MSC Nastran Reference Guide

MSC Nastran 2021.4

Reference Guide
Corporate Europe, Middle East, Africa
5161 California Ave. Suite 200 Am Moosfeld 13
University Research Park 81829 Munich, Germany
Irvine, CA 92617 Telephone: (49) 89 431 98 70
Telephone: (714) 540-8900 Email: [email protected]
Email: [email protected]

Japan Asia-Pacific
KANDA SQUARE 16F 100 Beach Road
2-2-1 Kanda Nishikicho, Chiyoda-ku #16-05 Shaw Tower
1-Chome, Shinjuku-Ku Singapore 189702
Tokyo 101-0054, Japan Telephone: 65-6272-0082
Telephone: (81)(3) 6275 0870 Email: [email protected]
Email: [email protected]

Worldwide Web
www.mscsoftware.com, www.hexagon.com

Support
https://simcompanion.hexagon.com/customers/s/article/support-contact-information-kb8019304

Disclaimer
Hexagon reserves the right to make changes in specifications and other information contained in this document without prior notice.
The concepts, methods, and examples presented in this text are for illustrative and educational purposes only, and are not intended
to be exhaustive or to apply to any particular engineering problem or design. Hexagon assumes no liability or responsibility to any
person or company for direct or indirect damages resulting from the use of any information contained herein.
User Documentation: Copyright 2021 Hexagon AB and/or its subsidiaries. Printed in U.S.A. All Rights Reserved.
This notice shall be marked on any reproduction of this documentation, in whole or in part. Any reproduction or distribution of this
document, in whole or in part, without the prior written consent of Hexagon is prohibited.
This software may contain certain third-party software that is protected by copyright and licensed from Hexagon suppliers. Additional
terms and conditions and/or notices may apply for certain third party software. Such additional third party software terms and
conditions and/or notices may be set forth in documentation and/or at http://www.mscsoftware.com/thirdpartysoftware (or successor
website designated by Hexagon from time to time).
PCGLSS 8.0, Copyright © 1992-2016, Computational Applications and System Integration Inc. All rights reserved. PCGLSS 8.0 is
licensed from Computational Applications and System Integration Inc.
MSC, Dytran, Marc, MSC Nastran, Patran, Hexagon logo, MSC Software logo, e-Xstream, Digimat, and Simulating Reality are
trademarks or registered trademarks of Hexagon AB and/or its subsidiaries in the United States and/or other countries.
NASTRAN is a registered trademark of NASA. FLEXlm and FlexNet Publisher are trademarks or registered trademarks of Flexera
Software. All other trademarks are the property of their respective owners.
Use, duplicate, or disclosure by the U.S. Government is subjected to restrictions as set forth in FAR 12.212 (Commercial Computer
Software) and DFARS 227.7202 (Commercial Computer Software and Commercial Computer Software Documentation), as
applicable.
U.S. Patent 9,361,413
December 3, 2021
NA:V2021.4:Z:Z:Z:DC-REF-PDF
Documentation Feedback
At MSC Software, we strive to produce the highest quality documentation and welcome your feedback.
If you have comments or suggestions about our documentation, write to us at: documentation-
[email protected].
Please include the following information with your feedback:
 Document name
 Release/Version number
 Chapter/Section name
 Topic title (for Online Help)
 Brief description of the content (for example, incomplete/incorrect information, grammatical
errors, information that requires clarification or more details and so on).
 Your suggestions for correcting/improving documentation
You may also provide your feedback about MSC Software documentation by taking a short 5-minute
survey at: http://msc-documentation.questionpro.com.

Note: The above mentioned e-mail address is only for providing documentation specific
feedback. If you have any technical problems, issues, or queries, please contact Technical
Support.
 Contents
MSC Nastran Reference Guide
Contents

Contents
Preface
How to Use this Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Prerequisites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Goals of this Guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
List of MSC Nastran Guides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
Using other Manuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Typographical Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Accessing MSC Nastran Manuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Downloading the PDF Documentation Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Navigating the PDF Files. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Printing the PDF Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Training and Internet Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
Technical Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Visit SimCompanion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
Help Us Help You . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1 nast20214 Command and Input Data

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
nast20214 Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
The NASTRAN Statement (Optional) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
File Management Section (FMS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Executive Control Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Case Control Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Case Control Section Requirements for Solutions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
Bulk Data Section. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Use of Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
Parameters in the Superelement Solution Sequences. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Main Index
 Contents v

2 Grid Points and Coordinate Systems

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Grid Point and Coordinate System Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
Grid Point Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3 Structural Elements
Element and Property Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Elements with Offsets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Axisymmetric Elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Material Properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Line Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Beam Element (CBEAM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
Three-Node Beam Element (CBEAM3). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Bar Element (CBAR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Bend Element (CBEND) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
Rod Elements (CROD, CONROD, CTUBE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Supplied Beam and Bar Libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
Adding Your Own Beam Cross-Section Library . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
Support for Offsets in Element Coordinates for CBAR & CBEAM Elements . . . . . . . . . . . . . . . . . . . . . . . . . 120
Surface Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Shear Panel Element (CSHEAR). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Shell Elements (CTRIA3, CTRIA6, CTRIAR, CQUAD4, CQUAD8, CQUADR) . . . . . . . . . . . . . . . . . . . . . . . . . . 125
QUADR and TRIAR Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
Conical Shell Element (RINGAX) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Shell Elements Corner Thickness Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
SHELL Structural Damping . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Solid Elements (CTETRA, CPENTA, CHEXA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Rigid Elements and Multipoint Constraints (R-type, MPC). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Rigid Element Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
Scalar Elements (CELASi, CMASSi, CDAMPi) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
Axisymmetric Solid Elements (CTRIAX6, CTRIAX, CQUADX). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
Crack Tip Elements (CRAC2D, CRAC3D) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
Gap and Line Contact Elements (CGAP, BCONP, BLSEG) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
Concentrated Mass Elements (CONM1, CONM2). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
General Element Capability (GENEL) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Direct Matrix Input on Grid Points (DMIG). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Hyperelastic Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 196
Bushing Elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200

Main Index
 vi MSC Nastran Reference Guide

Bush Element (CBUSH). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201

Bush1D Element (CBUSH1D) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Connector Elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
CWELD Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
Additional Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
Formulation Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
Connector Stiffness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Connector Control: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
CFAST Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
CSEAM Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
Limitations and Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235
CINTC Element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Element Geometry Checks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
ELSUM Case Control Command. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
Method and Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
Inputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252
Outputs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Guidelines and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
Example. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255

4 Material Properties
Isotropic Material. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
Orthotropic Material. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
Anisotropic Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
Nonuniform Material and Composite Properties for Shells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 264
Classical Lamination Theory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Transverse Shear Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
Data Recovery. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Failure Indices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Strength Ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
“LAM” Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Temperature-Dependent Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292
Hyperelastic Material. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
Reference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 297

5 Applied Loads
Static Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299
Point Loads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

Main Index
 Contents vii

Distributed Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299

Inertia Loads. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
Thermal and Enforced Element Deformation Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 300
Loads Due to Enforced Motion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
Frequency-Dependent Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
Time-Dependent Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 307
Simplified Static Loading Data in Dynamic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
Combining of Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
Summary of Nonlinear Static Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311
Summary of Nonlinear Transient Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316

6 Constraint and Set Notation

Degree of Freedom Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Degree-of-Freedom Set Definitions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Degree-of-Freedom Set Bulk Data Entries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Multipoint Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
Single Point Constraints. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
Rigid Body Supports. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331
Sets for Dynamic Reduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
Sets for Aerodynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 332
Rules of Sets for Undefined Degrees of Freedom . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
Output Selection Via Set Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 334

7 Solution Sequences
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
Output Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
MSC Nastran Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 340
Restart Procedures. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 360
General Operations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
Geometry Processing in SubDMAP PHASE0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362
Matrix Assembly Operations in SubDMAP SEMG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
Multipoint Constraint Operations in SubDMAPs SEKR and SEMR2 and SEMRB . . . . . . . . . . . . . . . . . . . . . 366
Single Point Constraint Operations in SubDMAPs SEKR, SEMR2 and SEMRB. . . . . . . . . . . . . . . . . . . . . . . 368
Constraint and Mechanism Problem Identification in SubDMAP SEKR . . . . . . . . . . . . . . . . . . . . . . . . . . . . 370
Static Condensation in SubDMAPs SEKR and SEMR2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 373
Static Solutions in SubDMAP SEKRRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 376

Main Index
 viii MSC Nastran Reference Guide

Static and Dynamic Load Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380

Dynamic Reduction and Component Mode Synthesis in SubDMAP SEMR3 . . . . . . . . . . . . . . . . . . . . . . . . 383
Formulation of Dynamic Equations in SubDMAP GMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 387
Data Recovery Operations in SubDMAP SEDISP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
Data Recovery Operations in SubDMAP SEDRCVR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 397
Summary of Solution Sequence Operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
Static Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 398
Real Eigenvalue Analysis in SubDMAPs SEMR3 and MODERS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 408
Complex Eigenvalue Analysis in SubDMAP CEIGRS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Frequency Response and Random Analysis in SubDMAP FREQRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423
Transient Response in SubDMAPs SEDTRAN and SEMTRAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
Buckling Analysis in SubDMAP MODERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 424
Superelement Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
Nonlinear Heat Transfer Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 438

8 Element Data Recovery Resolved at Grid Points

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
Stress Recovery at Grid Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 442
Description of Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
Topological Method : Plate Elements and Solid Elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444
Geometric Method : Surface Elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 447
Treatment of Exception Points : BRANCHing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
User Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 451
Grid Point Stress Output Description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452
General Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 453
Mesh Stress Discontinuities at Grid Points. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454

9 DMAP
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 460

10 Database Concepts
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462
Delivery Databases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
Dbset Deletion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
Database Autoassignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 464
Database Project and Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
Database Archival, Compression, and Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465
Database Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 465

Main Index
 Contents ix

Database Transfer Across Different Computer Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 466

11 Plotting
General Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
Superelement Plotting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 468
Structure Plotter Coordinate System and Orthographic Projection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 473
Perspective Projection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 476
Stereoscopic Projection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 478
Projection Plane to Plotter Transformations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 478
Deformation Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
Postprocessors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479

12 Interface With Other Programs

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 482
POST Case Control Command . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 482
PUNCH Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488
Patran Support of Heat Transfer Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519
Loads and Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 520
Results Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 520
Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 520
Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
The Nastran HDF5 Result Database (NH5RDB) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 521
Data Schema . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 522
Nastran Input and Output in NH5RDB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524
Nastran Solution Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 525
NH5RDB Optional Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526
NH5RDB DMAP modules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527
NH5RDB Data Access . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532
Schema document . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 581

A Numerical Accuracy Considerations

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
Linear Equation Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584
Eigenvalue Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 584

Main Index
 x MSC Nastran Reference Guide

Matrix Conditioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585

Definiteness of Matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 586
Numerical Accuracy Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 586
Sources of Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587
Sources of Nonpositive Definite Matrices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 588
Detection and Avoidance of Numerical Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589
Static Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 589
Eigensolutions Using the Inverse Iteration and Lanczos Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 590
Eigensolutions Using Transformation Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 590
Frequency Response Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 591
Transient Response Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 591
The Large Mass Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 591
The Lagrange Multiplier Method. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 592

B Theory of Eigenvalue Extraction

Sturm Sequence Theory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 594
Eigensolution Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 594

C Nonlinear Analysis
General Description of Nonlinear Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 609
Rudiments of User Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 612
Material Nonlinearity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623
Nonlinear Elastic Material ............................................................. 624
Hyperelastic Material. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 625
Elastic-Plastic Material ............................................................... 633
Viscoelastic Material in Nonlinear Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 636
Temperature-Dependent Material ....................................................... 637
Geometric Nonlinearity.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 638
Nonlinear Elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 638
Fully Nonlinear Hyperelastic Elements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655
Bushing Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 667
Adaptive Gap Element. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 671
3-D Slideline Contact Capability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 675
Nonlinear Static Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 684
Examples of Convergence Tolerance Changes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 690
Nonlinear Modal Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 698
Differential Stiffness and Follower Force Stiffness in Linear and Nonlinear Solution Sequences . . . . . . . . 701
Buckling Analysis Enhancement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 702
Nonlinear Transient Response Analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 712
Performance in SOL 129 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 725

Main Index
 Contents xi

Nonlinear Analysis With Superelements. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 732

Grid Point Force Balance and Element Strain Energy in Nonlinear Static Analysis . . . . . . . . . . . . . . . . . . . 732

D Acoustic Cavity Modeling

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 738
Assumptions and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 739
Acoustic Cavity Example Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 740

E Cyclic Symmetry
An Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746
Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 746
A-Frame (Dihedral Symmetry ). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 747
Pinwheel (Rotational Symmetry ). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 747
N-Gons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
Three-Dimensional Bodies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 748
Cyclic Symmetry versus Traditional Reflective Symmetry Techniques. . . . . . . . . . . . . . . . . . . . . . . . . . . . 749
Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 750
Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 751
Symmetrical Components. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 752
Transformation from Physical Components to Cyclic Components (ROT) . . . . . . . . . . . . . . . . . . . . . . . . . . 754
Transformation from Cyclic Components to Physical Components (ROT) . . . . . . . . . . . . . . . . . . . . . . . . . . 754
Boundary Conditions and Problem Solution (ROT) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 755
Summary of Transformations, Boundary Conditions, and Problem Solution (DIH). . . . . . . . . . . . . . . . . . . . 760
Transformation from Physical Components to Cyclic Components (DIH) . . . . . . . . . . . . . . . . . . . . . . . . . . 762
Transformation from Cyclic Components to Physical Components (DIH) . . . . . . . . . . . . . . . . . . . . . . . . . . 762
Boundary Conditions and Problem Solution (DIH) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762
Using Cyclic Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 764
Solution of Axisymmetric Problems. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 766

F Local Adaptive Mesh Refinement

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 771
Adaptive Mesh Refinement Loop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 771
Refinement by Regular Subdivision. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 773
Location of New Grid Points . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 776
Hanging Nodes and Multipoint Constraints on Hanging Nodes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 779
Selection of Elements to Refine. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 783
Refinement Criteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 784

Main Index
 xii MSC Nastran Reference Guide

Propagation of Refinement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 793

Transference of Analysis Data Between Unrefined and Refined Meshes . . . . . . . . . . . . . . . . . . . . . . . . . 798
Transference of Element Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 798
Transference of Distributed Loads and Concentrated Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 801
Transference of Displacement Coordinate System, Displacement Boundary Conditions, and Constraints . 801
Detection of Geometric Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804
Detection of Material, and Superelement Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 811

Main Index
 Preface

Preface


How to Use this Guide 2

List of MSC Nastran Guides 2

Using other Manuals 3

Typographical Conventions 4

Accessing MSC Nastran Manuals 5
 Training and Internet Resources 6

Technical Support 7

Main Index
 2 MSC Nastran Reference Guide
How to Use this Guide

How to Use this Guide

This guide is written for a any MSC Nastran user. No prior experience with commercial finite element software is assumed,
and no finite element-specific university coursework is required.

Prerequisites
It is assumed that you have a bachelor’s degree in any of the fields relevant to structural analysis: mechanical engineering,
civil engineering, engineering mechanics, or the equivalent.

Goals of this Guide

This guide provide you with the reference information regarding MSC Nastran input data, grind points and coordinate
system, structure elements, material properties, loads, constraints, solution sequence, DMA, data base concepts, and
plotting.
It also covers some basic theory regarding nonlinear analysis, local adaptive mesh refinement, accuracy considerations,
theory of Eigenvalue extraction, and acoustic cavity modeling.

List of MSC Nastran Guides

A list of some of the MSC Nastran guides is as follows:

Installation and Release Guides

 Installation and Operations Guide
 Release Guide
Reference Guides
 Quick Reference Guide
 DMAP Programmer’s Guide
 Reference Guide
 Utilities Guide
Demonstration Guides
 Linear Analysis
 Implicit Nonlinear (SOL 400)
 Explicit Nonlinear (SOL 700)
 MSC Nastran Verification Guide
User’s Guides
 Getting Started
 Linear Static Analysis
 Dynamic Analysis

Main Index
 Preface 3
Using other Manuals

 Embedded Fatigue
 Embedded Vibration Fatigue
 Thermal Analysis
 Superelements and Modules
 Design Sensitivity and Optimization
 Rotordynamics
 Implicit Nonlinear (SOL 400)
 Explicit Nonlinear (SOL 700)
 Aeroelastic Analysis
 User Defined Services
 Non Linear (SOL 600)
 High Performance Computing
 DEMATD

You may find any of these documents from MSC Software at:
http://simcompanion.hexagon.com

Using other Manuals

After reading the Getting Started Guide, we recommend that you go through either the Linear Static Analysis User's Guide,
the Dynamic Analysis User's Guide, or the Nonlinear User's Guide depending on the type of simulation you are interested in
performing. Details of some of the manuals are listed here for your reference.
This guide contains many excerpts from the Quick Reference Guide (QRG). Most of the excerpts have been edited-some
extensively-to eliminate material that is not relevant to the topics covered in this book.
 MSC Nastran Quick Reference Guide: The QRG contains a complete description of all the input entries for MSC
Nastran. It contains complete descriptions of all the finite element input data and options available in MSC
Nastran. Each entry provides a description, formats, examples, details on options, and general remarks. You will
find the full descriptions for all SOL 400 input entries in the QRG.
 MSC Nastran Reference Guide: It provides supporting information that relates to the theory of MSC Nastran inputs,
element libraries, and loads and boundary conditions.
 MSC Nastran Linear Static Analysis User’s Guide: It provides support information on the basic use of MSC Nastran
which can also be applied to SOL 400.
 Nonlinear User’s Guide SOL 400: This guide provides a complete background to MSC Nastran Implicit Nonlinear
(SOL 400) and describes how to use SOL 400 within the MSC Nastran environment. The theoretical aspects of
nonlinear analysis methods, types, and techniques are included along with thorough descriptions for nonlinear
material models, properties, and loads and constraints.

Main Index
 4 MSC Nastran Reference Guide
Typographical Conventions

 Dynamic Analysis User’s Guide: This guide describes the proper use of MSC Nastran for solving various dynamic
analysis problems. This guide serves as both an introduction to dynamic analysis for the new user and a reference
for the experienced user.
 MSC Nastran High Performance Computing User’s Guide: It describes the strategy for getting the best performance out of
MSC Nastran for SOL 101, 103, 108, 111, and 400. The guide also gives details on selecting the hardware system
that provides the optimal performance for your models.
 DMAP Programmer’s Guide: MSC Nastran provides a wide variety of solution sequences. Each solution sequence
consist of a series of DMAP (Direct Matrix Abstraction Program) statements. DMAP is a high level language with
its own compiler and grammatical rules. MSC Nastran allows you to modify the solution sequences or to write
your own solution sequences using DMAP.
This guide contains many highlighted links (in blue) to other MSC Nastran documents and all the documents were
delivered together as a collection. If you keep the collection together the links between documents will work.
Two ways of working with links are as follows:
 Use alt  to return back to the window your curser is in
 Open the other “linked to” documents in a new window from an Adobe Reader
a. Choose Edit  Preferences  Documents 
b. Open cross-document links in the same window.
c. Un-check the checkbox.
d. Select OK.

Typographical Conventions
The section provides a brief overview of the typographical conventions used in the document to help the user better follow
the MSC Nastran documentation.
This section describes some syntax that will help you in understanding text in the various chapters and thus in facilitating
your learning process. It contains stylistic conventions to denote user action, to emphasize particular aspects of a MSC
Nastran run or to signal other differences within the text.
Courier New Represents command-line options of MSC Nastran and results from f04/f06 files.

Example: nast20170 memorymax=16gb myjob.dat

Quoted Text Represents command-line options of MSC Nastran for in-line text.

Example: memorymax=16gb
Arial font To represent elements.

Example: RBE3 and RSPLINE are interpolation elements and are not rigid.
Red Text Represents items in the examples that we to emphasize.

Example: smp=16

Main Index
 Preface 5
Accessing MSC Nastran Manuals

Bold Text Represents items in the text that we want to emphasize.

Example: dmp=4
Italic Text Represents references to manuals/documents.

Note: Since there is no user interface in MSC Nastran, we shall use bold font to emphasize.

Accessing MSC Nastran Manuals

This section describes how to access the MSC Nastran documentation outside of MSC Software. MSC Nastran
documentation is available through PDF files. The PDF files can be obtained from the following sources:
 MSC Nastran documentation installer
 SimCompanion
 Combined documentation
The PDF documentation files are appropriate for viewing and printing with Adobe Acrobat Reader (version 5.0 or higher),
which is available for most Windows and Linux systems. These files are identified by a .pdf suffix in their file names.

Downloading the PDF Documentation Files

You can download the PDF documentation from SimCompanion (http://simcompanion.hexagon.com).

Navigating the PDF Files

For the purpose of easier online document navigation, the PDF files contain hyperlinks in the table of contents and index.
In addition, links to other guides, hyperlinks to all cross-references to chapters, sections, figures, tables, bibliography, and
index entries have been applied.

Printing the PDF Files

Adobe Acrobat PDF files are provided for printing all or part of the manuals. You can select the paper size to which you are
printing in Adobe Acrobat Reader by doing the following:
1. Click File.
2. Select the Print.... option. The Print dialog box is displayed.
3. Select Page Setup....
4. Choose the required paper size in the Page Setup menu.
The PDF files are recommended when printing long sections since the printout will have a higher quality.
If the page is too large to fit on your paper size, you can reduce it by doing the following:
1. Select the File -> Print.

Main Index
 6 MSC Nastran Reference Guide
Training and Internet Resources

2. Under Page Scaling, choose the Shrink to Printable Area option.

Training and Internet Resources

Information about MSC's products, services and latest events is available on our website www.mscsoftware.com.
The information about MSC Seminars is available on the training link http://www.mscsoftware.com/msc-training. You can also
use this link to schedule the seminars.
If you are a new MSC Nastran user, we recommend the following courses:

NAS 102A: Dynamic Analysis using MSC Nastran

Through this course you will learn the following:
 Fundamentals of dynamics
 Normal modes
 Direct and modal methods
 Frequency response
 Transient response
 Damping
 Enforced motion
 Response spectrum
 Random analysis
 Residual vectors

Main Index
 Preface 7
Technical Support

NAS 102B: Advanced Dynamic Analysis using MSC Nastran

This course will help you learn the following:
 Review and expand upon the basic dynamic analysis topics.
 Introduce model checkout tools for dynamics.
 Introduce advanced dynamic analysis capabilities.
 Show where to get technical support.
 Show the available training for MSC Nastran.

NAS 124: Thermal Analysis using MSC Nastran and Patran

This course will help you learn the following:
 Heat transfer
 Modes of heat transfer
 Types of analysis
 Thermal model elements and material properties
 Thermal model boundary conditions
 Thermal model loads
 Steady and transient state equation

NAS120: Linear Static Analysis using MSC Nastran and Patran

This seminar provides the foundation required for intermediate and advanced MSC Nastran applications:
 Linear static
 Normal modes
 Buckling analysis of structures with the use of MSC Nastran and Patran
In this course MSC Nastran data structure, the element library, modeling practices, model validation, and guidelines for
efficient solutions are discussed and illustrated with examples and workshops. Patran is an integral part of the examples and
workshops. It will be used to generate and verify illustrative MSC Nastran models, manage analysis submission requests,
and to visualize results.

Technical Support
If you encounter difficulties while using MSC Nastran, first please refer to the section(s) of the manual containing
information on the commands you are trying to use or the type of problem you are trying to solve.

Visit SimCompanion
The product documentation is available in SimCompanion (http://simcompanion.hexagon.com). The SimCompanion gives
you access to a wealth of resources for MSC Software products. You will find various information such as:
 Product documentations

Main Index
 8 MSC Nastran Reference Guide
Technical Support

 Knowledge base articles

 Product error lists (fixed and known issues for each release)
 SimAcademy webinars
 Product and support contact information
SimCompanion is a searchable database which allows you to find articles relevant to your inquiry. Valid MSC customer
entitlement and login is required to access the database and documents. It is a single sign-on that gives you access to product
documentation for complete list of products from MSC Software, allows you to manage your support cases, and participate
in our discussion forums.

Help Us Help You

Clients frequently call up the support engineers at MSC Software with enquiry regarding models that do not run correctly.
Our technical support staff can help you much more efficiently and effectively if you are working with a small model, since
debugging a small model is much easier, and the turnaround time to rerun a (hopefully) corrected test model is minutes
rather than hours.
 For information on the latest events, products and services for all products, refer to the MSC Software corporate
site (www.mscsoftware.com).
 For technical support phone numbers and contact information, please visit:
https://simcompanion.hexagon.com/customers/s/article/support-contact-information-kb8019304.

Main Index
 Chapter 1: nast20214 Command and Input Data

1 nast20214 Command and

Input Data


Introduction 10

nast20214 Command 13

The NASTRAN Statement (Optional) 13

File Management Section (FMS) 14

Executive Control Section 14
 Case Control Section 14

Bulk Data Section 31

Use of Parameters 31

Main Index
 10 MSC Nastran Reference Guide
Introduction

Introduction
nastran Command
The recommended procedure for invoking MSC Nastran is to use the nastran command. The nastran command is
described in Executing MSC Nastran in the MSC Nastran Quick Reference Guide.

Input Data
MSC Nastran input requires records that are 80 characters (or columns) in length. The input file is comprised of five
sections, which must be assembled in the sequence shown in Figure 1-1.

Figure 1-1 Structure of the MSC Nastran Input File

Main Index
 Chapter 1: nast20214 Command and Input Data 11
Introduction

The records of the first four sections are input in free-field format and only columns 1 through 72 are used for data. Any
information in columns 73 through 80 may appear in the printed echo, but will not be used by the program. If the last
character in a record is a comma, then the record is continued to the next record.
The NASTRAN statement is optional and is used in special circumstances (see the Executing MSC Nastran in the MSC Nastran
Quick Reference Guide).
The File Management Section (FMS) is optional and typically follows the NASTRAN statement(s). It ends with the
specification of an Executive Control statement. This section provides for database initialization and management, along
with job identification and restart conditions. The File Management statements are described in the File Management
Statements in the MSC Nastran Quick Reference Guide.
The Executive Control Section begins with the first Executive Control statement and ends with the CEND delimiter. It
identifies the job and the type of solution to be performed. It also declares the general conditions under which the job is to
be executed, such as maximum time allowed and type of system diagnostics desired. If the job is to be executed with a
solution sequence, the actual solution sequence is declared, along with any alterations to the solution sequence that may be
desired. If DMAP (Direct Matrix Abstraction Program) is used, the complete DMAP sequence must appear in the
Executive Control Section. The Executive Control statements and examples of their use are described in the Executive Control
Statements in the MSC Nastran Quick Reference Guide.
The Case Control Section follows CEND and ends with the BEGIN BULK delimiter. It defines the subcase structure for
the problem, defines sets of Bulk Data, and makes output requests for printing, punching and plotting. A general discussion
of the functions of the Case Control Section and a detailed description of the commands used in this section are given in
the Case Control Command Descriptions in the MSC Nastran Quick Reference Guide. The special requirements of the Case
Control for each solution sequence are discussed in Restart Procedures.
The Bulk Data Section follows BEGIN BULK and ends with the ENDDATA delimiter. It contains all of the details of the
model and the conditions for the solution. BEGIN BULK and ENDDATA must be present even though no new Bulk Data
is being introduced into the problem or if all of the Bulk Data is coming from an alternate source, such as user-generated
input. The BEGIN BULK entry is in free field format. The ENDDATA delimiter must begin in Column 1 or Column 2.
The input file may be formed by the insertion of other files, with the INCLUDE statement. This may be specified in any
of the five parts of the input file. Note, however, the data referenced by the INCLUDE statement must be appropriate for
that section of the input file.
Comments may be inserted in any of the parts of the input file. They are identified by a dollar sign ($) in any column
followed by any characters out to Column 80. Columns 2 through 72 may contain any desired text.
Table 1-1 gives valid characters for MSC Nastran.

Main Index
 12 MSC Nastran Reference Guide
Introduction

Table 1-1 ASCII Subset for MSC Nastran

Character ASCII (HEX) Character ASCII (HEX)
HT (Horizontal 09 @ 40
Tabulation) A 41
SP (Space “blank”) 20 B 42
C 43
! 21 D 44
" 22 E 45
# 23 F 46
$ 24 G 47
% 25 H 48
& 26 I 49
‘ 27 J 4A
( 28 K 4B
) 29 L 4C
* 2A M 4D
+ 2B N 4E
’ 2C O 4F
- 2D P 50
. 2E Q 51
/ 2F R 52
0 30 S 53
1 31 T 54
2 32 U 55
3 33 V 56
4 34 W 57
5 35 X 58
6 36 Y 59
7 37 Z 5A
8 38 [ 5B
9 39 \ 5C

Main Index
 Chapter 1: nast20214 Command and Input Data 13
nast20214 Command

Table 1-1 ASCII Subset for MSC Nastran (continued)

Character ASCII (HEX) Character ASCII (HEX)
: 3A ] 5D
; 3B ^ 5E
< 3C - 5F
= 3D
> 3E
? 3F

nast20214 Command
MSC Nastran is executed from the command line using the command nast20214. This command sets environment
variables in order for MSC Nastran to correctly execute on your system and it also has capabilities to predict memory and
solvers (as of MSC Nastran 2020 with solve=auto). The basic format of the nast20214 command is:
nast20214 input_data_file keywords
nast20214 input_data_file [keyword1=value1 keyword2=value2 …]
where input_data_file is the name of the file containing the input data and keyword=valuei is one or more optional keyword
assignment arguments. For example, to run an MSC Nastran job using the data file example1.dat, enter the following
command:
nast20214 example1.dat
Various options to the nast20214 command are available using keywords. Keyword assignments consist of a keyword,
following by an equal sign, followed by a keyword value. For example,
nast20214 example1.dat memorymax=16gb

Note: In Windows you can use a hash mark # instead of the equal sign. This is useful if
nast20214 example1.dat memorymax=16gb command is placed in .bat
file.

nast20214 example1.dat memorymax=16gb

The NASTRAN Statement (Optional)

The NASTRAN statement is used to change the default values for certain operational parameters (also called system cells).
See the Executing MSC Nastran in the MSC Nastran Quick Reference Guide for a description of the format and most often used
system cells. The NASTRAN statement is used for special circumstances and is therefore not needed in most runs. The
NASTRAN statement may also be specified in the Runtime Configuration (RC) files at the system, user, and job level. See
the MSC Nastran Installation and Operation Guide.

Main Index
 14 MSC Nastran Reference Guide
File Management Section (FMS)

File Management Section (FMS)

The File Management Section (FMS) is primarily intended for the attachment and initialization of Database sets (DBsets)
and FORTRAN files. The initialization of DBsets includes specification of their maximum size, member names, and
physical filenames. The initialization of FORTRAN files includes the specification of their filenames, FORTRAN unit
numbers, and FORTRAN attributes.
In most classes of problems that use MSC Nastran solution sequences, no File Management statements are required because
a default File Management Section is executed at the beginning of every run. The default File Management Section is
described in Database Concepts. For details, see File Management Statements in the MSC Nastran Quick Reference Guide.

Executive Control Section

The Executive Control statements, in general, select the solution sequences to be run and various diagnostics. The input
specifications are given in Executive Control Statements in the MSC Nastran Quick Reference Guide.
Most statements are order independent. Exceptions are the COMPILE, COMPILER, ALTER, ENDALTER, and LINK
statements. The LINK statement must appear after all COMPILE statements. The COMPILER statement (or equivalent
DIAGs) must appear before all COMPILE statements. This statement sets the default print control for subsequent COMPILE
statements.

Case Control Section

The Case Control Section defines the subcase structure for the problem, makes selections from the Bulk Data Section, and
makes output requests. Grid point stress and stress discontinuities output request are treated separately in Stress Recovery at
Grid Points and Mesh Stress Discontinuities at Grid Points, respectively. Structure plotter output requests and curve plot requests
are treated separately in Section 11.2 and Section 11.3, respectively.
The Case Control commands are described in the Case Control Command Descriptions in the MSC Nastran Quick Reference
Guide.
The Case Control Section has several basic functions:
 Selects loads and constraints.
 Requests printing, plotting, and/or punching of input and output data (plotter commands are described in
Plotting ).
 Defines the subcase structure for the analysis.

Case Control Section Requirements for Solutions

This section describes the options and requirements in the Case Control Section for the various solution sequences.
In general, a separate subcase is defined for each loading condition and/or each set of constraints. Subcases may also be used
in connection with output requests, such as requesting different output for each mode in a real eigenvalue problem. Only

Main Index
 Chapter 1: nast20214 Command and Input Data 15
Case Control Section

one level of subcase definition is provided. All items placed above the subcase level (ahead of the first subcase) will be used
for all following subcases unless overridden within the individual subcase.

Case Control Commands for Solutions

This section describes the Case Control structure for the various solution sequences.
Static Analysis (SOLutions 101, 144, 200, and 400)

 A separate subcase must be defined for each unique combination of constraints and static loads. Multiple
constraint sets are permitted in SOLutions 101, and 200 (for ANALYSIS = STATICS only) and for residual
structures only.
 A static loading condition must be defined for each subcase with a LOAD, TEMPERATURE(LOAD) or DEFORM
selection unless all loading is specified with grid point displacements on SPC entries. In SOLution 144, a loading
condition is optional, but a trim variable constraint must be selected with the TRIM command.
 An SPC set must be selected for each subcase, unless the model is a properly supported free body, or all constraints
are specified on GRID entries, scalar connection entries, or with general elements.
 REPCASE may be used to repeat subcases in order to allow multiple sets of the same output item.
 Direct input matrices may be selected with K2GG and/or M2GG above the subcase level. P2G is allowed in all the
above SOLutions.
 In SOLutions 101 and 200, the SUPORT1 command may be used to select SUPORT1 Bulk Data entries for inertia
relief analysis. See Superelement Subcase Structure for additional Case Control requirements for superelements.
 The SMETHOD command may be used to select parameters for an iterative solution technique.
In statics problems, provision has been made for the combination of the results of several subcases. This is convenient for
studying various combinations of individual loading conditions and for the superposition of solutions for symmetrical and
antisymmetrical boundaries.
The following examples of Case Control indicate typical ways of defining subcases:
 Static analysis with multiple loads and boundary conditions. Multiple boundary conditions are supported only in
SOLutions 101, 103, 105, 145, and 200. See the BC Case Control command.

Main Index
 16 MSC Nastran Reference Guide
Case Control Section

DISPLACEMENT = ALL
MPC = 3
SUBCASE 1
SPC = 2
TEMPERATURE(LOAD) = 101
LOAD = 11
SUBCASE 2
SPC = 2
DEFORM = 52
LOAD = 12
SUBCASE 3
SPC = 4
LOAD = 12
SUBCASE 4
MPC = 4
SPC = 4
Four subcases are defined in this example. The displacements at all grid points will be printed for all four subcases.
MPC = 3 will be used for the first three subcases and will be overridden by MPC = 4 in the last subcase. Since the
constraints are the same for subcases 1 and 2 and the subcases are contiguous, the static solutions will be performed
simultaneously. In SUBCASE 1, thermal load 101 and external load 11 are internally superimposed, as are the
external and deformation loads in SUBCASE 2. In SUBCASE 4, the static loading will result entirely from enforced
displacements of grid points.
 Linear combination of subcases
SPC = 2
SET 1 = 1 THRU 10,20,30
DISPLACEMENT = ALL
STRESS = 1
SUBCASE 1
LOAD = 101
OLOAD = ALL
SUBCASE 2
LOAD = 201
OLOAD = ALL
SUBCOM 51
SUBSEQ = 1.0,1.0
SUBCOM 52
SUBSEQ = 2.5,1.5
Two static loading conditions are defined in subcases 1 and 2. SUBCOM 51 defines the sum of subcases 1 and 2.
SUBCOM 52 defines a linear combination consisting of 2.5 times SUBCASE 1 plus 1.5 times SUBCASE 2. The
displacements at all grid points and the stresses for the element numbers in SET 1 will be printed for all four
subcases. In addition, the nonzero components of the static load vectors will be printed for subcases 1 and 2.
 Statics with one plane of symmetry. (Note that this is not supported in the superelement solution sequences.)

Main Index
 Chapter 1: nast20214 Command and Input Data 17
Case Control Section

SET 1 = 1,11,21,31,51
SET 2 = 1 THRU 10, 101 THRU
110
DISPLACEMENT = 1
ELFORCE = 2
SYM 1
SPC = 11
LOAD = 21
OLOAD = ALL
SYM 2
SPC = 12
LOAD = 22
SYMCOM 3
SYMCOM 4
SYMSEQ 1.0,-1.0
Two SYM subcases are defined in subcases 1 and 2. SYMCOM 3 defines the sum and SYMCOM 4 the difference of the
two SYM subcases. The nonzero components of the static load will be printed for SUBCASE 1 and no output is
requested for SUBCASE 2. The displacements for the grid point numbers in SET 1 and the forces for elements in
SET 2 will be printed for subcases 3 and 4.
 Use of REPCASE in statics problems
SET 1 = 1 THRU 10, 101 THRU 110,201 THRU 210
SET 2 = 21 THRU 30,121 THRU 130,221 THRU 230
SET 3 = 31 THRU 40,131 THRU 140,231 THRU 240
SUBCASE 1
LOAD = 10
SPC = 11
DISP = ALL
SPCFORCE = 1
FORCE = 1
REPCASE 2
FORCE = 2
REPCASE3
FORCE=3
This example defines one subcase for solution and two subcases for output control. The displacements at all grid
points and the nonzero components of the single-point forces of constraint along with forces for the elements in
SET 1 will be printed for SUBCASE 1. The forces for elements in SET 2 will be printed for REPCASE 2 and the
forces for elements in SET 3 will be printed for REPCASE 3.
Normal Modes (SOLutions 103, 111, 200, and 400)

 METHOD must be used to select an EIGR or EIGRL entry that exists in the Bulk Data Section. In SOLutions 103
and 200 multiple subcases may be specified with different sets of METHOD, MPC, SPC, and SUPORT1 commands.
In all other SOLs only one set may be specified. (See the BC Case Control command.)
 An SPC set must be selected unless the model is a free body or all constraints are specified on GRID entries, scalar
connection entries or with general elements.

Main Index
 18 MSC Nastran Reference Guide
Case Control Section

 Multiple subcases may be used to control output requests. A single subcase is sufficient if the same output is
desired for all modes. If multiple subcases are present, the output requests will be honored in succession for
increasing mode numbers. MODES may be used to repeat subcases in order to make the same output request for
several consecutive modes.
 Direct input matrices may be selected with K2GG and/or M2GG above the subcase level.
 In SOLutions 103 and 200, the SUPORT1 command may be used to select SUPORT1 Bulk Data entries for
computing rigid body modes.
 See Superelement Subcase Structure for additional Case Control requirements for superelements.
 Use of MODES in Eigenvalue Problems
METHOD = 2
SPC = 10
SUBCASE 1
DISP = ALL
STRESS = ALL
MODES = 2
SUBCASE 3
DISP = ALL
In this example, the displacements at all grid points will be printed for all modes. The stresses in all elements will
be printed for the first two modes. If subcases are not present, output requests act on all computed eigenvectors.
 Use of REPCASE in Normal Modes Analysis
In normal modes analysis the REPCASE command is applied only to the last mode in the previous subcase.

SET 200 = 1,7,10,15

DISP(PLOT) = ALL
SUBCASE 1
LABEL = MODE 1
REPCASE 2
LABEL = REPCASE FOR MODE 1
DISP = 200
SUBCASE 3
LABEL = MODE 2
REPCASE 4
LABEL = REPCASE FOR MODE 2
DISP = 200
This example requests the output of the first two eigenvectors. In the SUBCASEs, output is computed at all points
for plotting purposes. In the REPCASEs, output is printed only at selected points.
 Prestressed structures may be analyzed by specifying static subcases to define the pre-loads. Then the
STATSUB(PRELOAD) command that references the desired static subcases must be specified in the dynamic
response subcases to request the inclusion of differential stiffness in the dynamic formulation. By default, stiffness
due to follower forces is also included in the differential stiffness.
 For fluid-structure analysis, the METHOD(FLUID) and SDAMPING(FLUID) may be used to request eigensolution
and modal damping parameters for the fluid which are different from the structure.

Main Index
 Chapter 1: nast20214 Command and Input Data 19
Case Control Section

Buckling Analysis (SOLutions 105 and 200)

 The Case Control Section must contain at least two subcases. In SOL 105, multiple subcases may be specified
with different sets of METHOD, SPC, and MPC commands. See also the BC Case Control command. The buckling
subcase(s) may have a different boundary condition than the statics subcase(s).
 In SOL 105 and 200, the METHOD and STATSUB commands must be specified in one or more subcases. In
SOLution 105 and 200, two different STATSUB conditions may be selected:
• STATSUB(PRELOAD) for prestressing
• STATSUB(BUCKLE) for determining the buckling loading condition. Both may appear in the same subcase
with a METHOD command.
 A static loading condition must be defined in the first subcase with a LOAD, TEMPERATURE(LOAD), or
DEFORM selection, unless all loading is specified by grid point displacements on SPC entries.
 Output requests that apply to both the static solution and the buckling modes may be placed above the subcase
level.

Modal Formulation (SOLutions 110, 111, 112, 145, 146, 200, and 400)

 An SPC set must be selected above the subcase level unless the model is a free body or all constraints are specified
on GRID entries, scalar connection entries, or with general elements.
 SDAMPING must be used to select a TABDMP1 entry if modal damping is desired. Only one selection is allowed
in superelement solution sequences.
 For SOLutions 110, 111, and 112, STATSUB selects a static solution used to form differential stiffness.
 For SOLution 145, see also the BC Case Control command.
Complex Eigenvalue and Flutter Analysis (SOLutions 107, 110, 145, 200, and 400)

 At least one subcase must be defined for each unique set of direct input matrices (K2PP, M2PP, B2PP). Only one
set is allowed in SOLutions 107 and 110. More than one set may be selected in SOLution 145.
 Multiple subcases for each set of direct input matrices are used only to control output requests. A single subcase for
each set of direct input matrices is sufficient if the same output is desired for all modes. If consecutive multiple
subcases are present for a single set of direct input matrices, the output requests will be honored in succession for
increasing mode numbers. MODES may be used to repeat subcases in order to make the same output request for
several consecutive modes.
 CMETHOD must be used to select an EIGC entry in the Bulk Data Section for each set of direct input matrices.
Only one selection is allowed in SOLutions 107 and 110. In SOLution 145, CMETHOD is required only if
k-method flutter analysis is also selected.
 Constraints must be defined above the subcase level for SOLutions 107 and 110. Multiple boundary conditions
are allowed for SOLution 145 with a BC Case Control command.
 In SOLution 145, FMETHOD must be used to select a FLUTTER entry in the Bulk Data Section.

