 414f 234–235
next-generation cockpit for, 415–416,
415f Naisbitt, John, 422
helmet-mounted display, 416–417, NATOPS. See Naval air training and
416f operating procedures
panoramic cockpit display, 417, standardization (NATOPS)
417f Natural wind, 464
sensor suite of, 399–400, 403, 404f NAVAIR. See Naval Air Systems
AN/AAQ-37 electro-optical Command (NAVAIR)
distributed aperture system, Naval Air Station Patuxent River (PAX),
410–411, 411f 213–216, 237–239, 238f, 239f
AN/AAQ-40 electro-optical targeting detachments, 242t, 243f
system, 409–410, 409f, 410f flight sciences engineering, 239–240
AN/APG-81 radar, 404–405, test operations, 240–241
405f, 406f CV ship suitability testing, 243–247,
AN/ASQ-242 communications, 244f, 247f
navigation, and identification detachments planned and
system, 412–413, 412f accomplished, 241
AN/ASQ-239 electronic warfare/ instrumentation/data processing,
countermeasure system, 406–408, 247–248
407f, 408f STOVL ship suitability testing,
verification of, 418–419, 418f, 241–243
419f, 419t test pilots’ perspective, 248–249
734 INDEX

Naval Air Systems Command OTB-2 baseline, 39

(NAVAIR), 3 Outer mold line (OML), 80, 95
Naval air training and operating
procedures standardization PAIRs. See Production aircraft inspection
(NATOPS), 240 requirements (PAIRs)
NDI. See Nonlinear dynamic inversion Panoramic cockpit display (PCD),
(NDI) 417, 417f
Nimitz-class aircraft carriers, 454–455, PAO. See Polyalphaolephin (PAO)
456f PAUC. See Program acquisition unit cost
Nonlinear dynamic inversion (NDI), (PAUC)
290–291, 306f, 528, 534–535, 534f PBSs. See Performance-based
Nonpyrotechnic ejection, 200 specifications (PBSs)
Nunn-McCurdy breach, 38–41 PCD. See Panoramic cockpit display
(PCD)
OBIGGS. See Onboard inert gas generation PDM. See Product data management
system (OBIGGS) (PDM); Product data manager
OBM. See Onboard model (OBM) (PDM)
OBOGS. See Onboard oxygen generation PDR. See Preliminary design review (PDR)
system (OBOGS) PEO. See Program executive officer (PEO)
Observe, orient, decide, and act Performance-based specifications (PBSs),
(OODA), 425 254, 258
Off-ejector early trajectory, 632 Pilot eye
OHS. See Ordnance hoist system (OHS) to FRL angle, 466
OML. See Outer mold line (OML) to hook point length, 466
On-aircraft static ejection pit test, Pilot-to-vehicle interface (PVI), 479
601–602, 602f Pilot/vehicle interface (PVI), 415
On-aircraft video photogrammetry, 612 Pneumatic power source (PPS), 197,
aircraft target layout, 615 200–202, 581
cameras, 612–613 Polyalphaolephin (PAO), 369
flight test trajectory solutions, 617, 617f Power and thermal management system
one-time accuracy assessment, (PTMS), 202
616–617, 616f integration qualification effort/
store target layout, 613–614, 614f, 615f challenges of, 389–390, 390f
Onboard inert gas generation system overview of, 368–371, 369f, 370f
(OBIGGS), 46–47, 369 requirements of, 375–376
Onboard model (OBM), 534, 543–544 technology development of, 379–381,
Onboard oxygen generation system 380f, 381f
(OBOGS), 369 PPS. See Pneumatic power source (PPS)
On-ejector performance, 632 Preferred weapons-system concept
OODA. See Observe, orient, decide, and (PWSC), 7, 10–12
act (OODA) F-35 air vehicle configuration
OQLS. See Ordnance quick latch system development, 93–97
(OQLS) Preliminary design review (PDR), 18
Ordnance hoist system (OHS), 193–194 Product data management
Ordnance quick latch system (OQLS), 190, (PDM), 225
190f, 194, 582 Product data manager (PDM), 261
INDEX 735