Main Index
 20 MSC Nastran Reference Guide
Case Control Section

 Prestressed structures may be analyzed by specifying static subcases to define the pre-loads. Then the
STATSUB(PRELOAD) command that references the desired static subcases must be specified in the dynamic
response subcases to request the inclusion of differential stiffness in the dynamic formulation. By default, stiffness
due to follower forces is also included in the differential stiffness.
 For fluid-structure analysis, the METHOD(FLUID) and SDAMPING(FLUID) may be used to request eigensolution
and modal damping parameters for the fluid which are different from the structure.
 The SMETHOD command may be used to select parameters for an iterative solution technique.
Frequency, Random, and Aeroelastic Response (SOLutions 108, 111, 146, 200, and 400)

 At least one subcase must be defined for each unique set of direct input matrices (K2PP, M2PP, B2PP) or
frequencies. Only one set is allowed in SOLutions 108, 111, 118, 146, and 200.
 Consecutive subcases for each set of direct input matrices or frequencies are used to define the loading conditions -
one subcase for each dynamic loading condition.
 Constraints must be defined above the subcase level.
 DLOAD must be used to select a frequency-dependent loading condition DLOAD and RLOADi Bulk Data entries,
for each subcase. In SOLutions 108, 111, and 146, DLOAD may also select a time-dependent loading condition,
i.e., TLOADi entries.
 FREQUENCY must be used to select a set of FREQ, FREQ1, FREQ2, FREQ3, FREQ4, or FREQ5 entries from
the Bulk Data Section for each unique set of direct input matrices. Only one selection is allowed in SOLutions
118 and 146.
 OFREQUENCY may be used above the subcase level or within each subcase to select a subset of the solution
frequencies for output requests. The default is to use all solution frequencies.
 If Random Response calculations are desired, RANDOM must be used to select RANDPS and RANDTi entries
from the Bulk Data Section. Only one OFREQUENCY and FREQUENCY command can be used for each set of
direct input matrices.
 LOADSET may be used in the first subcase to select a set of LSEQ entries which define a set of static load vectors
to be used as DAREA entries in dynamic analysis.
 Prestressed structures may be analyzed by specifying static subcases to define the pre-loads. Then the
STATSUB(PRELOAD) command that references the desired static subcases must be specified in the dynamic
response subcases to request the inclusion of differential stiffness in the dynamic formulation. By default, stiffness
due to follower forces is also included in the differential stiffness.
 Multiple subcases with different SPC, MPC and SUPORT1 requests may be specified in SOLs 108 and 111.
Transient Response (SOLutions 109, 112, 200, and 400)

 One subcase must be defined for each dynamic loading condition. DLOAD or NONLINEAR must be used to
define a time-dependent loading condition for each subcase. Only one DLOAD and NONLINEAR command is
allowed in SOLution 200.
 Constraints must be defined above the subcase level.

Main Index
 Chapter 1: nast20214 Command and Input Data 21
Case Control Section

 TSTEP must be used to select the time-step intervals to be used for integration and output in each subcase. Only
one TSTEP selection is allowed in SOLution 200.
 If nonzero initial conditions are desired, lC must be used to select a TIC entry in the Bulk Data Section.
 LOADSET may be used in the first subcase to select a set of LSEQ entries which define a set of static load vectors
to be used as DAREA entries in dynamic analysis.
 Prestressed structures may be analyzed by specifying static subcases to define the pre-loads. Then the
STATSUB(PRELOAD) command that references the desired static subcases must be specified in the dynamic
response subcases to request the inclusion of differential stiffness in the dynamic formulation. By default, stiffness
due to follower forces is also included in the differential stiffness.
 At least one subcase must be defined for each unique set of direct input matrices (K2PP, M2PP, B2PP), dynamic
loading (DLOAD) and time step parameters (TSTEP). Only one set of direct input matrices is allowed.
Linear Steady State Heat Transfer (SOLution 101 and 400)

 A separate subcase must be defined for each unique combination of constraints and static loads.
 A static loading condition must be defined for (not necessarily within) each subcase with a LOAD selection, unless
all loading is specified with grid point temperatures on SPC entries.
 An SPC set must be selected for (not necessarily within) each subcase, unless all constraints are specified on GRID
entries or scalar connection entries.
 REPCASE may be used to repeat subcases in order to allow multiple sets of the same output item.
Nonlinear Steady State Heat Transfer (SOLution 153 and 400)

 A separate subcase must be defined for each unique combination of temperature loads (LOAD Case Control
command), temperature constraints (SPC and MPC command), and nonlinear iteration strategy (NLPARM
command). ANALYSIS = HEAT must be specified above the subcase level.
 The LOAD Case Control command references the static temperature load entries: QVECT, QVOL, QHBDY,
and QBDYi.
 An estimated temperature distribution must be defined on TEMP or TEMPD entries and selected by the
TEMPERATURE(INITIAL) Case Control command. Temperature constraints are specified on SPCi entries.
 Restarts are controlled by the PARAMeters SUBCASID, SUBID and LOOPID.
See the MSC Nastran Thermal Analysis User’s Guide for further details.
Nonlinear Transient Heat Transfer (SOLution 159 and 400)

 Only one set of temperature constraints (MPC and SPC Case Control command) may be requested and must be
specified above the subcase level. ANALYSIS = HEAT must be specified above the subcase level.
 A subcase must be defined for each unique combination of dynamic temperature load conditions (DLOAD
command) and nonlinear iteration strategy (TSTEPNL command).

Main Index
 22 MSC Nastran Reference Guide
Case Control Section

 The DLOAD and/or NONLINEAR command must be used to specify a time-dependent loading condition. The
static temperature load entries: QVECT, QVOL, QHBDY, and QBDYi may be used in defining a dynamic load
specified by the TLOADi entry. The set identification number on the static load entries, usually field 2, is
specified in the DAREA field of the TLOADi entry.
 The initial temperature condition is requested by the IC command which selects TEMP or TEMPD entries. The
IC command must be specified above the subcase level.
 Restarts are controlled by the PARAMeters STIME, LOOPID, and SLOOPID.
See the MSC Nastran Thermal Analysis User’s Guide for further details.
Cyclic Symmetry Statics (SOLution 114)

 Each loading condition must have a separate subcase.

 All constraints must be specified above the subcase level.
 A single set of direct input matrices may be selected with K2GG and/or M2GG above the subcase level. P2G is
allowed in SOLution 114.
For details, see Cyclic Symmetry.
Cyclic Symmetry Modes (SOLution 115)

 METHOD must be included above the subcase level to select an EIGR or EIGRL entry that exists in the Bulk
Data Section.
 All constraints must be specified above the subcase level.
 A single set of direct input matrices may be selected with K2GG and/or M2GG above the subcase level.
For details, see Cyclic Symmetry, 703.
Cyclic Symmetry Buckling (SOLution 116)

 Case Control requirements for cyclic symmetry are the same as in SOLution 105. However, only one static and
buckling subcase may be specified.
 For the static subcase requirements see the description under Cyclic Symmetry Statics (SOLution 114).
For details, see Cyclic Symmetry.
Design Sensitivity Analysis and Optimization (SOLution 200)

 SOLution 200 has the same case control requirements as the solution sequences identified on the ANALYSIS Case
Control command for SOL 200. The ANALYSIS command is specified in each subcase to identify the analysis to
which the subcase is to be applied. If superelements are present, then the ANALYSIS command need only be
specified in the residual structure subcases.
 Table 1-2 indicates the analysis types that are available in SOL 200 and whether the types support multiple subcases
and multiple boundary conditions.

Main Index
 Chapter 1: nast20214 Command and Input Data 23
Case Control Section

Table 1-2 Analysis Disciplines Supported in Multidisciplinary Analysis and Design of SOL 200
Multiple Boundary
Analysis SOL* Numbers Multiple Subcases Conditions
Statics 101 Y Y
Normal Modes 103 Y Y
Buckling 105 Y Y
Direct Complex 107 N N
Eigenanalysis
Direct Frequency 108 Y N
Modal Complex 110 N N
Eigenanalysis
Modal Frequency 111 Y N
Modal Transient 112 N N
Static Aeroelasticity 144 Y Y
Flutter 145 Y Y
*The column of Solution Numbers refers to the number MSC Nastran would use to perform the
separate analyses.

 The DESGLB and DESSUB commands select design constraints specified on DCONSTR and DCONADD Bulk Data
entries. The DESGLB selection is independent of subcase and the DESSUB applies only to the current subcase.
 The DESOBJ command selects a scalar design objective specified on a DRESP1, DRESP2, or DRESP3 Bulk Data
entry.
 The optional DESVAR command selects the set of design variables to be used. If this command is not specified, all
DESVAR Bulk Data entries are used.
 The MODTRAK command selects mode tracking parameters.
 The AUXCASE and AUXMODEL indicate subcases to be applied to auxiliary models.
 The DSAPRT command requests the output of design sensitivity coefficients.
See the Design Sensitivity and Optimization User’s Guide for further information.
Nonlinear Static Analysis (SOLution 106)

 A separate subcase must be defined for each unique combination of static loads (LOAD and
TEMPERATURE(LOAD)), constraints (SPC and MPC), and nonlinear iteration strategy (NLPARM and NLPCI).
 A static loading condition may be defined for each subcase with a LOAD and/or TEMPERATURE(LOAD)
selection unless all loading is specified through enforced displacements on SPC entries. If unloading or creep under
residual load is desired, then no static load is specified in the subcase.

Main Index
 24 MSC Nastran Reference Guide
Case Control Section

 If thermal loads are present, then TEMPERATURE(INIT) must be specified above the subcase level.
 An SPC command must be specified unless the model is constrained on the GRID entries.
 A single set of direct input matrices may be selected with K2GG, M2GG, and P2G.
 If upstream superelements have static loads, then LOADSET must be used to select LSEQ entries and, in the
residual structure subcases only, the CLOAD command must be used to assemble the upstream loads.
 Restarts are controlled by PARAMeters LOOPID and SUBID or SUBCASID. See Parameters in the MSC Nastran
Quick Reference Guide.
 The normal, complex, or unsymmetric modes due to the total nonlinear stiffness may be computed at the end of
each subcase as long as PARAM,NMLOOP,<>0 and a METHOD, CMETHOD, or UMETHOD command is
specified in the desired subcase and selects EIGB, EIGR, EIGRL, EIGC, or Bulk Data UEIG entry.
 The buckling modes due to the total nonlinear stiffness may be computed at the end of each subcase as long as
PARAM,BUCKLE,>0 and a METHOD command is specified in the desired subcase and selects an EIGB, EIGR, or
EIGRL Bulk Data entry.
See Nonlinear Analysis for further information.
Nonlinear Transient Response Analysis (SOLution 129)

 A subcase must be defined for each unique combination of dynamic loads (DLOAD) and nonlinear iteration
strategy (TSTEPNL).
 Only one set of constraints (SPC and MPC) may be selected and must be specified above the subcase level.
 LOADSET may be used to select LSEQ entries which reference static load sets to be used in the generation of
dynamic loads.
 Restarts are controlled by PARAMeters STIME, LOOPID, and SLOOPID. See Parameters in the MSC Nastran
Quick Reference Guide.
 Upstream superelements may be reduced dynamically using the METHOD commands.
See Nonlinear Analysis for further information.

Superelement Subcase Structure

Subcases must be defined for each superelement to specify load, constraint, and output selection. The syntax rules described
above also apply for superelements, i.e., requests above the subcase level are overridden by requests inside the subcase level.
See Superelement Case Control in this section for exceptions to the above examples. Static analysis with multiple loads, two
superelements.

DISP = ALL
ESE = ALL
SPC = 1000
SUBCASE 11
LOAD = 1000
SUBCASE 12
LOAD = 1002

Main Index
 Chapter 1: nast20214 Command and Input Data 25
Case Control Section

For most applications, a conventional subcase structure may be used in superelement analysis. The subcase command
SEALL = ALL and SUPER = ALL are the default in superelement solution sequences. SEALL = ALL requests the
program to run to completion. SUPER = ALL instructs the program to use the same subcases for all superelements. See
Restart Procedures, for use of commands alternative to the SEALL command for restarts and for applications where it is
necessary to use more than one SUPER command.
The specification of the METHOD command in a superelement subcase requests the calculation of component modes. In this
example, the same output, constraint, and load requests are used for all superelements and the residual structure. Two
loading conditions are analyzed.

Case Control Commands for Output

Printed output may be requested in either a SORT1 or SORT2 format. The SORT1 format is a tabular listing of the requested
output, sorted by grid point number or element number, for each requested loading condition, forcing frequency or time
step. The SORT2 format is a tabular listing of the requested output, sorted by loading condition, forcing frequency or time
step, for each requested grid point or element.
In any one run, the output requests must be all SORT1 or all SORT2. If both types of output format are desired, a restart
must be made to satisfy the second format request. Any request for deformed structure plots results in a request for the
SORT1 format. Any X-Y output request results in a request for the SORT2 format.
The default for all analyses except transient response is the SORT1 format. The default for transient response is the SORT2
format.
In the case of the SORT1 format for static or transient analysis, there is an additional option to sort the stresses on the basis
of magnitude and to print only the largest stresses. See the description of DTI,INDTA in the MSC Nastran Quick Reference
Guide, for details on preparing the user request.
This section describes the output options for the various solution sequences. MSC Nastran is designed to allow the solution
of large problems, which can generate very large amounts of output. For this reason, most types of output (e.g., stresses in
static analysis, or eigenvectors in modal analysis) are not automatically printed, but must be specifically requested by the
user. Options are provided which allow the user to restrict the printed output to only significant results. See Parameter
Descriptions for additional output options.
Static Analysis (SOLutions 101, 106, 114, 144, and 200)

 Displacements at selected grid points (DISPLACEMENT).

 Nonzero components of the applied static loads at selected grid points (OLOAD).
 Nonzero components of single point forces of constraint, including reactions on free-body supports due to applied
loads, at selected grid points or scalar points (SPCF).
 Forces and stresses in selected elements (FORCE and STRESS). In SOLution 144, forces and pressures on
aerodynamic elements (APRESSURE and AEROF).
 In all SOLutions except 4 and 106, grid point force balance at selected grid points (GPFORCE).
 In all SOLutions except 4 and 106, strain energy in selected elements (ESE).

Main Index
 26 MSC Nastran Reference Guide
Case Control Section

 Undeformed and deformed plots of the structural model (geometry for fundamental region in cyclic symmetry)
(OUTPUT(PLOT)).
 Multipoint forces of constraint for MPCs and rigid elements (MPCF).
 Contour plots for stresses in selected elements and displacements at selected grid points (geometry for fundamental
region in cyclic symmetry) (OUTPUT(PLOT)).
 Grid point stresses and strains in selected surfaces or volumes are available (GPSTRESS and GPSTRAIN).
 Nonlinear stress output is controlled by NLSTRESS. Its default is ALL.
Normal Modes (SOLutions 103, 115, and 200)

 The eigenvector for a list of selected GRID points (g-set) or SOLUTION points (a-set) along with the associated
eigenvalue for each mode (VECTOR or SVECTOR).
 Nonzero components of the single-point forces of constraint for selected modes at selected grid points (SPCF).
 Forces, stresses, and strains in selected elements for selected modes (FORCE, STRESS, and STRAIN).
 Grid point force balance at selected grid points (GPFORCE).
 Strain energy and kinetic energy in selected elements (ESE, EKE).
 Undeformed plot of the structural model and mode shapes for selected modes (geometry for fundamental region
in cyclic symmetry) (OUTPUT(PLOT)).
 Grid point stresses and strains in selected surfaces or volumes (GPSTRESS and GPSTRAIN).
 The application of response spectrum curves may be performed in SOLutions 103 and 115. See Local Adaptive Mesh
Refinement.
 Multipoint forces of constraint for MPCs and rigid elements (MPCF).
Buckling Analysis (SOLutions 105 and 200)

 Displacements and nonzero components of the static loads and single-point forces of constraint at selected grid
points for the static analysis (DISPLACEMENT, OLOAD and SPCF).
 Forces and stresses in selected elements for the static loading condition and selected buckling modes (FORCE and
STRESS).
 Mode shapes and nonzero components of the single-point forces of constraint at selected grid points for selected
modes (VECTOR and SPCF).
 Undeformed plot of the structural model and mode shapes for selected buckling modes (OUTPUT(PLOT)).
 Grid point force balance at selected grid points for the static loading condition (GPFORCE).
 Strain energy in selected elements for the static loading condition (ESE).
 Grid point stresses and strains in selected surfaces or volumes for the static loading condition (GPSTRESS and
GPSTRAIN).
 Multipoint forces of constraint for MPCs and rigid elements (MPCF).
Complex Eigenvalue Analysis (SOLutions 107, 110, 145, and 200)

Main Index
 Chapter 1: nast20214 Command and Input Data 27
Case Control Section

The following printed output, sorted by complex eigenvalue root number (SORT1), may be requested for any complex
eigenvalue extracted, as either real and imaginary parts or magnitude and phase angle (0 – 360 lead):
 The eigenvector for a list of PHYSICAL points (grid points and scalar points) or SOLUTION points (points used
in formulation of the general K system) (VECTOR or SVECTOR).
 Nonzero components of the single-point forces of constraint for a list of PHYSICAL points (SPCF).
 Forces, stresses, or strains in selected elements (FORCE and STRESS STRAIN).
 In addition, an undeformed plot of the structural model may be requested. (OUTPUT(PLOT))
 Multipoint forces or constraint for MPCs and rigid elements (MPCF).
 Strain energy, kinetic energy, and energy loss (ESE, EKE, EDE).
Frequency, Random, and Aeroelastic Response (SOLutions 108, 111, 118, 146, and 200)

The following printed output, sorted by frequency (SORT1) or by point number or element number (SORT2), is available,
either as real and imaginary parts or magnitude and phase angle (0 - 360), for the list of frequencies specified by
OFREQUENCY:
 Displacements, velocities, and accelerations for a list of PHYSICAL points (grid points and scalar points) or
SOLUTION points (points used in formulation of the generalized coordinate system) (DISPLACEMENT or
SDISPLACEMENT, VELOCITY or SVELOCITY, ACCELERATION or SACCELERATION).
 Nonzero components of the applied load vector and single-point forces of constraint for a list of PHYSICAL
points (OLOAD and SPCF).
 Stresses and forces in selected elements (ALL available only for SORT1) (STRESS and FORCE). In SOLution
146, forces on aerodynamic elements. (AEROF).
 The following plotter output is available for frequency response calculations:
• Undeformed and deformed structure plots of the structural model (OUTPUT(PLOT)).
• X-Y plot of any component of displacement, velocity, or acceleration of a PHYSICAL point or SOLUTION
point (OUTPUT(XYPLOT)).
• X-Y plot of any component of the applied load vector of single-point force of constraint
(OUTPUT(XYPLOT)).
• X-Y plot of any stress or force component for an element (OUTPUT(XYPLOT)).
 The following plotter output is available for random response calculations: The data used for preparing the X-Y
plots may be punched or printed in tabular form. Also, a printed summary is prepared for each X-Y plot which
includes the maximum and minimum values of the plotted function. The printing of the X-Y pairs is the only
form of printed output for Random Response. SORT2 output is required for X-Y plots. If both SORT1 and
SORT2 requests are made, only the SORT2 output will be printed.
• X-Y plot of the power spectral density versus frequency for the response of selected components for points or
elements (OUTPUT(XYPLOT)).
• X-Y plot of the autocorrelation versus time lag for the response of selected components for points or elements
(OUTPUT(XYPLOT)).

Main Index
 28 MSC Nastran Reference Guide
Case Control Section

 Multipoint forces of constraint for MPCs and rigid elements (MPCF).

 Strain energy, kinetic energy, and energy loss (ESE, EKE, EDE).
Transient Response (SOLutions 109, 112, 129, and 200)

The following printed output, sorted by point number or element number is available at selected multiples of the
integration time step:
 Displacements, velocities, and accelerations for a list of PHYSICAL points (grid points and scalar points) or
SOLUTION points (points used in formulation of the general K system) (DISPLACEMENT or SDISPLACEMENT,
VELOCITY or SVELOCITY, ACCELERATION or SACCELERATION).
 Nonzero components of the applied load vector and single-point forces of constraint for a list of PHYSICAL
points (OLOAD and SPCF).
 Nonlinear force vector for a list of SOLUTION points (NONLINEAR).
 Stresses and forces in selected elements (STRESS and FORCE). In SOLutions 109 and 129, forces are not
computed for nonlinear elements.
 Grid point stresses in selected surfaces or volumes (GPSTRESS).
 The following plotter output is available. The data used for preparing the X-Y plots may be punched or printed in
tabular form. Also, a printed summary is prepared for each X-Y plot which includes the maximum and minimum
values of the plotted function.
• Undeformed plot of the structural model (OUTPUT(PLOT)).
• Deformed shapes of the structural model for selected time intervals (OUTPUT(PLOT)).
• Contour plots for stresses in selected elements, and displacements, velocities or accelerations at selected grid
points for selected time intervals (OUTPUT(PLOT)).
• X-Y plot of any component of displacement, velocity, or acceleration of a PHYSICAL point or SOLUTION
point (OUTPUT(XYPLOT)).
• X-Y plot of any component of the applied load vector, nonlinear force vector, or single-point force of constraint
(OUTPUT(XYPLOT)).
• X-Y plot of any stress or force component for an element (OUTPUT(XYPLOT)).
 Multipoint forces of constraint for MPCs and rigid elements except SOLution 129.
Linear Steady State Heat Transfer (SOLutions 101 and 153)

 Temperatures (THERMAL) and nonzero components of static loads (OLOAD) and constrained heat flow
(SPCFORCE) at selected grid points or scalar points.
 The punch option of a THERMAL request to produce TEMP Bulk Data entries.
 Flux density (FLUX) in selected elements.
 Grid point heat balance (GPFORCE) at a selected set of grid points or scalar points (SOLution 101).
 Undeformed plots of the structural model and temperature profiles.
 Contour plots for temperatures at selected grid points.

Main Index
 Chapter 1: nast20214 Command and Input Data 29
Case Control Section

Steady State Heat Transfer (SOLution 153)

The following output may be requested for the last iteration in Heat Transfer Analysis:
 Temperature (THERMAL) and nonzero components of static loads (OLOAD) and constrained heat flow
(SPCFORCE) at selected grid points or scalar points.
 The punch option of a THERMAL request to produce TEMP Bulk Data entries.
 Flux density (FLUX) in selected elements. In the case of CHBDYi elements, a flux density summary is produced
that includes applied flux, radiation flux, and convective flux.
 Undeformed plots of the structural model and temperature profiles.
 Contour plots of temperatures at selected grid points.
See the MSC Nastran Thermal Analysis User’s Guide for further details.
Nonlinear Transient Heat Transfer (SOLution 159)

The following printed output, is available at selected multiples of the integration time step:
 Temperatures (THERMAL or SDISPLACEMENT) and derivatives of temperatures (VELOCITY or
SVELOCITY) for a list of PHYSICAL points (grid points and scalar points) or SOLUTION points (points used
in formulation of dynamic equation).
 Nonzero components of the applied load vector (OLOAD) and constrained heat flow (SPCFORCE) for a list of
PHYSICAL points.
 Nonlinear load vector (NLLOAD) for a list of SOLUTION points.
 Flux density (FLUX) in selected elements (ALL not allowed).
 Rate of change of enthalpy (HDOT).
The following plotter output is available for Transient Heat Transfer Analysis:
 Undeformed plot of the structural model.
 Temperature profiles for selected time intervals.
 X-Y plot of temperature (TEMP or SDISP) or derivative of temperature (VELO or SVELO) for PHYSICAL or
SOLUTION points.
 X-Y plot of the applied load vector (OLOAD), nonlinear load vector (NOLI), or constrained heat flow (SPCF).
 X-Y plot of flux density (FLUX) for an element.
 Contour plots of temperatures at selected grid points.
The data used for preparing the X-Y plots may be punched or printed in tabular form. Also, a printed summary is prepared
for each X-Y plot which includes the maximum and minimum values of the plotted function.

Case Control Looping

Boundary Condition Changes

Main Index
 30 MSC Nastran Reference Guide
Case Control Section

More than one set of MPC, SPC, and SUPORT1 selections is allowed in all of the subcases for the residual structure during
one run in SOLutions 101, 103, 105, 106, 108, 111 and 200. Only one boundary condition is allowed for superelements.
Frequency and Dynamic Load Set Changes in SOLutions 108, 111, 118, 146 and 200

More than one FREQUENCY Case Control command may be used per run except in SOLutions 146 and 200. Residual
structure subcases may be used to select different dynamic loads using the DLOAD command. Output requests may differ
between subcases.
Transient and Dynamic Load Set Changes in SOLutions 109, 112 and 200

More than one TSTEP and DLOAD Case Control command may be used per run except in SOLution 200. Only one
residual structure subcase is allowed.
Formulation of Dynamic Equations in All Dynamic Response SOLutions

Multiple subcases with different requests of K2PP, M2PP, B2PP, TFL, and SDAMPING Case Control commands may
be specified in SOLs 107 through 112, 129 and 145.
DMlG Bulk Data Entry Processing

Two forms of direct input matrix processing are provided. The Case Control commands B2GG, M2GG, K2GG, and P2G select
matrices added to the global matrices before constraints are applied and only one set may be selected for superelement or
residual structures. These terms may be added to any superelement or the residual structure. K2GG, B2GG and M2GG must
be real and symmetric (IF = 6). Their precision should be set equal to the precision of [K2GG] (TOUT = 0 will provide this).
The Case Control commands B2PP, M2PP and K2PP select matrices processed only in the residual structure. In the modal
formulations they are added after real eigenvalue extraction. These matrices need not be symmetric.
REPCASE, SUBCOM and SUBSEQ Case Control Commands

These commands may be used for superelement processing, with the provision that the SUPER Case Control command
must be used with REPCASE and SUBCOM. An example of a statics subcase structure follows:

SUBCASE 10
SUPER = 10,1
LOAD = 1
SET 1 = 1 THRU 10
DISP = 1
REPCASE 11
SUPER = 10,1
SET 2 = 11 THRU 100
DISP = 2
REPCASE 12
SUPER = 10,1
SET 3 = 101 THRU 200
DISP = 3
SUBCASE 20
SUPER = 10,2
...

Main Index
 Chapter 1: nast20214 Command and Input Data 31
Bulk Data Section

The REPCASEs must immediately follow the subcase they reference. They should contain the same SUPER = i,j command.
In this example, the two REPCASEs give output for the loading condition described in SUBCASE 10.

Bulk Data Section

The primary MSC Nastran input medium is the Bulk Data entry. These entries are used to define the model and data which
may be selected in the Case Control Section at execution time. For large problems the Bulk Data may consist of several
thousand entries.
The Bulk Data may be submitted with the entries in any order since a sort (see XSORT in the MSC Nastran DMAP
Programmer’s Guide) is performed prior to the execution of the Input File Processor (see IFP in the MSC Nastran Quick
Reference Guide). It should be noted that the machine time to perform this is minimized for input that is already sorted.
The sort time for badly sorted data will become significant for large numbers of entries. The user may obtain a printed copy
of either the unsorted or sorted Bulk Data by selection in the Case Control Section. A sorted echo is automatically provided
unless you specifically suppress it.
In order to minimize the handling of a large number of entries, provision has been made in the Structured Solution
Sequences (see Solution Sequences) to store the Bulk Data in the database, and it may be modified on subsequent runs. For
any cold start, the entire bulk data must be submitted. Thereafter, if the database from any previous run is saved, the Bulk
Data exists in the database in sorted form where it may be modified and reused on restart. Upon restart the Bulk Data
entries contained in the Bulk Data are added to the Bulk Data stored in the database. Entries are removed from the database
by using a “/” entry.
See the Bulk Data Entries in the MSC Nastran Quick Reference Guide, for a detailed description of the Bulk Data.

Use of Parameters
Parameters are used extensively in the solution sequences for input of scalar values and for requesting special features.
Parameters’ values are specified on PARAM Bulk Data entries or PARAM Case Control commands.
 The PARAM Bulk Data entry is described in the Bulk Data Entries in the MSC Nastran Quick Reference Guide.
 The PARAM Case Control command is described in the Case Control Command Descriptions in the MSC Nastran
Quick Reference Guide.
 The manner in which MSC Nastran processes PARAMeters is described in general terms in the MSC Nastran DMAP
Programmer’s Guide.
The specification of PARAMeters differs between the rigid formats and the superelement solutions sequences and SOL 4.
These differences require different user action in placing the PARAM entries in the MSC Nastran section, and in changing
PARAMeter values on restarts.

Parameters in the Superelement Solution Sequences

The program module reads the Case Control Section for PARAM entries. If PARAM entries exist, the VPS table is updated
with the values found in the Case Control Section. The PARAM values are updated at the beginning of the generation,

Main Index
 32 MSC Nastran Reference Guide
Use of Parameters

assembly, reduction, and the data recovery loops of the superelement solution sequences. This allows the values of some
user-supplied parameters to change for each superelement during the course of one run. In the nonlinear solution sequences
the PARAM values are updated at the beginning of the iteration loop. This allows the values of some user-set parameters
to vary between subcases. These parameters may appear in either the Case Control and/or Bulk Data Section. Other
PARAMeters must appear in the Bulk Data Section. If they appear in the Case Control Section, they will not cause a preface
error, but will be ignored. Other parameters may be reset to default values during the course of the run. This type should
be set only in the subcase of the Case Control Section. See Table 5-1 and Table 5-2 in the , for parameter applicability.
The PARAM Case Control command is treated in the same manner as constraint and output commands. Parameters set
above the subcase level are effective in every subcase. If there is also a parameter value set in the subcase, it takes precedence.
Only one value of a parameter may be current at any time during the execution of the run. These rules imply an order of
precedence if several forms are used in the same run. The rules are illustrated in the following example.
A partial subcase structure is listed on the following page. The subcases are listed in ascending superelement number order.
PARAM,GRDPNT is varied between them.

Subcase Structure, Sorted by SElD

SUBCASE 1
SUPER = 1
...
SUBCASE 2
SUPER = 2
PARAM,GRDPNT,100
...
SUBCASE 3
SUPER = 3
PARAM,GRDPNT,200
...
SUBCASE 4
SUPER = 4
...
SUBCASE 100 $ RESIDUAL STRUCTURE
PARAM,GRDPNT,-1
...
BEGIN BULK
PARAM,GRDPNT,0

Processing Order Value of GRDPNT

SEID
1 0
2 100
3 200
4 0
0 –1

Main Index
 Chapter 1: nast20214 Command and Input Data 33
Use of Parameters

The Bulk Data entries are processed at the beginning of the run. Any PARAM entries in the Bulk Data Section set the value
of that parameter until it is changed by a Case Control command. Superelement 1 is processed first. Since it has no
PARAM,GRDPNT entry the value of 0, as set by the Bulk Data entry PARAM,GRDPNT,0 is used. The value for
superelement 2 is set at 100 by a Case Control command. Superelement 4 is processed next. Its subcase contains no PARAM
entry, so that the value of the parameter that remains is set by the bulk data. The values for superelement 3 and the residual
structure are set by Case Control commands. An alternate method of parameter specification is to set the value most used
above the subcase level, and exceptional values within the subcase. For example, if PARAM,GRDPNT,0 were placed above
the subcase level and the values of 100, 200, and -1 placed in subcases 2, 3 and 100, respectively, the value used in each
superelement will be the same as described above, regardless of processing order.

Main Index
 Chapter 2: Grid Points and Coordinate Systems

Grid Points and Coordinate

2 Systems

Introduction 35

Grid Point and Coordinate System Definition 35
 Grid Point Properties 39

Main Index
 Chapter 2: Grid Points and Coordinate Systems 35
Introduction

Introduction
Analysis model geometry in MSC Nastran is represented by grid point locations. The physical behavior of the analysis
model is represented by the constraints and their associated coordinate system, and by the elements connecting the grid
points (elements are described in Structural Elements).

Grid Point and Coordinate System Definition

Geometric grid points are defined on GRID Bulk Data entries by specifying coordinates in either the basic coordinate
system or in a local coordinate system. The implicitly defined basic coordinate system is rectangular, except when using
axisymmetric elements. Local coordinate systems may be rectangular, cylindrical, or spherical and are related directly or
indirectly to the basic coordinate system. Figure 2-1 shows the coordinate system definition.

Main Index
 36 MSC Nastran Reference Guide
Grid Point and Coordinate System Definition

G2, (B1, B2, B3)

G3, (C1, C2, C3)

G13

G1, (A1, A2, A3)

G2
i
G1

G3
xb

zb

yb
Figure 2-1 Coordinate System Definition

Main Index
 Chapter 2: Grid Points and Coordinate Systems 37
Grid Point and Coordinate System Definition

where:

(i, j, k) = user defined coordinate system (x, y, z) or (r,  , z) or (r,  ,  )

( xb , yb , = basic coordinate system
zb )
k =  G2 – G1  normalized
G13 =  G3 – G1 
j =  k  G13  normalized
i =  j  k

The CORD1C, CORD1R and CORD1S entries are used to define cylindrical, rectangular, and spherical local coordinate
systems, respectively, in terms of three previously defined geometric grid points. The CORD2C, CORD2R and CORD2S entries
are used to define cylindrical, rectangular, and spherical local coordinate systems, respectively, in terms of the coordinates
of three points in a previously defined coordinate system.
Six rectangular displacement components (three translations and three rotations) are defined at each grid point. The
displacement coordinate system, which is used to define the directions of motion, may be different from the “location
coordinate system,” which is used to locate the grid point. Both the location coordinate system and the displacement
coordinate system are specified on the GRID entry for each geometric grid point. The orientation of displacement
components depends on the type of local coordinate system used to define the displacement components. If the defining
local system is rectangular, the displacement components are parallel to the local system and are independent of the grid
point location as indicated in Figure 2-2(a). If the local system is cylindrical, the displacement components are in the radial,
tangential, and axial directions as indicated in Figure 2-2(b). If the local system is spherical, the displacement components
are in the radial, meridional, and azimuthal directions as indicated in Figure 2-2(c).
If the displacements of a grid point are defined to be in a spherical or cylindrical coordinate system and the grid point is
located on, or relatively close to, the polar axis, then it will be replaced by the “defining” rectangular coordinate system. The
z-axis of the “defining” rectangular coordinate system is coincident with the polar axis and the x-axis is coincident with
 = 0 in cylindrical system, or  = 0 and  = 0 in the spherical system. It is recommended that a rectangular system
is explicitly specified by the user.
Each geometric grid point may have a unique displacement coordinate system associated with it. The collection of all
displacement coordinate systems is known as the global coordinate system. All matrices are formed and all displacements are
output in the global coordinate system. The symbols T1, T2 and T3 on the printed output indicate translations in the 1,
2, and 3-directions, respectively, for each grid point. The symbols R1, R2, and R3 indicate rotations about the three axes.
Provision is also made on the GRID entry to apply single-point constraints to any of the displacement components. Any
constraints specified on the GRID entry will automatically be used for all solutions. Constraints specified on the GRID
entry are usually restricted to those degrees of freedom that will not be elastically constrained and hence must be removed
from the model in order to avoid singularities in the stiffness matrix.
Element offsets are defined by a Cartesian coordinate system, which is also parallel to the displacement coordinate system
at the connecting grid point. However, the components of the offsets are always defined in units of translation, even if the
displacement coordinate system is cylindrical or spherical.

Main Index
 38 MSC Nastran Reference Guide
Grid Point and Coordinate System Definition

z u3
(a)Rectangular

u2

u1 Grid Point
Local System y

z
(b)Cylindrical u3 – z direction

r u2 –  direction

u1 – r direction

Local System Grid Point

u1 – r direction
(c)Spherical
u3 –  direction
q

u2 –  direction
Local System r Grid Point

Figure 2-2 Displacement Components

Main Index
 Chapter 2: Grid Points and Coordinate Systems 39
Grid Point Properties

The GRDSET Bulk Data entry is provided to avoid repeating the specification of location coordinate systems, displacement
coordinate systems, and single point constraints when many (or all) of the GRID entries have the same entries for these
items. When any of the three items are specified on the GRDSET entry, the entries are used to replace blank fields on the
GRID entry for these items. This feature is useful in the case of such problems as space trusses where one wishes to remove
all of the rotational degrees of freedom or in the case of plane structures where one wishes to remove all out-of-plane or all
in-plane motions.
Scalar points are defined either on an SPOINT entry or by reference on a connection entry for a scalar element. SPOINT
entries are used primarily to define scalar points appearing in constraint equations but to which no structural elements are
connected. A scalar point is implicitly defined if it is used as a connection point for any scalar element. Special scalar points,
called “extra points,” may be introduced for dynamic analyses. Extra points are used in connection with transfer functions
and other forms of direct matrix input used in dynamic analyses and are defined on EPOINT entries.

Grid Point Properties

Some of the characteristics of the model may be specified as properties of grid points rather than as properties of elements.
Mass, damping, and stiffness properties may be provided, in part or entirely, as properties of grid points through the use of
direct input matrices. The DMIG entry is used to define direct input matrices. These matrices may be associated with
components of geometric grid points, scalar points or extra points introduced for dynamic analysis. The TF entry is used
to define transfer functions that are internally converted to direct matrix input. The DMIAX entry is an alternate form of
direct matrix input that is used for hydroelastic problems (see Coupled Fluid-Structure Interaction).
Thermal fields are defined by specifying the temperatures at grid points. The TEMP entry is used to specify the temperature
at grid points for use in connection with thermal loading and temperature-dependent material properties. The TEMPD
entry is used to specify a default temperature in order to avoid a large number of duplicate entries on a TEMP entry when
the temperature is uniform over a large portion of the structure. The TEMPAX entry is used for conical shell problems.
Mass properties may be input as properties of grid points by using the concentrated mass element. The CONM1 entry is
used to define a 6  6 matrix of mass coefficients at a geometric grid point in any selected coordinate system. The
CONM2 entry is used to define a concentrated mass at a geometric grid point in terms of its mass, the three coordinates
of its center of gravity, the three moments of inertia about its center of gravity, and its three products of inertia, referred to
any selected coordinate system.
The SNORM Bulk Data entry and/or the PARAM,SNORM assigns the singularity normal of a shell element to a degree
of freedom. This singularity is then removed via AUTOSPC.

Main Index
 Chapter 3: Structural Elements

Structural Elements
3

Element and Property Definition 41
 Line Elements 53

Surface Elements 123

Solid Elements (CTETRA, CPENTA, CHEXA) 154

Rigid Elements and Multipoint Constraints (R-type, MPC) 159
 Scalar Elements (CELASi, CMASSi, CDAMPi) 178

Axisymmetric Solid Elements (CTRIAX6, CTRIAX, CQUADX) 179

Crack Tip Elements (CRAC2D, CRAC3D) 181

Gap and Line Contact Elements (CGAP, BCONP, BLSEG) 186

Concentrated Mass Elements (CONM1, CONM2) 192

General Element Capability (GENEL) 192

Direct Matrix Input on Grid Points (DMIG) 196

Hyperelastic Elements 196

Hyperelastic Elements 196

Bushing Elements 200
 Connector Elements 206

Connector Elements 206

Element Geometry Checks 241

ELSUM Case Control Command 252

Main Index
 Chapter 3: Structural Elements 41
Element and Property Definition

Element and Property Definition

Structural elements are defined on Bulk Data connection entries that identify the grid points to which the element is
connected. The mnemonics for all such entries have a prefix of the letter “C”, followed by an indication of the type of
element, such as CBAR and CROD. The order of the grid point identification defines the positive direction of the axis of a
one-dimensional element and the positive surface of a plate element. The connection entries include additional orientation
information when required. Some elements allow for offsets between its connecting grid points and the reference plane of
the element. The coordinate systems associated with element offsets are defined in terms of the grid point coordinate
systems described in Grid Points and Coordinate Systems. For most elements, each connection entry references a property
definition entry. If many elements have the same properties, this system of referencing eliminates a large number of
duplicate entries.
The property definition Bulk Data entries define geometric properties such as thicknesses, cross-sectional areas, and
moments of inertia. The mnemonics for all such entries have a prefix of the letter “P”, followed by some or all of the
characters used on the associated connection entry, such as PBAR and PROD. Other included items are the nonstructural
mass and the location of points where stresses will be calculated. For most elements, each property definition entry will
reference a material property entry.
In some cases, the same finite element can be defined by using different Bulk Data entries. These alternate entries have been
provided for user convenience. In the case of a rod element, the normal definition is accomplished with a connection entry
(CROD) which references a property entry (PROD). However, an alternate definition uses a CONROD entry which
combines connection and property information on a single entry. This is more convenient if a large number of rod elements
all have different properties.
Most of the elements may be used with elements of other types within the limitations of good modeling practice. Exceptions
are the axisymmetric elements, which are designed to be used by themselves. There are two types at present, linear and
nonlinear. The linear are described in the following section. The nonlinear elements used in SOL 106 and SOL 129 are
described under Hyperelastic Elements. The conical shell element (“CCONEAX”) describes a thin shell by sweeping a line
defined on a plane by two end points through a circular arc. Loads may be varied with azimuth angle through use of
harmonic analysis techniques. This element has a unique set of input entries, which may be used with a limited set of other
entries. These unique entries, if mixed with entries for other types of elements, cause a preface error. The advanced elements
used in nonlinear solution, SOL 400, are described in the MSC Nastran Nonlinear User’s Guide, Chapter 11: Element Library.
The elements used in explicit nonlinear solution, SOL 700, are described in the MSC Nastran Explicit Nonlinear (SOL 700)
User’s Guide, Chapter 6: Elements.

Elements with Offsets

The CBAR, CBEAM, CBEAM3, CQUAD4, CQUAD8, CQUADR, CTRIA3, CTRIA6, and CTRIAR elements allow
the user to apply offsets at the grid connections. The beam elements specify offset vectors at the connecting grids and the
shell elements use a ZOFF from the surface of grid points to the element reference plane.
MSC Nastran provides the bulk data entry: MDLPRM, OFFDEF, option; for control of offset behavior. For legacy
reasons, the default is MDLPRM, OFFDEF, ELMOFF, which imposes the following limitations:
• The differential stiffness is not supported. Therefore, they are not applicable in solution sequence where
differential stiffness is required, such as linear buckling analysis (SOL 105).

Main Index
 42 MSC Nastran Reference Guide
Element and Property Definition

• The effect of offsets is not included the mass matrix.

• The effect of offset is not included in computation of thermal loads, pressure loads, or gravity loads.
• For curved shell problems, the offset is defined in the z direction of the element and not the shell normal
direction, this results in gaps between elements when offset geometry is applied.
• The transformation is linear and therefore, it cannot be used in nonlinear analysis.
In general, the above limitations are removed if the recommended MDLPRM, OFFDEF, LROFF, is used. MDLPRM,
OFFDEF, LROFF and its options are supported in all solution sequences with the exception of SOL106, SOL129, and
SOL200.
The enhanced offset method has the following benefits:
• DIFFERENTIAL STIFFNESS - The differential stiffness is computed for the offset so that it is applicable to
solution sequences that required differential stiffness. These include, for example, SOL 103, SOL 105, and
SOL 400.
• MASS - The effect of offset is included in the mass matrix generation.
• LOADS - The effect of offset is included in the load generation.
• GEOMETRY COMPATIBILITY - For QUAD4, TRIA3, QUADR, TRIAR, QUAD8 and TRIA6 elements
the shell normal is used as the offset direction. Thus, the new offset will enhance solution of a model in two
aspects: there is no gap in the offset geometry, and the mass and stiffness matrices are computed based on the
offset geometry.
• NONLINEAR EFFECTS - The transformations are nonlinear so that it can be used in nonlinear analysis
such as SOL 400
In SOL103, if the case control ADAMSMNF FLEXBODY=YES or AVLEXB EXBBODY=YES is present, then
MDLPRM, OFFDEF, LROFF must be replaced by MDLPRM, OFFDEF, NOMASS.
For shell elements, there are two ways to account for an offset from the surface of grid points to the element reference plane.
The ZOFF on the shell connection entry and Z0 on the PCOMP or PCOMPG entry. The MDLPRM, OFFDEF, option
is only valid for the connection ZOFF method. Further, the Z0 composite option will in general lead to incorrect loading
results for SOL105.

Main Index
 Chapter 3: Structural Elements 43
Element and Property Definition

If there are offsets on shells or beams, MDLPRM, OFFDEF, LROFF should be used except in SOL106, SOL129, and
SOL200. Remember, however, that SOL105, a linear buckling analysis, is an attempt to approximate a nonlinear instability,
so it can be sensitive to stiffness and large rotation of offset vectors.
In SOL105, K6ROT must be thought of as a model imperfection for CQUAD4 and CTRIA3 elements. For these elements:

Main Index
 44 MSC Nastran Reference Guide
Element and Property Definition

1. LROFF uses the shell normal at each grid

2. LROFF scales that normal by the element offset to get an offset vector
3. LROFF moves the mid plane to the offset position.(internal m-set, n-set, GmnT Kmm Gmn)
4. K6ROT applies a numerical drilling stiffness to this moved normal to form Kgg
5. For differential stiffness Kdd, LROFF does a 1-3 above
6. Then LROFF rotates that offset vector based on preload rotations
7. The differential stiffness matrix Kdd is based on this rotated offset vector
In SOL105, spurious modes often occur when rigid elements such as RBE2 or RBE3 are present because they generate
Lagrange multipliers. When rigid elements are present RIGID=LGELIM or LAGRAN must be used in SOL101, SOL103
or SOL105 to account for rigid element rotation. In SOL105:
1. Spurious modes are not caused by LROFF. LROFF never produces Lagrange multipliers in any solution sequence.
2. Spurious modes may be caused by the rigid elements such as the RBE2 element which generates Lagrange
multipliers.
3. In SOL105, RIGID=LAGRAN & LGELIM should get same result if for RIGID=LAGRAN the parameters
LMFACT & PENFN are set properly.
4. LMFACT and PENFN are the scale factor and penalty function for the Lagrange rigid elements and the contact
analysis.
5. The purpose of LMFACT and PENFN is to make the values of stiffness matrix of the Lagrange rigid elements
and/or the contact components about the same relative magnitude as those of the other elements in the model.
6. The default value is 1.0E+5 for all solution sequences except SOL 400 and SOL 101 linear contact. In SOL 400
and SOL 101 linear contact, MSC Nastran will compute the appropriate default values for LMFACT and PENFN.
7. This computed SOL400 default is good in general. But when the material/element stiffness (e.g., Young’s Modulus,
Spring Stiffness) used in the model vary in a wide range, users may have to adjust the computed LMFACT and
PENFN.
8. For some SOL105 models, reducing LMFACT and PENFN to 1.0E+4 allows the RIGID=LAGRAN to remove
spurious modes. But “rotation” modes are still obtained.
9. Because of the above discussion, it is recommended in SOL105 if rigid elements are present that RIGID= LGELIM
be used to avoid the pain of LMFACT and PENFN adjustment. (Note RIGID=LGELIM is not valid is SOL400).
10. It is recommended in SOL105 to start with a value of K6ROT=100. as a low value such as K6ROT=1. is in the
noise of the solution. It may be necessary to increase the value of K6ROT > 100 in order to provide higher drilling
stiffness to counter large rotation values and their impact on rigid elements and LROFF.
11. Models containing, RBE2 and RBE3 elements, BCONTACT surfaces and LRFOFF and may be extremely flexible.
Thus rotations are large. To get control of the rotations for the CQUAD4/CTRIA3 elements K6ROT values as
high as K6ROT=1800 may be required.
12. Always us the lowest possible converged K6ROT value, too high a value may be just as detrimental as to low a value.

Main Index
 Chapter 3: Structural Elements 45
Element and Property Definition

13. To alleviate the K6ROT problem, NASTRAN QRMETH=5 may be tried. However, if spurious modes occur with
this method, there is no way to add rotational stiffness such as K6ROT to correct them. Also, some of its spurious
modes are difficult to detect.
14. In buckling analysis, always plot and examine the mode shapes.

Axisymmetric Elements
The axisymmetric elements CTRIAX6, CTRIAX, and CQUADX define a solid ring by sweeping a surface defined on a plane
through a circular arc. Loads are constant with azimuth for these elements; that is, only the zeroth harmonic is considered.
There may be innovative modeling techniques that allow coupling this class of axisymmetric element with other elements,
but there are no features to provide correct automatic coupling.
Although the axisymmetric elements CCONEAX and CTRIAX6 provide useful functions, the user should be aware that
their usage precludes the use of many features available for other elements (see Tables 3-1, 3-2, and 3-3). An alternative
modeling technique is to use conventional elements with the cyclic symmetry solution sequences to provide equivalent
models. This technique generally provides the features not available with axisymmetric elements. The CTRIAX and
CQUADX axisymmetric elements are finite deformation, hyperelastic elements with fully nonlinear characteristics (see
Hyperelastic Elements).

Material Properties
The material property definition entries are used to define the properties for each of the materials used in the structural
model. The MAT1 entry is used to define the properties for isotropic materials and may be referenced by any of the
structural elements. The MATT1 entry specifies table references for isotropic material properties that are temperature
dependent. The TABLEM1, TABLEM2, TABLEM3, and TABLEM4 entries define four different types of tabular
functions for use in generating temperature dependent material properties. More information is given in Material Properties.
The MAT2 entry is used to define the properties of anisotropic materials for triangular or quadrilateral membrane and
bending elements. The MAT2 entry specifies the relationship between the in-plane stresses and strains. It may also be used
for anisotropic transverse shear. The angle between the material coordinate system and the element coordinate system is
specified on the connection entries. The MATT2 entry specifies table references for anisotropic material properties that are
temperature dependent. This entry may reference any of the TABLEM1, TABLEM2, TABLEM3, or TABLEM4 entries.
The MAT3 entry is used to define the properties for orthotropic materials used in the modeling of axisymmetric shells. This
entry may only be referenced by CTRlAX6 entries. The MATT3 entry specifies table references for use in generating
temperature-dependent properties for this type of material. More information is given in Isotropic Material.
The MAT8 entry is used to define the properties of orthotropic materials used in the modeling of quadrilateral and
triangular shell elements for composite structures.
The MAT9 entry is used to define the properties of anisotropic materials for the CHEXA, CPENTA, and CTETRA elements.
The MATT9 entry specifies table references for use in generating temperature-dependent properties for this type of
material. More information is given in Orthotropic Material.

Main Index
 46 MSC Nastran Reference Guide
Element and Property Definition

The MATS1 entry specifies table references for isotropic material properties that are stress and/or temperature dependent.
The TABLES1 entry defines a tabular stress-strain function for use in creep analysis (CREEP entry) or material nonlinear
analysis. The TABLEST entry specifies TABLES1 references for nonlinear elastic temperature-dependent material
properties. The MATS1 entry may be combined with a corresponding MAT1 to define the following types of nonlinear
material analysis: (1) plastic (2) nonlinear elastic and (3) creep. More information is given in the MSC Nastran Handbook
for Nonlinear Analysis.
The MATHP entry specifies hyperelastic material properties and must be referenced by one of the PLPLANE or PLSOLID
property entries, specifying a finite deformation nonlinear analysis, including large strain and large rotation. Experimental
data may be input on TABLES1 entries for automatic curve fitting of material parameters (see TABLES1 (Ch. A)).
In linear analysis, temperature-dependent material properties are computed once only at the beginning of the analysis. In
nonlinear static analysis (SOLution 106), temperature-dependent material properties for linear (MAT1, MAT2, and MAT9
entries) and nonlinear elastic materials (MAT1 and MATS1 entries) may be updated many times during the analysis.
The advanced materials used in nonlinear solution, SOL 400, are described in the MSC Nastran Nonlinear User’s Guide,
Chapter 10: Materials. The elements used in explicit nonlinear solution, SOL 700, are described in the MSC Nastran Explicit
Nonlinear (SOL 700) User’s Guide, Chapter 6: Elements.

The General Element (GENEL)

The GENEL entry is used to define general elements whose properties are defined in terms of deflection influence
coefficients or stiffness matrices which can be connected between any number of grid points. One of the important uses of
the general element is the representation of part of a structure by means of experimentally measured data. No output data
is prepared for the general element. Detailed information on the general element is given in General Element Capability (GENEL).

Dummy, Mass, and Damping Elements

Dummy elements are provided in order to allow the user to investigate new structural elements with a minimum
expenditure of time and money. A dummy element is defined with a CDUMi (i = index of element type, 1  i  9 ) entry
and its properties are defined with the PDUMi entry. The ADUMi entry is used to define the fields on the connection and
property entries.
Mass elements have inertia properties, but do not have stiffness or damping. The CONM2 entry provides convenient input
of mass attached to a grid point whereas the CONM1 entry provides a more general description of mass at a grid point.
Damping elements have damping properties only. The CVISC entry provides damping along a line between two grid
points.

Constraint Elements
There are two classes of constraint elements. They generate constraint equations that allow elimination of some variables,
rather than add coefficients to the stiffness matrix. However, they do have other element-like properties, such as the
satisfaction of equilibrium conditions. The RBAR, RBE1, RBE2, RROD, RJOINT, RSSCON, and RTRPLT entries
describe rigid elements. The RBE3 and RSPLINE entries describe interpolation elements. (See also Rigid Elements and
Multipoint Constraints (R-type, MPC)).

Main Index
 Chapter 3: Structural Elements 47
Element and Property Definition

Other Element Types

Other types of elements are described in different subsections of Additional Topics, 511. They include elements with fluid,
acoustics, and nonlinear analysis capability. Heat transfer elements are described in the MSC Nastran Thermal Analysis User’s
Guide.
Details of the finite elements in MSC Nastran are given in the following sections. A summary of the finite elements and
their characteristics is given in Table 3-1. An X in the table indicates the existence of an item.
Element identification numbers must be unique across all element types.

Table 3-1 Element Summary -- Conventional MSC Nastran Elements

Structural Matrices Materials Static Load Heat Transfer

Element Deformation
Geometric Nonlinear

Material Nonlinear

Viscous Damping
Differential Stiffness

Heat Conduction

Heat Capacity

Thermal Load
Axisymmetric

Anisotropic

Orthotropic

Pressure
Stiffness

Isotropic

Thermal

Gravity
Mass

Element
Type
CAXIFi X
CBAR X LC X X EB X X X X X X
CBEAM X LC X X X X EB X X X X X X
CBEAM3 X LC X X X X EB X X X
CBUSH FD X FD
CBUSH1D X X X X X X
CBEND X C X X EB X X X X X
CCONEAX X L X X X X E X X
CONMi LC X
CONROD CS LC X X X X E X X X X X
CRAC2D I LC X X E X
CRAC3D I LC X X E X
CDAMPi X X
CELASi X X X
CFAST X X X
CFLUIDi X

Main Index
 48 MSC Nastran Reference Guide
Element and Property Definition

Table 3-1 Element Summary -- Conventional MSC Nastran Elements (continued)

Structural Matrices Materials Static Load Heat Transfer

Element Deformation
Geometric Nonlinear

Material Nonlinear

Viscous Damping
Differential Stiffness

Heat Conduction

Heat Capacity

Thermal Load
Axisymmetric

Anisotropic

Orthotropic

Pressure
Stiffness

Isotropic

Thermal

Gravity
Mass

Element
Type
CGAP X X
CHBDYi X X X
CHEXA I LC X X* X* X X E X X X X X
CMASSi L X
CPENTA I LC X X* X* X X E X X X X X
CQUAD4 I LC X X X X X X EB X X X
CQUAD8 I LC X X X X EB X X X X X
CQUADR I LC X X X X EB X X X X X
CROD CS LC X X X X E X X X X X
CSEAM X X X
CSHEAR CS L X X E X X
CSLOTi X
CTETRA I LC X X X X X E X X X X X
CTRIA3 I LC X X X X X X EB X X X X X
CTRIA6 I LC X X X X EB X X X X X
CTRIAR I LC X X X X EB X X X X X
CTRIAX6 I LC X X X E X X X X X
CTUBE CS LC X X X X E X X X X X
CVISC X
CWELD X X X
†For the fully nonlinear hyperelastic elements, see Table 3-1.

*With the exception of hyperelastic elements, no midside grid points may be defined with the nonlinear stiffness formulation.

Main Index
 Chapter 3: Structural Elements 49
Element and Property Definition

Table 3-2 Element Data Recovery Summary - Part 1

Data Recovery

Force Sum Element Edges

Grid Point Stresses

Force Sum Global
Stress Complex*

Force Complex*

Structure Plot
Element Type

Contour Plot
Stress Real*

Force Real*
CAXlFi 12 23
CBAR 16 19 9 17 X X X
CBEAM 89 111 89 177 X X
CBEND 23 23 17 31 X X
CBUSH 7 13 7 13 X
CBUSH1D 8 X
CCONEAX** 18 7 X
CONMi
CONROD 5 5 3 5 X X X
CRAC2D 8 X
CRAC3D 10
CDAMPi 2 3
CELASi 2 3 2 3 X
CFLUIDi X
CGAP 8 X
CHBDYP, CHBDYG X
CHEXA 193 121 X X X X
CMASSi
CPENTA 151 95 X X X X
CQUAD4 17 15 9 17 X X X X X
CQUAD8 87 77 47 87 X X X

Main Index
 50 MSC Nastran Reference Guide
Element and Property Definition

Table 3-2 Element Data Recovery Summary - Part 1 (continued)

Data Recovery

Force Sum Element Edges

Grid Point Stresses

Force Sum Global
Stress Complex*

Force Complex*

Structure Plot
Element Type

Contour Plot
Stress Real*

Force Real*
CQUADR 87 77 47 87 X X X
CROD 5 5 3 5 X X X
CSHEAR 8 5 17 33 X X X X
CSLOTi 7 13 X
CTETRA 109 69 X X X
CTRIA3 17 15 9 17 X X X X X
CTRIA6 70 62 38 70 X X X
CTRIAR 70 62 38 70 X X X X
CTRIAX6 30 34 X X X
CTUBE 5 5 3 5 X X X X
CVISC 5 X
CWELD 9 17
*The integers represent the number of output words per element, useful for storage requirement calculations.

**This requires the presence of an AXISYM Case Control command.

Main Index
 Chapter 3: Structural Elements 51
Element and Property Definition

Table 3-3 Element Data Recovery Summary - Part 2

Data Recovery

Composite* Failure Indices

Strain Energy Density

Composite* Stresses

Strain Complex*
Element Type

Strain Energy

Heat Transfer
MSGSTRESS

Strain Real*
SORT2
CAXlFi X
CBAR X X X X
CBEAM X X X X
CBEND X X X X X
CCONEAX**
CBUSH X X
CBUSH1D X X 8
CONMi
CONROD X X X X X
CRAC2D
CRAC3D
CDAMPi X
CELASi X X X
CFLUIDi X
CGAP X X
CHBDYi X X
CHEXA X X X X X 193 121
CMASSi
CPENTA X X X X X 151 95
CQUAD4 X X X X X 9 11 17 15
CQUAD8 X X X X 9 11 87 77

Main Index
 52 MSC Nastran Reference Guide
Element and Property Definition

Table 3-3 Element Data Recovery Summary - Part 2

Data Recovery

Composite* Failure Indices

Strain Energy Density

Composite* Stresses

Strain Complex*
Element Type

Strain Energy

Heat Transfer
MSGSTRESS

Strain Real*
CQUADR X X X SORT2 X 9 11 87 77
CROD X X X X
CSHEAR X X X
CSLOTi X
CTETRA X X X X 109 69
CTRIA3 X X X X X 9 11 17 15
CTRIA6 X X X X 9 11 70 62
CTRIAR X X X X 9 11 70 62
CTRIAX6 X X X X X
CTUBE X X X X
CVISC
CWELD X X X
*The integers represent the number of output words per element, useful for storage requirement calculations.

**This requires the presence of an AXISYM Case Control command.

Stiffness Matrix Mass Matrix

CS -- Constant strain element L -- Lumped mass only
LS -- Linear strain element C -- Coupled mass only
I -- Modified isoparametric element LC -- Lumped mass or coupled mass
FD -- Frequency dependent

Main Index
 Chapter 3: Structural Elements 53
Line Elements

Thermal Load
E -- Extension load only
EB -- Both extension and bending load

Integer values indicate number of items output.

Line Elements
The line elements consist of the following:
 CBEAM
 CBEAM3
 CBAR
 CBEND
 CROD
 CONROD
 CTUBE

Beam Element (CBEAM)

The beam element is defined with a CBEAM entry and its properties are defined with a PBEAM, PBCOMP, or PBEAML
entry. The beam element includes extension, torsion, bending in two perpendicular planes, the associated shears, and the
features listed below:
1. Distributed mass polar moment of inertia.
2. Separate shear center, neutral axis, and nonstructural mass center of gravity.
3. Arbitrary variation of the section properties (A, I1, 12, I12, J) and of the nonstructural mass (NSM) along the beam
(PBEAM only).
4. Shear relief due to taper (PBEAM only).
5. The ability to apply either concentrated or distributed loads along the beam, using the PLOAD1 entry.
6. The effect of cross-sectional warping on torsional stiffness (PBEAM only).
7. The ability to model a beam made up of offset rods, using the PBCOMP entry.
8. Nonlinear material properties: elastic-perfectly plastic only (see TYPE = PLASTIC on MATS1 entry).
Any five of the six forces at either end of the element may be set equal to zero by using the pin flags on the CBEAM entry.
The integers 1 to 6 represent the axial force, shearing force in Plane 1, shearing force in Plane 2, axial torque, moment in
Plane 2 and moment in Plane 1, respectively. The structural and nonstructural mass of the beam are lumped at the ends of

Main Index
 54 MSC Nastran Reference Guide
Line Elements

the elements, unless coupled mass is requested with the PARAM entry, COUPMASS (see Parameters in the MSC Nastran
Quick Reference Guide).
The beam element coordinate system and nomenclature are shown in Figure 3-1. End A is offset from grid point GA an
amount measured by vector w a , and end B is offset from grid point GB an amount measured by vector w b . The vectors

w a and w b are measured in the global coordinates of the connected grid point. The x-axis of the element coordinate
system is defined by a line connecting the shear center of end A to that at end B of the beam element. The orientation of
the beam element is described in terms of two reference planes. The reference planes are defined with the aid of vector v .
This vector may be defined directly with three components in the global system at end A of the beam, or by a line drawn
from end A parallel to the line from GA to a third referenced grid point. The first reference plane (Plane 1) is defined by
the x-axis and the vector v . The second reference plane (Plane 2) is defined by the vector cross product x  v and the
x-axis. The subscripts 1 and 2 refer to forces and geometric properties associated with bending in Planes 1 and 2,
respectively. The reference planes are not necessarily principal planes. The coincidence of the reference planes and the
principal planes is indicated by a zero product of inertia  l 12  on the PBEAM entry. When pin flags and offsets are used,
the effect of the pin is to free the force at the end of the element x-axis of the beam, not at the grid point.

z elem y ma

y na z ma
z na Nonstructural Mass
y elem Center of Gravity
Plane 2
Plane 1
v
End A (0, 0, 0)
Neutral Axis
w a offset x elem v y mb
Grid Point GA Shear Center y nb z mb
z nb

End B  x b 0 0 
w b offset

Grid Point GB
Figure 3-1 CBEAM Element Coordinate System

The positive directions for element forces are shown in Figure 3-2. The following element forces, either real or complex
(depending on the solution sequence), are output on request at both ends and at intermediate locations defined on the
PBEAM entry:

Main Index
 Chapter 3: Structural Elements 55
Line Elements

Beam element internal forces and moments

 Bending moments in the two reference planes at the neutral axis.
 Shear forces in the two reference planes at the shear center.
 Axial force at the neutral axis.
 Total torque about the beam shear center axis.
 Component of torque due to warping.
The following real element stress data are output on request:
 Real longitudinal stress at the four points prescribed for each cross section defined along the length of the beam on
the PBEAM entry.
 Maximum and minimum longitudinal stresses.
 Margins of safety in tension and compression for the element if the user enters stress limits on the MAT1 entry.

z elem y elem

Plane 1

Plane 2

M1

V2 M2
Neutral Axis
v1

Fx

x elem Shear Center

Tx
Figure 3-2 CBEAM Element Internal Forces and Moments

Main Index
 56 MSC Nastran Reference Guide
Line Elements

Tensile stresses are given a positive sign and compressive stresses a negative sign. Only the longitudinal stresses are available
as complex stresses. The stress recovery coefficients on the PBEAM entry are used to locate points on the cross section for
stress recovery. The subscript 1 is associated with the distance of a stress recovery point from Plane 2. The subscript 2 is
associated with the distance from Plane 1. Note that if zero value stress recovery coefficients are used, the axial stress is
output.
The PBCOMP entry allows the input of offset rods to define the beam’s section properties. A program automatically
converts the data to an equivalent PBEAM entry. The input options that allow efficient descriptions of various symmetric
cross sections are shown in Figure 3-3.

Z ref Z ref Z ref

1  0 2K z 
5 1 1 2
8 2  K y K z 
3
6 2
Y ref Y ref 4 Y ref
7
3 0 2K y 0 
8
8 4
6 4 7

7 3 5 6
5

SECTION = 0 (default) SECTION = 1 SECTION = 2

Symmetric about Y ref and Z ref (w/continuation entry) Symmetric about Y ref
Symmetric about Y ref and Z ref
I zz I yy 1 Y1 = Y3 = -Y5 = -Y7 Y1 = Y5
Ky = ------ , K z = ------, C1 = --- Z1 = -Z3 = Z5 = -Z7, etc. Z1 = -Z5, etc.
A A 8
I zz - moment of inertia about z-axis
I yy - moment of inertia about y-axis

Main Index
 Chapter 3: Structural Elements 57
Line Elements

Z ref Z ref Z ref

2 1 1 2 3 4
5 1
3 5
6 2
Y ref 4 Y ref 6 Y ref

3 8 7
7
8 4 5 6 7 8

SECTION = 3 SECTION = 4 SECTION = 5

Symmetric about Z ref Mirror Symmetry about Y ref No symmetry
and Z ref
Y1 = -Y5, Z1 = Z5, etc.
Y1 = -Y5, Z1 = -Z5, etc.

Figure Notes:

1. Integration points (lumped area) are numbered 1 through 8.

2. User-specified points are denoted by and the program default point denoted by .

Figure 3-3 PBCOMP Entry SECTION Types

Element Properties on the PBEAM Entry

The cross-sectional properties of the element which may be defined by the user and the manner in which these properties
are interpolated along the x-axis of the element are listed in the following table:

Locations at Which Method of

Quantity Definition User Specifies Properties Interpolation
A Cross-sectional Area Ends, Interior points Linear between points
I1,I2,I12 Moments and product of Ends, Interior points Linear between points
inertia about neutral axis for
planes 1 and 2
J Torsional stiffness parameter Ends, Interior points Linear between points
NSM Nonstructural mass per unit Ends, Interior points Linear between points
length
K1,K2 Shear factors in (K)AG for One value Constant
planes 1 and 2

Main Index
 58 MSC Nastran Reference Guide
Line Elements

Locations at Which Method of

Quantity Definition User Specifies Properties Interpolation
S1,S2 Shear relief factors for planes One Value Constant
1 and 2
CW Warping stiffness parameter Ends Linear between ends
NSI Nonstructural mass moment Ends Linear between ends
of inertia per unit length

Mass Matrix
The inertia properties of the CBEAM element include the following terms:
 Structural mass per unit length, RHO  A , on the neutral axis.
 Nonstructural mass per unit length, NSM.
 Moment of inertia of structural mass per unit length RHO   I1 + I2  , about neutral axis.
 Moment of inertia of nonstructural mass, NSI.
where RHO is the density defined on the MAT1 entry.

Cross Sectional Warping

Open section members such as channels will undergo torsion as well as bending when transverse loads act anywhere except
at the shear center of a cross-section. This torsion produces warping of the cross-section so that plane sections do not remain
plane and, as a result, axial stresses are produced. This situation can be represented in the following differential equation
for the torsion of a beam about the axis of shear centers:

d d d2 d2
G ------  J ------ – E --------  C w --------- = m (3-1)
dx  dx dx 2  dx 2

E = Young’s modulus of elasticity

C w = Warping constant
G = Shear modulus
J = Torsion constant
 = Angle of rotation at any cross-section
m = Applied torsional moment per unit length

Note that C w , the warping constant, has units of (length)6. The development of the above differential equation and
methods for the numerical evaluations of the warping constant are available in the literature. (See, for example, Timoshenko

Main Index
 Chapter 3: Structural Elements 59
Line Elements

and Gere, Theory of Elastic Stability, McGraw Hill Book Company, 1961.) An example that demonstrates the use of the
warping constant is the section titled “Example Problem of Channel Section”.

Shear Relief
The shear relief factor accounts for the fact that in a tapered flanged beam the flanges sustain a portion of the transverse
shear load. This situation is illustrated in Figure 3-4:

P 

MB V hB hA V MA

P 
Figure 3-4

Here, the net transverse shear, Q , at the cross-section of depth h B is

Q = V – 2P sin  (3-2)
and, if the entire bending moment is carried by the flanges,
2 tan 
Q = V --------------- M B (3-3)
hr

Equation (3-3) is represented in MSC Nastran as:

S1
Q 1 = V 1 – ----- M 1 (3-4)
l

S2
Q 2 = V 2 – ----- M 2
l
where l is the length of the CBEAM element in question and the subscripts refer to plane 1 and plane 2, respectively. The
terms S 1 and S 2 are denoted as the shear relief coefficients. The value of the shear coefficient for a tapered beam with
heavy flanges that sustain the entire moment load may then be written as:
2  hA – hB 
S 1 = --------------------------
 hA + hB 

Main Index
 60 MSC Nastran Reference Guide
Line Elements

where:

h A = depth at end A of CBEAM element (GA on CBEAM entry)

h B = depth at end B of CBEAM element (GB on CBEAM entry)

Example Problem of Channel Section

The simply supported beam illustrated in Figure 3-5 is modeled with five CBEAM elements. Since the beam is an open
section channel with a single plane of symmetry, buckling failure can occur either through a combination of torsion and
bending about the element x-axis or the lateral bending about the y-axis. The effect of cross-sectional warping coefficients,
CW(A) and CW(B), on the PBEAM entry are necessary to capture these effects. Since the column is of uniform
cross-section only the PBEAM entry illustrated below is required.

1 2 3 4 5 6 7 8 9 10
PBEAM 1 1 .986 1.578465 .1720965 .0094985
NO 1.
.3010320 .3010320
.7659450 .769450
Warping requires seven degrees of freedom for the beam element. Thus, for warping, each grid must have an associated
scalar point.

Main Index
 Chapter 3: Structural Elements 61
Line Elements

y x
e

l = 72. z

y
Section Properties

2
3b
e = --------------- = .4690484
n 6b + h
3
e t
=h = 3.17 x J = ----  2b + h  = .0094985
3
Neutral
Shear 3 2
Axis tb h 3b + 2h
Center C w = ------------- ------------------- = .3010320
12 6b + h
t = .17
A = .986
n = .2968966
E = 3.  107 b = 1.315 I x = 1.579465
I y = .1720965
X o =  e + n  = .7659450

Figure 3-5 Simply-Supported Channel

Example of Tapered Beam

The cantilevered tapered flanged beam illustrated in Figure 3-6 is utilized to demonstrate the use of shear relief factors, S1
and S2, with the CBEAM element. As noted earlier in the section, shear relief provides for the fact that in a tapered flanged
beam some of the transverse shear load is carried by the flanges. All of the bending load is carried by the flanges. The shear
relief coefficient for the tip cell is computed by the following formula:
2  hA – hB  2  10. – 20. 
S 1 = -------------------------- = ----------------------------- = – .66667
 hA + hB   10. + 20. 

Main Index
 62 MSC Nastran Reference Guide
Line Elements

The PBEAM entry for this tip cell is illustrated below:

1 2 3 4 5 6 7 8 9 10
PBEAM 1 1 1. 60. 1.
5. -5.
YES 1. 2. 240.
10. -10.
-.666667

Note: A unique PBEAM entry is required for each CBEAM element as the shear relief factor among other properties
vary from element to element. Also note that the default values of 1. are accepted for the shear stiffness factors.

Flange Area = .6

WEB, tw = .1
V = 103 10.
60.

Flange Area = .6

100.

5 CBEAM Elements @ 20.

Figure 3-6 Tapered Cantilevered Beam With Shear Relief

Three-Node Beam Element (CBEAM3)

The Three-Node Beam Element is defined with a CBEAM3 entry and its properties are defined with a PBEAM3 entry. The
CBEAM3 element is described in the MSC Nastran Implicit Nonlinear User’s Guide. See the MSC Nastran Linear Static Analysis
User’s Guide, CBEAM3 Three-Node Beam Element. The material presented here will only be formulations and equations
not found in the other guides.

Main Index
 Chapter 3: Structural Elements 63
Line Elements

Formulation

Strain-Displacement Relations
The formulation of a three-node beam element is based on Timoshenko beam theory. The kinematical assumption
hereafter is that the displacement vector at an arbitrary point of the beam cross section is uniquely defined by the
displacement vector at the beam reference axis and the rotations of the cross section, superposed by a longitudinal warping
displacement. For small deformation theory, the rotation of the cross section is a vector. The displacement vector at point
(s, y, z) of a beam cross section is
u  s y z  = U x e x + U y e y + U z e z

where:
U x  s y z  = u  s  + z y  s  – y z  s  +   s   y z 
U y  s y z  =   s  – z x  s 
U z  s y z  = w  s  + y x  s 

in which u  s  , v  s  and w  s  are three translative displacement components at the beam axis;  x  s  ,  y  s  and
 z  s  are three rotational degrees of freedom of the beam cross section; and   y z  is a given warping function and
  s  the warping degree of freedom. The assumption of the warping displacement is based on the Saint-Venant torsion
theory of straight bars.
The Green-Lagrangian normal and shear strains,    y  z  , are

 =  0 + z y – y z +  s
 y =  y0 – z x +  y 
 z =  z0 + y x +  z 

where:

Main Index
 64 MSC Nastran Reference Guide
Line Elements

RB t
-----------------------
-
2 2
r 1–
0 = u  s + y w – z v
 y0 =   s –  z +  z u –  x w
 z0 = w  s +  y –  y u +  x 
 x =  x s –  z  y +  y  z
 y =  y s +  z  x
 z =  z s –  y  x

 y = ------- +  z 
y

 z = ------- –  y 
z
in which    s =      s  , the derivative with regard to the arc-length of beam axis.

Constitutive Relation
The general stress-strain relations are given, including thermal loads, at an arbitrary point of a beam cross-section, as

 C 11 C 12 C 13      T 
     
 =  y  = C 22 C 23    y  –  0 
     
 z  Sym C 33    z   0 

where C ij are material elastic constants,  the thermal expansion coefficient and T the variation of temperature.

Element Forces and Moments

The axial and shear forces as well as torsion and bending moments of the cross-section in the local coordinate system are

Main Index
 Chapter 3: Structural Elements 65
Line Elements

Nx =  dA
A

Vy =  y dA
A

Vz =  z dA
A

Mx =   yz – zy  dA
A

My =  z dA
A

Mz =   –y  dA
A
and the bi-shear force and bi-moment, which are associated with the warping degree of freedom and its derivative, are

Vb =   y y + z z  dA
A

Mb =   dA
A

Bar Element (CBAR)

The bar element is defined with a CBAR entry and its properties are defined with a PBAR entry. The bar element is a
one-dimensional bending element which is prismatic, and for which the elastic axis, gravity axis, and shear center all
coincide. The use of pin flags and the orientation of the bar element are similar to the beam element as shown in Figure 3-7.
If shearing deformations are included in the bar element, the reference axes (Planes 1 and 2) and the principal axes must
coincide.
The positive directions for element forces are shown in Figure 3-7. The following element forces, either real or complex
(depending on the rigid format), are output on request:

Main Index
 66 MSC Nastran Reference Guide
Line Elements

y V1

M 1a M 1b
T
x
Fx Fx T
a Plane 1 b

V1

z V2

M 2a M 2b
x

Plane 2

V2

Figure 3-7 CBAR Element Forces

where M 1a , M 1b , M 2a , and M 2b are the bending moments at both ends in the two reference planes, V 1 and V 2 are
the shear forces in the two reference planes, F x is the average axial force, and T is the torque about the x-axis.

The following real element stresses are output on request:

 Average axial stress.
 Extensional stress due to bending at four points on the cross section at both ends. (Optional, calculated only if the
user enters stress recovery points on the PBAR entry.)
 Maximum and minimum extensional stresses at both ends.

Main Index
 Chapter 3: Structural Elements 67
Line Elements

 Margins of safety in tension and compression for the whole element. (Optional, calculated only if the user enters
stress limits on the MAT1 entry.)
Tensile stresses are given a positive sign and compressive stresses a negative sign. Only the average axial stress and the
extensional stresses due to bending are available as complex stresses. The stress recovery coefficients on the PBAR entry are
used to locate points on the cross section for stress recovery. The subscript 1 is associated with the distance of a stress
recovery point from Plane 2. The subscript 2 is associated with the distance from Plane 1.
CBAR element force and stress data recovery with distributed loads (PLOAD1) and distributed mass (coupled mass) effects
included may be obtained at intermediate as well as end points from the dynamic solution sequences. The following items
must be in the input data:
 A LOADSET in Case Control which selects an LSEQ entry referencing PLOAD1 entries in the Bulk Data Section.
 Use PARAM,COUPMASS to select the coupled mass option for all elements.
The use of the BAROR Bulk Data entry avoids unnecessary repetition of input when a large number of bar elements either
have the same property identification number or have their reference axes oriented in the same manner. BAROR defines
default values on the CBAR entry for the property identification number and the orientation vector for the reference axes.
The default values are used only when the corresponding fields on the CBAR entry are blank.

Bend Element (CBEND)

The bend element is defined with a CBEND entry and its properties are defined with a PBEND entry. The BEND element
is a one-dimensional bending element with a constant radius of curvature. The CBEND element may be used to analyze
either curved beams or pipe elbows. The bend element includes extension, torsion, bending in two perpendicular planes,
and the associated transverse shear. The bend element cross section properties are assumed to be constant along the length
of the element and the principal axes coincide with the element coordinates shown in Figure 3-8. Structural and
nonstructural mass for the bend element is obtained only as consistent mass. The lumped mass option is not available.

Main Index
 68 MSC Nastran Reference Guide
Line Elements

Arc of the
Geometric Centroid
Center of
Curvature

z elem
Arc of the
Neutral Axis
Plane 2 B
End A GA
GB
R elem
v
 elem N
ZC
Plane 1
End B RC RB

(Note that Plane 1 is parallel to the

plane defined by GA, GB, and v.)

Figure 3-8 CBEND Element Coordinate System

Offsets of the ends of the element from the grid points are the same at both ends. The offsets are measured in the element
coordinate system as shown in Figure 3-8. The element coordinate system is defined by one of four methods that the user
specifies in the GEOM field on the CBEND entry. The Z-direction of the element coordinate system is defined by the cross
product v ab  v of the vector v ab connecting grid point GA to grid point GB and the vector v for GEOM = 1. For

GEOM = 2, 3, or 4, the Z-direction is defined by the cross product v  v ab . The center of curvature and intersection of
the tangent lines from end A and end B are located using the data required for each of the four options. The R-direction is
obtained by the vector extending from the center of curvature to end A. The  -direction is the cross product of Z  R .
When  = 0 , end A of the element is indicated and  =  B represents end B. Plane 1 of the element lies in the R

plane of the element coordinates. Plane 1 is parallel to the plane defined by GA, GB and the vector v , but it is offset by
ZC in the Z-direction. Plane 2 lies in the Z plane and is offset from GA and GB by RC in the R-direction. The subscripts
1 and 2 refer to forces and geometric properties associated with bending in Planes 1 and 2, respectively. These reference
planes are the principal planes of the element cross section.
The neutral axis radial offset shown in Figure 3-8 from the geometric centroid due to bending of a curved beam with a
constant radius of curvature is defined as follows:

Main Index
 Chapter 3: Structural Elements 69
Line Elements

RB
N = -------------------- (3-5)
2
AR R
1 + ----------
Z
where:

R B = the bend radius

A = the cross section area

2
r dA -
Z =  ---------------
R
1 + ------
RB
r = a local variable aligned with R elem direction

The user can use the default provided with the general format or he can calculate and input a value using the above formula.
For the circular section format, the neutral axis offset is automatically calculated with analytical expressions for hollow and
solid circular cross-sectional elements.
The flexibility factors which multiply the bending terms of the flexibility matrix are selected by the FSI field on the hollow
circular section format of the property entry. The options available are
for FSI = 1:
k p = 1.0

for FSI = 2:

2
1.65r 1
k p = --------------- --------------------------------------------------------- (3-6)
RB t Pr r 4  3  R 13
1 + 6 ------  -  ------B
Et  t   r

for FSI = 3:
1.66 1
k p = ---------- ------------------------------------------------------------------------- (3-7)
R B t 1 + 1.75 4  3
exp  – 1.5
–1  4

where:

Main Index
 70 MSC Nastran Reference Guide
Line Elements

 =
 = 2
PR
----------B
Ert
r = mean cross-sectional radius
t = wall thickness
P = internal pressure
E = Young’s modulus
 = Poisson’s ratio

The positive sign conventions for internal element forces are shown in Figure 3-9. The following element forces, either real
or complex (depending on the rigid format), are output on request at both ends:
 Bending moments in the two reference planes, M 1 and M 2 .

 Shears in the two reference planes, V 1 and V 2 .

 Average axial force, F  .

 Torque about the bend axis, M  .

z elem

Plane 2
GA
M1
R elem GB
V2
v
 elem
Plane 1 V1
F
M2 M

Figure 3-9 CBEND Element Internal Forces and Moments

Main Index
 Chapter 3: Structural Elements 71
Line Elements

The following real element stress data are output on request:

 Real longitudinal stress at the four points which are the same at both ends for the general cross-sectional property
entry format. If the circular cross-sectional property and format is used, the stress points are automatically located
at the points indicated on the following figure.

r elem
C

D
z elem
F

 Maximum and minimum longitudinal stresses.

 Margins of safety in tension and compression for the element if the user enters stress limits on the MAT1 entry.
 Whenever the pipe format is used, the stress data are modified to account for stress intensification resulting from
internal pressurization and curvature of the element. The internal pressure is prescribed on the property entry. The
methods used to calculate the stress intensification factor are selected through the FSI parameters; see the Remarks
under the PBEND in the MSC Nastran Quick Reference Guide.
Tensile stresses are given a positive sign and compressive stresses a negative sign. Only the longitudinal stresses are available
as complex stresses. The stress recovery coefficients on the general form of the PBEND entry are used to locate points on
the cross section for stress recovery. The subscript 1 is associated with the distance of a stress recovery point from Plane 2.
The subscript 2 is associated with the distance from Plane 1. If zero value stress recovery coefficients are used, the axial stress
is output.

Rod Elements (CROD, CONROD, CTUBE)

The rod element is defined with a CROD entry and its properties with a PROD entry. The rod element includes extensional
and torsional properties. The CONROD entry is an alternate form that includes both the connection and property
information on a single entry. The tube element is a specialized form that is assumed to have a circular cross section. The
tube element is defined with a CTUBE entry, and its properties with a PTUBE entry. The structural and nonstructural
mass of the rod are lumped at the adjacent grid points unless coupled mass is requested with the PARAM,COUPMASS
(see Parameters in the MSC Nastran Quick Reference Guide). Theoretical aspects of the rod element are treated in Section 5.2
of The NASTRAN Theoretical Manual.
The x-axis of the element coordinate system is defined by a line connecting end a to end b, as shown in Figure 3-10. The
axial force and torque are output on request in either real or complex form. The positive directions for these forces are
indicated in Figure 3-10.
The following real element stresses are output on request:

Main Index
 72 MSC Nastran Reference Guide
Line Elements

 Axial stress
 Torsional stress
 Margin of safety for axial stress
 Margin of safety for torsional stress
X

G2

G1

Figure 3-10 Rod Element Coordinate System and Element Forces

Positive directions are the same as those indicated in Figure 3-10 for element forces. Only the axial stress and the torsional
stress are available as complex stresses.
Another kind of rod element is the viscous damper. It has extensional and torsional viscous damping properties rather than
stiffness properties. The viscous damper element is defined with a CVISC entry and its properties with a PVISC entry. This
element is used in the formulation of dynamic matrices. The mode displacement method (PARAM,DDRMM,-1) must be
selected for element force output.