Production aircraft inspection propulsion integration and integrated

requirements (PAIRs), 226 flight propulsion control ground
Product lifecycle manager, 23 testing, 358–359
Program acquisition unit cost (PAUC), 40 overview of, 314–316
Program executive officer (PEO), 3 STOVL, 314–315, 314f, 315f
Propulsion system, F-35 and X-35, 313 Propulsion technologies, F-35 air vehicle,
ASTOVL/LSPM of, 316, 316f 139–157
capabilities and design features for, diverterless supersonic inlet,
316–317 140–146, 141f
verification testing for, 317–319, 318f, background, 140–142
319f conceptual development,
CTOL/CV, 314–315, 314f 142–143, 143f
feasibility of, 316, 316f flight demonstration program, 145,
capabilities and design features for, 145f
316–317 integrated aircraft design and
verification testing for, 317–319, validation, 143–145, 144f
318f, 319f JSF program, transition to, 146
functional feasibility of, 319–322, LOAN, 146–151
319f, 320f aircraft aerodynamic integration,
aircraft flight testing for, 333–339 148–149
aircraft propulsion ground testing for, background, 146–147, 146f
333–339 ground and flight testing,
air induction system, 323 149–150, 149f
air vehicle integration, 326 JSF program, transition to, 150–151
AIS development testing for, radio frequency testing, 150, 150f
330–333, 331f, 332f, 333f rapid prototype evaluation,
developmental testing for, 327–330, 147–148, 148f
328f, 329f overview, 139–140, 139f
hot gas ingestion, 323 STOVL lift system, 151–157
IFPC development testing for, 330 background, 151–152, 151f, 152f
performance, 322–323 3BSM, 156–157
STOVL integrated flight propulsion JSF program, transition to, 157
control and conversion LiftFan system, 153
development, 323–326, 325f roll post nozzle development, 156
operational deployment of, 339 shaft/clutching functionality,
aircraft propulsion ground and flight 153–154
testing, 359–361, 359f, 360f technology demonstration
air vehicle integration, program, 157
348–350, 349f variable area vane box nozzle
AIS, 341–343, 341f, 342f development, 154–156, 155f
developmental inlet testing, 351, PTMS. See Power and thermal
352–357, 352f–357f management system (PTMS)
integrated flight propulsion control PVI. See Pilot-to-vehicle interface (PVI);
development and failure immunity, Pilot/vehicle interface (PVI)
343–348, 343f, 346f, 347f PWSC. See Preferred weapons-system
performance, 339–340 concept (PWSC)
736 INDEX