Supplied Beam and Bar Libraries

The Bulk Data entries PBARL and PBEAML provide libraries of cross sections. For open sections and the HAT1, thin-
walled theory is assumed.
The following pages provide equations that define the PBEAM and PBAR geometric property entries in terms of entries
on the PBEAML or PBARL. Symbols absent for a particular cross section are normally set to zero.

Main Index
 Chapter 3: Structural Elements 73
Line Elements

A Cross-sectional area.
yc Distance to centroid along Y element axis.
zc Distance to centroid along Z element axis.
ys Distance to shear center along Y element axis.
zs Distance to shear center along Z element axis.
I1 Moment of inertia about the Z element axis at the centroid. I 1 = I  zz 
elem
I2 Moment of inertia about the Y element axis at the centroid. I 2 = I  YY 
elem
I 12 Product moment of inertia at the centroid. I 12 = I  ZY 
elem
J Torsional Stiffness Constant.
C, D, E, F Location of the stress recovery points in the element coordinate system relative to the shear
center. For the PBARL the locations must be changed to be relative to the centroid. This can
be done by adding y na , z na to the listed equations.
K1 , K2 Shear stiffness factor for plane 1 and plane 2.
Iw Warping coefficient for the cross section relative to the shear center.
y na , z na Coordinates of the centroid relative to the shear center.

y elem

C
DIM1

F D z elem r = DIM1

TYPE = “ROD”

A = r 2

Main Index
 74 MSC Nastran Reference Guide
Line Elements

I1  
 = --4- r 4
I2 


J = --- r 4
2

C = r, 0
D = 0, r
E = – r, 0
F = 0, –r

K1 
 = 0.9
K2 

y elem

C
DIM1
r 0 = DIM1
D
F z elem
DIM2 r i = DIM2

E
TYPE = “TUBE”

Figure 3-11 Geometric Property Formulas for a ROD

A =   r 02 – r i2 

Main Index
 Chapter 3: Structural Elements 75
Line Elements

I1  
 = --4-  r 04 – r i4 
I2 


J = ---  r 04 – r i4 
2

C = r0 , 0
D = 0, r 0
E = –r0 , 0
F = 0, –r 0

K1 
 = 0.5
K2 

y elem
F C

b = DIM1
DIM2 z elem
h = DIM2

E D
DIM1

TYPE = “BAR”
Figure 3-12 Geometric Property Formulas for a TUBE Cross Section

A = bh

bh 3
I 1 = ---------
12

Main Index
 76 MSC Nastran Reference Guide
Line Elements

b3h
I 2 = ---------
12

I 12 = 0.

1 h h4
J = bh 3 --- – .21 ---  1 – ------------
3 b 12b 4

h b
C = --- , ---
2 2
h b
D = – --- , ---
2 2
h b
E = – --- , – ---
2 2
h b
F = --- , – ---
2 2

K1  5
 = --6-
K2 

y elem
DIM3
F C
b = DIM1
h = DIM2
DIM2 z elem t 1 = DIM3
t 2 = DIM4

E D b i = b – 2t 2
DIM4 h i = h – 2t 1
DIM1

TYPE = “BOX”
Figure 3-13 Geometric Property Formulas for a BAR

Main Index
 Chapter 3: Structural Elements 77
Line Elements

A = bh – b i h i

3
bh 3- b----------
-------- i hi
I1 = –
12 12

hb 3 h i b i3
I 2 = --------- – ----------
12 12

I 12 = 0.

2 2
2t 2 t 1  b – t 2   h – t 1 
J = ------------------------------------------------------
bt 2 + ht 1 – t 22 – t 12

h b
C = --- , ---
2 2
h b
D = – --- , ---
2 2
h b
E = – --- , – ---
2 2
h b
F = --- , – ---
2 2

hi t2
K 1 = ---------  2
A
bi t1
K 2 = ---------  2
A

Main Index
 78 MSC Nastran Reference Guide
Line Elements

y elem

DIM3
C a = DIM2
F
b = DIM3
h = DIM1
z elem
DIM6 ta = DIM5
DIM4 DIM1 t b = DIM6
DIM5 t w = DIM4
hw = h –  ta + tb 
E D h f = h – 0.5  t a + t b 
DIM2

TYPE = “I”
Figure 3-14 Geometric Property Formulas for a BOX

A = t a a + h w t w + bt b

1 1
y c = ---  --- h w  h w + t a t w + h f t b b
A2 
tb hf b 3
ys = -------------------------------
-
 tb b 3 + ta a 3 

y na = y c – y s

bt b3 at a3 t w h w3 2 2
I 1 = -------- + -------- + ------------ +  h f – y c  bt b + y c2 at a +  y c – 0.5  h w + t a   h w t w
12 12 12

b 3 t b t a a 3 h w t w3
I 2 = ---------- + ---------- + ------------
12 12 12

Main Index
 Chapter 3: Structural Elements 79
Line Elements

I 12 = 0.

1
J = ---  t b3 b + t a3 a + t w3 h f 
3
Note that I c and y s are relative to the center of flange defined by Dim1 and Dim4.

C = h f – y s + .5t b , .5b
D = – y – .5t a , .5a
E = – y – .5t a , – .5a
F = h f – y s + .5t b , – .5a

tw hw
K 1 = ------------
A

5  at a + bt b 
K 2 = ------------------------------
6A

h 2 tb ta b 3 a 3
I w = --------------------------------------
12  t b b 3 + t a a 3 

Note that I w and y s are based on thin walled formulations.

Main Index
 80 MSC Nastran Reference Guide
Line Elements

y elem

DIM1

F C D
d = DIM1
z elem
h = DIM2 – 0.5t f
DIM3 t f = DIM3
DIM2 t w = DIM4
h w = DIM2 – t f
DIM4

E
TYPE = “T”
Figure 3-15 Geometric Property Formulas for an I Section

A = dt f + h w t w

hw tw  hw + tf 
y na = – ---------------------------------
2A

dt f3 t w h w3 2 2
I 1 = -------- + ------------ + h w t w  y na + .5  h w + t f   + dt f y na
12 12

t f d 3 h w t w3
I 2 = --------- + ------------
12 12

I 12 = 0.

1
J = ---  t f3 d + t w3 h 
3

Main Index
 Chapter 3: Structural Elements 81
Line Elements

C = .5t f , 0
D = .5t f , .5d
E = – h, 0
F = .5t f , –.5d

hw tw
K 1 = ------------
A

tf d
K 2 = ------
A

y elem

F C b = DIM1 – 0.5t 2
DIM4 h = DIM2 – 0.5t 1
DIM2
DIM3 t 1 = DIM3
z elem t 2 = DIM4
E D h 2 = DIM2 – t 1
DIM1 b 1 = DIM1 – t 2

TYPE = “L”
Figure 3-16 Geometric Property Formulas for a T Section

A =  b + .5t 2 t 1 + h 2 t 2

t2 h2  h2 + t1 
y c = --------------------------------
2A
t1 b1  b1 + t2 
z c = --------------------------------
2A

t 13  b + .5t 2  t 2 h 23 2
I 1 = ---------------------------- + t 1  b + .5t 2 y c2 + ---------- + h 2 t 2  0.5  h 2 + t 1  – y c 
12 12

Main Index
 82 MSC Nastran Reference Guide
Line Elements

3
t 23 h 2 t 1  b + .5t 2  2
I 2 = ---------- + t 2 h 2 z c2 + ------------------------------ + t 1  b + .5t 2   .5b 1 – z 2 
12 12

I 12 = z c y c t 1 t 2 – b 1 t 1 y c  .5  b 1 + t 2  – z c  – h 2 t 2 z c  .5  h 2 + t 1  – y c 

1
J = ---  t 33 b + t 23 h 
3
C = h, .5t 2
D = – .5 t 1 , b
E = .5t 1 , – .5t 2
F = h, – .5t 2

h2 t2
K 1 = ----------
A
l
b1 t1
K 2 = ----------
A

y na = y c
z na = z c

Main Index
 Chapter 3: Structural Elements 83
Line Elements

y elem
DIM4
F C

b = DIM1 – .5t w
h = DIM2 – t f
DIM2 t w = DIM3
z elem
DIM3 t f = DIM4
b f = DIM1 – t w
h w = DIM2 – 2.t f
E D
DIM1

TYPE = “CHAN”
Figure 3-17 Geometric Property Formulas for an L Section

A = 2t f b f +  h + t t t w

bf tf  bf + tw 
z c = -----------------------------
A
b 2 tf
z s = ---------------------------------
 2bt + 1--- ht 
 w 3 f

3
h 2 t f b f b f t f3  h + t f  t w
I1 = -------------
- + --------- + -------------------------
2 6 12
(continued)

 h + t f t w3 b f3 t f 2
I 2 = ----------------------- + --------- + .5  b f + t w  b f t f – z c2 A
12 6

1
J = ---  2bt f3 + ht w3 
3

Main Index
 84 MSC Nastran Reference Guide
Line Elements

t f b 3 h 2  2t w h + 3t f b
I w = ---------------  ----------------------------
12  t w h + 6t f b 

C = .5  h + .5t f , z s + b
D = – .5  h + .5t f , z s + b
E = – .5  h + .5t f , z s – .5t w
F = .5  h + .5t f , z s – .5t w
tw hw
K 1 = ------------
A
2t f b f
K 2 = -----------
A

z na = z c + z s

Note that z c , z s are distances measured relative to an origin positioned at the center of the web.

y elem

DIM3 b 1 = DIM1
E F
h = DIM2
b 2 = DIM3
DIM4
DIM2 t 2 = DIM4
t 1 = DIM5
DIM5
z
C DIM1 D

TYPE=”HAT1”

Figure 3-18 Geometric Property Formulas for a Closed-hat Section

Main Index
 Chapter 3: Structural Elements 85
Line Elements

A = b 1 t 1 + t 2  b 1 – b 2  + 2t 2  h – t 1  + t 2  b 2 – 2t 2 

2 2 2
 b 1 t 1  2  + t 2  b 1 – b 2   t 1 + t 2  2  + t 2  h – t 1  + t 2  b 2 – 2t 2   h – t 2  2 
y c = -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
A

 t2
y s = ------ + ----
1 2

where:

1 = 3 2 2 3 3
t2  b1 – b2 + t2  t 2  b1 – b2 + t2   b1 + b2 – t2  t 2  h – t1 – t2   b2 – t2  t 2  b2 – t2  t1 b1
---------------------------------------- + ------------------------------------------------------------------------
- + --------------------------------------------------------
- + ---------------------------
- + ---------
-
6 16 2 12 12
= 2
 t1  b1 – b2 + t2   b1 + b2 – t2  t 1 b1
 b 2 – t 2   h – t 1  2 – t 2  2  ----------------------------------------------------------------------- + ---------- +  3
24 12
3
t2  b1 – b2 + t2   b1 + b2 – t2  t2  h – t1 – t2   b2 – t2  t1 t2 b1
+  b 2 – t 2   h – t 1 – t 2  ----------------------------------------------------------------------- + ------------------------------------------------------ +  3 + ---------------
8 4 12
2 3 2
t1 t2  b1 – b2 + t2   b1 + b2 – t2   b2 – t2  t 2 b1 t1 t2 b1  b2 – t2 
– ------------------------------------------------------------------------------------------------ + ---------- – ------------------------------------
8 12 12
2 2 2
t1 t2  b2 – t2   b1 – b2 + t2   b1 + b2 – t2  t 2  b2 – t2   b1 – b2 + t2   b1 + b2 – t2  t 2 b1  b2 – t2 
+ ------------------------------------------------------------------------------------------------ – -------------------------------------------------------------------------------------------- – -------------------------------
12 8 12
2
t2  b2 – t2   b1 – b2 + t2   b1 + b2 – t2 
+ --------------------------------------------------------------------------------------------
12

and
t1 t2 2
 3 = ------------------------------------------------------------------------------------
2t 1  h – t 1 – t 2  +  b 2 – t 2   t 1 + t 2 

where:

Main Index
 86 MSC Nastran Reference Guide
Line Elements

2 = 2 3
 b1 – b2 + t2   b1 + b2 – t2   h – t1 – t2  h – t 1 – t2   b2 – t2   b 2 – t2 
----------------------------------------------------------------------------------------------- + ---------------------------------------------------- + -----------------------
4 2 12
2 2
 b1 – b2 + t2   b1 + b2 – t2   b2 – t2  h – t 1 – t2   b2 – t2  b1  b2 – t2 
+ --------------------------------------------------------------------------------------- + ---------------------------------------------------- + --------------------------
6 2 12

y na = y s –  h – y c 

3 3
b1 t1 2  b 1 – b 2 t 2
I 1 = ---------- + b 1 t 1  y c – t 1  2  + --------------------------
12 12
2
+  b 1 – b 2 t 2  y c –  t 1 + t 2  2  
3
t2  h – t1  h – t1 2
+ ------------------------- + 2t 2  h – t 1   y c –  ------------- + t 1
6   2  
3
 b 2 – 2t 2 t 2 2
+ ---------------------------- +  b 2 – 2t 2 t 2  y c –  h – t 2  2  
12
3 3 2
t1 b1 t2  b1 – b2  t2  b1 – b2   b1 + b2 
I 2 = ---------- + ----------------------------- + ----------------------------------------------------
12 48 16
3 2
 h – t 1 t 2  h – t 1 t 2  b 2 – t 2 
+ ---------------------- + ----------------------------------------------
6 2
3
2t 2  b 2  2 – t 2 
+ ------------------------------------
3

I 12 = 0

2 2
4t 1 t 2  b 2 – t 2   h – t 1  2 – t 2  2  3
J = ------------------------------------------------------------------------------------ +  b 1 – b 2   t 1 + t 2   3
 b 2 – t 2   t 1 + t 2  + t 1  2h – t 1 – t 2 

C = ys – h , – b1  2

Main Index
 Chapter 3: Structural Elements 87
Line Elements

D = ys – h , b1  2

E = ys , –b2  2

F = ys , b2  2

n
1 2  h – t 1 t 2
K 1 = ---  -------------------------
-
n mA
i=1

n
1 1 – 2  h – t 1 t 2
K 2 = ---  ----------------------------------
-
n mA
i=1

where n is the number of sections and m = 2 for end A or end B and m = 1 for any intermediate station.

Adding Your Own Beam Cross-Section Library

The standard cross sections provided by MSC should be adequate in the majority of cases. If these standard sections are not
adequate for your purposes, you can add your own library of cross sections to suit your needs. To add your own library, you
need to write few simple subroutines in FORTRAN and link them to MSC Nastran through inter-process
communications.
The MSC Nastran Utilities Guide will describe to you the current server requirements and provide you with the location
of starter subroutines as described below.
This process requires writing and/or modifying up to eight basic subroutines:
1. BSCON -- Defines the number of dimensions for each of the section types.
2. BSBRP -- Calculates section properties based on section dimensions.
3. BSGRQ -- Defines NSECT, the number of section types, and NDIMAX, the maximum number of dimensions
(including nonstructural mass) required by any of the sections.
4. BSBRT -- Provides the name, number of dimensions and number of design constraints for each section type.
5. BSBRI -- Provides information for the calculation of gradients of section properties with respect to section
dimensions.
6. BSBRG -- Calculates any nonlinear gradients of section properties with respect to section dimensions.
7. BSBRC -- Defines constraints in the design of section dimensions.
8. BSMS -- A utility routine; handles errors that occur in the beam library.
BSCON and BSBRP are always required. BSGRQ, BSBRT, and BSBRI are required if you wish to perform sensitivity
and/or optimization tasks using the beam library. BSBRG is required if you are providing nonlinear analytical sensitivities

Main Index
 88 MSC Nastran Reference Guide
Line Elements

in the design task, and BSBRC is an optional routine that can be provided to help the optimizer to stay within physical
design constraints. BSMSG handles any error messages you feel are appropriate.

BSCON SUBROUTINE
This routine provides the number of fields in the continuation lines to be read from the Bulk Data entries PBARL and
PBEAML for each cross section in the library. The value of the ENTYP variable may be 0, 1, or 2. When ENTYP = 0, the
value returned is the number of DIMi. When ENTYP = 1, the value returned includes both the DIMi and NSM fields.
The value of 1 is used for PBARL only. When ENTYP = 2, the value returned includes the DIMi, NSM, SO, and XIXB
fields for 11 different stations. The value of 2 applies to PBEAML only.
The calling sequence and example routine for the standard MSC library is given below.
SUBROUTINE BSCON(GRPID,TYPE,ENTYP,NDIMI,ERROR)
C ----------------------------------------------------------------------
C Purpose
C To get the number of maximum fields in continuation entries for
C each section in the library.
C
C Arguments:
C
C GRPID input integer Integer id of this group or group name.
C Not used, reserved for future use.
C TYPE input character*8 Name of cross section
C ENTYP input integer O: dimensions only without NSM
1: PBARL, total # of data items for 2:PBEAML
C NDIMI output integer Number of dimi fields for the 'ENTYP'
C section
C ERROR output integer Error code
C
C Called by BCCON
C ----------------------------------------------------------------------
C=== Argument Type Declaration
INTEGER GRPID,ENTYP,NDIMI,ERROR
CHARACTER*8 TYPE

C=== Default to 'nothing wrong'

ERROR = 0

C=== Dimensions vary with section type

IF ( TYPE.EQ.'ROD ') THEN
NDIMI = 1
ELSEIF( TYPE.EQ.'TUBE ' .OR. TYPE.EQ.'BAR ') THEN
NDIMI = 2
ELSEIF( TYPE.EQ.'HEXA ') THEN
NDIMI = 3
ELSEIF( TYPE.EQ.'BOX ' .OR. TYPE.EQ.'T ' .OR.
+ TYPE.EQ.'L ' .OR.
+ TYPE.EQ.'CHAN ' .OR. TYPE.EQ.'CROSS ' .OR.
+ TYPE.EQ.'H ' .OR.
+ TYPE.EQ.'I1 ' .OR. TYPE.EQ.'T1 ' .OR.
+ TYPE.EQ.'CHAN1 ' .OR.
+ TYPE.EQ.'Z ' .OR. TYPE.EQ.'CHAN2 ' .OR.
+ TYPE.EQ.'T2 ' .OR. TYPE.EQ.'HAT ') THEN

Main Index
 Chapter 3: Structural Elements 89
Line Elements

NDIMI = 4
ELSEIF( TYPE.EQ.'I ' .OR. TYPE.EQ.'BOX1 ') THEN
NDIMI = 6
ELSE
C=== Set error code if invalid name for the section
ERROR = 5150
RETURN
ENDIF

C=== Number of data items to be read on PBARL entry is DIMi

C=== plus the NSM field

IF (ENTYP.EQ.1) NDIMI = NDIMI+1

C=== Number of data items to be read on PBEAML entry is DIMi

C=== plus the NSM, SO and X/XB fields for eleven different
C=== stations.

IF (ENTYP.EQ.2) NDIMI = (NDIMI+3)*11

C--------------------------------------------------------------
RETURN
END

BSBRP Subroutine
Finite element analysis requires section properties such as area, moment of inertia, etc., instead of section dimension.
Therefore, the dimensions specified on PBARL and PBEAML need to be converted to equivalent properties usually
specified on PBAR and PBEAM entries. The images of all these entries are stored in EPT data block as records.
BSBRP subroutine is the interface of your properties evaluator with MSC Nastran. You may use your own naming
convention for the subroutines that calculate the cross-section properties from the dimensions. The calling tree used for the
MSC standard library is shown in Figure 3-19.

BSBRP

MEVBR MEVBM

OTHER
BRTUP BRBXP SECTIONS BMTUP BMBXP

Main Index
 90 MSC Nastran Reference Guide
Line Elements

Figure 3-19 Calling Tree Generate Property Data

The BSBRP calls the bar evaluator routine MEVBR for the BAR element and the beam evaluator routine MEVBM for the
BEAM element. The evaluators in turn call the routines for each section. The routines are named BRXXP and BMXXP,
where XX is a two-letter identifier for the section. For example, the routines for the TUBE section are called BRTUP and
BMTUP. The details for various routines are given in Listing 3-1.

Listing 3-1 BSBRP Subroutines

SUBROUTINE BSBRP(GRPID,ENTYP,TYPE,IDI,NID,IDO,NIDO,DIMI,NDIMI,
+ DIMO,NDIMO,ERROR)
C=======================================================================
C PURPOSE: This is the interface for the properties evaluator with
C MSC Nastran.
C
C ARGUMENTS
C GRPID input integer The ID of group name
C ENTYP input integer 1: PBARL, 2: PBEAML
C TYPE input character*8 Arrays for cross section types
C IDI input integer Array containing the integer words
C in the PBARL or PBEAML EPT record
C NID input integer Dimension of the IDI array. It is
C equal to two.
C IDO output integer Array containing the integer words
C in the PBAR or PBEAM EPT record.
C NIDO output integer Dimension of the IDO array. It is
C equal to two for PBAR and four for
C the PEBAM EPT record.
C DIMI input real Array containing the floating words
C in the PBARL or PBEAML EPT record
C for the 'TYPE' section
C NDIMI input integer Dimension of the DIMI array.
C DIMO output real Array containing the real words for
C the PBAR or PBEAM EPT record.
C NDIMO input integer Dimension of the DIMO array. It is
C equal to 17 for the PBAR and 193 for
C the PBEAM EPT record.
C ERROR output integer Error code
C
C-----------------------------------------------------------------------
C CALLED BY:
C BCBRP
C-----------------------------------------------------------------------
C CALLS:
C MEVBR ,MEVBM
C-----------------------------------------------------------------------
C IMPLICIT DECLARATIONS
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)
C.......................................................................
C EXPLICIT DECLARATIONS
INTEGER ENTYP,ERROR,GRPID
CHARACTER*8 TYPE
C-----------------------------------------------------------------------
C DIMENSION STATEMENTS

Main Index
 Chapter 3: Structural Elements 91
Line Elements

INTEGER IDI(NID), IDO(NIDO)

DOUBLE PRECISION DIMI(NDIMI), DIMO(NDIMO)
C=======================================================================
C=== ENTYP=1, FOR PBAR1; 2, FOR PBEAM1
IF (ENTYP.EQ.1) THEN
CALL MEVBR(GRPID,TYPE,IDI,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO
+ ,ERROR)
ELSE IF (ENTYP.EQ.2) THEN
CALL MEVBM(GRPID,TYPE,IDI,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO
+ ,ERROR)
END IF
C-----------------------------------------------------------------------
RETURN
END
MEVBR and MEVBM Subroutines

MEVBR and MEVBM are the branched routines for the various sections, and convert the section dimensions to section
properties for Bar and Beam elements. You may rename these routines as you like or move the function of these routines to
BSBRP. These routines call the BRXXP routines where XX is the two-letter keyword for various section types. The MEVBR
routine for the MSC standard library is given in Listing 3-2.

Listing 3-2 MEVBR Subroutine

SUBROUTINE MEVBR(GRPID,TYPE,ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,
+ NDIMO,ERROR)

C=======================================================================
C Purpose
C Call the default type subroutine to convert PBAR1 to PBAR
C
C Arguments
C
C GRPID input int ID of group
C TYPE input char Type of cross section
C ID input int Array of values PID, MID contained in PBAR1
C entries
C NID input int Size of ID array, NID=2 for PBAR1 entry
C IDO output int Array of integer values contained in PBAR
C entries
C NIDO output int Size of IDO array, NIDO=2 for PBAR entry
C DIMI input flt Dimension values of cross section
C NDIMI input int Size of DIMI array
C DIMO output flt Properties of cross section
C NDIMO output flt Size of DIMO array
C ERROR output int Type of error
C
C Method
C Call the subroutine with respect to the section type
C
C Called by
C BSBRP
C
C CALLS
C BRRDPD,BRTUP,BRBRPD,BRBXP,BRIIPD,BRTTPD,BRLLPD,BRCHPD
C-----------------------------------------------------------------------

Main Index
 92 MSC Nastran Reference Guide
Line Elements

IMPLICIT INTEGER (I-N)

IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER ERROR,GRPID,ID(NID),IDO(NIDO)
CHARACTER*8 TYPE
DOUBLE PRECISION DIMI(NDIMI),DIMO(NDIMO)

C=======================================================================

C Clear the output array before usage

CALL ZEROD ( DIMO, NDIMO )
CALL ZEROI ( IDO, NIDO )
ERROR = 0
C
IF ( TYPE.EQ.'ROD ') THEN
CALL BRRDPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'TUBE ') THEN
CALL BRTUP(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'BAR ') THEN
CALL BRBRPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'BOX ') THEN
CALL BRBXP(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'I ') THEN
CALL BRIIPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'T ') THEN
CALL BRTTPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'L ') THEN
CALL BRLLPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'CHAN ') THEN
CALL BRCHPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'CROSS ') THEN
CALL BRCRPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'H ') THEN
CALL BRHHPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'T1 ') THEN
CALL BRT1PD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'I1 ') THEN
CALL BRI1PD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'CHAN1 ') THEN
CALL BRC1PD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'Z ') THEN
CALL BRZZPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'CHAN2 ') THEN
CALL BRC2PD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'T2 ') THEN
CALL BRT2PD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'BOX1 ') THEN
CALL BRB1PD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'HEXA ') THEN
CALL BRHXPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSEIF( TYPE.EQ.'HAT ') THEN
CALL BRHTPD(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
ELSE
ERROR = 5150
END IF
C-----------------------------------------------------------------------

Main Index
 Chapter 3: Structural Elements 93
Line Elements

RETURN
END
BRXXP and BMXXP Subroutines

The purpose of the BRXXP and BMXXP routines is to calculate the properties from the section dimensions. For each cross
section, subroutines are required to convert the images of PBARL and PBEAML records to the images of PBAR and
PBEAM records in the EPT data block.
BRTUP Subroutine

BRTUP is an example routine that shows how to convert PBARL EPT record to PBAR EPT record for the Tube section.
First, the details of the PBARL and the PBAR record are shown, and then the routine itself is given.
PBARL Record

The PBARL record in the EPT data block is a derived from the PBARL Bulk Data entry and is given below.

Table 3-4 PBARL (9102, 91, 52)

Word Name Type Description
1 PID I Property identification number.
2 MID I Material identification number.
3 Group Char Group Name.
4 Group Char Group Name.
5 TYPE Char4 Cross-section Type.
6 TYPE Char4 Cross-section Type.
7 Dim1 RS Dimension 1.
8 Dim2 RS Dimension 2.
n+7-1 Dim n RS Dimension n (note that the final dimension is the nonstructural
mass).
n+7 Flag I -1. Flag indicating end of cross-section dimensions.

PBAR Record

The PBAR record in the EPT data block is derived from the PBAR Bulk Data entry and consists of 19 words. It is a replica
of the Bulk Data entry, starting with PID field. The word 8 in the record is set to 0.0 since the field 9 in the first line of the
PBAR Bulk Data entry is not used. The details of the PBAR record are given in Table 3-5.

Main Index
 94 MSC Nastran Reference Guide
Line Elements

Table 3-5 PBAR (52, 20, 181)

Word Name Type Description
1 PID I Property identification number.
2 MID I Material identification number.
3 A RS Area of cross-section.
4 I1 RS Area moment of inertia for bending in plane 1.
5 I2 RS Area moment of inertia for bending in plane 2.
6 J RS Torsional constant.
7 NSM RS Nonstructural mass per unit length.
8 FE RS Not used. Set to 0.0.
9 C1 RS Stress recovery location.
10 C2 RS Stress recovery location.
11 D1 RS Stress recovery location.
12 D2 RS Stress recovery location.
13 E1 RS Stress recovery location.
14 E2 RS Stress recovery location.
15 F1 RS Stress recovery location.
16 F2 RS Stress recovery location.
17 K1 RS Area factor of shear for plane 1.
18 K2 RS Area factor of shear for plane 2.
19 I12 RS Area product of inertia.

Listing 3-3 BRTUP Subroutine

SUBROUTINE BRTUP(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
C=======================================================================
C Purpose:
C Convert PBAR1(entity type : TUBE) to PBAR
C
C Arguments:
C
C DIMI input flt Array of dimension values for cross-section
C (SEE FIG. 5 IN MEMO SSW-25, REV. 4, DATE 8/16/94)
C NDIM input int Size of DIMI array
C DIMO output flt Array of property values for cross-section
C NDIMO output int Size of DIMO array
C ERROR output int Type of error
C

Main Index
 Chapter 3: Structural Elements 95
Line Elements

C DISCRIPTION FOR DIMO ARRAY:

C DIMO (1) = A
C DIMO (2) = I1
C DIMO (3) = I2
C DIMO (4) = J
C DIMO (5) = NSM
C DIMO (6) = FE
C DIMO (7) = C1
C DIMO (8) = C2
C DIMO (9) = D1
C DIMO (10) = D2
C DIMO (11) = E1
C DIMO (12) = E2
C DIMO (13) = F1
C DIMO (14) = F2
C DIMO (15) = K1
C DIMO (16) = K2
C DIMO (17) = I12
C
C Method
C
C Called by:
C MEVBR
C-----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER ID(NID),IDO(NIDO),ERROR
DOUBLE PRECISION DIMI(NDIMI),DIMO(NDIMO)
C=======================================================================
C
C=== WRITE THE PART OF INTEGER
DO 30 II = 1,NID
IDO(II) = ID(II)
30 CONTINUE
C
DIM1 = DIMI(1)
DIM2 = DIMI(2)
DIMO(1) = PI*(DIM1*DIM1-DIM2*DIM2)
DIMO(2) = PI*(DIM1**4-DIM2**4)/4.D0
DIMO(3) = DIMO(2)
DIMO(4) = PI*(DIM1**4-DIM2**4)/2.D0
DIMO(5) = DIMI(3)
DIMO(6) = 0.D0
DIMO(7) = DIM1
DIMO(8) = 0.D0
DIMO(9) = 0.D0
DIMO(10) = DIM1
DIMO(11) = -DIM1
DIMO(12) = 0.D0
DIMO(13) = 0.D0
DIMO(14) = -DIM1
DIMO(15) = 0.5D0
DIMO(16) = 0.5D0
DIMO(17) = 0.D0

Main Index
 96 MSC Nastran Reference Guide
Line Elements

IF ( DIMI(1).LE.DIMI(2) ) ERROR = 5102

C-----------------------------------------------------------------------
RETURN
END
BMTUP Subroutine

BMTUP is an example routine that shows how to convert PBEAML EPT record to PBEAM EPT record for the Tube
section. First, the details of the PBEAML and the PBEAM records are shown, and then the routine itself is given.
PBEAML Record

The PBEAML record in the EPT data block is a derived from the PBEAML Bulk Data entry and is given in Table 3-6.

Table 3-6 PBEAML (9202, 92, 53)

Word Name Type Description
1 PID I Property ID
2 MID I Material ID
3 Group Char Group Name
4 Group Char Group Name
5 TYPE Char4 Cross-section Type
6 TYPE Char4 Cross-section Type
7 SO RS Stress output request flag, 1=yes,=no
8 XXB RS X/XB - parametric location of the station
9 Dim1 RS Dimension 1
10 Dim2 RS Dimension 2
n+9-1 Dim n RS Dimension n (note that the final dimension is the nonstructural
mass)
n+9 Flag I -1. Flag indicating end of cross-section dimensions
Words 7 through ndim+11 repeat 11 times.

PBEAM Record

The PBEAM record in EPT data block consists of 197 words. The first five words and the last 16 words are common to all
the 11 stations. Each of the 11 stations have their own 21 unique words. The details of the PBEAM record are given in
Table 3-7.

Main Index
 Chapter 3: Structural Elements 97
Line Elements

Table 3-7 PBEAM (5402, 54, 262)

Word Name Type Description
1 PID I Property identification number.
2 MID I Material identification number.
3 N I Number of intermediate stations.
4 CCF I Constant cross-section flag. 1 = constant, 2 = variable.
5 X RS Unused.
6 SO RS Stress output request. 1.0 = yes, 0.0 = no.
7 XXB RS Parametric location of the station. Varies between 0. and 1.0.
8 A RS Area.
9 I1 RS Moment of inertia for bending in plane 1.
10 I2 RS Moment of inertia for bending in plane 2.
11 I12 RS Area product of inertia.
12 J RS Torsional constant.
13 NSM RS Nonstructural mass.
14 C1 RS Stress recovery location.
15 C2 RS Stress recovery location.
16 D1 RS Stress recovery location.
17 D2 RS Stress recovery location.
18 E1 RS Stress recovery location.
19 E2 RS Stress recovery location.
20 F1 RS Stress recovery location.
21 F2 RS Stress recovery location.
Words 6 through 21 repeat 11 times.
182 K1 RS Area factor for shear for plane 1.
183 K2 RS Area factor for shear for plane 2.
184 S1 RS Shear-relief coefficient for plane 1.
185 S2 RS Shear-relief coefficient for plane 2.
186 NSIA RS Nonstructural mass moment of inertia at end A.
187 NSIB RS Nonstructural mass moment of inertia at end B.
188 CWA RS Warping coefficient for end A.

Main Index
 98 MSC Nastran Reference Guide
Line Elements

Table 3-7 PBEAM (5402, 54, 262) (continued)

Word Name Type Description
189 CWB RS Warping coefficient for end B.
190 M1A RS Y-coordinate of center of gravity for nonstructural mass at end A.
191 M2A RS Z-coordinate of center of gravity for nonstructural mass at end A.
192 M1B RS Y-coordinate of center of gravity for nonstructural mass at end B.
193 M2B RS Z-coordinate of center of gravity for nonstructural mass at end B.
194 N1A RS Y-coordinate for neutral axis at end A.
195 N2A RS Z-coordinate for neutral axis at end A.
196 N1B RS Y-coordinate for neutral axis at end B.
197 N2B RS Z-coordinate for neutral axis at end B.

Listing 3-4 BMTUP Subroutine

SUBROUTINE BMTUP(ID,NID,IDO,NIDO,DIMI,NDIMI,DIMO,NDIMO,ERROR)
C=======================================================================
C Purpose
C Convert PBEAM1(entity type : TUBE) to PBEAM
C
C Arguments
C
C ID input int Contain the integer information PID, MID
C NID input int Size of ID array, NID = 2
C DIMI input flt Dimension values of cross section
C ( See FIG.5 IN MEMO SSW-25, REV. 4, DATE 8/16/94)
C NDIMI input int Size of DIMI array
C IDO output int Contain the integer information PID,MID,N,CCF
C NIDO output int Size of IDO array, NIDO = 4
C DIMO output flt Properties of cross section
C NDIMO output int Size of DIMO array
C ERROR output int Type of error
C
C Description for DIMO array
C DIMO (1) = X
C DIMO (2) = SO
C DIMO (3) = XXB
C DIMO (4) = A
C DIMO (5) = I1
C DIMO (6) = I2
C DIMO (7) = I12
C DIMO (8) = J
C DIMO (9) = NSM
C DIMO (10) = C1
C DIMO (11) = C2
C DIMO (12) = D1
C DIMO (13) = D2
C DIMO (14) = E1

Main Index
 Chapter 3: Structural Elements 99
Line Elements

C DIMO (15) = E2
C DIMO (16) = F1
C DIMO (17) = F2
C DIMO(2) thru DIMO(17) repeat 11 times
C DIMO (178) = K1
C DIMO (179) = K2
C DIMO (180) = S1
C DIMO (181) = S2
C DIMO (182) = NSIA
C DIMO (183) = NSIB
C DIMO (184) = CWA
C DIMO (185) = CWB
C DIMO (186) = M1A
C DIMO (187) = M2A
C DIMO (188) = M1B
C DIMO (189) = M2B
C DIMO (190) = N1A
C DIMO (191) = N2A
C DIMO (192) = N1B
C DIMO (193) = N2B
C
C Method
C Simply calculate the properties and locate that data
C to the image of PBEAM entries
C Called by
C MEVBM
C-----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER ERROR,ID(NID),IDO(NIDO)
DOUBLE PRECISION DIMI(NDIMI),DIMO(NDIMO)

C Local variables
INTEGER NAM(2)

C NASTRAN common blocks

COMMON /CONDAD/PI

C Local data
DATA NAM/4HBMTU,4HPD /
C=======================================================================
C
C=== WRITE THE PART OF INTEGER
DO 30 II = 1,NID
IDO(II) = ID(II)
30 CONTINUE
C-----------------------------------------------------------------------
C=== DETECT HOW MANY STATION , CONSTANT OR LINEAR BEAM.
C-----------------------------------------------------------------------
ISTATC = NDIMI/11
DO 35 II = 0,10
NW = II*ISTATC
IF (DIMI(3+NW).EQ.0.D0) THEN
IDO(3) = II-2
IDO(4) = 1

Main Index
 100 MSC Nastran Reference Guide
Line Elements

IF (DIMI(3).NE.DIMI(3+NW-ISTATC)) IDO(4)=2
GO TO 40
END IF
35 CONTINUE
40 DIMO(1) = 0.D0
DO 100 L1 = 0,10
LC = 16*L1
NW = L1*ISTATC
IF (DIMI(3+NW+ISTATC).EQ.0.D0) LC = 160
DIM1 = DIMI(3+NW)
DIM2 = DIMI(4+NW)

IF ( DIM1.LE.DIM2 ) ERROR = 5102

DIMO(2+LC) = DIMI(1+NW)
DIMO(3+LC) = DIMI(2+NW)
DIMO(4+LC) = PI*(DIM1*DIM1-DIM2*DIM2)
DIMO(5+LC) = PI*(DIM1**4-DIM2**4)/4.D0
DIMO(6+LC) = DIMO(5+LC)
DIMO(7+LC) = 0.D0
DIMO(8+LC) = PI*(DIM1**4-DIM2**4)/2.D0
DIMO(9+LC) = DIMI(5+NW)
DIMO(10+LC) = DIM1
DIMO(11+LC) = 0.D0
DIMO(12+LC) = 0.D0
DIMO(13+LC) = DIM1
DIMO(14+LC) = -DIM1
DIMO(15+LC) = 0.D0
DIMO(16+LC) = 0.D0
DIMO(17+LC) = -DIM1
IF (LC.EQ.160) GO TO 110
100 CONTINUE
110 DIMO(178) = 0.5D0
DIMO(179) = 0.5D0
C-----------------------------------------------------------------------
300 RETURN
END

BSGRQ Subroutine
For optimization of PBARL entries, you need to provide overall information such as number of cross sections in your library
and the maximum number of fields in the continuation lines. BSGRQ provides the information and is required only if you
wish to perform sensitivity or optimization with the section dimensions. The calling sequence and example routine is given
in Listing 3-5.

Listing 3-5 BSGRQ Routine

SUBROUTINE BSGRQ(GRPID,NSECT,NDIMAX,ERROR)
C
C ======================================================================
C PURPOSE:
C PROVIDE OVERALL CHARACTERISTICS OF A BEAM SECTION LIBRARY
C ----------------------------------------------------------------------
C ARGUMENTS:
C

Main Index
 Chapter 3: Structural Elements 101
Line Elements

C GRPID input integer Group name. Not used, reserved for future
c use.
C NSECT output integer Number of different section types
C NDIMAX output integer Maximum number of dimension for any
C section type
C ERROR output integer Indicates if an error has occurred. The
C code returned indicates the type of error
C ----------------------------------------------------------------------
C CALLED BY:
C BCGRQ routine
C-----------------------------------------------------------------------
C
C=== Set number of section in the library
nsect = 19
C=== Set the maximum number of DIMi fields (including 1 for nonstructured mass)
required by any one section
ndimax = 7
C-----------------------------------------------------------------------
RETURN
END
Note that the arguments GRPID and ERROR are not used. GRPID is reserved for future use. You may use the ERROR
argument to send an error code which could later be used to print an error message.

BSBRT Subroutine
The BSBRT routine provides the name, number of fields in the continuation line in the PBARL Bulk Data entry, and the
number of constraints for each section in the library. As an example, the name of the tube section, shown in Figure 3-20, in
the MSC standard library is “TUBE”, the number of dimensions for the tube section is three (OUTER RADIUS, INNER
RADIUS, and NSM), and there is one physical constraint.
The physical constraint is that the inner radius (DIM2) can not be greater than outer radius (DIM1). It is necessary to
specify the constraints so that the optimization of the section dimension in SOL200 does not result into an inconsistent
shape.

Y
r 0 = DIM1
r i = DIM2
C
DIM1

D
F Z
DIM2

Main Index
 102 MSC Nastran Reference Guide
Line Elements

Figure 3-20 Tube Section

This routine is called during optimization process. The calling sequence and example routine is given in Listing 3-6.

Listing 3-6 BSBRT Subroutine

SUBROUTINE BSBRT(GRPID,ENTYP,TYPE,NDIM,NCONST,NSECT,ERROR)
C ======================================================================
C Purpose
C Get the name, number of fields in the continuation line and
C number of constraints
C to be used by optimization routines
C
C Arguments
C
C GRPID input integer The ID of group name
C ENTYP input integer 1: PBARL, 2: PBEAML
C TYPE output character*8 Arrays for cross section types
C NDIM output integer Number of dimensions for isect
C section type, including 1 for
nonstructural mass
C NCONST output integer Number of dimensional constraints
C imposed by isect section type
C NSECT input integer Number of sections
C ERROR output integer Error code
C
C Called by
C BCBRT routine
C ----------------------------------------------------------------------
C=== Argument Type Declaration
INTEGER GRPID,ENTYP,NDIM(NSECT),NCONST(NSECT)
CHARACTER*8 TYPE(NSECT)

C=== Local variables

INTEGER NAM(2)
C ======================================================================
C=== Currently, only PBARL is supported for optimization. Based on
C=== ENTYP the library may have different number of DIMi fields and
C=== number of constraints. Currently, GRPID and ENTYP are not being
C=== used. So just set them to default values, even though they are
C=== input type.
GRPID = 1
ENTYP = 1

C ----------------------------------------------------------------------
C=== Set the name, number of fields in the continuation line for the PBARL entry,
C=== and number of constraints in the TYPE, DIMI and NCONST arrays,
respectively.
C=== Note that the value in DIMI is one more (for NSM field) than DIMi fields
C=== Make sure names are all capitals, even if they are lower case in the
input data file.
TYPE(1) = 'ROD'
NDIM(1) = 2
NCONST(1) = 0
C
TYPE(2) = 'TUBE'

Main Index
 Chapter 3: Structural Elements 103
Line Elements

NDIM(2) = 3
NCONST(2) = 1
C
TYPE(3) = 'BAR'
NDIM(3) = 3
NCONST(3) = 0
C
TYPE(4) = 'BOX'
NDIM(4) = 5
NCONST(4) = 2
C
TYPE(5) = 'I'
NDIM(5) = 7
NCONST(5) = 3
C
TYPE(6) = 'T'
NDIM(6) = 5
NCONST(6) = 2
C
TYPE(7) = 'L'
NDIM(7) = 5
NCONST(7) = 2
C
TYPE(8) = 'CHAN'
NDIM(8) = 5
NCONST(8) = 2
C
TYPE(9) = 'CROSS'
NDIM(9) = 5
NCONST(9) = 1
C
TYPE(10) = 'H'
NDIM(10) = 5
NCONST(10) = 1
C
TYPE(11) = 'T1'
NDIM(11) = 5
NCONST(11) = 1
C
TYPE(12) = 'I1'
NDIM(12) = 5
NCONST(12) = 1
C
TYPE(13) = 'CHAN1'
NDIM(13) = 5
NCONST(13) = 1
C
TYPE(14) = 'Z'
NDIM(14) = 5
NCONST(14) = 1
C
TYPE(15) = 'CHAN2'
NDIM(15) = 5
NCONST(15) = 2
C
TYPE(16) = 'T2'
NDIM(16) = 5

Main Index
 104 MSC Nastran Reference Guide
Line Elements

NCONST(16) = 2
C
TYPE(17) = 'BOX1'
NDIM(17) = 7
NCONST(17) = 2
C
TYPE(18) = 'HEXA'
NDIM(18) = 4
NCONST(18) = 1
C
TYPE(19) = 'HAT'
NDIM(19) = 5
NCONST(19) = 2

C
C ======================================================================
RETURN
END

BSBRI Subroutine
The BSBRI subroutine is required if optimization is to be performed. Its function is to provide information required in the
calculation of the sensitivities (gradients) of the bar properties with respect to the bar dimensions.
Two basic types of information are provided. The first is the SENTYP array, which indicates how each section property
varies as a function of each dimension. Values in the SENTYP array can be either: 0 for no variation; 1 for a linear variation;
2 for a nonlinear variation; or 3 for an unknown variation. The SENTYP = 3 option is to be used when you know that the
property is a function of the design dimension, but analytical gradient information is not being provided using the BSBRG
subroutine. In this case, MSC Nastran will calculate the gradients for you using central differencing techniques.
The second piece of information is the ALIN array. This array provides any linear sensitivity data. For example, the C1
stress recovery location for the TUBE section is 1.0  DIM1 so that this sensitivity of this stress recovery point with respect
to the first dimension is 1.0.
You may use your own naming convention for the subroutines that specify the section sensitivity data. The calling tree used
for the MSC standard library is shown in Figure 3-21.

Main Index
 Chapter 3: Structural Elements 105
Line Elements

BSBRI

MSBRI

OTHER
BRTUI BRBXI SECTIONS

Figure 3-21 Calling Tree to Generate Sensitivity Data

BSBRI calls the bar evaluator routine MSBRI/S for the BAR element while the corresponding beam evaluator routine is
absent since the PBEAM1 dimensions cannot be designed. The evaluator in turn calls the routines for each section. The
routines are named BRXXID, where XX is a two-letter identifier for the section. For example, the routine for the TUBE
section is called BRTUI. The details for various routines are given in Listing 3-7.

Listing 3-7 BSBRI Subroutine

SUBROUTINE BSBRI(GRPID,ENTYP,SECTON,SENTYP,ALIN,NDIM,NPROP,
1 ERROR)

C
C =====================================================================
C Purpose
C set up section dependent information for a particular cross
C section type
C
C Arguments
C
C GRPID input integer ID of the group
C ENTYP input integer 1: PBAR1, 2: PBEAM1
C SECTON input character*8 Section type
C SENTYP output integer Type of sensitivity, 0: invariant,
C 1: linear, 2: nonlinear, 3: calculated by
finite difference
C ALIN output double Matrix providing the linear
C factors for sensitive relationships
C NDIM input integer Number of dimensions
C NPROP input integer Number of properties
C ERROR output integer Type of error
C
C Method

Main Index
 106 MSC Nastran Reference Guide
Line Elements

C Simply transfer control based on entry type

C
C Called by
C BCBRID
C
C Calls
C MSBRI, MSBMID
C ----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER ENTYP,ERROR,GRPID,SENTYP(NPROP,NDIM)
CHARACTER*8 SECTON
DOUBLE PRECISION ALIN(NPROP,NDIM)
C ======================================================================
GRPID = 1
IF(ENTYP.EQ.1) THEN
CALL MSBRI(SECTON,SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE
ERROR = 5400
END IF

C
RETURN
END
MSBRI Subroutine

MSBRI is a branched routine for providing information on the calculation of sensitivities for each of the bar types. You may
rename this routine as you like or move its function to BSBRI. MSBRI calls the BRXXID routines, where XX is the two-
letter keyword for various section types. The routine MSBRI for the MSC standard library is given in Listing 3-8.

Listing 3-8 MSBRI Subroutine

SUBROUTINE MSBRI(SECTON,SENTYP,ALIN,NDIM,NPROP,ERROR)
C
C =====================================================================
C Purpose
C To set up section dependent information for PBAR1 cross section
C
C Arguments
C
C SECTON input character*8 Name of section type
C SENTYP output integer Type of sensitivity, 0: invariant,
C 1: linear, 2: nonlinear
C ALIN output double Matrix providing the linear factors
C for sensitive relationships
C NDIM input integer No. of dimensions
C NPROP input integer No. of properties in EPT datablock
C ERROR output integer Type of error
C
C Method
C Simply transfer the section dependent information of
C the 19 kinds
C Called by

Main Index
 Chapter 3: Structural Elements 107
Line Elements

C BSBRI
C
C Calls
C BRRDID, BRBRID, BRTUI, BRBXI, BRIIID, BRLLID, BRTTID,
C BRCHID, BRCRID, BRHHID, BRT1ID, BRI1ID, BRC1ID, BRZZID,
C BRC2ID, BRT2ID, BRB1ID, BRHXID, BRHTID
C ZEROI, ZEROD (Nastran utility)
C ----------------------------------------------------------------------

IMPLICIT INTEGER (I-N)

IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

CHARACTER*8 SECTON
INTEGER ERROR,SENTYP(NPROP,NDIM)
DOUBLE PRECISION ALIN(NPROP,NDIM)
C ======================================================================
C
CALL ZEROI( SENTYP(1,1), NPROP*NDIM )
CALL ZEROD( ALIN(1,1), NPROP*NDIM )
ERROR = 0

C
IF(SECTON.EQ.'ROD') THEN
CALL BRRDID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'BAR') THEN
CALL BRBRID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'TUBE') THEN
CALL BRTUI(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'BOX') THEN
CALL BRBXI(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'I') THEN
CALL BRIIID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'L') THEN
CALL BRLLID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'T') THEN
CALL BRTTID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'CHAN') THEN
CALL BRCHID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'CROSS') THEN
CALL BRCRID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'H') THEN
CALL BRHHID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'T1') THEN
CALL BRT1ID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'I1') THEN
CALL BRI1ID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'CHAN1') THEN
CALL BRC1ID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'Z') THEN
CALL BRZZID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'CHAN2') THEN
CALL BRC2ID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'T2') THEN
CALL BRT2ID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'BOX1') THEN
CALL BRB1ID(SENTYP,ALIN,NDIM,NPROP,ERROR)

Main Index
 108 MSC Nastran Reference Guide
Line Elements

ELSE IF(SECTON.EQ.'HEXA') THEN

CALL BRHXID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE IF(SECTON.EQ.'HAT') THEN
CALL BRHTID(SENTYP,ALIN,NDIM,NPROP,ERROR)
ELSE
ERROR = 5410
END IF

C
RETURN
END
BRTUI Subroutine

BRTUI is an example routine that shows how to define the sensitivity type of each of the bar properties for the tube and
the subset of the sensitivities that are linear.

Listing 3-9 BRTUI Subroutine

SUBROUTINE BRTUI(SENTYP,ALIN,NDIM,NPROP,ERROR)
C
C ======================================================================
C Purpose
C To set up section dependent information for rod cross section
C
C Arguments
C
C SENTYP output integer Type of sensitivity, 0: invariant,
C 1: linear, 2: nonlinear
C ALIN output double Matrix providing the linear factors for
C sensitive relationships
C NDIM input integer No. of dimensions
C NPROP input integer No. of properties in EPT datablock
C ERROR output integer Type of error
C
C Method
C Simply provides the information
C
C Called by
C MSBRI
C ----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER ERROR,SENTYP(NPROP,NDIM)
DOUBLE PRECISION ALIN(NPROP,NDIM)

C Local variables
INTEGER NAM(2)

C Local data
DATA NAM/4HBRTU,4HID /
C ======================================================================
C
ALIN( 7,1) = 1.0D0

Main Index
 Chapter 3: Structural Elements 109
Line Elements

ALIN(10,1) = 1.0D0
ALIN(11,1) = -1.0D0
ALIN(14,1) = -1.0D0
ALIN( 5,3) = 1.0D0
C
SENTYP( 1,1) = 2
SENTYP( 1,2) = 2
SENTYP( 2,1) = 2
SENTYP( 2,2) = 2
SENTYP( 3,1) = 2
SENTYP( 3,2) = 2
SENTYP( 4,1) = 2
SENTYP( 4,2) = 2
SENTYP( 7,1) = 1
SENTYP(10,1) = 1
SENTYP(11,1) = 1
SENTYP(14,1) = 1
SENTYP( 5,3) = 1

C
RETURN
END

BSBRG Subroutine
The BSBRG subroutine is required if optimization is to be performed and analytical sensitivities are needed (SENTYP = 2
in subroutine BSBRI). Its function is to provide the nonlinear gradients of the bar properties with respect to the bar
dimensions. You may use your own naming convention for the subroutines that calculate the section gradients from the
dimensions. The calling tree used for the MSC standard library is shown in Figure 3-22.

BSBRG

MSBRG

OTHER
BRTUG BRBXG SECTIONS

Figure 3-22 Calling Tree to Generate Nonlinear Sensitivity Data

Main Index
 110 MSC Nastran Reference Guide
Line Elements

BSBRG calls the bar evaluator routine MSBRG/S for the BAR element. The evaluator, in turn, calls the routines for each
section. The routines are named BRXXGD, where XX is a two-letter identifier for the section. For example, the routine for
the TUBE section is called BRTUG. The details for various routines are given in Listing 3-10.

Listing 3-10 BSBRG Subroutine

SUBROUTINE BSBRG(GRPID,ENTYP,TYPE,DIMI,NDIMI,ANONL,NPROP,ERROR)
C
C ======================================================================
C Purpose
C To get the nonlinear factors of sensitivities for default sections
C
C Arguments
C
C GRPID input integer ID of group name
C ENTYP input integer 1: PBAR1, 2: PBEAM1
C TYPE input character*8 Type name of cross-section
C DIMI input double Array from EPT record
C NDIMI input integer Number of dimensions (Plus NSM)
C ANONL output double Array providing the nonlinear factors
C for sensitivity relationships
C NPROP input integer Number of properties in PBAR entries
C ERROR output integer Type of error
C
C Method
C Simply transfer control based on entry types
C Called by
C BCBRGD
C
C Calls
C MSBRG
C ----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER GRPID,ENTYP,NDIMI,NPROP,ERROR
CHARACTER*8 TYPE
DOUBLE PRECISION ANONL(NPROP,NDIMI), DIMI(NDIMI)

C Local variables
INTEGER NAM(2)

C Local data
DATA NAM/4HBSBR,4HGD /
C ======================================================================

GRPID = 1
IF(ENTYP .EQ. 1) THEN
CALL MSBRG(TYPE,DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE
ERROR = 5300
END IF
C ----------------------------------------------------------------------
RETURN
END

Main Index
 Chapter 3: Structural Elements 111
Line Elements

MSBRG Subroutine

MSBRG is a branched routine for providing information on the calculation of nonlinear gradients for each of the bar types.
You may rename this routine as you like or move its function to BSBRG. MSBRG calls the BRXXGD routines, where XX
is the two-letter keyword for various section types. The routine MSBRG for the MSC standard library is given in Listing 3-11.

Listing 3-11 MSBRG Subroutine

SUBROUTINE MSBRG(TYPE,DIMI,NDIMI,ANONL,NPROP,ERROR)
C
C ======================================================================
C Purpose
C To get the nonlinear factors of sensitivities for PBAR1 entries
C
C Arguments
C
C TYPE input character*8 Type name of cross-section
C DIMI input double Array from EPT record
C NDIMI input integer Number of dimensions (Plus NSM)
C ANONL output double Array providing the nonlinear factors
C for sensitivity relationships
C NPROP input integer Number of properties in PBAR entries
C ERROR output integer Type of error
C
C Method
C Simply transfer information based on cross-section type
C Called by
C BSBRG
C
C Calls
C BRRDGD,BRBRGD,BRBXG,BRTUG,BRIIGD,BRTTGD,BRLLGD,BRCHGD
C BRCRGD,BRHHGD,BRT1GD,BRT2GD,BRI1GD,BRC1GD,BRC2GD,BRZZGD
C BRHXGD,BRB1GD,BRHTGD,ZEROD(Nastran utility)
C ----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER NDIMI,NPROP,ERROR
CHARACTER*8 TYPE
DOUBLE PRECISION ANONL(NPROP,NDIMI), DIMI(NDIMI)
C ======================================================================

CALL ZEROD( ANONL(1,1), NPROP*NDIMI )

ERROR = 0

IF(TYPE.EQ.'ROD') THEN
CALL BRRDGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'BAR') THEN
CALL BRBRGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'BOX') THEN
CALL BRBXG(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'TUBE') THEN
CALL BRTUG(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'I') THEN
CALL BRIIGD(DIMI,NDIMI,ANONL,NPROP,ERROR)

Main Index
 112 MSC Nastran Reference Guide
Line Elements

ELSE IF(TYPE.EQ.'L') THEN

CALL BRLLGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'T') THEN
CALL BRTTGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'CHAN') THEN
CALL BRCHGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'CROSS') THEN
CALL BRCRGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'H') THEN
CALL BRHHGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'T1') THEN
CALL BRT1GD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'I1') THEN
CALL BRI1GD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'CHAN1') THEN
CALL BRC1GD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'Z') THEN
CALL BRZZGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'CHAN2') THEN
CALL BRC2GD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'T2') THEN
CALL BRT2GD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'BOX1') THEN
CALL BRB1GD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'HEXA') THEN
CALL BRHXGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE IF(TYPE.EQ.'HAT') THEN
CALL BRHTGD(DIMI,NDIMI,ANONL,NPROP,ERROR)
ELSE
ERROR = 5310
ENDIF

C
RETURN
END
BRTUG Subroutine

BRTUG is an example routine that shows how to define the nonlinear gradients of the TUBE section as a function of the
outer and inner radius of the tube.

Listing 3-12 BRTUG Subroutine

SUBROUTINE BRTUG(DIMI,NDIMI,ANONL,NPROP,ERROR)
C
C ======================================================================
C Purpose
C To get the nonlinear factors of sensitivities for TUBE section
C
C Arguments
C
C DIMI input double Array of EPT records (Dimi+NSM)
C NDIMI input integer Number of dimensions (Plus NSM)
C ANONL output double Array providing the nonlinear factors
C for sensitivity relationships
C NPROP input integer Number of properties in PBAR entries

Main Index
 Chapter 3: Structural Elements 113
Line Elements

C ERROR output integer Type of error

C
C Method
C Simply calculates the nonlinear factors
C Called by
C MSBRG
C ----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER NDIMI,NPROP,ERROR
DOUBLE PRECISION ANONL(NPROP,NDIMI), DIMI(NDIMI)

C Nastran common blocks

COMMON /CONDAD/ PI
C ======================================================================

DIM1 = DIMI(1)
DIM2 = DIMI(2)
PDIM13 = PI*DIM1**3
PDIM23 = PI*DIM2**3
C
ANONL(1,1) = 2*PI*DIM1
ANONL(1,2) = -2*PI*DIM2
ANONL(2,1) = PDIM13
ANONL(2,2) = -PDIM23
ANONL(3,1) = PDIM13
ANONL(3,2) = -PDIM23
ANONL(4,1) = 2*PDIM13
ANONL(4,2) = -2*PDIM23

C
RETURN
END

BSBRC Subroutine
The BSBRC subroutine allows you to place constraints on values the beam dimensions can take during a design task. It is
not needed unless optimization is used and, even then, is available only to impose conditions on the dimensions to keep
the optimization process from selecting physically meaningless dimensions.
For example, the optimizer might select a TUBE design with the inner radius greater than the outer radius because this
allows for a negative area and therefore a negative weight (something a weight minimization algorithm loves!) These
constraints are not the same as the PMIN and PMAX property limits that are imposed on the DVPREL1 entry. Instead,
these are constraints that occur between or among section dimensions. A DRESP2 entry could be used to develop the same
design constraints, but the subroutine reduces the burden on the user interface, the primary goal of the beam library project.
The calling tree used for the MSC standard library, which is shown in Figure 3-23.

Main Index
 114 MSC Nastran Reference Guide
Line Elements

BSBRC

MSBRC

OTHER
BRTUC BRBXC SECTIONS
Figure 3-23 Calling Tree to Generate Beam Dimension Constraints

BSBRC calls the bar evaluator routine MSBRC. The evaluator in turn calls the routines for each section that requires
constraints. The routines are named BRXXCD, where XX is a two-letter identifier for the section. For example the routine
for the TUBE section is called BRTUC. The details for various routines are given in Listing 3-13.

Listing 3-13 BSBRC Subroutine

SUBROUTINE BSBRC(GRPID,ENTYP,TYPE,AFACT,NCONST,NDIMI,ERROR)
C
C ======================================================================
C Purpose
C To get constraint information for default types
C
C Arguments
C
C GRPID input integer ID of the group name
C ENTYP input integer 1: PBAR1, 2: PBEAM1
C TYPE input character*8 Section type
C NCONST input integer Number of constraints
C for the section type
C NDIMI input integer Number of dimensions
C for the section type
C AFACT output double The factor for the NDIMI dimension in
C the constraint relation. Dimensions are
C NCONST by NDIMI.
C ERROR output integer type of error
C
C Method
C Simply transfers control based on PBAR1 or PBEAM1 entries
C
C Called by
C BCBRCD
C

Main Index
 Chapter 3: Structural Elements 115
Line Elements

C Calls
C MSBRC
C ----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

CHARACTER*8 TYPE
INTEGER GRPID,ENTYP,NCONST,NDIMI,ERROR
DOUBLE PRECISION AFACT(NCONST,NDIMI)
C ======================================================================
C
GRPID = 1
IF(ENTYP.EQ.1) THEN
CALL MSBRC(TYPE,AFACT,NCONST,NDIMI,ERROR)
END IF
C ----------------------------------------------------------------------
RETURN
END
MSBRC Subroutine

MSBRC is a branched routine for providing information on the calculation of gradients for each of the bar types. You may
rename this routine as you like or move its function to BSBRC. MSBRC calls the BRXXCD routines, where XX is the
two-letter keyword for various section types. The routine MSBRC for the MSC standard library is given in Listing 3-14.

Listing 3-14 MSBRC Subroutine

SUBROUTINE MSBRC(TYPE,AFACT,NCONST,NDIMI,ERROR)
C=================================================================
C Purpose
C To get constraint information for PBAR1 types
C
C Arguments
C
C TYPE input character*8 Section type
C AFACT output double The factor for the NDIMI
C dimension in the constraint
C relation. Dimensions are NCONST
C by NDIMI.
C NCONST input integer Number of constraints
C for the section type
C NDIMI input integer Number of dimensions
C for the section type
C ERROR output integer type of error
C
C Method
C Simply transfers control based on group name
C
C Called by
C BSBRCT
C
C Calls
C BRTUC, BRBXC, BRIICD, BRTTCD, BRLLCD, BRCHCD,
C BRCRCD, BRHHCD, BRT1CD, BRI1CD, BRC1CD, BRZZCD,
C BRC2CD, BRT2CD, BRB1CD, BRHXCD, BRHTCD, ZEROD

Main Index
 116 MSC Nastran Reference Guide
Line Elements

C (Nastran utility)
C -----------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

CHARACTER*8 TYPE
INTEGER NCONST,NDIMI,ERROR
DOUBLE PRECISION AFACT(NCONST,NDIMI)

C =================================================================

CALL ZEROD( AFACT(1,1), NCONST*NDIMI )

ERROR = 0

IF(TYPE.EQ.'TUBE') THEN
CALL BRTUC(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'BOX') THEN
CALL BRBXC(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'I') THEN
CALL BRIICD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'L') THEN
CALL BRLLCD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'T') THEN
CALL BRTTCD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'CHAN') THEN
CALL BRCHCD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'CROSS') THEN
CALL BRCRCD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'H') THEN
CALL BRHHCD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'T1') THEN
CALL BRT1CD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'I1') THEN
CALL BRI1CD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'CHAN1') THEN
CALL BRC1CD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'Z') THEN
CALL BRZZCD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'CHAN2') THEN
CALL BRC2CD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'T2') THEN
CALL BRT2CD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'BOX1') THEN
CALL BRB1CD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'HEXA') THEN
CALL BRHXCD(AFACT,NCONST,NDIMI,ERROR)
ELSE IF(TYPE.EQ.'HAT') THEN
CALL BRHTCD(AFACT,NCONST,NDIMI,ERROR)
ELSE
ERROR = 5500
END IF

Main Index
 Chapter 3: Structural Elements 117
Line Elements

C
RETURN
END
BRTUC Subroutine

BRTUC is an example routine that shows how to define the constraints for a bar section. This example routine is for the
TUBE section and imposes a single constraint that – DIM1 + DIM2  0.0 , where DIM1 is the outer radius and DIM2
is the inner radius of the tube. The constraints should always be specified so that the specified linear combination of
dimensions is less or equal to zero when the constraint is satisfied.

Listing 3-15 BRTUC Subroutine

SUBROUTINE BRTUC(AFACT,NCONST,NDIMI,ERROR)
C
C ======================================================================
C Purpose
C To get constraint information for TUBE type
C
C Arguments
C
C AFACT output double The factor for the NDIMI dimension in the
C constraint relation. Dimensions are NCONST
C by NDIMI.
C NCONST input integer Number of constraints for the section type
C NDIMI input integer Number of dimensions for the section type
C ERROR output integer type of error
C
C Method
C Simply transfers constraint information
C
C Called by
C MSBRC
C ----------------------------------------------------------------------
IMPLICIT INTEGER (I-N)
IMPLICIT DOUBLE PRECISION (A-H,O-Z)

C Calling sequence arguments

INTEGER NCONST,NDIMI,ERROR
DOUBLE PRECISION AFACT(NCONST,NDIMI)
C ======================================================================
C
IF(NCONST.NE.1) THEN
ERROR = 5502
RETURN
END IF
C
AFACT(1,1) = -1.0D0
AFACT(1,2) = 1.0D0

C
RETURN
END

Main Index
 118 MSC Nastran Reference Guide
Line Elements

BSMSG Subroutine
The Error handling is performed by a subroutine called bsmsg.f. This routine has the following parameters:
SUBROUTINE BSMSG(GRPID,ERRCOD,MXLEN,Z,ERROR)
C ======================================================================
C Purpose
C To handle the error messages for User Defined Group.
C ----------------------------------------------------------------------
C Arguments
C
C GRPID input int ID of group or group name - not used
C ERRCOD input int Error message number if any found
C MXLEN input int Maximum length of the message that can be
C passed
C Z output char Array to contain the message return
C ERROR output int The code returned indicates the type of error
C
C Called by
C BCMSG routine
C---------------------------------
The purpose of this subroutine is to return an error String associated with an error code. The error codes are to be returned
by the other 7 “BS...” routines and as such the “BSMSG” routine is used as the repository of all of the error messages for
the Beam Library applications.
For example, suppose that “BSBRP” application returns an error code of 5103 when a certain error condition occurs. Then,
the Beam Library Client routines will expect that there will be a String returned from the “BSMSG” routine, which
corresponds to this error code. These error messages will be printed in the “*.f06” file to guide the user as to what the error
could have been and how to fix it. The string may be as long as 160 characters for this release of MSC Nastran.
The following is an example of BSMSG code construct:
IF(ERRCOD .EQ. 5103) THEN
Z(1:MXLEN) = 'This is a User Specified Error Message ....'//
& 'Messages could be 160 characters long ...'
ERROR = ERRCOD
.....
Again, it is highly recommended that you use the example Beam Server files as a template to generate your BSMSG routines.

Linking Your Library to MSC Nastran

Once you have created the eight “BS...” routines, these routines may be linked with MSC Nastran Beam Server Library to
build a beam server executable. It is highly recommended that you study, build, and use the example “Beam Server” before
you build your own version of the beam server.
The MSC Nastran special library contains a main routine as well as the communications routines that allow MSC Nastran
to communicate with the user-defined beam server.
The MSC Nastran user may connect up to 10 beam servers in a single job execution. This connection is made using the
concept of evaluator groups described in the remainder of this section. For each group, the user specifies on the

Main Index
 Chapter 3: Structural Elements 119
Line Elements

PBEAML/PBARL entry referring to an external beam server, MSC Nastran will start and communicate with the beam
server.
You may define as many beam evaluators as required using the “CONNECT” FMS commands. Only 10 of these
evaluators, however, may be referenced in the groups on the PBARL or PBEAML Bulk Data entries.
The PBARL/PBEAML entries specifies a “Group” name on the fourth field of the first entry. This group name is associated
with an “Evaluator” class using a “Connect” command in the FMS section. Finally, the “Evaluator” class is associated with
an executable using the MSC Nastran configuration file specified via the “gmconn” key word on the nastran command line.
The following example shows the mappings mentioned:
1. Group is referenced on the PBARL/PBEAML entry (or entries).
PBARL,39,6,LOCSERV,I_SECTION (Specify the “Group” name.)
2. The “Group” is associated with an “Evaluator” class.
CONNECT,BEAMEVAL,LOCSERV,EXTBML (Associate the “Group” name with an “Evaluator” class.)
3. The external evaluator connection file associates the “Evaluator” class with a server executable. The following
statement must be specified in the connection file:
EXTBML,-,beam_server_pathname
4. Refer to the external evaluator connection file on the command line using the “gmconn” keyword
nastran myjob ... gmconn=external_evaluator_pathname.
5. In the example Beam Library section that standard input (FORTRAN unit 5) or standard output (FORTRAN unit
6) are not used as these I/O channels are reserved by the Inter-Process Communications (IPC) subsystem.

Example of Building and Linking a Beam Server

As an example of building and linking a beam server executable, the sample beam server will be modified. Complete
instructions on building and using a beam server are provided in the MSC Nastran Utilities Guide for your system.
 Make a copy of the beam server sample source.
 Edit the source for the BRTUP subroutine; this routine describes the equations that convert the PBARL
dimensions into the standard PBAR dimensions for a tube cross section.
 Add an extra multiplication of 3.0 to the DIMO(2) equation to increase the calculated moments of inertia.
Since the formulation of this bar section has been changed, the sensitivities for optimization will also change. Rather than
calculate what the new sensitivities should be, the MSC Nastran can calculate them using central differencing techniques.
To permit this, edit the source file for the BRTUI subroutine and change all occurrences of SENTYP = 2 to SENTYP = 3.
Build your new beam server using the instructions detailed in the MSC Nastran 2016 Utilities Guide. Once you have built
the beam server executable, you must create an external evaluator connection to point to your executable. Typically, this file
would be kept in the user's home directory, but for this example it will remain in the current directory. Edit the new file
bmconfig.fil. Put the following line in the file:
LOCBMLS,-,pathname

Main Index
 120 MSC Nastran Reference Guide
Line Elements

where LOCBMLS is the evaluator referenced in the SAMPLE data file included with the beam server. Remember, this file
can contain references to any number of beam servers.
To run the sample job, type in the following command:
nastran sample scr=yes bat=no gmconn=bmconfig.fil
Common problems which may occur when attempting to run an external beam library job are generally indicated in the
F06 by USER FATAL MESSAGE 6498. If this message includes the text “No such group defined,” the PBARL/PBEAML
selected a group not defined on a CONNECT entry. If UFM 6498 includes the text “No such evaluator class,” either the
“gmconn” keyword was not specified or the CONNECT entry selected an evaluator not defined in the configuration file.
If the job was successful, you can look at the Design Variable History and see that the results for the variable mytubeor are
different than the results for tubeor. These variables refer to the outer radius of tube sections from equivalent models. One
model used the MSC tube section while the other used the tube section in your modified beam server.

Support for Offsets in Element Coordinates for CBAR & CBEAM Elements
Bar elements (CBAR and CBEAM Bulk Data entries) have been used for many years to model structural members that can
transmit axial loads, shear forces and bending moments. The offset feature of these elements allows the analyst to position
the beam axis relative to the line connecting the two grid points used in the definition of the element.

Theory
Structural stiffeners such as I-beams, wide flange beams, channel sections, z-sections and hat sections are typically modeled
using the bar and beam finite elements available in MSC Nastran. When these members are used to support flat or curved
panel structures, it is often desirable to include the offset effects of the member neutral/shear axis location when modeling
the structure. These effects are most generally prevalent when the grid points are located at the mid-plane or outer surface
of the panels. Under these circumstances, when a beam element is attached to these points, the line connecting the beam
element end grid points does not correctly represent the element neutral axis location relative to the panel surface. To handle
this situation, the CBAR and CBEAM Bulk Data entries provide for the specification of offset vectors for each end of the
element that locate the positions of the element end points with respect to the locations of the element connection grid
points.
Until now, the components of the offset vectors had to be specified in the principal directions of the output coordinate
systems of the connection grid points. The program would then transform each of the offset vectors so that the components
were measured in the basic coordinate reference system of the model. The locations of the end grid points are also converted
to this basic system if necessary. Adding the offsets to the coordinates of the grid points produces the element end point
locations. The vector between the two end points together with the orientation vector can be used to generate the element
coordinate system. Thus, when defining element offset vectors, the analyst must keep in mind the fact that the vectors must
be specified in the output coordinate system directions of the connection grid points. For example, in the following figure,
the local x direction at grid point GA is in the opposite direction of the local y direction at grid point GB and the local y
at grid point GA and the local x at grid point GB are in the same direction. Let us assume that the element x-axis is offset
0.5 units in the local x direction at grid point GA from the line connecting grid points GA and GB. Then, in order to define
the neutral axis of the element properly, the analyst must define the End A offset vector as Wa = {0.5, 0.0, 0.0} and at the
same time specify the End B offset vector as Wb = {0.0, -0.5, 0.0}.

Main Index
 Chapter 3: Structural Elements 121
Line Elements

X GA
End A End B
Wa Wb

GA GB
Y GA X GB
Y GB
Figure 3-24

If we use the concept of the local element coordinate system, then no discussion of the local output coordinate systems
employed at the connection grid points is necessary. For this case, the element x-axis is from GA to GB and the orientation
vector is in the direction of X GA , defining the direction of the element- y-axis and the x-y plane. The offset vector
components will be the same at both ends and will be specified as Wa = Wb = {0.0, 0.5, 0.0}. The program will take care
of any necessary transformations between the element system, the basic system and the local systems at the end grid points.

Inputs
The specification of the offset vectors for the two ends of the bar and beam elements is accomplished using fields four
through nine of the first continuation entry for the CBAR and CBEAM Bulk Data entries respectively. The orientation
(reference) vector is specified using fields 6, 7 and 8 of the bulk data entries.
How MSC Nastran interprets the contents of these fields depends upon the offset and reference vector specification codes
supplied in field nine of the main (parent) bulk data entry. Each of the entries is described by one character. There must be
three characters in the string if it is present: The first character is associated with the orientation vector . The second
character is associated with the end A offset vector. The third character is associated with the end B offset vector.
The character “G” signifies that the input is specified in the global coordinate reference system. The character “B” signifies
that the input is specified in the basic coordinate reference system. The character “O” signifies that the input is specified
local offset reference system.

String Orientation Vector End A Offset End B Offset

GGG Global Global Global
BGG Basic Global Global
GGO Global Global Offset
BGO Basic Global Offset
GOG Global Offset Global
BOG Basic Offset Global
GOO Global Offset Offset
BOO Basic Offset Offset

Main Index
 122 MSC Nastran Reference Guide
Line Elements

For the example given in the theory section, for Wa = Wb = {0.0, 0.5, 0.0}, the OFFT code used would be GOO. Invalid
codes will give a User Fatal Message.

Outputs
There are no special output considerations when the offset vectors are specified in the local element coordinate system.

Example
A simple model is presented to demonstrate the specification of the offset vectors in the element coordinate system. The
input file contains two CBAR elements that use offsets. One of the CBARs uses the original method of specifying the offsets
in the global system components at the connection grid points. The second of the CBARs uses the new technique of
specifying the offsets in the element coordinate system. The displacements and element forces should be the same for the
two elements if the offset vectors are “equivalent”.

Model Description
Two CBAR elements are used to demonstrate the use of the offset vectors in the element coordinate system. The grid points
that define the ends of the elements are situated along the basic y-axis. The output coordinate system at each grid point is
the basic coordinate system. The global system is equivalent to the basic system. The orientation vector for both elements
places the element x-y plane at a thirty degree cant to the basic y-z plane toward the negative x-direction. Element 101
defines the offsets to be -0.50 units in the global (basic) x-direction and 0.866 units in the global (basic) z-direction. This
is standard legacy usage of the offset vectors. Element 111 defines the offsets to be 1.0 units in the y-direction of the local
element coordinate system. One end of each bar is completely constrained resulting in a cantilever beam model. Three
subcases are defined to apply unit loads to the free end of the beam in each of the basic directions. Element forces are
recovered for the CBAR elements for each subcase.

Input File
id test,cbar
sol 101
cend
$
title = v2004 release notes demonstration problem
subtitle = test cbar offset vector specification options
label= two equivalent cbars
disp=all
spcf=all
elfo=all
spc = 100
$
$ unit load in basic x-direction
$
subcase 100
load=100
$
$ unit load in basic y-direction
$
subcase 200
load = 200

Main Index
 Chapter 3: Structural Elements 123
Surface Elements

$
$ unit load in basic z-direction
$
subcase 300
load = 300
begin bulk
$
$============================================================
$ offset bar example.
$ grid output cstm is basic
$ offset in element coordinate system is {0.0,1.0,0.0}(t)
$ which is equivalent to {-0.5,0.0,0.866025}(t) in basic
$ element force results should be the same for both elements
$
grid,1001,,0.0,11.0,0.0
grid,1002,,0.0,20.0,0.0
$
$ Standard legacy input.
$ Element 101 specifies offsets and orientation vector in the
$ global coordinate system.
$
cbar,101,1101,1001,1002,-0.50,0.0,.866025,,+br101
+br101,,,-0.50,0.0,0.866025,-0.50,0.0,0.86602
$
$ New feature option input.
$ Element 111 specifies offsets in the local element system and
$ orientation vector in the global coordinate system.
$
cbar,111,1101,1001,1002,-0.50,0.0,.866025,goo,+br111
+br111,,,0.0,1.0,0.0,0.0,1.0,0.0
$
force,100,1002,,1000.0,1.0,0.0,0.0
force,200,1002,,1000.0,0.0,1.0,0.0
force,300,1002,,1000.0,0.0,0.0,1.0
spc1,100,123456,1001
$
$============================================================
$ common data
$
pbar,1101,1,1.0,1.0,1.0,1.0
mat1,1,10.7+6,,.33
enddata

Surface Elements
The surface elements consist of the following:
 Shear panel
 Shell
 Conical shell

Main Index
 124 MSC Nastran Reference Guide
Surface Elements

Shear Panel Element (CSHEAR)

The shear panel is defined with a CSHEAR entry and its properties with a PSHEAR entry. A shear panel is a
two-dimensional structural element that resists the action of tangential forces applied to its edges, and the action of normal
forces if effectiveness factors are used on the alternate form of the PSHEAR Bulk Data entry. The structural and
nonstructural mass of the shear panel is lumped at the connected grid points. Details of the shear panel element are
discussed in Section 3.0 of The NASTRAN Theoretical Manual.
The element coordinate system for a shear panel is shown in Figure 3-25. The labels G1, G2, G3, and G4 refer to the order
of the connected grid points on the CSHEAR entry. The element forces are output on request in either real or complex
form.
G3

G4

Y elem

X elem
G2
G1

Figure 3-25 Shear Panel Connection and Coordinate System

The positive directions for these forces are indicated in Figure 3-26. These forces consist of the forces applied to the element
at the corners in the direction of the sides, kick forces at the corners in a direction normal to the plane formed by the two
adjacent edges, and shear flows (force per unit length) along the four edges. The shear stresses are calculated at the corners
in skewed coordinates parallel to the exterior edges. The average of the four corner stresses and the maximum stress are
output on request in either the real or complex form. A margin of safety is also output when the stresses are real.

Main Index
 Chapter 3: Structural Elements 125
Surface Elements

K4 F 41
K3
F 32
F 43 q3
G4 G3
F 34

q4
K2 q2
K1 F 21

G2

q1
F 12 G1
F 23

F 14

Figure 3-26 CSHEAR Element Corner Forces and Shear Flows

Shell Elements (CTRIA3, CTRIA6, CTRIAR, CQUAD4, CQUAD8, CQUADR)

MSC Nastran includes two different shapes of isoparametric shell elements (triangular and quadrilateral) and two different
stress systems (membrane and bending). There are in all a total of six different forms of shell elements that are defined by
connection entries as follows:
 CTRIA3 -- Isoparametric triangular element with optional coupling of bending and membrane stiffness.
 CTRIA6 -- Isoparametric triangular element with optional coupling of bending and membrane stiffness and
optional midside nodes.
 CTRIAR -- Isoparametric triangular element with no coupling of bending and membrane stiffness; the membrane
stiffness formulation includes rotation about the normal to the plane of the element.
 CQUAD4 -- Isoparametric quadrilateral element with optional coupling of bending and membrane stiffnesses.
 CQUAD8 -- Isoparametric quadrilateral element with optional coupling of bending and membrane stiffness and
optional midside nodes.
 CQUADR -- Isoparametric quadrilateral element with no coupling of bending and membrane stiffnesses; the
membrane stiffness formulation includes rotation about the normal to the plane of the element.

Main Index
 126 MSC Nastran Reference Guide
Surface Elements

The CQUAD4, CQUAD8, CTRIA6, and CTRIA3 elements may be used to model hyperelastic materials; see CQUAD4 and
CQUADR Coordinate Systems, 127.
The properties for the above elements are defined on the PSHELL entry. Anisotropic material may be specified for all shell
elements. Transverse shear flexibility may be included for all bending elements on an optional basis. Structural mass is
calculated from the membrane density and thickness. Nonstructural mass can be specified for all shell elements. Lumped
mass procedures are used unless coupled mass is requested with the parameter COUPMASS. Differential stiffness matrices
are generated for all shell elements. Plane strain analysis may be requested for all shell elements.
The element coordinate systems for triangular and quadrilateral shell elements are shown in Figure 3-27 through Figure 3-30.
The integers 1 through 8 refer to the order of the connected grid points on the connection entries defining the elements.
The angle, THETA, is the orientation angle for material properties. The element coordinate system for the CQUAD8 and
CTRlA6 elements is a Cartesian system defined locally for each point (  ,  ) with the following orientation:

 The plane containing x elem and y elem is tangent to the surface of the element.

 For the CQUAD8 element, x elem and y elem are obtained by doubly bisecting the lines of constant  and  .

 For the CTRIA6 element, x elem is tangent to the line of constant  .

 x elem increases in the general direction of increasing  and y elem of  .

G3
y elem

x material

THETA

x elem
G1 G2

Figure 3-27 CTRIA3 and CTRIAR Coordinate Systems

Main Index
 Chapter 3: Structural Elements 127
Surface Elements

y elem
G3

G4

+
 = ------------ x elem
2

 x material

z elem 
THETA 

G1 G2

Figure 3-28 CQUAD4 and CQUADR Coordinate Systems

Main Index
 128 MSC Nastran Reference Guide
Surface Elements

yl
Lines of Constant  or 

G3 xl

 = 1,  = 1
 = constant

y material
yl
x material
 = 0.5 G5
THETA
G6
xl
yl
yl
 = constant
xl G4  = 0.5
THETA
 = 0
G1
G2
xl

Figure 3-29 CTRIA6 Coordinate System

Main Index
 Chapter 3: Structural Elements 129
Surface Elements

yl yl
G7 yl
G3 x material
G4 xl
yl

xl   + THETA 
4
xl 
 Gi/4
i=1 xl
Center
G8
G6

yl yl

xl
G5
THETA G2 xl
G1
Stress output at each
Gi local system

Figure 3-30 CQUAD8 Coordinate System

Main Index
 130 MSC Nastran Reference Guide
Surface Elements

e
yl



e
B
A
2 


 = 45 – ---
45 2

Gi
xl

Figure 3-31 Figure 3-29 CQUAD8 Coordinate System (continued)

The CQUAD4, CQUAD8, CTRIA3, and CTRIA6 elements may be offset relative to the mean plane of their connected grid
points. There are three commonly used techniques to define these offsets in MSC Nastran. These are generically denoted
as the ZOFFS, MID4, and RBAR methods and are described below.
It is generally recommended the ZOFFS be used if one has a sufficiently fine mesh in the region in which offsets are to be
defined. In regions of a finite element model in which the analyst has determined that a coarse mesh is satisfactory, the use
of RBARs to define offsets is generally more accurate.
Irrespective of which of the three methods is used to define the offset for the element, it is required that values be specified
for both MID1 and MID2 on the PSHELL entry referenced by the offset element. The mass properties of an offset element
are not modified to reflect the existence of the offset when the ZOFFS and MID4 methods are used. Therefore, if the weight
or mass properties of an offset element are to be used in an analysis, the RBAR method must be used to represent the offset.
The CQUADR and CTRIAR elements may be offset by the RBAR method only, but no membrane-bending coupling is
computed.
For the CQUAD4 and the CTRIA3, the forces are evaluated at the centroid of the element. For the CQUAD8, CTRIA6,
CQUADR, and CTRIAR, the forces are evaluated at the centroid and at the vertices. The positive directions for shell element
forces in the element coordinate system are shown in Figure 3-32 and Figure 3-33. The following element forces per unit of
length, either real or complex, are output on request:
 Bending moments on the x and y faces.
 Twisting moment.
 Shear forces on the x and y faces.
 Normal forces on the x and y faces.