QRC. See Quick reaction capability (QRC) maximum engaging speed on-center,
Quick disconnect (QD) fittings, 202 482–483
Quick reaction capability (QRC), 223 mean pitch high sink, 475–477
nose down high sink, 477–478
RCA. See Root cause analysis (RCA) roll, 478–479, 478f
Recovery crosswind, 465 roll/yaw opposite, 480–481
Recovery headwind (RHW), 465 roll/yaw same, 481
RHI. See Right-hand inceptor (RHI) tail down high sink, 477
RHW. See Recovery headwind (RHW) yaw arrestment, 479–480
Right-hand inceptor (RHI), 347 Short takeoff and vertical landing
Roll post nozzle, 156 (STOVL), 1, 5–7, 11
Roll-post-nozzle actuators, 44 aircrafts, 79, 79f
Root cause analysis (RCA), 283 DARPA programs, 79
performance verification, 640, 645,
SAE. See Senior acquisition executive 645f, 646t
(SAE) background, 643, 644f, 645
SAMDC. See Structural analysis methods overview, 642–643, 642f
and design criteria (SAMDC) short takeoff, 646–659
SBA. See Simulation-based acquisition timeline, 645f
(SBA) vertical landing bring-back, 659–679
Schedule B award fee, 17–18 propulsion system, 81–82, 82f
SCUBA. See Self-contained underwater Short takeoff (STO) performance
breathing apparatus (SCUBA) verification, 646–659
SDB. See Small diameter bomb (SDB) operational considerations,
SDC. See Structural design criteria (SDC) 657–659, 658f
SDD. See System development and results, 656–657, 656f
demonstration (SDD) ship-based short takeoff testing,
SDLF. See Shaft-driven lift fan (SDLF) 650–655, 651f, 653f–655f
Self-contained underwater breathing shore-based short takeoff testing,
apparatus (SCUBA), 646–650, 647f, 649f, 650f
235–236 SIL. See System Integration Laboratory
Senior acquisition executive (SAE), 3 (SIL)
Sensor fusion, 422, 424–425 Simulation-based acquisition (SBA),
Sensor management, 433–435 21–22, 22f
Shaft-driven lift fan (SDLF), 81 Single hoist ordnance loading system
Ship board testing, 444 (SHOLS), 193
Ship suitability flight test, 442 Sled test, 397f
Ship’s wind, 465 Small business innovation research (SBIR)
SHOLS. See Single hoist ordnance loading program, 179–180, 179f, 180f
system (SHOLS) Small diameter bomb (SDB), 184, 224
Shore-based arrestments, 443–444, SMS. See Stores management system
472–485, 475f (SMS)
free flight or high pitch/low angle, Spin recovery chute (SRC), 547–548
483–485, 484f canister, 547f
maximum engaging speed ground deployment test, 548f
off-center, 483 installation, 548f
INDEX 737

SRC. See Spin recovery chute (SRC) Structural analysis methods and design
S&RE. See Suspension and release criteria (SAMDC), 260–261
equipment (S&RE) Structural coupling, 299–301, 300f
Static rack ejection tests, 599–601, 600f Structural design criteria (SDC), F-35,
STINs. See System track information needs 257–262
(STINs) Structures certification process, F-35, 258f
STM. See Surface target management Supply chain, F-35, 161–164, 163f
(STM) Surface target management (STM), 413
Store separation flight test, F-35 Surge, weapons, 223–224, 224f
instrumentation, 602–603 Suspension and release equipment
for aerodynamics model (S&RE), 198–203, 200t, 201t
validation, 606 SUU-96 pylon, 197, 198f
6DOF/IMU TM Kit analysis, SWAT. See STOVL Weight Attack Team
604–606, 604f (SWAT)
photogrammetric analysis, 603–604, System development and demonstration
603f (SDD), 15–16, 161, 184, 215
validation, 606–608, 607f alternate engine program, 52–53
aircraft response, 610–612, 611f certified weapon loadouts, 187f
6DOF/IMU TM Kit, 617–621 challenges
on-aircraft video photogrammetry, CV arresting hook redesign, 45
612–617 fuel tank inerting redesign, 46–47
planning, 608–609 helmet and seat redesigns, 47–49
preflight measurements, 609–610 over target baseline, 33–41
store aerodynamics, 621–626, 623f, STOVL probation, 42–45, 42f
624f, 625f, 626f weight growth, 25–33
time considerations, 626–627 of F-35 air vehicle configuration
vignettes, 627–628 development, 97–102, 98f, 101f
aerodynamics effects on S&RE flight tests, 49–50
performance and initial conditions, full-scale ground tests, 50–51
628–630 over target baseline, 33–41
aircraft maneuver in posttest analysis, mid-course update, 35–38
635–637 Nunn-McCurdy breach, 38–41
GBU-12 tailkit lanyard post-SWAT reset, 34–35
reconfiguration, 633–635, 634f, technical baseline review, 38–41
635f, 636f Pratt & Whitney F135 in, 52–53
multiple store separation events per program, 16–19, 17f, 18f
flight, 630–633 strategies, 20–25
Stores management system (SMS), affordability and best value, 20
206–209, 207f collaborative environment, 22–23
Store vibration test, 601 concurrent engineering, 23
STOVL. See Short takeoff and vertical digital thread, 22–23
landing (STOVL) enterprise-guiding principles, 20
STOVL Weight Attack Team (SWAT), integrated product teams, 23
27–31 KSDIs, 24–25
Strength and summary operating performance-based specifications,
restrictions (SSOR) reports, 283 20–21
738 INDEX