Main Index
 Chapter 3: Structural Elements 131
Surface Elements

Vy
Fy

yelem

Fxy

zelem
Fxy
Vx
Fx Fx
Fxy

Vx

xelem
Fxy

Fy
Vy

Figure 3-32 Forces in Shell Elements

Main Index
 132 MSC Nastran Reference Guide
Surface Elements

yelem

Mxy My

Mx
zelem

Mxy Mxy

Mx

My xelem

Mxy

Figure 3-33 Moments in Shell Elements

For the CQUAD4 and CTRIA3 elements, the stresses are evaluated at the centroid of the element. For the CQUAD4, stress
at the corner can be requested if STRESS(CORNER) is requested. For the CQUAD8, CTRIA6, CQUADR, and CTRIAR
elements, the stresses are evaluated at the centroid and at the vertices. The positive directions for the stresses are shown in
Figure 3-34. The stresses are calculated in the element coordinate system. The following real stresses are output on request:
 Normal stresses in the x and y directions.
 Shear stresses on the x face in the y direction.
 Angle between the x axis and the major principal axis.
 Major and minor principal stresses.
 von Mises equivalent stress if STRESS(VONM) is requested or maximum shear stress if STRESS(SHEAR) is
requested.

Main Index
 Chapter 3: Structural Elements 133
Surface Elements

y

 xy

 xy
x x

 xy

 xy

y
Figure 3-34 Stresses in Shell Elements

The von Mises equivalent stress for plane strain analysis is defined as follows:
12
 2 2 2 2
 v = 1  2    x –  y  +   y –  z  +   z –  x   + 3 xy (3-8)
 

For plane stress analysis,  z = 0 . Only the normal stresses and shearing stresses are available in the complex form.

The von Mises equivalent strain is defined as

4 2 2 1 2 12
 v = ---   x +  y –  x  y  + ---  xy (3-9)
9 3
where the strain components are defined as
u v v u
 x = ----- ;  y = ----- ;  xy = ----- + ----- (3-10)
x y x y
and the curvatures are defined as

Main Index
 134 MSC Nastran Reference Guide
Surface Elements

2 2 2
 w  w  w
 x = ---------- ;  y = ---------- ;  xy = 2 ----------- (3-11)
x
2
y
2  x y

The maximum shear stress is

x – y 2 2
12
̂ =  ------------------ +  xy (3-12)
 2 
The maximum shear strain is

2 2 12
̂ =    x –  y  +  xy  (3-13)

The stresses are calculated at two specified points on the cross section. The distances to the specified points are given on the
property entries. The default distance is one-half the thickness. The positive directions for these fiber distances are defined
according to the right-hand sequence of the grid points specified on the connection entry.
In addition, interpolated grid point stresses and mesh stress discontinuities are calculated in user-specified coordinate
systems for grid points which connect the shell elements. See Element Data Recovery Resolved at Grid Points, 441 for a description
of the Case Control Section to request grid point stresses. Only real stresses are available at the grid points. Mesh stress
discontinuities are available in linear static analysis only.
Grid point stresses are computed by
Ne

g =  W ge  ge (3-14)
e=1

where  ge , a grid point stress component, multiplied by W ge , the interpolation factor, and summed for all elements, N e ,
connected to the grid point.
The stress discontinuity for one component and one element is
 ge =  ge –  g (3-15)

The discontinuity from all elements for one component is then obtained by

Ne
2
g =   W ge  ge  (3-16)
e=1

The total discontinuity at a grid point from all elements and all components N c defines the error and is obtained by

Main Index
 Chapter 3: Structural Elements 135
Surface Elements

Nc
1-
-----
2
g =
Nc   gc (3-17)
c=1

 gc and  g are requested and printed with the GPSDCON Case Control command.

The total discontinuity at an element for one component over all of its grid points N g is

Ng
1-
-----
2
 ec =
Ng  g (3-18)
g=1

and the error for an element is

Ng
1-
-----
2
e =
Ng  g (3-19)
g=1

 ec and  e are requested and printed with the ELSDCON Case Control command.
The CQUAD4 element is intended for use when the surfaces are reasonably flat and the geometry is nearly rectangular. For
these conditions, the quadrilateral elements eliminate the modeling bias associated with the use of triangular elements and
the quadrilaterals give more accurate results for the same mesh size. If the surfaces are highly warped, curved, or swept,
triangular elements should be used. Under extreme conditions, quadrilateral elements will give results that are considerably
less accurate than triangular elements for the same mesh size. Quadrilateral elements should be kept as nearly square as
possible, because the accuracy tends to deteriorate as the aspect ratio of the quadrilateral increases. Triangular elements
should be kept as nearly equilateral as practicable because the accuracy tends to deteriorate as the triangle becomes obtuse
and the ratio of the longest to the shortest side increases.
The CQUADR and CTRIAR elements are improved shell elements that, when compared to the CQUAD4 and CTRIA3,
respectively, are much less sensitive to high aspect ratios and values of Poisson’s ratio near 0.5. In these elements, a rotational
stiffness is computed about the normal to the element at the vertices and used in the formulation of the element stiffness.
It should be noted that this degree of freedom must not be constrained unless at a prescribed boundary. The CQUADR and
CTRIAR element stiffness formulations do not include coupling of membrane and bending stiffness. These elements are
not implemented for differential and nonlinear stiffness and heat transfer analysis.
The CQUAD8 and CTRIA6 elements should not be used with all midside nodes deleted since they are excessively stiff in
this configuration. A user warning message is produced when all midside nodes are deleted. Use CQUAD4 and CTRIA3
for this condition.

Main Index
 136 MSC Nastran Reference Guide
Surface Elements

QUADR and TRIAR Elements

Traditional Mindlin shell elements, such as QUAD4, have 5 degrees of freedom per grid point--three translations and two
bending rotations. The stiffness for the rotational degree of freedom normal to the element (the drilling degree of freedom)
is zero. This creates modeling difficulties which may eventually lead to poor solutions. Many methods--such as AUTOSPC
and K6ROT--have been devised to remedy this deficiency. None of them, however, supply physical stiffness to the drilling
degree of freedom to resist the drilling moments.

Benefits
The results for the new QUADR are compared with the QUAD4 and the old QUADR.
Comparing with the QUAD4, the new QUADR has the following benefits:
 It has six degrees of freedom at each grid point as opposed to the QUAD4, which has five degrees of freedom. This
removes the necessity of using such methods as the AUTOSPC and K6ROT--which sometimes cast doubts on
results--to account for the normal rotational DOF in the QUAD4.
 When using AUTOSPC or K6ROT for the QUAD4, the stiffness at the drilling degree of freedom is not an actual
physical stiffness. Therefore, drilling loads cannot be correctly transferred between elements, which lead to
modeling difficulties. For QUADR, the stiffness at the drilling degree of freedom is an actual physical stiffness
and, therefore, the drilling loads can be transferred correctly between elements. The ability of the QUADR to
resist the drilling moments correctly is demonstrated in Example 3.
 The membrane behaviors of the new QUADR is far superior than those of the QUAD4 and approaches that of
the QUAD8, as demonstrated by Examples 2 and 6.
Many capabilities that are available to the QUAD4 are not included in the old QUADR. For the new QUADR, all the
capabilities in QUAD4 have been incorporated into the QUADR. In comparison with the old QUADR, the current
QUADR has the following benefits:
 An improved theoretical formulation to make the current QUADR a more robust element (see the Theory
Section).
 Coupled with the shell normal, the current QUADR is well suited for modeling shell problems (see the Section on
Shell Models and Shell Normal).
 The rotational mass is implemented for the drilling degree of freedom (see the Section Mass Properties.)
 The consistent surface loads and edge loads are implemented (see the Section Consistent Surface and Edge
Loads).
 The differential stiffness matrix is implemented for the new QUADR so that buckling and other analyses requiring
the differential stiffness can be performed.
 Coupling between bending and membrane is implemented.
 Composite material is implemented for layered stress/strain output.
 Offset is allowed for the new QUADR.
 The QUADR is implemented into the solution sequence SOL 200, optimization.
 QUADR is supported in heat transfer analysis.

Main Index
 Chapter 3: Structural Elements 137
Surface Elements

 A System cell 370 is available for converting between QUAD4/TRIA3 and QUADR/TRIAR.
 Nonlinear material and geometric analysis will be implemented in a future version.

Theory
Allman and Cook were the first to propose a successful membrane element with drilling DOFs. On the other hand, Hughes
and Brezzi presented a variational formulation, which was first suggested by Reissner. This variational principle uses the
drilling degrees of freedom as independent rotation field; i.e., it is based on the separate kinematic variables of displacements
and rotations. Ibrahimbegavic, Taylor and Wilson combined the Allman type of interpolation and the variational principle
to construct a quadrilateral shell element. In MSC Nastran, membrane elements with drilling degrees of freedom were first
introduced in 1988. A summary of the theory for the membrane portion of the shell element is given below.
Allman type of interpolation fields assumes the normal displacements to be quadratic and the tangential displacements to
be linear. For example, along the edge from grid 1 to 2 (see Figure 3-35), the displacement in the middle of the edge is
1 1
u m = ---  u 1 + u 2  + --- l   2 –  1 n 1 (3-20)
2 81
where all bold letters indicate vectors with x and y-components. u 1 and u 2 are the displacements at grids 1 and 2,
respectively; l 1 is the length of the edge;  1 and  2 are the drilling rotations at grid 1 and 2, respectively; n 1 is the
vector normal to the edge in the x-y plane; and u m is the interpolated displacement vector in the middle of the edge.

Figure 3-35 In-Plane displacements and drilling rotation

Hughes and Brezzi presented two variational formulations: one is the mixed type and the other is the displacement type,
which is the one adopted here. The displacement type variational formulation states that the function to be minimized for
elements with drilling degrees of freedom is
1 T 1 2
 = ---   Gt dA + --- 'G    –   t dA (3-21)
2 2
A A

Main Index
 138 MSC Nastran Reference Guide
Surface Elements

1 u y u x
 = ---  -------- – -------- (3-22)
2  x y 
where:

 = non-dimensional constant, for most problems  = 1.0

G = the shear modulus, A and t are the area and thickness
 = the rotational field given by Equation (3-22)
 = the interpolated drilling field
 = the strain vector
G = the stress-strain relationship defined by  = G

The first term is the normal strain energy and the second term is the contribution due to the drilling degrees of freedom.
In the variational formulation given by Equation (3-21), the drilling field and the displacement field are interpolated
independently. For the drilling field, the interpolation function used is the linear QUAD4 shape function. For displacement
field, the Allman type interpolation functions given by Equation (3-20) are used. They are given by the following equations:
4
     =  Ni    i
i=1

4 4
li
u     =  N i    u i +  N i + 4     ---   i + 1 –  i n i c r + N 0    u 0 (3-23)
8
i=1 i=1
where:

i = 1, 2, 3, 4
5 = 1

The Nis in above equations are shape functions for QUAD4 and QUAD8 elements. They are defined by:

1
N i = ---  1 +  i    1 +  i   for i = 1, 2, 3, 4
4
1 2
N i = ---  1 –    1 +  i   for i = 5, 7 (3-24)
2
1 2
N i = ---  1 +  i    1 –   for i = 6, 8
2

Main Index
 Chapter 3: Structural Elements 139
Surface Elements

The N 0 in Equation (3-23) is a bubble function or an auxiliary strain function. The u 0 is the displacement corresponding
to the bubble function or the auxiliary function. The n i is the outward normal vector to element boundary and l i is its
length.
With the shape functions defined by and Equation (3-23), and the variational principle defined by Equation (3-21), the standard
finite element theory can be used to obtain the element forces and stiffness.
With the formulation given above, the singularity in the 6th degree of freedom of element such as QUAD4 is replaced with
a true stiffness. In addition, the drilling rotation becomes a true load carrying degree of freedom. We will show that the
membrane behavior is significantly improved. In fact, the accuracy of QUADR approaches that of the QUAD8.

Shell Models and Shell Normal

The shell element described in the previous Theory Section is for a flat element. In using flat shell elements for modeling
a curved shell, local normals of adjacent shell elements have different directions. In MSC Nastran, an algorithm has been
developed to generate a unique normal at a grid point. The unique normal is the average normal at each grid point so that
the angles between the average normal and all the local normals of the adjacent shell elements are minimized. This unique
normal is called the Shell Normal, which is requested by the PARAM, SNORM or the Bulk Data entry SNORM.
The new QUADR/TRIAR, formulation is more robust, is less sensitive to the shell normal, and is well suited for solving
shell problems. The following list illustrates the effect of the shell normal to the solution of shell problems. Some of the
problems are the MacNeal/Hard test problems.
 Twist beam (MacNeal/Harder test, see example below) -- no effect, excellent solution.
 Scordelis-Lo roof (MacNeal/Harder test, see example below) -- no effect, excellent solution.
 Spherical shell (MacNeal/Harder test, see example below) -- the solution is slightly stiffer with the shell normal
than without the shell normal. The solution for the coarse model with shell normal converges rapidly to that of the
fine mesh without shell normal.
 Cylindrical shell with internal pressure (see example below) and without the shell normal, solution converges with
the fine mesh. Withe the shell normal, the solution is excellent even with a coarse mesh.
 Pinched cylinder (not presented) -- almost no effect, excellent solution. The pinched cylinder is a cylinder with
two rigid end diaphragms. This cylinder is subject to a pair of concentrated forces, which pinch the cylinder.
 Raasch hook (see example below) -- no effect, excellent solution.
 We can conclude from the results that the shell normal is beneficial for membrane-dominant shell problems, and
unfavorable effect for bending-dominant shell problems. To provide good solutions for a wide class of shell
problems, the default value for PARAM, SNORM is set to 20.0. Also from the above results, we can conclude
that, with or without the shell normal, the new QUADR/TRIAR is well suited for modeling shell problems.
Shell normal yields wrong results for elements with offsets. Therefore, the shell normal should be turned off globally
(PARAM,SNORM,0.0) or turned off locally with the Bulk Data entry, SNORM.

Main Index
 140 MSC Nastran Reference Guide
Surface Elements

Mass Properties
In MSC Nastran, two mass formulations are available--the lumped mass and the coupled mass. Lumped mass is the default
formulation. Coupled mass formulation is requested by the parameter PARAM, COUPMASS, 1.
For the QUADR/TRIAR, the mass properties are as follows:
 Lumped mass -- only the translational masses are computed. No rotational mass is computed for the drilling
degrees of freedom.
 Coupled mass -- in addition to the translational masses, the rotational masses are computed for the drilling degrees
of freedom.

Consistent Surface and Edge Loads

PLOAD4, the most commonly used surface load, may have both normal and tangential components relative to the surface
of an element. For the QUADR/TRIAR element, the consistent loads due to the surface load contain both concentrated
forces at the connected grid points and the drilling moments at the drilling degrees of freedom. For the new
QUADR/TRIAR, the consistent surface loads have been implemented.
For QUADR/TRIAR--because of the presence of drilling degrees of freedom--it is important to compute the consistent
grid point forces for any distributed load along the edge of the element. This point is best illustrated by the following
example. A single QUADR element is loaded with uniform tensile edge load along the opposite edge as shown in
Figure 3-36A. If we lump the uniform load into grid point forces, as shown in Figure 3-36B, then we will get the non-uniform
displacement solution, instead of the correct uniform displacement solution. This error can be corrected if we compute the
consistent grid point forces for the distributed edge loads. The consistent grid point forces consist of the lumped grid forces
and a pair of drilling moments, as shown in Figure 3-36C.
For QUADR/TRIAR only, the consistent edge loads have been implemented. User can request the consistent edge loads
using the PLOAD4 Bulk Data entry.

Main Index
 Chapter 3: Structural Elements 141
Surface Elements

Figure 3-36 Consistent Edge Loads

User Interfaces
Parameters SNORM and Bulk Data Entry SNORM

The parameter SNORM is used for requesting unique shell normals for all grid points which are connected to the shell
elements, such as QUADR, in a model. The default shell normal value is PARAM, SNORM, 20.0. For relationships
between a shell model and the shell normal, please see the discussions in the Shell Normal and Shell Models Section.
The Bulk Data entry, SNORM requests a local shell normal at a grid point. It will override the shell normals requested by
PARAM, SNORM, and can also be used to turn off the local default shell normal.
Bulk Data Entries CQUADR, CTRIAR, and PSHELL

The QUADR/TRIAR elements are requested by the existing CQUADR, CTRIAR, and PSHELL Bulk Data entries. For the
new QUADR/TRIAR, there is no modification to the PSHELL entry. For the CQUADR and CTRIAR entries, two new
fields are added:
 ZOFFS -- this entry is used to input the offsets from the surface of the grid points to the element plane.
 TFLAG -- this flag allows user to input the relative values for the variable grid point thickness, instead of absolute
values. See Section for further details.
Bulk Data Entry PLOAD4

The PLOAD4 entry is used to define a pressure load for a 3-D or 2-D element. For the QUADR/TRIAR only, this entry
is extended to accommodate consistent edge load. See the Consistent Surface and Edge Loads Section for further details.
Two new fields are added to PLOAD4 for this purpose:
 SORL -- selects whether the load is a surface or an edge load.
 SORL=SURF selects the surface load and SORL=LINE selects the consistent edge load. SORL=SURF is the
default.
 LDIR -- defines the direction of the consistent edge load (SORL=LINE). It has the value of X, Y, Z, TANG, or
NORM. See PLOAD4 for their meanings. The default is NORM.
For example, the uniform extensile edge loads shown in Figure 3-36 can be specified by the following PLOAD4 entries:

1 2 3 4 5 6 7 8 9 10
PLOAD4 10 1 20.0
LINE

The edge loads have a magnitude of 20.0 and the direction is NORM, i.e., it is in the mean plane of the element, normal
to the edge, and pointing outward from the element.
NASTRAN System (370) -- QRMETH

The new default QUADR/TRIAR formulation is recommended. However, a new system cell 370 (QRMETH) is available
for reverting back to the old formulation.

Main Index
 142 MSC Nastran Reference Guide
Surface Elements

A companion version of the QUAD4/TRIA3 is also developed along with the new QUADR/TRIAR. We will call this
version the new formulation for QUAD4/TRIA3, and the existing QUAD4 and TRIA3 the old formulation. The
performances of the old and the new formulations are about the same.
The default for the QUADR/TRIAR uses the new formulation. The default for the QUAD4/TRIA3 uses the old
formulation.
In addition to the selection of the new or old formulation for QUADR/TRIAR, QRMETH provides the following options:
 0 -- selects the new QUADR/TRIAR formulation.
 1 -- selects the old QUADR/TRIAR formulation.
 2 -- converts QUADR/TRIAR into QUAD4/TRIA3 using the new QUAD4/TRIA3 formulation.
 3 -- converts QUADR/TRIAR into QUAD4/TRIA3 using the old QUAD4/TRIA3 formulation.
 4 -- selects the new QUAD4/TRIA3 formulation.
 5 -- converts QUAD4/TRIA3 into QUADR/TRIAR using the new QUADR/TRIAR formulation.
The default is for QRMETH is 0.
Converting a QUAD4 model to QUADR

The following examples will illustrate that the QUADR is superior to the QUAD4. User may want to convert a QUAD4
model to QUADR by using system cell QRMETH. However, the following should be noted:
 Using QRMETH=5 will convert all QUAD4/TRIA3 in a model to QUADR/TRIAR.
 At the boundaries of the model, the drilling degrees of freedom must be constrained if the user wants fixed
boundaries. For example, in the example Static Analysis of Straight Beam with Coarse Mesh, the drilling degrees
of freedom must be constrained at the left end.
 At internal grid points of a model, user must remove any SPC or the PS field on the GRID Bulk Data entries
which constrain the drilling degrees of freedom because the QUADR/TRIAR supply stiffness for these degrees of
freedom. On the other hand, if user uses K6ROT or AUTOSPC to manage the drilling degrees of freedom, these
will be automatically converted with QRMETHD=5. No user action is required.

Examples
Example 1 - Static Analysis of Straight Beam with Coarse Mesh

For membrane behavior, the performances of the QUADR element are substantially better than those of the QUAD4
element. This can best be illustrated by the static analysis of a straight beam with coarse mesh, shown in Figure 3-37. The
load is an in plane shear load acting at the tip of the beam. This test is taken from the MacNeal-Harder tests and is designed
to exhibit in plane behavior of shell elements for different element shapes. The results for this test are shown in Table 3-8.
The displacements shown are at the tip of the beam, normalized to the theoretical value.

Main Index
 Chapter 3: Structural Elements 143
Surface Elements

Figure 3-37 Straight Beam with Coarse Mesh

Table 3-8 Normalized Displacements for In-plane Shear Load at Tip

Element Rectangle Trapezoid Parallelogram
QUAD4 0.9929 0.0515 0.6232
QUADR 0.9926 0.9613 0.9491

From the results, we see that the QUADR provides excellent results for both regular and irregular element shapes, while
QUAD4 fails for the irregular shapes. We may argue that, in real life, we don’t use element mesh as coarse as the one in this
test. Next, we will look at the results of a fine element mesh, which is discussed in next sub-example.
Example 2 - Normal Modes Analysis of Straight Beam with Fine Mesh

For this example, a refined mesh of Example 1 is used in a normal mode analysis. Table 3-9 summarizes the results of the
first mode using various element shapes.

Main Index
 144 MSC Nastran Reference Guide
Surface Elements

Figure 3-38 Straight beam with fine mesh

Table 3-9 In-Plane Bending Frequency of Straight Beam with Fine Mesh
Element Rectangle (hz) Trapezoid (hz) Parallelogram (hz)
QUAD4 9.39 15.89 9.41
QUADR 9.39 9.42 9.39

Once again, the QUAD4 element exhibits locking effects for trapezoidal shapes. Even for finer meshes, the results of the
QUADR elements are still noticeable better than the results of the QUAD4 element in the case of distorted element shapes.
Example 3 - Static Analysis of Straight Beam with Drilling Moments

This example shows that the QUADR element provides the true physical stiffness for the drilling degrees of freedom and
hence able to resist the actual drilling loads. A straight beam is modeled by six QUADR elements and is subject to a pair of
end drilling moments as shown in Table 3-9.

Figure 3-39 Straight Beam Subject to Tip Drilling Moments

The tip displacements of the straight beam for a static analysis are
 Theory = -5.40e-2
 Finite element solution = -5.46e-2

Main Index
 Chapter 3: Structural Elements 145
Surface Elements

The above results show that the QUADR can predict correctly the response due to applied drilling moments.
Example 4 - Static Analysis of Cylindrical Shell with Internal Pressure

The purpose of this example is to illustrate the benefits of the shell normal in a shell structure modeled by QUADR
elements. The shell normal is discussed in the Shell Models and Shell Normal Section. A cylinder of height 1000.0, radius
1000.0, and thickness 10.0, is subject to internal pressure of 1.0, as shown in Figure 3-40. The cylinder is free at both edges.
Due to symmetry, only half of the height and a quarter of the cylinder is modeled. The radial displacement at the free edge
is given in Table 3-10.

Figure 3-40 Cylindrical Shell with Internal Pressure

Table 3-10 Radial Displacement at Free Edge for Pressurize Cylinder

Mesh Theory With SNORM No SNORM
4x4 0.4762 0.4670 0.6733
8x8 0.4762 0.4739 0.5251
16x16 0.4762 0.4756 0.4882

From Table 3-10, we can see that the results with the shell normal are substantially better than those without the shell normal.
Example 5 - Raasch Hook

The Raasch hook is a curved strip clamped at one end and loaded at the other end with an in-plane shear load as shown in
Figure 3-41. A static analysis is performed on the Raasch hook for cases with and without the shell normal. The tip

Main Index
 146 MSC Nastran Reference Guide
Surface Elements

displacements in the direction of the shear force are summarized in Table 3-11 using the new QUADR with various mesh
size.
From Table 3-11, we can see that, with or without the shell normal, the results for the new QUADR are excellent. Using the
old QUADR, the solution converges to the wrong results if shell normals are not used.

Figure 3-41 Raasch Hook

Table 3-11 Tip Displacement for Raasch Hook

Mesh Theory With SNORM No SNORM
1x9 5.012 4.600 4.601
3x17 5.012 4.890 4.888
5x34 5.012 4.916 4.915
10x68 5.012 4.970 4.971
20x136 5.012 5.011 5.012

Example 6 - MacNeal and Harder Tests

MacNeal and Harder suggested a series of test problems for evaluating the accuracy of finite element. We have run these
test problems for QUADR, QUAD4 and QUAD8 for comparisons. The results are summarized in Table 3-12 and a score
is given to each test problem based on how well it compares to the theoretical solution. The scoring criteria is shown:

Main Index
 Chapter 3: Structural Elements 147
Surface Elements

 A -- error is less than and equal to 2%.

 B -- error is between 2% and 10%.
 C -- error is between 10% and 20%.
 D -- error is between 20% and 50%.
 F -- error is greater than 50%.
From Table 3-12, we see that QUADR got the best scores among the three elements. It does not have a score below B.
Between QUADR and QUAD4, the results of QUADR are better than those of the QUAD4. Between QUADR and
QUAD8, the accuracy of the QUADR approaches that of QUAD8 for membrane problems and is better than that of
QUAD8 for bending problems.
For QUADR, the tests are run with the default value of the shell normal equal to 20.0. Except for the spherical shell (14),
the shell normal has no effect on the tests. For the spherical shell, it will score an A if shell normal is turned off (PARAM,
SNORM, 0.0.)

Table 3-12 MacNeal and Harder Tests for QUADR, QUAD4, and QUAD8
Element Loading
MacNeal/Harder Test Element
Description In-Plane Out-of-Plane Shapes QUADR QUAD4 QUAD8
1. Patch Test X Irregular A A C
2. Patch Test X Irregular A A D
3. Straight Beam, X All A A A
Extension
4. Straight Beam, X Regular A A A
Bending
5. Straight Beam, X Irregular B F B
Bending
6. Straight Beam, X Regular A A A
Bending
7. Straight Beam, X Irregular A A A
Bending
8. Straight Beam, Twist X X All B B B
9. Curved Beam X Regular A C A
10. Curved Beam X Regular B B B
11. Twisted Beam X X Regular A A A

Main Index
 148 MSC Nastran Reference Guide
Surface Elements

Table 3-12 MacNeal and Harder Tests for QUADR, QUAD4, and QUAD8
Element Loading
MacNeal/Harder Test Element
Description In-Plane Out-of-Plane Shapes QUADR QUAD4 QUAD8
12. Rectangular Plate X Regular A B B
(N=4)
13. Scordelis-Lo Roof X X Regular B B A
(N=4)
14. Spherical Shell (N=8) X X Regular B A C
15. Thick Walled Cylinder X Regular B B B

Conical Shell Element (RINGAX)

The properties of the conical shell element are assumed to be symmetrical with respect to the axis of the shell. However, the
loads and deflections need not be axisymmetric because they are expanded in Fourier series with respect to the azimuthal
coordinate. Due to symmetry, the resulting load and deformation systems for different harmonic orders are independent,
a fact that results in a large time saving when the use of the conical shell element is compared with an equivalent model
constructed from plate elements. Theoretical aspects of the conical shell element are provided in Section 5.9 of The
NASTRAN Theoretical Manual.
The conical shell element cannot be combined with other types of elements. This is primarily because the Fourier
coefficients are stored on internally generated pseudo grid points. The unconventional nature of this element results in its
capabilities being limited, as summarized in Table 3-1. The existence of a conical shell problem is defined by the AXIC entry.
This entry also indicates the number of harmonics desired in the problem formulation. Only a limited number of Bulk
Data entries are allowed when using conical shell elements. The list of allowable entries is given on AXIC (p. 1234) in the .
This element is not allowed in the superelement solution sequences.
The geometry of a problem using the conical shell element is described with RINGAX entries instead of GRID entries. The
RINGAX entries describe concentric circles about the basic z-axis, with their locations given by radii and z-coordinates as
shown in Figure 3-42. The degrees of freedom defined by each RINGAX entry are the Fourier coefficients of the motion
with respect to angular position around the circle. For example, the radial motion u r at any angle  is described by the
equation
N N
2 2*
ur    =  u r cos n +  u r sin n (3-25)
n=0 n=0

n n*
where u r and u r are the Fourier coefficients of radial motion for the n-th harmonic. For calculation purposes, the series
is limited to N harmonics as defined by the AXIC entry. The first sum in the above equation describes symmetric motion

Main Index
 Chapter 3: Structural Elements 149
Surface Elements

with respect to the  plane. The second sum with the “starred” (*) superscripts describes the antisymmetric motion. Thus
each RINGAX entry will produce six times N degrees of freedom for each series.

z
Uz - Displacement Coordinates

U z
 Ur

r
RB

RA

V
W RINGAX B

U - Element Coordinates

RINGAX A

Figure 3-42 Geometry for Conical Shell Element

The selection of symmetric or antisymmetric solutions is controlled by the AXISYM command in the Case Control Section.
For general loading conditions, a combination of the symmetric and antisymmetric solutions must be made, using the Case
Control command, SYMCOM (Case) in the MSC Nastran Quick Reference Guide. If the AXISYM command is not present,
stress and element force requests will be ignored.
Since the user is rarely interested in applying loads in terms of Fourier harmonics and interpreting data by manually
performing the above summations, special data entries which automatically perform these operations are provided. The
POINTAX entry is used like a GRID entry to define physical points on the structure for loading and output. Sections of
the circle may be defined by a SECTAX entry which defines a sector with two angles and a referenced RINGAX entry. The
POINTAX and SECTAX entries define six degrees of freedom each. The basic coordinate system for these points is a
cylindrical system ( r ,  , z ), and their applied loads must be described in this coordinate system. Since the displacements
of these points are dependent on the harmonic motions, they may not be constrained in any manner.
The conical shell element is connected to two RINGAX rings with a CCONEAX entry. The properties of the conical shell
element are described on the PCONEAX entry. The RINGAX points must be placed on the neutral surface of the element,

Main Index
 150 MSC Nastran Reference Guide
Surface Elements

and the points for stress calculation must be given on the PCONEAX entry relative to the neutral surface. Up to fourteen
angular positions around the element may be specified for stress and force output. These values will be calculated midway
between the two connected rings.
The structure defined with RINGAX and CCONEAX entries must be constrained in a special manner. All harmonics may
be constrained for particular degree of freedom on a ring by using permanent single-point constraints on the RINGAX
entries. Specified harmonics of each degree of freedom on a ring may be constrained with an SPCAX entry. The entry is
the same as the SPC entry except that a harmonic must be specified. The MPCAX, OMITAX, and SUPAX data entries
correspond to the MPC, OMIT, and SUPORT data entries except that harmonics must be specified. SPCADD and
MPCADD entries may be used to combine constraint sets in the usual manner.
The stiffness matrix includes five degrees of freedom per grid circle per harmonic when transverse shear flexibility is
included. Since the rotation about the normal to the surface is not included, the sixth degree of freedom must be
constrained to zero when the angle between the meridional generators of two adjacent elements is zero. Since only four
independent degrees of freedom are used when the transverse shear flexibility is not included, the fifth and sixth degrees of
freedom must be constrained to zero for all rings. These constraints can be conveniently specified on the RINGAX entry.
The conical shell structure may be loaded in various ways. Concentrated forces may be described by FORCE and
MOMENT entries applied to POINTAX points. Pressure loads may be input by the PRESAX data entry which defines an
area bounded by two rings and two angles. Temperature fields are described by a paired list of angles and temperatures
around a ring as required by the TEMPAX entry. Direct loads on the harmonics of a RINGAX point are given by the
FORCEAX and MOMAX entries. Since the basic coordinate system is cylindrical, the loads are given in the r ,  , and z
directions. The value of a harmonic load F n is the total load on the whole ring of radius r . If a sinusoidal load-per-unit
length of maximum value is given, the value on the FORCEAX entry must be
F n = 2 r a n n = 0 (3-26)

Fn =  r an n = 0 (3-27)

Displacements of rings and forces in conical shell elements can be requested in two ways:
 The harmonic coefficients of displacements on a ring or forces in a conical element
 The displacements at specified points or the average value over a specified sector of a ring. The forces in the
element at specified azimuths or average values over specified sectors or a conical element.
Harmonic output is requested by ring number for displacements and conical shell element number for element forces. The
number of harmonics that will be output for any request is a constant for any single execution. This number is controlled
by the Case Control command, HARMONICS (Case) in the MSC Nastran Quick Reference Guide.
The following element forces per unit of width are output either as harmonic coefficients or at specified locations on request
 Bending moments on the u and v faces
 Twisting moments
 Shearing forces on the u and v faces.

Main Index
 Chapter 3: Structural Elements 151
Surface Elements

The following element stresses are calculated at two specified points on the cross section of the element and output as
harmonic coefficients or at specified locations on request:
 Normal stresses in u and v directions
 Shearing stress on the u face in the v direction
 Angle between the u -axis and the major principal axis
 Major and minor principal stresses
 Maximum shear stress

Shell Elements Corner Thickness Specification

The CQUAD4, CQUAD8, CQUADR, CTRIA3, CTRIA6, and CTRIAR elements allow the user to specify either corner
element thickness or a single thickness on the PSHELL entry. A TFLAG field has been added to the connection entries of
each of the above shell elements that allows the user to specify the corner thickness as a fraction relative to the T value of
the PSHELL entry.

Benefits
In forming shell like structures, quite often the taper at the corners is really just a function of some average thickness. With
this option, the average thickness can be specified as T on the PSHELL and the corner thickness will be computed as a
fraction of the T on the PSHELL. In SOL 200, the T on the PSHELL can be specified on the DVPREL1 with average
thickness specified using a DESVAR entry and the shell will be tapered accordingly.

Theory
If TFLAG=1 the following calculation is used for each corner:
Ti = Ti  T

Inputs
If the fraction relative to the T value of the PSHELL is desired, TFLAG=1 should be specified on the element connection
entry.

Outputs
Standard shell element output is obtained.

Guidelines and Limitations

This option should be useful in design optimization.

Main Index
 152 MSC Nastran Reference Guide
Surface Elements

Example
The following example input demonstrates the form and use of the new entries:

Input File
diag 8,15,56
sol 200
cend
$
title = Cantilever Plate: Optimization
subtitle = Relative Thickness
$
analysis = statics
$
desobj = 1000
dessub = 2000
$
dsaprt
$
load = 10
spc = 20
$
disp = all
stress = all
$
begin bulk
$
param, grdpnt, 0
$
dresp1, 1000, weight, weight
$
dresp1, 2000, disp, disp,,, 2,, 5
dconstr, 2000, 2000, -0.5
$
doptprm, desmax, 15
$
$ Design Model: Relative Thickness
$
desvar, 1, t1, 1., .001
desvar, 2, t2, 1., .001
desvar, 3, t3, 1., .001
$
dvprel1, 100, pshell, 1, t
, 1, 0.1
dvprel1, 200, pshell, 2, t
, 2, 0.2
dvprel1, 300, pshell, 3, t
, 3, 0.1
$
$ Test Plate: Sheet Metal Thickness Variation
$
$ Relative Thicknesses on CQUAD4/CTRIA3 Entries
$
$ ---------------------------------------------------------
$
grid, 1,, 0., 0., 0.

Main Index
 Chapter 3: Structural Elements 153
Surface Elements

= ,*(1),, *(1.), ==
=(3)
grid, 11,, 0., 1., 0.
= ,*(1),, *(1.), ==
=(2)
grid, 21,, 0., 2., 0.
= ,*(1),, *(1.), ==
=(3)
$
cquad4, 1, 1, 1, 2, 12, 11,
, , 1, 1., 1.1, 1.1, 1.
cquad4, 2, 1, 2, 3, 13, 12,
, , 1, 1.1, 1.2, 1.3, 1.1
cquad4, 3, 1, 3, 4, 14, 13,
, , 1, 1.2, 1.1, 1.6, 1.3
$
pshell, 1, 1, 0.1, 1
$
cquad4, 11, 2, 11, 12, 22, 21
, , 1, 1.0, 1.1, 1.1, 1.0
cquad4, 12, 2, 12, 13, 23, 22
, , 1, 1.1, 1.0, 1.2, 1.1
cquad4, 13, 2, 13, 14, 24, 23
, , 1, 1.0, 0.8, 1.1, 1.2
$
pshell, 2, 1, 0.2, 1
$
ctria3, 21, 3, 4, 5, 14
, , 1, 1.1, 1.0, 1.6
ctria3, 22, 3, 5, 25, 14
, , 1, 1.0, 2.0, 1.6
ctria3, 23, 3, 14, 25, 24
, , 1, 1.6, 2.0, 2.2
$
pshell, 3, 1, 0.1, 1
$
mat1, 1, 10000.,, 0.3, 1.
$
force, 10, 5,, 10., 0., -1.
force, 10, 25,, 10., 0., -1.
$
spc1, 20, 12, 1, 11, 21
spc1, 20, 3456, 1, thru, 25
$
enddata
Partial Output File

Main Index
 154 MSC Nastran Reference Guide
Solid Elements (CTETRA, CPENTA, CHEXA)

S U M M A R Y O F D E S I G N C Y C L E H I S T O R Y
***************************************************************

(HARD CONVERGENCE ACHIEVED)

(SOFT CONVERGENCE ACHIEVED)

NUMBER OF FINITE ELEMENT ANALYSES COMPLETED 5

NUMBER OF OPTIMIZATIONS W.R.T. APPROXIMATE MODELS 4

OBJECTIVE AND MAXIMUM CONSTRAINT HISTORY

---------------------------------------------------------------------------------------------------------------
OBJECTIVE FROM OBJECTIVE FROM FRACTIONAL ERROR MAXIMUM VALUE
CYCLE APPROXIMATE EXACT OF OF
NUMBER OPTIMIZATION ANALYSIS APPROXIMATION CONSTRAINT
---------------------------------------------------------------------------------------------------------------

INITIAL 1.299167E+00 -4.583377E-02

1 1.088284E+00 1.088204E+00 7.350594E-05 -1.224685E-02

2 1.067696E+00 1.067703E+00 -6.364063E-06 1.266837E-03

3 1.065613E+00 1.065610E+00 2.349259E-06 2.949357E-03

4 1.065610E+00 1.065610E+00 0.000000E+00 2.949357E-03

* **************************************************************
---------------------------------------------------------------------------------------------------------------
1 CANTILEVER PLATE: OPTIMIZATION MARCH 18, 2003 MSC Nastran 3/18/03 PAGE 40
RELATIVE THICKNESS
0
DESIGN VARIABLE HISTORY
-------------------------------------------------------------------------------------------------------------------------------
---
INTERNAL | EXTERNAL | |
DV. ID. | DV. ID. | LABEL | INITIAL : 1 : 2 : 3 : 4 : 5 :
-------------------------------------------------------------------------------------------------------------------------------
---
1 | 1 | T1 | 1.0000E+00 : 1.2538E+00 : 1.3250E+00 : 1.3398E+00 : 1.3398E+00 :
2 | 2 | T2 | 1.0000E+00 : 8.2646E-01 : 7.5902E-01 : 7.5085E-01 : 7.5085E-01 :
3 | 3 | T3 | 1.0000E+00 : 3.8964E-01 : 3.8075E-01 : 3.7388E-01 : 3.7388E-01 :
*** USER INFORMATION MESSAGE 6464 (DOM12E)
RUN TERMINATED DUE TO HARD CONVERGENCE TO AN OPTIMUM AT CYCLE NUMBER = 4.

SHELL Structural Damping

4
If PARAM,SHLDAMP,DIFF is present in the run, a structural damping matrix K is formed for each shell element based
on a relationship of the following form:
4
K = ge mid1  membrane  stiffness
+ ge mid2  bending  stiffness
+ ge mid3  transverse  stiffness
+ ge mid4  combined  stiffness

Solid Elements (CTETRA, CPENTA, CHEXA)

MSC Nastran includes three different solid polyhedron elements which are defined on the following Bulk Data entries:
1. CTETRA -- Four-sided solid element with 4 to 10 grid points. See Figure 3-43.

Main Index
 Chapter 3: Structural Elements 155
Solid Elements (CTETRA, CPENTA, CHEXA)

2. CPENTA -- Five-sided solid element with 6 to 15 grid points. See Figure 3-44.
3. CHEXA -- Six-sided solid element with 8 to 20 grid points. See Figure 3-45.
These elements can be used with all other MSC Nastran elements except the axisymmetric elements. Connections are made
only to displacement degrees of freedom at the grid points.
G4

G8
G10

G9

G3
G7

G1
G6

G5
G2
Figure 3-43 The CTETRA Element

Three types of integration are available for the CPENTA and CHEXA elements: reduced shear integration with and without
the use of bubble functions and standard isoparametric integration. Only standard isoparametric integration is available for
the CTETRA element. The type of integration is selected on the PSOLID entry.
Reduced shear integration minimizes shear locking and will not cause zero energy modes. If reduced shear integration is
selected with bubble functions, it minimizes Poisson’s ratio locking which occurs in nearly incompressible materials and in
elements under bending. In plastic analysis, bubble functions are necessary to reduce locking caused by the plastic part in
the material law. Therefore, reduced shear integration with bubble functions is the default option for the CPENTA and
CHEXA elements. However, the use of bubble functions requires more computational effort.

Main Index
 156 MSC Nastran Reference Guide
Solid Elements (CTETRA, CPENTA, CHEXA)

G6
y
z

G15
G14
G4
G13

G5
G12

G11
G10 x

G3 G8
G9

G7 G2
G1

Figure 3-44 The CPENTA Element

Main Index
 Chapter 3: Structural Elements 157
Solid Elements (CTETRA, CPENTA, CHEXA)

G18
G7
G6

G19
G15 G17
G14
G20
G8
G2
G10 G5
G16 G3

G11 G13 G9

G4 G12

G1
Figure 3-45 The CHEXA Element

In standard isoparametric integration, the number of Gauss points, or integration network, may be changed to
underintegrate or overintegrate the solid elements. Underintegration may cause zero energy modes and overintegration
results in an element which may be too stiff. Standard isoparametric integration is more suited to nonstructural problems.
The element coordinate systems are described under the description of the CHEXA, CPENTA, and CTETRA entries,
respectively. These elements may use anisotropic materials as defined on the MAT9 Bulk Data entry. The material reference
and integration network is specified on the PSOLID entry.
Structural mass is calculated for all solid elements. The default mass procedure is lumped mass. The coupled mass procedure
may be requested with PARAM,COUPMASS in the MSC Nastran Quick Reference Guide.
The solid elements can be used to define fluid elements for coupled fluid-structural analysis. See Coupled Fluid-Structure
Interaction.
The solid elements can be used in differential stiffness analysis and buckling analysis. A geometric and material nonlinear
stiffness formulation is available for these elements only if there are no midside nodes.
The following stresses and strains are output on request:
 Normal:  x ,  y ,  z , and  x ,  y ,  x

Main Index
 158 MSC Nastran Reference Guide
Solid Elements (CTETRA, CPENTA, CHEXA)

Shear:  xy ,  yz ,  zx , and  xy ,  yz ,  zx

Principal with magnitude and direction

1
Mean pressure p o = ---   x +  y +  z 
3
Octahedral shear stress or
1 2 2 2 2 2 2 12
 = ---    x –  y  +   y –  z  +   z –  x  + 6 yz + 6 zx + 6 xy  (3-28)
3
the von Mises equivalent stress
3
 v =  -------  o (3-29)
 2

 Octahedral shear strain or

12
1 2 2 2 1 2 2 2
 o = ---    x –  y  +   y –  z  +   z –  x   + ---   xy +  yz +  zx  (3-30)
9  6

the von Mises equivalent strain

12
2 2 2 2 1 2 2 2
 v = ---    x –  y  +   y –  z  +   z –  x   + ---   xy +  yz +  zx  (3-31)
9  3

where the strain components are defined as

u v w
 x = ----- ;  y = ----- ; z = ------ (3-32)
x y z

v u w v w u
 xy =  ----- + ----- ;  yz =  ------ + ----- ;  zx =  ------ + ----- (3-33)
 x y  y z  x z 

In nonlinear nonhyperelastic (small strain) analysis, the mechanical strain   E is output rather than the total strain
defined above. The hyperelastic elements output total strains. See Fully Nonlinear Hyperelastic Elements.
By default, the stresses and strains are evaluated in the basic coordinate system at each of the corner points and the centroid
of the element. The stresses and strains may also be computed in the material coordinate system as defined in the CORDM
field on the PSOLID entry.
In addition, interpolated grid point stresses and mesh stress discontinuities are calculated in user-specified coordinate
systems for grid points which connect solid elements. See Element Data Recovery Resolved at Grid Points for a description of the
Case Control Section to request grid point stresses and mesh stress discontinuities. Only real stresses are available at the grid

Main Index
 Chapter 3: Structural Elements 159
Rigid Elements and Multipoint Constraints (R-type, MPC)

points. Mesh stress discontinuities are available in linear static analysis only. For further discussion see the end of Shell
Elements (CTRIA3, CTRIA6, CTRIAR, CQUAD4, CQUAD8, CQUADR).

Rigid Elements and Multipoint Constraints (R-type, MPC)

The multipoint constraint, or MPC entry, provides the capability to model rigid bodies and to represent other relationships
which can be treated as rigid constraints. The MPC entry provides considerable generality but lacks user convenience.
Specifically, the user must supply all of the coefficients in the equations of constraint defined through the MPC entry.
To enhance user convenience, nine rigid body elements are available in MSC Nastran. See Table 3-13. These elements require
only the specification of the degrees of freedom that are involved in the equations of constraint. All coefficients in these
equations of constraint are calculated internally in MSC Nastran.

Table 3-13 Rigid Element and MPC Entries

Finite m = Dependent
Name Description Rotation Degrees of freedom
RROD A open-ended rod which is rigid in extension Y m = 1
RBAR Rigid bar with six degrees of freedom at each end. Y 1m6
RJOINT Rigid joint with six degrees of freedom at each Y 1m6
end.
RTRPLT Rigid triangular plate with six degrees of freedom Y 1  m  12
at each vertex.
RBE2 A rigid body connected to an arbitrary number of Y m1
grid points. The independent degrees of freedom
are the six components of motion at a single grid
point. The dependent degrees of freedom at the
other grid points all have the same user-selected
component numbers.
RBE1 A rigid body connected to an arbitrary number of Y m1
grid points. The independent and dependent
degrees of freedom can be arbitrarily selected by
the user.

Main Index
 160 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

Table 3-13 Rigid Element and MPC Entries (continued)

Finite m = Dependent
Name Description Rotation Degrees of freedom
RBE3 Defines a constraint relation in which the motion Y 1m6
at a “reference” grid point is the least square
weighted average of the motions at other grid
points. The element is useful for “beaming” loads
and masses from a “reference” grid point to a set
of grid points.
RSPLINE Defines a constraint relation whose coefficients N m1
are derived from the deflections and slopes of a
flexible tubular beam connected to the referenced
grid points. This element is useful in changing
mesh size in finite element models.
RSSCON Define a multipoint constraint relation which N m5
models a clamped connection between shell and
solids.
MPC Rigid constraint that involves user-selected N m = 1
degrees of freedom at both grid points and at
scalar points. The coefficients in the equation of
constraint are computed and input by the user.

Any combination of the above elements may be used in an MSC Nastran analysis in any of the structural solution
sequences, however, these elements should be used with care in geometric nonlinear analysis (see the MSC Nastran
Handbook for Nonlinear Analysis). The nine rigid elements are ignored in the heat transfer solution sequences. See Rigid
Element Enhancements for nonlinear enhancements to the rigid element.

User Input
A Bulk Data entry is associated with each of the nine rigid elements (see the Bulk Data Entries in the MSC Nastran Quick
Reference Guide). Before proceeding with a discussion of input requirements, it is important to understand how MPC and
rigid element entries are treated.
Each MPC entry generates a single equation of the form
A 1 u 1 + A 2 u 2 + A n u n = 0 (3-34)

where u 1 , u 2 , ..., u n are user-designated degrees of freedom (grid point plus component) and A 1 , A 2 , ..., A n are
coefficients which are user-supplied. The first named degree of freedom is placed in the u m set (degrees of freedom
eliminated by multipoint constraints). A rigid element (RROD, RBAR, RTRPLT, RBE1, RBE3, RSPLINE, RBE2,
RJOINT, or RSSCON) generates a number of equations of the form of (3-34). The user designates the degrees of freedom

Main Index
 Chapter 3: Structural Elements 161
Rigid Elements and Multipoint Constraints (R-type, MPC)

to be placed in the u m set (one for each equation of constraint). In some cases (RBAR and RTRPLT), the designation can
be accomplished by a default option. The coefficients in (3-34) are computed automatically from geometric data and for the
RBE3 element. For the RSSCON, the M-set degrees of freedom are automatically assigned. The complete u m set consists
of the first named terms on the MPC entries plus the designated degrees of freedom on the rigid element entries. The user
has complete control over the membership of this set.
Input Requirements

The following table lists the procedural requirements and rules that the user must follow when using MPCs and the (RBAR,
RTRPLT, RBE1, RBE2, RBE3, RSSCON, and RSPLINE) rigid elements in an analysis.
1. A member of the u m set cannot also be a member of any other user defined set. User-defined sets include:

u s -- degrees of freedom eliminated by SPCi entries, AUTOSPC, and PS field on GRID entries

u o -- degrees of freedom specified on OMITi Bulk Data entries

u r -- degrees of freedom specified on SUPORT and SUPORT1 Bulk Data entries

u a -- members of the analysis set specified on ASETi Bulk Data entries or exterior degrees of freedom in
superelement analysis
2. A degree of freedom cannot be designated as a member of the u m set more than once.

A fatal error will result if, for example, the same degree of freedom is designated as dependent by two rigid elements
or it the first-named degree of freedom on any MPC entry is also a designated member of u m on a rigid element
entry.
3. The user-selected independent degrees of freedom u n , for the RBAR, RTRPLT, RBE1, RBE2, RBE3 and
RSPLINE rigid elements must be sufficient to define any general rigid body motion of the element.
These degrees of freedom are independent only for the particular element, and they may be declared dependent by
other rigid element, MPCs, SPCs, OMITs, or SUPORTs. As far as a particular rigid element is concerned, it is
always acceptable to select all six independent degrees of freedom at one grid point. This may not, however, be a
good choice when the total problem requirements are considered.
For six of the elements under discussion (RBAR, RTRPLT, RBE1, RBE3, RSPLINE, and RBE2), the user lists the
degrees of freedom in u m and u n . The remaining degrees of freedom at the grid points to which the rigid element
is jointed are not involved with the rigid element. This lack of connection represents either a sliding or rotating joint
release, or both. The rigid rod element (RROD) is an exception because once a component of translation is placed
in u m , all of the five remaining components of translation will automatically be placed in u n . The rotational
degrees of freedom are not involved in the RROD element.
4. The user must avoid over-constraining the structure when two or more rigid elements are used.

Main Index
 162 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

A structure is over-constrained when the degrees of freedom, which remain after the members of u m have been
selected, are insufficient to represent a general rigid body motion of the structure as a whole.
Consider, for example, a number of RBAR elements connected together to form a rigid ring. Let the grid points be
numbered from 1 to N and assume that the u m degrees of freedom for each rigid element are placed at the higher
numbered grid point so that the only degrees of freedom which remain independent as each element is added to the
ring are those at grid point 1. The addition of the last rigid element between grid points N and 1 will remove even
those independent degrees of freedom and thereby over-constrain the structure.
5. For the RSSCON, the shell degrees of freedom are placed in u m . The translational degrees of freedom of the shell
edge are connected to the translational degrees of freedom of the upper- and lower-edge of the solid. The shell’s two
rotational degrees of freedom are also connected to the translational degrees of freedom of the upper and lower edge
of the solid. The RSSCON only impresses a rigid constraint on the shell’s two rotational degrees of freedom.
6. Nonlinear forces in dynamic analysis cannot be applied to members of the u m set.

Rigid Element Enhancements

Rigid elements, such as RBAR, RBE1, … etc., have been available in MSC Nastran since the inception of MSC Nastran.
However, they contain limitations, which are discussed below.
In the “Theory” section, first, an introduction is given to the formulation of large rotation theory used by the Lagrange
rigid element, followed by a summary of theoretical formulation for each type of the rigid element is given.
Finally, the “User Interface” and “Examples” sections demonstrate the usage of the new Lagrange rigid elements.

Types of Rigid Element

In MSC Nastran, two types of rigid elements are available -- the linear rigid element and the Lagrange rigid element.
Linear Rigid Elements

Linear rigid elements have been available since the inception of MSC Nastran. The linear rigid elements use the linear
elimination technique and are not “real” finite elements. Instead, they are internally represented by a set of MPC equations.
By using these MPC equations, the dependent degrees of freedom (the m-set) are eliminated from the solution set (the l-
set). These types of rigid elements have the following limitations. They:
 Do not compute thermal load.
 Do not have differential stiffness matrix so that the solutions are incorrect for the buckling analysis or other
solution sequences where the differential stiffness matrix is required.
 Use the small rotation theory in the geometrical nonlinear analysis so that the solutions are incorrect in this type of
analysis.
 Use the elimination method for solution resulting in very dense stiffness matrices. These dense matrices cannot
take advantages of the sparse matrix algorithm.

Main Index
 Chapter 3: Structural Elements 163
Rigid Elements and Multipoint Constraints (R-type, MPC)

The Lagrange rigid element, which is discussed in next sub-section, does not have the above limitations.

Lagrange Rigid Elements

These Lagrange rigid elements are “real” finite elements, similar to, for example, QUAD4 element. Instead of using MPC
equations, the element stiffness matrix is computed for each Lagrange rigid element. All the limitations for the linear rigid
element are removed, i.e., they
 Include thermal load effect.
 Include the differential stiffness.
 Support for both the small rotation and the large rotation. Use the large rotation theory in the geometrical
nonlinear analysis (PARAM,LDISP,1).
 Take advantages of the sparse matrix algorithm if the augmented Lagrange multiplier method, defined below, is
used for solution.
For each Lagrange rigid element, a number of Lagrange multiplier degrees of freedom are created by MSC Nastran
internally. For example, one to six Lagrange multiplier degrees of freedom are created for RBAR and one Lagrange
multiplier degree of freedom is created for RROD. For each Lagrange rigid element, the independent degrees of freedom,
the dependent degrees of freedom, and the Lagrange multiplier degrees of freedom are first placed in the solution set (the
l-set). Two methods are available for solution:
Augmented Lagrange Multiplier Method

For this method, the solution is obtained with the independent degrees of freedom, the dependent degrees of freedom, and
the Lagrange multipliers degrees of freedom in the solution set. Thus, the sparse characteristic of the stiffness matrix is
maintained and the sparse matrix algorithm can be utilized.
Lagrange Elimination Method

For this method, additional operations of eliminating both the dependent and Lagrange multiplier degrees of freedom from
the solution set are performed after the global stiffness matrix is assembled. The solution is performed on the independent
set (the n-set), creating dense matrices after elimination. Once again, these dense matrices cannot take advantage of the
efficiency features of the sparse method.
Method Selection

For most problems, the augmented Lagrange multiplier method is the preferred method. The tangent stiffness matrix using
this method, however, is not positive definite, and can potentially pose numerical difficulties. It is for this reason that the
Lagrange eliminator method is introduced as a backup method.
For linear rigid elements, the dependent degrees of freedom are eliminated and placed in the mp-set. For the augmented
Lagrange multiplier rigid elements, both the dependent degrees of freedom and the internally created Lagrange multipliers
are placed in the l-set. For the Lagrange elimination rigid elements, both the dependent degrees of freedom and the
Lagrange multipliers are eliminated and placed in the mr-set. The mp-set and the mr-set are subsets of the m-set. Please see
Degree-of-Freedom Set Definitions in the MSC Nastran Quick Reference Guide.

Main Index
 164 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

The Lagrange elements are only available for linear static analysis (SOL 101), normal modes analysis (SOL 103), buckling
analysis (SOL 105), and general nonlinear analysis (SOL 400). Furthermore, the Lagrange elimination method is not
implemented for the following reasons:
 For nonlinear analysis, the tangent stiffness matrix must be updated at each nonlinear iteration for the Lagrange
elimination method. (See The Lagrange Elimination Method.) This implies that only the full Newton-Raphson method
can be used, which conflicts with the general iteration solution scheme employed in MSC Nastran.
 The denseness of the tangent stiffness matrix after elimination poses an inefficiency problem for the sparse solver.
 For nonlinear analysis, only stiffness matrix decomposition is required and the eigenvalue analysis is not required.
For stiffness matrix decomposition, the method for solving a non-positive definite matrix is well established, and
hence removes the need for a backup method for the Lagrange multiplier method.
There are some additional limitations for SOL 400. Please see Limitations on SOL 400.

Theory
In this section, we give introduction to the theoretical formulation of the rigid elements. Rigid elements use large rotation
theory and hence warrant a discussion of large rotation. In the following sections, the method for representing large rotation
for rigid elements is presented first, followed by the solution methods for the linear elimination method, the Lagrange
multiplier method, and the Lagrange multiplier method with elimination. Also, only the solution methods for the
nonlinear analysis are discussed.
Rotation Vector

Mathematically, there are many ways of representing large rotation; the most commonly used methods are Euler angles,
Gimbal angles, Rodrigues parameters, quaternion (Euler parameters), and rotation vector methods. MSC Nastran
implemented the Gimbal angles and the rotation vector method into the geometric nonlinear analysis. The Gimbal angles
have inherent limitations. They are non-invariant parameters, which means that their values are intimately associated with
the coordinate system used. On the other hand, the rotation vector gives a simple geometric meaning of rotations and it is
an invariant parameter. Thus, the Lagrange elements use the rotation vector for representing large rotations, which is
described in the next paragraph. Please note that the Lagrange elements do not have an option to use the Gimbal angles.
The most general motions of a rigid body can be represented by a translation of a base point and rotation about the base
point. Therefore, the motion of a GRID point rigid body can be decomposed into a translation of the GRID point and a
rotation defined by:
 = p
where:

Main Index
 Chapter 3: Structural Elements 165
Rigid Elements and Multipoint Constraints (R-type, MPC)

1---
T 2
 =  =  

 px 
  
p = ---- =  p y 
  
 pz 
(3-35)

Thus  is the magnitude or the angle of the rotation and p is the direction of the rotation.
It is well known that large rotations cannot be treated as vectors. Consequently, the assumption that rotations are linear,
which is one of the basic assumptions of linear analysis, does not provide a unique solution when a series of large rotations
is applied in different sequences. However, the effect of large rotation can be expressed in terms of the rotation matrix
R    . Any vector v can be rotated into new position v r given by the equation:

v r = R   v

2
cos  + p x  1 – cos   – p z sin  + p x p y  1 – cos   p y sin  + p x p z  1 – cos  
R = 2 (3-36)
p z sin  + p x p y  1 – cos   cos  + p y  1 – cos   – p x sin  + p y p z  1 – cos  
2
– p y sin  + p x p z  1 – cos   p x sin  + p y p z  1 – cos   cos  + p z  1 – cos  

The rotational matrix is an orthogonal matrix, which has the property:

T –1 T
RR = I or R = R

A very important matrix associated with the rotation vector  is the skew-symmetric matrix  , which is defined by:

0 –z y
ˆ
 = z 0 –x (3-37)
–y x 0

The rotation vector  and its skew-symmetric matrix have the relationship:
ˆ ˆ
  = 0 and  v =   v for any vector v

Main Index
 166 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

ˆ
 is the axial vector of the skew-symmetric matrix  . In term of the skew-symmetric matrix and the unit vector p of
rotation, the rotation matrix can be rewritten as:
T T ˆ
R    = pp + cos   I – pp  + sin p (3-38)

where p̂ is the skew-symmetric matrix of the unit vector p .

Composite Rotations -- Left Rotation or Right Rotation

A composite rotation is successive application of two or more rotations. For large rotation, this series of rotations is no
longer additive. In addition, they are not commutative, i.e., the order of application of the rotation is important. For
example, in a geometric nonlinear analysis, an incremental rotation  g is computed for a load increment or a nonlinear
iteration. The composite rotation of  g and the previous rotation  i can be expressed by the equation:

R i + 1   i + 1  = R l   g R i   i  (3-39)

Let us examine the coordinate system used for (3-39). Both  i + 1 and  i are computed in terms of the global coordinate
system at the GRID point. Note that the global coordinate system at any GRID point is fixed in space. Here we must be
very careful about the coordinate system used for the incremental rotation  g . In order for (3-39) to be true, we must
compute the incremental rotation in terms of the same coordinate system used by  i + 1 and  i , i.e., the global
coordinate system at the GRID point. The corresponding incremental rotation matrix R l computed is called the left
rotation.
There is another way to compute the incremental rotation matrix. Let us rotate the global coordinate system at the GRID
point by  i , i.e.

m x = R i   i g x (3-40)

where g x , x = 1 2 3 , are the unit vectors of the global coordinate system g at the GRID point and m x are the unit
vectors of the rotated coordinate system m. The coordinate system m is called the material coordinate system at the GRID
point. Please note that the material coordinate system is not fixed in the space and it is rotated with the GRID point. Now,
let us compute the incremental rotation  m in terms of the material coordinate system. Then, the composite rotation of
the incremental rotation and the previous rotation can be expressed by:
R i + 1   i + 1  = R i   i R r   m  (3-41)

The rotation matrix R r so computed is called the right rotation. It can be shown that the right rotation and the left
rotation are related by the equations:

Main Index
 Chapter 3: Structural Elements 167
Rigid Elements and Multipoint Constraints (R-type, MPC)

T
 m = R i  g (3-42)

T
R r = R i R l R i (3-43)

The right rotation has the following advantages:

 The representation of inertia forces is greatly simplified.
 The stiffness properties of each elastic member may be described in a rigorous manner, including the geometric
stiffening effects.
To improve the geometric nonlinear analysis, the rigid elements are formulated in terms of the right rotation approach.
However, all the existing elements, CQUAR4, CTRIA3, and CBEAM, are formulated in terms of the left rotation
approach. In order to make the existing elements compatible with the rigid elements, the existing elements are modified in
terms of the right rotation approach.
The Linear Elimination Method

For the linear elimination method, all rigid elements are reduced using MPC constraint equations given by:
um = Gm un (3-44)

where u m are the dependent degrees of freedom, u n are the independent degrees of freedom, and G m is the constraint
matrix. During the solution process, the dependent degrees of freedom are eliminated from the solution set (the l-set). This
method is valid only for small rotation because the relationship given by (3-44) is linear. Also, the thermal loads and the
differential stiffness matrix are not computed. For large rotation, the relationships between dependent degrees of freedom
and independent degrees of freedom are nonlinear. This elimination method can be modified to solve the nonlinear rigid
element problem. However, we use the augmented Lagrange multiplier method instead, which is discussed in next sub-
section.
The Augmented Lagrange Multiplier Method

The equilibrium equation for a static nonlinear solution of a constrained structural model using the Lagrange multiplier
method is given by:
T
F+H  = p (3-45)
u = q

where F is the internal force vector,  is a vector of Lagrange multipliers, p is the external force vector, u is the displacement
vector,  is a vector of nonlinear constraints on the displacements generated by the rigid elements, q is the external
enforced displacements or effect due to thermal loads, and H is the displacement gradient given by:
 
H = ------    u  – q  = ------- (3-46)
u u

Main Index
 168 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

The explicit form for  and H are dependent on the individual rigid element type. For example, the forms are different
for the RBAR and the RBE3.
The equilibrium for (3-45) is obtained by minimizing the functional:

T 1 T
 =  k + k   – q   + ---   – q  p   – q  (3-47)
2
The first term  k of (3-47) contains the regular strain energy and external work done to the system. For rigid elements, this
term is zero. The second term is contribution due to the Lagrange multipliers, where k is a scale factor introduced to
balance the magnitude of the Lagrange multipliers with other terms in the stiffness matrix. The last term is the added
penalty function term and p is the penalty function. The reason for adding the penalty function term is to reduce
numerical difficulty during matrix decomposition.
The constrained equilibrium equation can be also solved by the penalty function method. For this method, the magnitude
of the penalty function must be very large in comparison to the tangential stiffness matrix in order to obtain an
approximation solution of the original constrained equation. When the penalty function approaches infinity, the solution
becomes the exact solution of the original constrained equation. The shortcoming of this method is that, when the penalty
function approaches infinity, the tangential stiffness matrix becomes ill conditioned. On the other hand, in the augmented
Lagrangian method as given by (3-47), the penalty function p needs not to be large. It only needs to be a similar magnitude
as the other terms in the tangential stiffness matrix for enforcing the positive definiteness of the stiffness matrix.
* *
For the nonlinear analysis, the next step is to derive the Newton-Raphson iteration equation from (3-47). Let  u    be
an approximation solution to (3-47). By using the Newton-Raphson method, a corrected solution can be obtained in the
form:
* *
 u,   =  u + u, +   (3-48)

 u   eu 
K ee   =   (3-49)
    e 

where K ee is the tangential stiffness matrix and the right hand side is the error vector. The first and second variations of
(3-47) are

T T T
 = u H  p   – q  + k  +  k   – q  (3-50)

T T T T  T
d    = u H  pHdu + kd  +  kHdu + u ------  H  p   – q  + k  du (3-51)
u
From (3-51), we obtain the corresponding element tangential stiffness matrix as:

Main Index
 Chapter 3: Structural Elements 169
Rigid Elements and Multipoint Constraints (R-type, MPC)

T T - T
-----
+ u  H  p   – q  + k   0
e d
K ee = K ee + K ee = H pH kH (3-52)
kH 0 0 0
In the above equation, the first term is the element stiffness and the second term is the differential stiffness for the rigid
element. Next, we need to compute the error vector for the rigid element in a nonlinear iteration. For the approximate
* *
solution  u    , the error vector can be obtained from (3-50) as
T * * *
 eu   – H  u   k + p    u  – q   
  =   (3-53)
 e   kq – u 
* 

With the tangential stiffness matrix and the error vector defined, (3-48) and (3-49) are used to obtain a better approximation.
This process continues until the solution converges.
The Lagrange Elimination Method

The stiffness matrix for the Lagrange multiplier method given by (3-52) is not positive definite. For most problems, this
method is appropriate. However, for certain type of problems, this method may create numerical difficulties. Therefore, the
Lagrange multiplier method with elimination, called the Lagrange elimination method, is implemented as a backup method
for the Lagrange multiplier method. For reasons given in Types of Rigid Element, 162, the Lagrange elimination method is
implemented only for the linear static analysis, the normal modes analysis, and the buckling analysis. However, for
completeness, the theory given below is still based on the nonlinear analysis.
The Lagrange elimination method is summarized as follows:
 Compute the rigid element tangential stiffness matrix and error vector by using the Lagrange multiplier method as
given by (3-52) and (3-53) with the penalty function p=0.
 Assemble the element stiffness and error vector into global stiffness and error vector, which will create a system
equation similar to (3-49).
 Eliminate the Lagrange multipliers and the dependent degrees of freedom from the tangential stiffness matrix and
the error vector.
 Construct and solve the Newton-Raphson equation in terms of the independent degrees of freedom.
The method of elimination is derived as follows. Let g denote the total displacement degrees of freedom for the structural
model, m the dependent degrees of freedom, and n the independent degrees of freedom. Then the displacement gradient
defined by (3-46) can be partitioned into m-set and n-set as

 u m 
H mg u g = H mm H mn   (3-54)
 u n 
And the constraint equation for the dependent degrees of freedom can be computed as

Main Index
 170 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

u m = G mn u n
–1
(3-55)
G mm = – H mm H mn

Now, the global tangential stiffness matrix in g-set, similar to (3-49), (3-52), and (3-53) can be expanded into m-set and n-set
as

T  p – F  u *  – G T  u *  * 
K nn K nm G mn  u n   n n g mn g m   en 
   * *   
T  u m  =  p – F  u  +   =  em  (3-56)
K nm K mm – I   
m m g m
  

G mn – I 0  m  –H–1  q –   u*     qm 
 mm m g 
The right hand side of above equation is the error vector. Elimination of m-set displacements and the Lagrange multipliers
from above equation yields the tangent stiffness matrix in terms of independent degrees of freedom, the n-set, as:
K nn u n = e n (3-57)

T T T
K nn = K nn + K nm G mn + G mn K nm + G mn K mm G mn (3-58)

T T
e n = e n + G mn e m +  K nm + G mn K mm q m (3-59)

u m = G mn u n – q m (3-60)

T
 m = K nm u n + K mm u m – e m (3-61)

(3-57), (3-58), and (3-59) can be used in the Newton-Raphson iterations, and (3-60) and (3-61) can be used to recover the
displacements in m-set and the Lagrange multipliers. In general, the error vector given by (3-57) is dependent on the
tangential stiffness matrix. In order to compute the correct error vector, the stiffness matrix must be updated at each
nonlinear iteration. This means only the full Newton-Raphson method can be used to obtain the solution for the Lagrange
elimination method.

User Interfaces
Case Control Command RIGID
Control Case command RIGID selects the type of rigid element. It has the following format:

Main Index
 Chapter 3: Structural Elements 171
Rigid Elements and Multipoint Constraints (R-type, MPC)

 LINEAR 
 
RIGID =  LAGR 
 
 LGELIM 
LINEAR will select the linear rigid elements, LAGR will select the Lagrange rigid element with the Lagrange multiplier
method, and LGELIM will select the Lagrange element with the Lagrange elimination method.
If RIGID command does not exist in the user Case Control file, the linear rigid element will be used for all solution
sequences except SOL 400. For SOL 400, the default rigid element type is the Lagrange rigid element with the Lagrange
multiplier method.
Parameters LANGLE, LMFACT, PENFN, and ORIGID

The parameter LANGLE selects the method to represent the lager rotations, which has the following values:
LANGLE,1 -- use the Gimbal angle method.
LANGLE,2 -- use the left rotation method.
LANGLE,3 -- use the right rotation method.
See sub-section, Composite Rotations -- Left Rotation or Right Rotation, 166 for a description of the rotation methods. LANGLE=1
or 2 is the existing method. LANGLE=3 is the new method implemented for the Lagrange rigid elements and is the only
method available for the Lagrange rigid elements. Therefore, if there are Lagrange rigid elements in a structural model,
MSC Nastran will automatically use LANGLE=3, and ignore user input on the LANGLE parameter.
The parameters LMFACT and PENFN are the scale factor k and penalty function p (see (3-47)) for the Lagrange rigid
element with the Lagrange multiplier method. The purpose of LMFACT and PENFN is to make the values of the stiffness
matrix for the Lagrange rigid elements about the same relative magnitude as those of the other elements in a structural
model. Too small a value produces inaccurate results and too large a value poses numerical difficulties. Usually, the same
value is assign to both LMFACT and PENFN. Under special requirement, user may assign different values to LMFACT
and PENFN. For example, if LMFACT  0.0 and PENFN = 0.0, then the solution method becomes the pure Lagrange
multiplier method instead of the augmented Lagrange multiplier method. However, user must exercise caution when
assigning different values to LMFACT and PENFN. The default value is 1.0e+5 for all solution sequences except SOL 400.
For the nonlinear solution sequence SOL 400, MSC Nastran will compute the appropriate default value in nonlinear
iterations.
The parameter ORIGID is the offset for the identification number of the Lagrange multiplier. MSC Nastran internally
generates the Lagrange multiplier degrees of freedom for the Lagrange elements. The identification number for the
Lagrange multiplier is ORIGID+MGID, where MGID is the grid point identification number of the dependent grid point
of the rigid element. The default value for ORIGID is 101000000 if there are no other internally generated grid points.
This value is modifiable using system cell 178
Bulk Data Entries

Main Index
 172 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

Existing rigid elements -- RBAR, RBE1, RBE2, RBE3, RROD, and RTRPLT -- can be used as either the linear rigid
element or the Lagrange element, which is selected by the Case Control command, RIGID. In addition, there are three new
Bulk Data entries, RBAR1, RTRPLT1, and RJOINT, which are discussed in next two sub-sections.
For each Lagrange rigid element, a number of Lagrange multiplier degrees of freedom are created internally by MSC
Nastran. The number of the Lagrange multiplier degrees of freedom is equal to the dependent degrees of freedom defined
by the rigid element. The identification numbers of Lagrange multiplier are determined by the parameter ORIGID.
All of the rigid elements described above can either be the linear rigid element or the Lagrange rigid element. However, the
input rules are not the same for these two types of rigid element. The following paragraphs discuss the differences.
Except for the RBE3 and RROD, the major difference in the input format between the linear rigid element and the
Lagrange rigid element is the selection of the independent degrees of freedom. This can be illustrated by the Bulk Data
entry for RBAR:

RBAR EID GA GB CNA CNB CMA CMB ALPHA

The independent degrees of freedom are selected by CNA and CNB. For the linear rigid element, the independent degrees
of freedom can be assign to both CNA and CNB: for example CNA=1236, CNB=34, as long as the total number equals
to six and they can jointly represent any general rigid body motion. However, for the Lagrange rigid element, all six
independent degrees of freedom must be assigned to a single grid point, i.e., “CNA=123456, CNB=blank”, or
“CNA=blank, CNB=123456”. The same rule applies to the RBE1, RBE2, and RTRPLT elements.
For the RBE3, the REFC degrees of freedom can be any combination of the integer 1 through 6 for the linear rigid element.
For the Lagrange rigid element, RFEC must be 123, 456, or 123456.
For the RROD element, the user must select one dependent degree of freedom by inputting either CMA or CMB for the
linear rigid element. However, for Lagrange rigid element, user can leave both fields blank and let MSC Nastran to select
the best component as the dependent degree of freedom. In fact, this is the recommended method.
For all rigid elements, there is a new field ALPHA. This is the thermal coefficient of expansion. For the Lagrange rigid
elements, if ALPHA is given and the thermal loads are requested by the Case Control command
TEMPERATURE(INITIAL) and TEMPERATURE(LOAD), the thermal load effect will be computed for the rigid
elements. The temperature loads are taken as the average temperature given by the independent grid point and the
dependent grid point. For example, the temperature load for the BAR element is taken as the average temperature of grid
points GA and GB. For the linear rigid element, no temperature effect is computed and the ALPHA field is ignored.
RBAR1 and RTRPLT1

RBAR1 and RTRPLT1 are two new Bulk Data entries and are alternative input formats for RBAR and RTRPLT. The Bulk
Data entry for RBAR1 is shown below:

RBAR1 EID GA GB CMB ALPHA

GA is the independent grid point and all six degrees of freedom are independent. The dependent degrees of freedom are
selected by CMB of the dependent grid point GB. This is a much simpler input format than RBAR and is preferred input
format for the Lagrange rigid element.
The format for RTRPLT1 is shown below:

Main Index
 Chapter 3: Structural Elements 173
Rigid Elements and Multipoint Constraints (R-type, MPC)

RTRPLT1 EID GA GB GC CMB CMC ALPHA

RTRPLT1 defines one independent grid point GA and two dependent grid points GB and GC with dependent degrees of
freedom CMB and CMC, respectively. Between RTRPLT1 and RTRPLT, RTRPLT1 is the preferred input format for the
Lagrange rigid element.
RJOINT

RJOINT is a new rigid element. The format for RJOINT is as follows:

RJOINT EID GA GB CB

GA is the independent grid point and all six degrees of freedom are independent. GB is the dependent grid point. The
length between points GA and GB must be zero. Since its length is zero, the thermal load effect is not applicable.
If CB=123456 or blank, then the grid point GB is constrained to move with the grid point GA and two the grid points
move as a single point. If any degree of freedom is released on CB, then RJOINT becomes a mechanical joint. A mechanical
joint is a mechanical system that has two bodies jointed at a point. The two bodies can rotate relatively about one, two or
three axis of a local coordinate system at the point. For example:
 Hinge - A hinge is a mechanical joint that rotates freely about one axis about the local coordinate system. It can be
simulated by RJOINT with one rotational degree of freedom released, i.e., CB = 12356, 12346, or 12345.
 Universal joint - A universal joint is a mechanical joint that rotates freely in two axes. It can be simulated by
RJOINT with two rotational degrees of freedom released, i.e., CB = 1234, 1235, or 1236.
 Spherical joint - A spherical joint is a mechanical joint that rotates freely about all three axes. It can be simulated
by a zero length RJOINT with all rotational degrees of freedom released, i.e. CB = 123.
 Prismatic joint - A prismatic joint is a mechanical system with two blocks that are constrained to have the same
rotations, but translate relative with each other along a local axis. It can be simulated by RJOINT with one
translational degree of freedom released, i.e., CB = 23456, 13456, 12456.
 Cylindrical joint - A cylindrical joint is a mechanical system that allows two grid points have relative translation
along a moving axis and, at the same, have relative rotation about the same axis. It can be simulated by RJOINT
with one translational degree of freedom and one rotational degree of freedom released, i.e. CB = 2356, 1346,
1245.
Guidelines for the RJOINT element:
 The theory for the RJOINT is formulated such that a consistent mechanical joint is created even if user requests
different global coordinate systems at grid points GA and GB. If different global coordinate systems are used, the
degrees of freedom to be released is determined by the coordinate system at GB.
 For linear or nonlinear static analysis, the joints must constrained by other elements in the structural model.
Otherwise a singular stiffness matrix may be produced, leading to failure in decomposition.

Limitations on SOL 400

As described in sub-section, Composite Rotations -- Left Rotation or Right Rotation, the Lagrange rigid elements are formulated in
terms of the right rotation method and the existing elements, CQUAD4, CTRIA3, and CBEAM, are formulated in the left

Main Index
 174 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

rotation method. In order to make these existing elements compatible with the Lagrange rigid elements, the existing
elements must be modified to formulate in the right rotation method. This work is under development for the initial release
of MSC Nastran 2004. Therefore, for the initial release, SOL 400 has the following limitations:
 If Lagrange rigid elements are used in a model that contains CQUAD4, CTRIA3, or CBEAM elements, then the
problem will converge only for truly 2D motion. If there are follower force effects, the out-of-plane motion may
need to be constrained in this case. Full 3D motion is supported with the Lagrange rigid elements if the model
does not contain the above elements.

Examples
In this section, four simple examples are included to illustrate the usage of the Lagrange rigid elements. These four examples
are
 Linear static analysis with thermal loads
 Nonlinear geometric analysis with thermal loads
 Buckling analysis
 Normal modes analysis with differential stiffness
Example 1: Linear Static Analysis with Thermal Loads (nlrgd11a.dat)

 To illustrate the effect of thermal loads on the Lagrange rigid elements in a static analysis, a simple beam is
modeled by a RBAR1 element, which has a length of 1.0. At left end of the beam, a very soft rotational spring is
connected to the normal rotation and other five degrees-of freedom are fixed. At right end, all six degrees of
freedom are free and a vertical force of 100.0 is acting on it. This beam is subjected to a thermal load, which will
double its length to 2.0. This problem is solved by the Lagrange multiplier method (RIGID=LAGR) with the
linear static analysis (SOL 101).
 The input file for this problem is shown below.
ID RIGID,NLRGD10A
SOL 101
CEND
TITLE = LINEAR STATIC ANALYSIS - THERMAL LOAD ON RBAR
SUBTIT= LAGRANGE MULTIPLIER METHOD
RIGID = LAGR
TEMP(INIT) = 10
SUBCASE 1
TEMP(LOAD) = 20
LOAD = 100
SPC = 10
DISP = ALL
SPCF = ALL
FORC = ALL
BEGIN BULK
FORCE,100,2,0,500.0,0.0,1.0,0.0
CELAS2,101,1000.0,1,6
GRID,1,,1.0,0.0,0.0
GRID,2,,2.0,0.0,0.0

Main Index
 Chapter 3: Structural Elements 175
Rigid Elements and Multipoint Constraints (R-type, MPC)

RBAR1,3,1,2,123456,2.0-2
TEMP,10,1,0.0
TEMP,10,2,0.0
TEMP,20,1,50.0
TEMP,20,2,50.0
SPC1,10,12345,1
ENDDATA
Results for nlrgdlla.dat:

Description Value
Vertical displacement 0.5
Horizontal displacement 1.0
Angle of rotation 28.6o
Initial beam length 1.0
Final beam length 2.06

These results show that horizontal thermal expansion is correctly computed as 1.0. However, because of the soft rotational
spring on the left end of beam, this problem is a large displacement problem; therefore, the vertical and horizontal
displacements, angle of rotation, and the final beam length are incorrect.
Example 2: Geometric Nonlinear Analysis with Thermal Loads (nlrgd11b.dat)

This problem is to illustrate the use of the Lagrange rigid elements in geometric nonlinear analysis. The model and loading
for this problem is identical to nlrgd11a, except that the geometric nonlinear analysis (SOL 400 and LGDISP,1) is used.
Note that the RIGID Case Control command in the is not needed for SOL 400 as the default for SOL 400 is
RIGID=LAGR.
The input file for this problem follows.
ID RIGID,NLRGD11B
SOL 400
CEND
TITLE = GEOMETRIC NONLINEAR ANALYSIS - THERMAL LOADS ON RBAR
SUBTIT= LAGRANGE MULTIPLIER METHOD
TEMP(INIT) = 10
NLPARM = 10
SUBCASE 1
TEMP(LOAD) = 20
LOAD = 100
SPC = 10
DISP = ALL
SPCF = ALL
FORC = ALL
BEGIN BULK
NLPARM,10,10

Main Index
 176 MSC Nastran Reference Guide
Rigid Elements and Multipoint Constraints (R-type, MPC)

PARAM,LGDISP,+1
.
.
.
ENDDATA
Results for nlrgd11b.dat:

Description Value
Vertical displacement 1.344
Horizontal displacement 0.481
Angle of rotation 42.22o
Initial beam length 1.0
Final beam length 2.0

Now the final beam length is 2.0, which is expected, and all above results are correct when large displacement theory is used.
Example 3: Buckling Analysis (nlrgd10d.dat)

To illustrate the use of the Lagrange rigid elements in a buckling analysis, two simple columns are modeled by RBAR
elements and subject to compressive loads of 100.0 at top of the column. At bottom of the columns, they are connected to
a single grid point and are constrained by very soft rotational spring in both horizontal directions. Also, at bottom, the three
translations and axial rotation are fixed. The buckling analysis (SOL 105) and the Lagrange elimination method
(RIGID=LGELIM) are used to obtain the solution.
The input file for this problem is shown below.
ID RIGID,NLRGD10D
SOL 105
CEND
TITLE = BUCKLING ANALYSIS - RBAR
SUBTI = LAGRANGE ELIMINATION METHOD
DISP = ALL
SPC = 10
RIGID = LGELIM
SUBCASE 1
$ STATIC PRELOAD CASE
LOAD = 100
SUBCASE 2
$ BUCKLING CASE
METHOD = 10
BEGIN BULK
EIGRL,10,,,10
FORCE,100,4 ,0,-100.0,1.0,0.0,0.0
FORCE,100,41,0,-100.0,1.0,0.0,0.0
CELAS2,101,100.0,3,5

Main Index
 Chapter 3: Structural Elements 177
Rigid Elements and Multipoint Constraints (R-type, MPC)

CELAS2,102,120.0,3,6
GRID, 3 ,,2.0,0.0,0.0
GRID, 4 ,,4.0,0.0,0.0
GRID, 41,,4.0,0.0,0.0
RBAR, 4,3,4
RBAR, 3,3,41,123456, , ,123456
SPC1,10,1234,3
ENDDATA
The buckling load factors calculated are 0.25 and 0.30.
Example 4: Normal Modes Analysis with Differential Stiffness (nlrgd14c.dat)

This problem illustrates the effect of differential stiffness matrix with the Lagrange rigid elements for normal modes
analysis. The model of this problem is similar to that of nlrgd10a, except that concentrated masses are added to the top of
the columns and tensile forces of 100.0 units are applied. The normal modes analysis (SOL 103) using the Lagrange
multiplier method (RIGID=LAGR) are used to obtain the solution.
Two problems are investigated--one without and one with the effects of the differential stiffness matrix included.
The input file for the problem including the differential stiffness effect is shown below. For the problem without the
differential stiffness, just remove subcase 1.
ID RIGID,NLRGD14C
SOL 103
CEND
TITLE = NORMAL MODES ANALYSIS WITH DIFFERENTIAL STIFFNESS - RBAR
SUBTI = LAGRANGE MULTIPLIER METHOD
DISP = ALL
SPC = 10
RIGID = LAGR
SUBCASE 1
$ STATIC PRELOAD CASE
LOAD = 100
SUBCASE 2
$ NORMAL MODES CASE
STATSUB = 1
METHOD = 10
BEGIN BULK
EIGRL,10,,,10
CONM2,201,4 ,,10.
CONM2,202,41,,10.
FORCE,100,4 ,0,+100.0,1.0,0.0,0.0
FORCE,100,41,0,+100.0,1.0,0.0,0.0
CELAS2,101,100.0, 3,5
CELAS2,102,120.0, 3,6
GRID, 3 ,,2.0,0.0,0.0
GRID, 4 ,,4.0,0.0,0.0
GRID, 41,,4.0,0.0,0.0
RBAR, 4,3,4
RBAR, 3,3,41,123456, , ,123456
SPC1,10,1234,3

Main Index
 178 MSC Nastran Reference Guide
Scalar Elements (CELASi, CMASSi, CDAMPi)

ENDDATA
Results for nlrgd11b.dat:

Description Value
Eigenvalues without preload 1.25, 1.50
Eigenvalues with preload 6.25, 6.50

From the above results, we can see clearly that the differential stiffness can affect the results substantially.