System development and demonstration detachments planned and

(SDD) (Continued) accomplished, 241
risk management, 23–25 instrumentation/data processing,
simulation-based acquisition, 247–248
21–22, 22f STOVL ship suitability testing,
systems engineering and 241–243
requirements management, 23 test pilots’ perspective, 248–249
total system performance Test pilots, 235–237, 236f, 237f, 248–249,
responsibility, 20–21 463–464
system requirements deployment, 19 TFLIR. See Targeting forward-looking
weapons carriage, 579–580, 580f infrared (TFLIR)
weapon separation test program, 579 Three-bearing swivel module (3BSM), 140,
weight growth, challenge, 25–33 156–157
BRAT, 27 Three-bearing swivel nozzle (3BSD), 5
post-SWAT, 31–33, 32f Three-dimensional models (3D models),
SWAT, 27–31 169–171
wind tunnel testing, 506, 507t Thrust Vector Lever (TVL), 347
System Integration Laboratory (SIL), 391 Time compliance technical directives
System track information needs (TCTDs), 226
(STINs), 413 Time difference of arrival (TDOA),
435–437, 436f
Tactical situation models (TSMs), Total system performance responsibility
413–414 (TSPR), 20–21
Targeting forward-looking infrared Translational rate command (TRC), 301
(TFLIR), 409 Transonic roll-off (TRO), 544
TBR. See Technical baseline review (TBR) TRC. See Translational rate command
TCTDs. See Time compliance technical (TRC)
directives (TCTDs) TRO. See Transonic roll-off (TRO)
TDOA. See Time difference of arrival TSMs. See Tactical situation models
(TDOA) (TSMs)
Technical baseline review (TBR), 41 TSPR. See Total system performance
Telescoping vectoring exhaust nozzle responsibility (TSPR)
(TVEN), 10 TVEN. See Telescoping vectoring exhaust
Test operations nozzle (TVEN)
EDW, 229 TVL. See Thrust vector lever (TVL)
communication, 231 Two-sting-rig (TSR) system, 594
planning, scheduling, and execution,
232–233 UAI. See Universal armament interface
requirements, 231–232 (UAI)
roles and responsibilities, 231 ULFD. See Upper liftfan door (ULFD)
teamwork and training, Uncertainty analysis, 515–516
229–230, 230f Unit recurring flyaway (URF) costs, 203
Naval Air Station Patuxent River (PAX), Universal armament interface (UAI),
240–241 207, 210
CV ship suitability testing, 243–247, Up-and-away (UA) control laws,
244f, 247f 535–536, 536f
INDEX 739

Upper liftfan door (ULFD), 360 PTMS

U.S. fighter procurement profile, 2f integration qualification effort/
USAF F-35A IOC, 67–69, 68f challenges of, 389–390, 390f
USMC F-35B IOC, 65–67, 66f overview of, 368–371, 369f, 370f
USN F-35C IOC, 69–70 requirements of, 375–376
technology development of, 379–381,
Variable area vane box nozzle (VAVBN), 380f, 381f
10, 90, 152, 154–156, 155f, 315, Vehicle system integration facility (VSIF),
346, 346f 386, 386f
VAVBN. See Variable area vane box nozzle Verification cross reference matrix
(VAVBN) (VCRM), 257–258
VCRM. See Verification cross reference Verification requirements, F-35, 257–262
Matrix (VCRM) Vertical landing bring-back (VLBB)
Vehicle management computers (VMCs), performance verification, 642,
371, 446, 448 659–679, 660f
Vehicle subsystems, F-35, 365–366 HGI mechanisms, 662f
EHAS operational considerations, 669
integration qualification effort/ derotation, 673–675, 674f, 675f
challenges of, 386–388, 387f descent rate, 670–673, 671f–673f
overview of, 367–368, 367f, 368f hover capability, 669–670, 670f
requirements of, 375 pilot perspective, 678–679
technology development of, 378–379 ship operations, 675–678, 676f,
ejection system 677f, 678
integration qualification effort/ pattern profile, 659f
challenges of, 395–397, 396f, 397f results, 668–669, 669f
overview of, 374–375, 374f ship-based testing, 666
requirements of, 378 hover capability, 666–668, 667f
technology development of, 383–384, vertical landing, 668
384f, 385f shore-based testing, 663
EPS hover capability, 663–665, 664f
integration qualification effort/ pattern fuel, 665–666, 666f
challenges of, 390–394, 391f, 392f vertical landing, 665
overview of, 371–372, 372f, 373f VIMs. See Virtual interface models (VIMs)
requirements of, 376–377 Virtual interface models (VIMs), 427–428
technology development of, VMCs. See Vehicle management
381–382 computers (VMCs)
federated vs integrated, 366, 366f VSIF. See Vehicle system integration
HUA facility (VSIF)
integration qualification effort/
challenges of, 394, 395f Water tunnel testing, 117f
overview of, 372–374, 373f WBDD. See Weapon bay door drive
requirements of, 377–378, 377t (WBDD)
technology development of, 382–383, WDA. See Weapons delivery accuracy
383f (WDA)
integration qualification effort/ Weapon bay door drive (WBDD),
challenges, 386–397, 386f 378, 396f
740 INDEX

Weapons delivery accuracy (WDA), 6DOF/IMU TM Kit analysis,

224, 224f 604–606, 604f
Weapons design integration, F-35, photogrammetric analysis,
183–212 603–604, 603f
external weapon bay design, 195–197, LAU-148, 582–583, 583f
195f, 196f, 198f LAU-151, 582–583, 583f
future growth, 209–210, 210f, 211f LAU-147 missile ejector rack, 581, 581f
general arrangement, 184–187, 185f, modeling, 584–585
186f, 187f rail-launched store example,
gun system, 203–206, 204f, 205f, 206f 586–587, 586f
internal weapon bay design, 187–195, released/ejected store example,
189f, 190f, 192f, 193f, 194f 585–586, 585f
overview, 183–184 nomenclature, 575–576
SDD certified weapon loadouts, 187f on-aircraft static ejection pit test,
station capacity, 186f 601–602, 602f
stores management system, 206–209, overview, 576–579, 577f
207f requirements to, 578f
suspension and release equipment SDD, 579–584, 580f
design, 198–203, 200t, 201t simulation event, 587
weapons suite, 184–187 static rack ejection tests, 599–601, 600f
Weapons separation flight testing, store separation path to certification, 578f
575, 576f store vibration test, 601
aerodynamics models, 592–593, 593f validation, store separation flight test,
attached loads, 593, 593f 606–608, 607f
CFD, 596–598, 596f–598f aircraft response, 610–612, 611f
flowfield grid, 594, 595f 6DOF/IMU TM Kit, 617–621
flowfield trajectory simulation, on-aircraft video photogrammetry,
594–595, 595f 612–617
store-only freestream, 594, 594f planning, 608–609
store validation, 606–627, 607f preflight measurements, 609–610
BRU-61, 582f store aerodynamics, 621–626, 623f,
BRU-67/BRU-68 ejector rack, 624f, 625f, 626f
580f, 583f time considerations, 626–627
dynamic system model, 587–588, 588f vignettes, store separation flight test,
full dynamics M&S, 589–592, 627–628
591f, 592f aerodynamics effects on S&RE
initial conditions motion M&S, performance and initial conditions,
589, 590f 628–630
validation of, 599–602 aircraft maneuver in posttest analysis,
F-35 lightning II, 579–584 635–637
GBU-39, 582f GBU-12 tailkit lanyard
ground vibration tests, 601 reconfiguration, 633–635, 634f,
instrumentation, store separation flight 635f, 636f
test, 602–603 multiple store separation events per
for aerodynamics model flight, 630–633
validation, 606 Weapons surge, 223–224, 224f
INDEX 741

Weapons System Acquisition Reform Act CTOL/CV and STOVL propulsion

of 2009, 40 systems, 320, 320f
Weight as an independent variable DSI development timeline for,
(WAIV) curve, 105–106, 106f 330–331, 331f
Weight management, F-35, 510f, 511 flight testing of, 333–339
Wheel and Tire (W&T) assemblies, 228 IGE wind tunnel oil flow vs. flight test
Wilson, E. O., 422 demonstration, 333, 333f
Wind over deck (WOD), 465 LiftFan inlet distortion pattern,
Wind tunnel model 332, 332f
installations, 506f LiftFan installation in, 335f
testing, 115f, 116f nose gear soft support, 336f
Wing commonality approach, 85f propulsion ground testing of,
WOD. See Wind over deck (WOD) 333–339, 336f
soft support system, 337f
X-35 aircrafts, 12–15, 12f, 14f, 15f, 121,
289–290 YAK-38 aircraft, 5
conversion propulsion, 325–326, 325f YAK-141 aircraft, 5
SUPPORTING MATERIALS

A complete listing of titles in the Progress in Astronautics and Aeronautics series

is available from AIAA’s electronic library, Aerospace Research Central (ARC) at
arc.aiaa.org. Visit ARC frequently to stay abreast of product changes, corrections,
special offers, and new publications.
AIAA is committed to devoting resources to the education of both practicing
and future aerospace professionals. In 1996, the AIAA Foundation was founded.
Its programs enhance scientific literacy and advance the arts and sciences of aero-
space. For more information, please visit www.aiaafoundation.org.
ABOUT THE BOOK
In the years since the F-35 Joint Strike Fighter development program was awarded to
Lockheed Martin, countless articles have appeared in the popular media discussing the
merits and challenges of the program. F-35 information has also occasionally appeared
in the scientific literature through AIAA or other technical venues. However, an integrated
story of the F-35’s technical development has heretofore never been presented in a
single work.
The purpose of The F-35 Lightning II: From Concept to Cockpit is to tell that story – a
full-spectrum history of the design, development and verification of the F-35 Lightning
II as described by the engineers, scientists and managers intimately involved throughout
the development program. The journey to achieve the F-35’s extraordinary capability
was long with untold challenges encountered on the path. The reader will find many of
the key technical challenges, and the innovative solutions that resulted, discussed herein.
The reader should realize this work is not intended as an overall assessment of the F-35
or the Joint Strike Fighter program. Rather, it is an engineering development story from
the perspective of insiders, many of whom dedicated the best part of their careers to the
F-35 and are justifiably very proud of their work.
The book is based on 18 technical papers presented in a two-day F-35 track at the
2018 AIAA AVIATION Forum. The “From Concept to Cockpit” subtitle appropriately
summarizes the contents, with the chapters grouped into three major sections:
• F-35 Program Overview, which begins with an overall history of the program and
further presents discussion of aircraft configuration design, weapons integration, key
technologies brought into the program, and overviews of aircraft manufacturing and
the flight test program.
• F-35 Air Vehicle Design, which presents discussion of key aircraft systems, including
airframe structure, flight controls, propulsion, subsystems and mission systems.
• F-35 Test and Verification, which presents discussion of the flight and ground test
programs undertaken to verify the F-35’s airworthiness, flight performance and
mission capability.
The chapters are in a logical progression but are not interdependent, and so one may
read the contents in any order without confusion.

ABOUT THE EDITOR

Jeffrey W. Hamstra, Lockheed Martin Corporation. Jeff Hamstra is a Sr.
Fellow at Lockheed Martin Corporation with a career spent in combat
aircraft program management, air vehicle design, propulsion system
integration and technology development. Jeff is an AIAA Fellow, former
member of the AIAA Board of Directors and graduate of the University
of Michigan.

I S B N 978-1-62410-566-1
$119.95
11995
American Institute of Aeronautics and Astronautics

aiaa.org

ISBN: 978-1-62410-566-1

9 781624 105661