Scalar Elements (CELASi, CMASSi, CDAMPi)

Scalar elements are connected between pairs of degrees of freedom (at either scalar or geometric grid points) or between one
degree of freedom and ground. Scalar elements are available as springs, masses, and viscous dampers. Scalar spring elements
are useful for representing elastic properties that cannot be conveniently modeled with the usual structural elements
(elements whose stiffnesses are derived from geometric properties). Scalar masses are useful for the selective representation
of inertia properties, such as occurs when a concentrated mass is effectively isolated for motion in one direction only. The
scalar damper is used to provide viscous damping between two selected degrees of freedom or between one degree of
freedom and ground. It is possible, using only scalar elements and constraints, to construct a model for the linear behavior
of any structure. However, use of scalar elements with offsets will cause incorrect results in buckling analysis and differential
stiffness because the large displacement effects are not calculated. Offsets will also cause internal constraints in linear
analysis, i.e., hidden constraints to ground. Therefore, it is expected that these elements will be used only when the usual
structural elements are not satisfactory. Scalar elements are useful for modeling part of a structure with its vibration modes
or when trying to consider electrical or heat transfer properties as part of an overall structural analysis. The reader is referred
to Sections 5.5 and 5.6 of The NASTRAN Theoretical Manual for further discussions on the use of scalar elements.
The most general definition of a scalar spring is given with a CELAS1 entry. The associated properties are given on the
PELAS entry. The properties include the magnitude of the elastic spring, a damping coefficient, and a stress coefficient to
be used in stress recovery. The CELAS2 defines a scalar spring without reference to a property entry. The CELAS3 entry
defines a scalar spring that is connected only to scalar points and the properties are given on a PELAS entry. The CELAS4
entry defines a scalar spring that is connected only to scalar points and without reference to a property entry. No damping
coefficient or stress coefficient is available with the CELAS4 entry. Element force F is calculated from the equation.
Static:
F = k  u1 – u2 

Frequency:
·
F =  1 + i  g + g e  kU ; U =  u 1 – u 2  real + i· u 1 – u 2  img (3-62)

Transient:

Main Index
 Chapter 3: Structural Elements 179
Axisymmetric Solid Elements (CTRIAX6, CTRIAX, CQUADX)

g ge
F =  u 1 – u 2 k +  -------- + -------- k  u· 1 – u· 2  (3-63)
 W3 W4

where k is the stiffness coefficient for the scalar element and u 1 is the displacement of the first degree of freedom listed
on its connection entry. Element stresses are calculated from the equation
 = SF (3-64)

where S is the stress coefficient on the connection or property entry and F is as defined above.
Scalar elements may be connected to ground without the use of constraint entries. Grounded connections are indicated on
the connection entry by leaving the appropriate scalar identification number blank. Since the values for scalar elements are
not functions of material properties, no references to such entries are needed.
The CMASS1, CMASS2, CMASS3, and CMASS4 entries define scalar masses in a manner similar to the scalar spring
definitions. The associated PMASS entry contains only the magnitude of the scalar mass.
The CDAMP1, CDAMP2, CDAMP3, CDAMP4, and CDAMP5 entries define scalar dampers in a manner similar to the
scalar spring definitions. The associated PDAMP entry contains only a value for the scalar damper. The mode displacement
method (PARAM,DDRMM,-1) must be selected for element force output.
For the CDAMP, the force equations are:
Frequency:
F = iB  u 1 – u 2 

Transient:
F = B  u· 1 – u· 2 

Axisymmetric Solid Elements (CTRIAX6, CTRIAX, CQUADX)

The axisymmetric elements CTRIAX6, CTRIAX, and CQUADX define a solid ring by sweeping a surface defined on a plane
through a circular arc. The CTRIAX and CQUADX elements are described in Hyperelastic Elements. The CTRIAX6 element
is described as follows.
The triangular ring element (CTRIAX6) is a linear isoparametric element with an axisymmetric configuration that is
restricted to axisymmetric applied loading. It is used for the modeling of axisymmetric, thick-walled structures of arbitrary
profile. This element is not designed to be used with any other elements. Otherwise, CTRIAX6 is used in a conventional
manner, and except for its own connection entry (and a pressure load entry, PLOADX1) it does not require special bulk
data entries.
The coordinate system for the CTRIAX6 element is shown in Figure 3-46. Cylindrical anisotropy is optional. Orientation
of the orthotropic axes in the ( r , z ) plane is specified by the angle  . Deformation behavior of the element is described

Main Index
 180 MSC Nastran Reference Guide
Axisymmetric Solid Elements (CTRIAX6, CTRIAX, CQUADX)

in terms of translations in the r and z directions at each of the six grid points. All other degrees of freedom should be
constrained.
z

z = zbasic
zm
rm

G5

axial G4
G3
G6
TH

G2
G1
r=
radial

Figure 3-46 CTRIAX6 Element Coordinate System

The isoparametric triangular ring element is defined with a CTRIAX6 entry. No property entry is used for this element.
The material property reference is given on the connection entry.
The following stresses, evaluated at the three vertex grid points and at the centroid of the element, are output on request:
  r -- stress in r m direction of material coordinate system.

Main Index
 Chapter 3: Structural Elements 181
Crack Tip Elements (CRAC2D, CRAC3D)

   -- stress in azimuthal direction.

  z -- stress in z m direction of material coordinate system.

  rz -- shear stress in material coordinate system.

 Maximum principal stress.
 Maximum shear stress.
 von Mises equivalent or octahedral shear stress.

Crack Tip Elements (CRAC2D, CRAC3D)

MSC Nastran crack tip elements include both two-dimensional (CRAC2D) and three-dimensional (CRAC3D) types.
ADUM8 and ADUM9 Bulk Data entries are needed with these elements. The formats of these entries are as follows:

1 2 3 4 5 6 7 8 9 10
ADUM8 18 0 5 0 CRAC2D

ADUM9 64 0 6 0 CRAC3D

The two geometries of the CRAC2D element are shown in Figure 3-47 and Figure 3-48. For the quadrilateral crack element,
corner and crack grid points 1 through 10 are required while midside grid points 11 through 18 are optional. The CRAC3D
element geometry is depicted in Figure 3-49. For the brick crack element, grid points 1 through 10 and 19 through 28 are
required; whereas grid points 11 through 18 and 29 through 64 are optional.

Main Index
 182 MSC Nastran Reference Guide
Crack Tip Elements (CRAC2D, CRAC3D)

G3 G12 G4 G13 G5

2 3
G11 G14
1
4
G2 G1 G6
x
G10
5
8
G18 G15
7 6

G9 G17 G8 G16 G7
Figure 3-47 Quadrilateral Crack Element

G3 G12 G4 G13 G5

2 3
G11 G14

1 4
x
G2 G1 G7 G6
Figure 3-48 Symmetric Half-Crack Option

Main Index
 Chapter 3: Structural Elements 183
Crack Tip Elements (CRAC2D, CRAC3D)

The CRAC2D and CRAC3D entries are used to define the element geometry, while properties are defined on PRAC2D
and PRAC3D entries. Both elements can be modeled with temperature-independent anisotropic materials. The 2-D
element may be either plane stress or plane strain. The element will generate either coupled or lumped mass matrices.
Figure 3-47 depicts the quadrilateral and symmetric half options for the CRAC2D element. The element is divided, by the
program, into basic triangular elements 1 through 8, whereas the symmetric half-crack option is divided into triangles 1
through 4. Grid points 1 through 7 are required for the symmetric half-crack option, while grid points 11 through 14 are
optional. Stresses and the local coordinates of these stresses for the quadrilateral option are computed at the origin of the
natural coordinates of triangles 4 and 5. The stresses and coordinates are then averaged and reported. Stress intensity factors,
K I and K II , are computed for triangles 1 and 8, averaged and reported. For the symmetric half-crack option, coordinates
and stresses are reported at the origin of the natural coordinates of triangle 4 while the stress intensity factor K I only is
reported for triangle 1.
Interpretation of CRAC2D Element Stress Output (Dummy Element Format)

S1 S2 S3 S4 S5 S6 S7 S8 S9
x y x y  xy KI KII 0 0

where x and y are the element coordinates where stresses are reported. K I and K II are stress intensity factors.

Figure 3-49 depicts the 3-D brick and symmetric half-crack options for the CRAC3D element. The brick element is divided,
by the program, into basic wedge elements labeled 1 through 8, whereas the symmetric half-crack option is divided into
wedges 1 through 4. Grid points 1 through 7 and 19 through 25 are required for symmetric crack option. Stresses and the
local coordinates of these stresses for the brick element option are computed at the origin of the natural coordinates of
wedges 4 and 5. The stresses and coordinates are then averaged and reported. Stress intensity factors, K I and K II , are
computed for wedges 1 and 8, averaged, and reported. For the symmetric half-crack option, coordinates and stresses are
reported at the origin of the natural coordinates of wedge 4, whereas the stress intensity factor K I only is reported for wedge
1.

Main Index
 184 MSC Nastran Reference Guide
Crack Tip Elements (CRAC2D, CRAC3D)

G21 G22
G23

G20
G19 G24

G28

G27 G26 G25

y

(a) Brick Crack Option with Required

Connection Points Only.
G4 G5
G3
2 3
G21 G22 G23

G2 1 4 G6
x
G10 G1
8 5 G25
G20 G24
7 6
G19
G9 G8 G7

G4 G5 (b) Symmetric Half-Crack Option

G3 with Required Connection Points
2 3 Only.
1 4
x
G2 G1 G7 G6

Figure 3-49 The CRAC3D Solid Crack Tip Element

Main Index
 Chapter 3: Structural Elements 185
Crack Tip Elements (CRAC2D, CRAC3D)

G21 G30 G22 G31 G23

YB G57 G58 G59

G29 G32
G56
G20 G19 G60
G24
G39 G64
G40 G28
G41 G33
y G36 G63
Y1 X1 G62 G61

G25
G38 G27 G35 G26 G34
XB
G37 G42
G46
G12 G4 G13
G3 G5

G48 G50 (a) Brick Crack Option

G11 G14 G45 G44 G43
G49
G1
G47 G51 G21 G30 G22 G31 G23
G2 x
G10 G55 G6
G54 G52 G29 G57 G58 G59
G18 G15
G53
G20 G19 G25 G24

G9 G17 G8 G16 G7 G56

G40
G39 G41

y G32
G38

G37 G42

G3 G12 G4 G13
G5

G48 G50
G11 G14 (b) Symmetric Half-Crack Option
G49

x
G2 G47 G1 G7 G6

Figure 3-50 The CRAC3D Solid Crack Tip Element (continued)

Main Index
 186 MSC Nastran Reference Guide
Gap and Line Contact Elements (CGAP, BCONP, BLSEG)

Interpretation of CRAC3D Element Stress Output (Dummy Element Format)

S1 S2 S3 S4 S5 S6 S7 S8 S9
x y z xy yz zx KI KII KIII

Stresses are reported at the average of the origin of the natural coordinate of wedges 4 and 5, whereas for the symmetric
crack option they are reported at the origin of the natural coordinate of wedge 4.
The CRAC2D element is based upon a 2-D formulation, but may be used in three-dimensional structures. However, the
element should be planar. Any deviation from a planar element is checked, and if significant deviations arise, error messages
will be issued.
The CRAC3D element is based upon a 3-D formulation. Both the faces (formed by grid points 1 through 18 and grid
points 19 through 36) and the midplane (grid points 37 through 46) should be planar. Any significant deviation will be
checked and error messages will be issued.

Gap and Line Contact Elements (CGAP, BCONP, BLSEG)

Gap Element
The gap and friction elements are specified on a CGAP entry. The element coordinate system and nomenclature are shown
in Figure 3-51. CID is required, if it is used to define the element coordinate system. Otherwise, the X-axis of the element
coordinate system, x elem , is defined by a line connecting GA and GB of the gap element. The orientation of the gap

element is determined by vector v similar to the definition of the beam element, which is in the direction from grid points
GA to GO or defined by (X1, X2, X3).

Figure 3-51 Gap Element Coordinate Systems

The properties for the gap elements are defined on the PGAP entry. The initial gap opening is defined by U O . If the gap
is closed ( U A – U B  U O ), the axial stiffness ( KA ) has a very large value (relative to the adjacent structure). When the
gap is open, there is a small stiffness KB in the axial direction.
MSC Nastran includes two types of gap elements: nonadaptive and adaptive. When the nonadaptive GAP element is used,
the anisotropic coefficients of friction (  1 and  2 ) are specified for the frictional displacements. Also, the anisotropic
coefficients of friction are replaced by the coefficients of static and dynamic friction  s and  k . On the PGAP
continuation entry, the allowable penetration limit T max should be specified because there is no default. In general, the
recommended allowable penetration T max is about 10% of the element thickness for plates or the equivalent thickness for

Main Index
 Chapter 3: Structural Elements 187
Gap and Line Contact Elements (CGAP, BCONP, BLSEG)

z elem

y elem
KA and KB in this figure
are from the PGAP entry.

GA

x elem v

KB KA - KB

GB

other elements that are connected by GA and GB. When T max is set to zero, the penalty values will not be adjusted
adaptively.
Gap element forces (or stresses) and relative displacements are requested by the STRESS or FORCE Case Control command
and computed in the element coordinate system. A positive axial force F x indicates compression. For the element with
friction, the magnitude of the slip displacement is always less than the shear displacement after the slip starts. For the
element without friction, the shear displacements and slip displacements have the same value.

3-D Slideline Contact

Definition and Acronyms

The terminology associated with the slideline capability is as follows:

Main Index
 188 MSC Nastran Reference Guide
Gap and Line Contact Elements (CGAP, BCONP, BLSEG)

Two deformable bodies that come into contact are called the primary and the secondary. The modeling of
interactions between the primary and secondary bodies requires you to define the contact regions in terms of
slidelines. A primary line is a list of grid points in the topological order on the primary body and a secondary line
is a list of grid points in the topological order on the secondary body. A slideline contact region consists of a
primary line and a secondary line as shown in Figure 3-52. The grid points on the primary line are called the
primary nodes and those on the secondary line are called the secondary nodes.

k - 1 Secondary Segment
Secondary Line
k
k+1 k-1 k - 1 Secondary Segment Normal
l -1 lth Primary Segment Normal
l+ 1 Primary Line
y Normals Must Face
Each Other
lth Primary Segment
x

z Slideline Plane Vector Direction

Figure 3-52 Typical Slideline Contact Region

 Arrows show positive direction for ordering nodes. Counterclockwise from primary line to secondary line.
 X-Y plane is the slideline plane. Unit normal in the Z-direction is the slideline plane vector.
 Secondary and primary segment normals must face each other; i.e., the normals must be in the opposite directions.
 A line segment joining two consecutive primary nodes is called a primary segment. The primary and the secondary
nodes in a contact region must lie in a plane called the slideline plane (see Figure 3-52). The slideline plane is
defined by a vector normal to the plane called the slideline plane vector. The primary and the secondary can have
large relative motions within the slideline plane. However, relative motion outside the slideline plane is ignored,
and therefore, motion must be small compared to a typical primary segment. Thus, slideline contact is ideal for
modeling interactions between two bodies that may come in contact in specified planes. MSC Nastran checks to
make sure that in the initial geometry all the nodes of a contact region lie in the slideline plane and issues the
appropriate messages if they are not. However, no such check is made during the analysis.

Main Index
 Chapter 3: Structural Elements 189
Gap and Line Contact Elements (CGAP, BCONP, BLSEG)

In general, contact is determined only for secondary nodes and the primary line. This may result in primary nodes
penetrating the secondary line. However, the error involved depends only on the mesh discretization. At the
expense of increase in computer runtime, you may wish to determine contact between primary nodes and the
secondary line in addition to determining contact between secondary nodes and the primary line. This is called
symmetric penetration. Thus for symmetric penetration, no distinction is made between secondary and primary.
Initial penetration between two bodies is not allowed. The initial gaps between the secondary nodes and the
primary line are calculated based on the coordinates specified for the secondary and the primary nodes. If the grid
coordinates are not specified accurately enough (i.e., MSC Nastran will determine whether initial penetration
exists), then MSC Nastran will adjust the coordinates automatically. In the case where the calculated initial
penetration is less than ten percent of the length of the primary segment to which it projects, MSC Nastran issues
a warning message and automatically adjusts the coordinates. However, in the case where the initial penetration is
larger than ten percent, the analysis terminates and a fatal message is issued. If the element normal of the primary
and secondary segments are not pointing towards each other, MSC Nastran will stop and issue a fatal message
similar to one issued in the case of initial penetration, even though the initial gaps may be acceptable.
Guidelines

MSC Nastran creates a local coordinate system for each primary segment to determine contact between the secondary nodes
and the primary line. The tangent direction for the primary segment is given by the vector from its first node to the second
node. It is very important that you specify the primary nodes in the topological order. MSC Nastran then determines the
normal direction to the primary segment by the cross product between the slideline plane vector and the tangent vector.
The normal direction must point from the primary line to the secondary line because MSC Nastran determines the contact
between the secondary nodes and the primary line by measuring how close the secondary nodes are to the primary segment
in that normal direction (see Figure 3-52).
For symmetric penetration, tangent and normal directions are also defined for the secondary segment similar to the primary
segment. The normals for the primary segment must point toward the secondary and the normals for the secondary must
point toward the primary (see Figure 3-52). This is generally accomplished by traversing from primary line to secondary line
in a counter-clockwise fashion or clockwise fashion depending on whether the slideline plane vector forms a right-hand or
a left-hand coordinate system with the slideline plane. It is your responsibility to make sure that the normals face each other.
MSC Nastran checks indirectly to make sure that the normals face each other by calculating the initial gap between the two
bodies. This is not a foolproof check, especially when there is no initial clearance.
Once MSC Nastran detects a contact between a node and a body, a three node slideline element as shown in Figure 3-53 is
automatically created. The first node of this element is the secondary node and the last two nodes are the primary nodes of
the primary segment with which the secondary node is in contact. The secondary node is forced to slide on the primary
line after contact and must remain on the primary line until a tensile force develops.

Main Index
 190 MSC Nastran Reference Guide
Gap and Line Contact Elements (CGAP, BCONP, BLSEG)

x2 – x1

6
5
n
m2

2
gn
1
4 S
t 3
m1

a x2 – x1
gt

a 0 x2 – x1

 S, m 1 , , m 2 = secondary, primary node 1 and primary node 2, respectively

 a, a 0 = current and previous surface coordinate

 g n = penetration of secondary node into the primary segment

 g t = sliding of the secondary node on the primary segment

 n = normal direction for the primary segment

Figure 3-53 Geometry and Definition of a Typical Three-Node Slideline Element

Main Index
 Chapter 3: Structural Elements 191
Gap and Line Contact Elements (CGAP, BCONP, BLSEG)

In MSC Nastran the contact and traction forces/stresses are always associated with secondary nodes. In order to compute
the contact and traction stresses, an area is associated with each secondary node, based on the contributory length and the
width/thickness from the adjacent secondary segments. Therefore, you need to specify widths associated with secondary
lines in a three-dimensional model or thickness associated with secondary lines in a two dimensional model. You can also
request output results for any number of slideline contact regions. The output for a secondary node consists of (a) the
slideline contact region identification number, (b) the primary segment to which it projects, (c) the parametric surface
coordinate to identify the exact projection of secondary nodes relative to the primary nodes of the primary segment, and
(d) a slip ratio to indicate whether the secondary node is sticking, slipping or sliding.
Text Input

Table 3-14 shows how the slideline input data is given to MSC Nastran through executive and case control as well as the bulk
data entries. For a complete list of these input entries, please refer to the MSC Nastran Quick Reference Guide.

Table 3-14 MSC Nastran Input for Nonlinear Slideline Contact

Slideline Input MSC Nastran Input
Analysis Type and Diagnostics Executive Control Parameters:
Request
Nonlinear Statics SOL 106
Nonlinear Dynamics SOL 129
Slideline Diagnostics DIAG 35 -- Provides extended slideline diagnostics output.
Slideline Output Request Case Control Parameters:
BOUTPUT -- Selects the contact regions for history output.
Bulk Data Entries:
Contact Parameters BCONP -- Defines the parameters for contact between two bodies.
Boundary Line Segments BLSEG -- Defines the curve, which consists of a number of line segments via grid
point numbers that may come in contact with other body.
Boundary Line Segments BWIDTH -- Defines width/thickness for line segments in 3-D/2-D slideline contact
defined in the corresponding BLSEG Bulk Data entry.
Width/Thickness
Contact Friction BFRIC -- Defines frictional properties between two bodies in contact.
Contact Output BOUTPUT -- Defines the secondary nodes at which the output is requested.
User Control and Results ADPCON -- (Default = 1.0); Automatically calculated penalty values can be scaled
Processing; Parameters by specifying a real value for the ADPCON. This is useful for changing the penalty
values on restart. This will change the penalty value to be equal to a number
calculated by MSC Nastran times SFAC value specified in BCONP Bulk Data entry
times |ADPCON|. ADPCON is applied to all the contact regions.

Main Index
 192 MSC Nastran Reference Guide
Concentrated Mass Elements (CONM1, CONM2)

The slideline contact formulation is based on the penalty method, which uses an artificial stiffness value. MSC Nastran will
automatically calculate the appropriate numbers based on local stiffness. You have the option to scale the calculated penalty
values. The automatically calculated penalty values may sometimes cause convergence difficulties. However, you can
overcome the convergence problems by restarting the analysis from any previous converged solution and scaling down the
penalty value. Generally, we recommend decreasing the penalty values by one order of magnitude at a time.

Concentrated Mass Elements (CONM1, CONM2)

The concentrated mass elements are used to define a concentrated mass at a grid point. Two forms of input of concentrated
mass are supported, the CONM1 and CONM2 elements.
The CONM1 allows a general 6 × 6 symmetric mass matrix in a specified coordinate system to be assigned to a geometric
grid point. In addition, a less general form, the CONM2 element allows a concentrated mass about its center of gravity to
be specified. CONM2 allows for the specification of the offset of the center of gravity of the concentrated mass relative to
grid point location, a reference coordinate system, the mass and a 3 × 3 symmetric matrix of mass moments of inertia
measured from its center of gravity. For additional information see the Bulk Data Entries in the MSC Nastran Quick Reference
Guide.

General Element Capability (GENEL)

The input required for the general element is the GENEL Bulk Data entry. The description of GENEL in the MSC Nastran
Quick Reference Guide contains a complete description of the input options and an example.
The principal application of the GENEL element is to represent the stiffness of a substructure which has an arbitrary
number of connected grid point components and/or scalar points. The input data may be obtained either from another
computer run or from test data.
The general element is a structural stiffness element connected to any number of degrees of freedom, as specified by the
user. In defining the form of the externally generated data on the stiffness of the element, two major options are provided
to the user.
1. Instead of supplying the stiffness matrix for the element directly, the user provides the deflection influence
coefficients for the structure supported in a nonredundant manner. The associated matrix of the restrained rigid
body motions may be input or may be generated internally by the program.
2. The stiffness matrix of the element may be input directly. This stiffness matrix may be for the unsupported body,
containing all the rigid body modes, or it may be for a subset of the body’s degrees of freedom from which some or
all of the rigid body motions are deleted. In the latter case, the option is given for automatic inflation of the stiffness
matrix to reintroduce the restrained rigid body terms, provided that the original support conditions did not
constitute a redundant set of reactions. An important advantage of this option is that, if the original support
conditions restrain all rigid body motions, the reduced stiffness matrix need not be specified by the user to high
precision in order to preserve the rigid body properties of the element.
The defining equation for the general element when written in the flexibility form is

Main Index
 Chapter 3: Structural Elements 193
General Element Capability (GENEL)

 ui  Z S  fi 
  =   (3-65)
T
 fd  –S 0  ud 
where:

Z = is the matrix of deflection influence coefficients for coordinates  u  when coordinates  u 
i d
are rigidly restrained.
S = is a rigid body matrix whose terms are the displacements  u  due to unit motions of the
i
coordinates  u d  , when all f i = 0 .

 f i  = are the forces applied to the element at the  u i  full coordinates.

 f d  = are the forces applied to the element at the  u d  coordinates. They are assumed to be statically
related to the  f i  forces, i.e., they constitute a nonredundant set of reactions for the element.

The defining equation for the general element when written in the stiffness form is

 fi  k –k S  ui 
  = T T   (3-66)
 fd  –S k S k S  ud 

–1
where all symbols have the same meaning as in (3-65) and  k  =  Z  , when  k  is nonsingular. Note, however, that it
is permissible for  k  to be singular. (3-66) derivable from (3-65) when  k  is nonsingular.

Input data for the element consists of lists of the u i and u d coordinates, which may occur at either geometric or scalar grid
points; the values of the elements of the  Z  matrix, or the values of the elements of the  k  matrix; and (optionally) the
values of the elements of the  S  matrix.

The user may request that the program internally generate the  S  matrix. If so, the  u i  and coordinates occur only at
geometric grid points, and there must be six or less  u d  coordinates that provide a nonredundant set of reactions for the
element as a three-dimensional body.
The  S  matrix is internally generated as follows. Let  u b  be a set of six independent motions (three translations and
three rotations) along coordinate axes at the origin of the basic coordinate system. Let the relationship between  u d  and
 ub  .

 ud  =  Dd   ub  (3-67)

Main Index
 194 MSC Nastran Reference Guide
General Element Capability (GENEL)

The elements of  D d  are easily calculated from the basic (x,y,z) geometric coordinates of the grid points at which the
elements of  u d  occur, and the transformations between basic and global (local) coordinate systems. Let the relationship
between  u i  and  u b  be

 ui  =  Di   ub  (3-68)

where  D i  is calculated in the same manner as  D d  . Then, if  D d  is nonsingular,

–1
 S  =  Di   Dd  (3-69)

Note that, if the set  u d  is not a sufficient set of reactions,  D d  is singular and  S  cannot be computed in the manner
shown. When  u d  contains fewer than six elements, the matrix  D d  is not directly invertible but a submatrix  a  of
rank r , where r is the number of elements of  u d  , can be extracted and inverted.

A method which is available only for the stiffness formulation and not for the flexibility formulation will be described. The
flexibility formulation requires that  u d  have six components. The method is as follows. Let  u d  be augmented by 6-r
displacement components  u d   which are restrained to zero value. We may then write

 ud  Dd
  =  ub  =  D   ub  (3-70)
 dd   Dd 

The matrix  D d  is examined and a nonsingular subset  a  with r rows and columns is found.  u b  is then reordered
to identify its first r elements with  u d  . The remaining elements of  u b  are equated to the elements of  u d  . The

complete matrix  D  then has the form

a b
D = (3-71)
0 I
with an inverse

–1 –1
–1 a –a b
D = (3-72)

0 I

Main Index
 Chapter 3: Structural Elements 195
General Element Capability (GENEL)

Since the members of  u d   are restrained to zero value,

 ub  =  Dr   ud  (3-73)

where  D r  is the ( 6  r ) partitioned matrix given by

–1
a
 Dr  = (3-74)

The  D i  matrix is formed as before and the  S  matrix is then

 S  =  Di   Dr  (3-75)

Although this procedure will replace all deleted rigid body motions, it is not necessary to do this if a stiffness matrix rather
than a flexibility matrix is input. It is, however, a highly recommended procedure because it will eliminate errors due to
nonsatisfaction of rigid body properties by imprecise input data.
The stiffness matrix of the element written in partitioned form is

K ii K id
 K ee  = (3-76)
T
K id K dd

When the flexibility formulation is used, the program evaluates the partitions of  K ee  from  Z  and  S  as follows:

–1
 K ii  =  Z  (3-77)

–1
 K id  =  Z   S  (3-78)

T –1
 K dd  =  S   Z   S  (3-79)

If a stiffness matrix,  k  , rather than a flexibility matrix is input, the partitions of  K ee  are

 K ii  =  k  (3-80)

 K id  = –  k   S  (3-81)

Main Index
 196 MSC Nastran Reference Guide
Direct Matrix Input on Grid Points (DMIG)

T
 K dd  = –  S   k   S  (3-82)

No internal forces or other output data are produced for the general element.

Direct Matrix Input on Grid Points (DMIG)

The Bulk Data entry DMIG can be used to input a stiffness (or mass) matrix which connects specified degrees of freedom.
The matrix so defined will be added to the stiffness (or mass) matrix computed from finite element properties.
The DMIG entry includes provisions for unsymmetric terms and complex values, both of which are useful in dynamic
analysis. These provisions should not be used in static or normal modes. Note that an entry in the Case Control Section is
required (K2GG = NAME for a stiffness matrix or M2GG = NAME for a mass matrix). See Case Control Commands in the
MSC Nastran Quick Reference Guide for general instructions regarding the Case Control Section.
The primary application of the DMIG Bulk Data entry is to enter stiffness and mass data for parts of the structure which
are obtained from another computer run. The format is cumbersome (two matrix terms per continuation entry) and the
matrix should be input to high precision (see Use of Parameters for a discussion of double-field entries). For stiffness matrices
only, the GENEL Bulk Data entry is an alternative for manually inputting data.

Hyperelastic Elements
The hyperelastic elements are intended for fully nonlinear (finite deformation) analysis using SOL 106 or SOL 129
including the effect of large strain and large rotation (SOL 400 is recommended). Geometric nonlinearity is a subset of this
type of analysis. In addition, the elements are especially designed to handle nonlinear elastic materials at the nearly
incompressible limit. Volumetric locking avoidance is provided through a mixed formulation, based on a three field
variational principle, with isoparametric displacement and discontinuous pressure and volumetric strain interpolations.
Shear locking avoidance is provided through the use of second order elements. See Small Strain Elements.
The hyperelastic elements are defined on the same connection entries as the other shell and solid elements and are
distinguished by their property entries. A PLPLANE or PLSOLID entry defines a hyperelastic element. The hyperelastic
material, which is characterized by a generalized Rivlin polynomial form of order 5, applicable to compressible elastomers,
is defined on the MATHP entry. See Hyperelastic Material for a more detailed discussion.

Hyperelastic Solid Elements

The following elements are available:
1. CTETRA -- Four-sided solid element with 4 to 10 nodes.
2. CPENTA -- Five-sided solid element with 6 to 15 nodes.
3. CHEXA -- Six-sided solid element with 8 to 20 nodes.

See Figure 3-43, Figure 3-44, and Figure 3-45 on solid elements.

Main Index
 Chapter 3: Structural Elements 197
Hyperelastic Elements

There is no element coordinate system associated with the hyperelastic solid elements. All output is in the basic coordinate
system. The following quantities are output at the Gauss points:
 Cauchy stresses
 x ,  y ,  z ,  xy ,  yz ,  zx
 Pressure
1
p = ---   x +  y +  z 
3
Logarithmic strains
3
T
 =  lnl Nl Nl
l=1
 Volumetric strain
dV – dV 0
J – 1 = ----------------------
dV 0
See Small Strain Elements for definitions of these quantities.

Hyperelastic Plane Elements

These are plane strain elements defined on the following connectivities:
 CQUAD -- Quadrilateral element with 4 to 9 nodes. When the center node is missing, this element may also be
specified on a CQUAD8 connectivity entry. When all edge nodes are missing, the CQUAD4 connectivity may be
used.
 CTRIA3 -- Triangular element with 3 nodes.
 CTRIA6 -- Triangular element with 3 to 6 nodes.
Figure 3-54 shows the element connectivity for the CQUAD element.

Main Index
 198 MSC Nastran Reference Guide
Hyperelastic Elements

 = 1 G7
G4
G3
 = constant
Lines of
Constant  or 
 = 0  = 1
 = 0  = constant
G8 G9
G6

 = –1
G5 G2
G1
Figure 3-54 CQUAD Element

See Figure 3-27 through Figure 3-30 for the other elements. Note, however, that there is no element coordinate system
associated with the hyperelastic plane elements. All output is in the CID coordinate system. Cauchy stresses  x ,  y ,  z ,
 xy , pressure p = 1  3   x +  y +  z  , logarithmic strains and volumetric strain are output at the Gauss points.
The plane of deformation is the XY plane of the CID coordinate system, defined on the PLPLANE property entry. The
model and all loading must lie on this plane, which, by default, is the XY plane of the basic coordinate system. The
displacement along the Z axis of the CID coordinate system is zero or constant.

Hyperelastic Axisymmetric Elements

The axisymmetric hyperelastic elements are defined on the following connectivity entries:
 CQUADX -- Quadrilateral axisymmetric element with 4 to 9 nodes.
 CTRIAX -- Triangular axisymmetric element with 3 to 6 nodes.

Main Index
 Chapter 3: Structural Elements 199
Hyperelastic Elements

z = y basic

G5

axial G4
G3
G6

G2
G1
radial r = x basic

Figure 3-55 The CTRIAX Hyperelastic Coordinate System Element

Main Index
 200 MSC Nastran Reference Guide
Bushing Elements

z = y basic

G3
G4 G7

G6
axial G9
G8 G2

G5
G1

radial r = x basic
Figure 3-56 The CQUADX Hyperelastic Coordinate System Element

The plane of deformation is the XY plane of the basic coordinate system, and Y is the axisymmetry axis. The model and all
loading must lie on this plane.
All output is in the basic coordinate system. Cauchy stresses  x (radial), pressure  y (axial),  z (circumferential),  xy ,
pressure p = 1  3   x +  y +  z  , logarithmic strains and volumetric strain are output at the Gauss points.

Pressure loads, specified on the PLOAD4 and PLOADX1 Bulk Data entries respectively, with follower force characteristics,
are available for the solid and axisymmetric elements. A pressure load may not be specified on the plane strain elements.
Temperature loads may be specified for all hyperelastic elements on the TEMP and TEMD entries. The hyperelastic
material, however, may not be temperature dependent. Temperature affects the stress-strain relation as indicated in
Hyperelastic Material.
GPSTRESS and FORCE (or ELFORCE) output is not available for hyperelastic elements.

Bushing Elements
The bushing (generalized spring and damper) elements consist of the following:
 CBUSH
 CBUSH1D

Main Index
 Chapter 3: Structural Elements 201
Bushing Elements

Bush Element (CBUSH)

The generalized spring-damper element CBUSH is a structural scalar element connecting two noncoincident grid points,
or two coincident grid points, or one grid point with an associated PBUSH entry. This combination is valid for any
structural solution sequence. To make frequency dependent the PBUSH need only have an associated PBUSHT Bulk Data
entry. The PBUSHT entry, for frequency dependency, is only used in SOL 108 and SOL 111. The PBUSHT entry can be
used to define load-displacement dependency in SOL 106 and SOL 400.
In modal frequency response, the basis vectors (system modes)    will be computed only once in the analysis and will be
based on nominal values of the scalar frequency dependent springs. In general, any change in their stiffness due to frequency
will have little impact on the overall contribution to the structural modes.
The stiffness matrix for the CBUSH element takes the diagonal form in the element system:

ku
kv
kw
Ko =
k
x

k
y

k
z

For the B matrix replace the k terms with b.

When transformed into the basic system, there is coupling between translations and rotations, thus ensuring rigid body
requirements.
The element axes are defined by one of the following procedures:
1. If a CID is specified then the element x-axis is along T1, the element y-axis is along T2, and the element z-axis is
along T3 of the CID coordinate system. The options GO or (X1,X2,X3) have no meaning and will be ignored.
Then  T ab  is computed directly from CID.

2. For noncoincident grids ( GA  GB ), if neither options GO or (X1,X2,X3) is specified and no CID is specified,
then the line AB is the element x-axis. No y-axis or z-axis need be specified. This option is only valid when K1 (or
B1) or K4 (or B4) or both on the PBUSH entry are specified but K2, K3, K5, K6 (or B2, B3, B5, B6) are not
specified. If K2, K3, K5, K6 (or B2, B3, B5, B6) are specified, a fatal message will be issued. Then  T ab  is
computed from the given vectors like the beam element.
Direct Frequency Response

Nominal Values

Main Index
 202 MSC Nastran Reference Guide
Bushing Elements

The following matrices are formed only once in the analysis and are based on the parameter input to EMG of /’ ’/ implying
that the nominal values only are to be used for frequency dependent springs and dampers.
1 + K 2x + iK 4
K dd =  1 + ig K dd dd dd

1 + M2
M dd = M dd dd

1 + B 2x
B dd = B dd dd

ESTF

The following matrices are formed at each frequency in the analysis and are based on the parameter input to EMG of
/’ESTF’/
f =  1 + ig K 1f
K dd dd

4f = g f  k f 
K dd j = 16
e j

f = B 1f
B dd dd

ESTNF

The following matrices are formed at each frequency in the analysis and are based on the parameter input to EMG of
/’ESTNF’/
4f nominal =  g f k f nominal 
K dd j = 16
e j

and j runs through the stiffness values specified for the CBUSH element or j = 1 for the CELAS1 and CELAS3 elements.
Then at each frequency form:
f + iK 4f + iK 4f nominal
K dd  K dd + K dd dd dd

f
B dd  B dd + B dd

Then the equation to be solved is:

 –  2 M dd + i B dd + K dd   u d  =  P d 

Modal Frequency Response

Basis Vector and Nominal Values

Main Index
 Chapter 3: Structural Elements 203
Bushing Elements

The basis vector matrix    (system modes) is formed only once in the analysis using nominal values for frequency
dependent elements.
The following matrices are formed only once in the analysis and are based on the parameter input to EMG of /’ ’/ implying
that the nominal values only are to be used for frequency dependent springs and dampers.
2 = K 2x + igK 1 + iK 4
K dd dd dd dd

2 = B 1 + B 2x
B dd dd dd

ESTF

The following matrices are formed at each frequency in the analysis and are based on the parameter input to EMG of
/’ESTF’/
2f = igK 1f
K dd dd

4f = g f  k f 
K dd j = 16
e j

2f = B 1f
B dd dd

ESTNF

The following matrices are formed at each frequency in the analysis and are based on the parameter input to EMG of
/’ESTNF’/
4f nominal =  g f k f nominal 
K dd j = 16
e j

and j runs through the stiffness values specified for the CBUSH element or j = 1 for the CELAS1 and CELAS3 elements.
Then at each frequency form:
2  K 2 + K f + iK 4f + iK 4f nominal
K dd dd dd dd dd

2  B 2 + B f
B dd dd dd

The GKAM module will then produce:

T 2 
K hh =  k  +     K dd

T 2 
B hh =  b  +     B dd

T 2 
M hh =  m  +     M dd

Main Index
 204 MSC Nastran Reference Guide
Bushing Elements

Then the equation to be solved is:

 –  2 M hh + iB hh + K hh   u h  =  P h 

Element force calculation:

Frequency:
F e =   1 + ig t K e + iB e U e

where
gt = g + ge

Static:
Fe = Ke Ue

Transient:

g ge ·
F e = K e U e  t  + B e +  ------- + ------- K e U e  t 
 w3 w4
CBUSH Stiffness and Structural Damping Matrices

If both stiffness K and structural damping ge is specified each with their own frequency dependent tables, matrix terms
of the following form are created:
f f f
K j + i  ge j  K j

If stiffness K with frequency dependent tables and non frequency dependent structural damping ge are specified, matrix
terms of the following form are created:
f 0 f
K j + i  ge j  K j

If non frequency dependent stiffness K with frequency dependent structural damping ge are specified, matrix terms of the
following form are created:
0 f 0
K j + i  ge j  K j

In the above three expressions, the superscript f denotes frequency dependent and the superscript 0 denotes a nominal
value. The subscript j implies the j-th degree of freedom of the CBUSH element. The real term goes into the element
4
stiffness matrix and the imaginary term goes into the element K matrix.

Main Index
 Chapter 3: Structural Elements 205
Bushing Elements

Bush1D Element (CBUSH1D)

The BUSH1D is a one dimensional version of the BUSH element (without the rigid offsets) and supports large
displacements.
The element is defined with the CBUSH1D and a PBUSH1D entry. The user may define several spring or damping values
on the PBUSH1D property entry. It is assumed that springs and dampers work in parallel. The element force is the sum of
all springs and dampers.
The BUSH1D element has axial stiffness and axial damping. The element includes the effects of large deformation. The
elastic forces and the damping forces follow the deformation of the element axis if there is no element coordinate system
defined. The forces stay fixed in the x-direction of the element coordinate system if the user defines such a system. Arbitrary
nonlinear force-displacement and force-velocity functions are defined with tables and equations. A special input format is
provided to model shock absorbers.

Benefits
The element damping follows large deformation. Arbitrary force deflection functions can be modeled conveniently. When
the same components of two grid points must be connected, we recommend using force-deflection functions with the
BUSH1D element instead of using NOLINi entries. The BUSH1D element produces tangent stiffness and tangent
damping matrices, whereas the NOLINi entries do not produce tangent matrices. Therefore, BUSH1D elements are
expected to converge better than NOLINi forces.

Output
The BUSH1D element puts out axial force, relative axial displacement and relative axial velocity. It also puts out stress and
strain if stress and strain coefficients are defined. All element related output (forces, displacements, stresses) is requested
with the STRESS Case Control command.

Guidelines
The element is available in all solution sequences. In static and normal modes solution sequences, the damping is ignored.
In linear dynamic solution sequences, the linear stiffness and damping is used. In linear dynamic solution sequences, the
BUSH1D damping forces are not included in the element force output.
In nonlinear solution sequences, the linear stiffness and damping is used for the initial tangent stiffness and damping. When
nonlinear force functions are defined and the stiffness needs to be updated, the tangents of the force-displacement and
force-velocity curves are used for stiffness and damping. The BUSH1D element is considered to be nonlinear if a nonlinear
force function is defined or if large deformation is turned on (PARAM,LGDISP,1). For a nonlinear BUSH1D element, the
element force output is the sum of the elastic forces and the damping forces. The element is considered to be a linear
element if only a linear stiffness and a linear damping are defined and large deformation is turned off.

Main Index
 206 MSC Nastran Reference Guide
Connector Elements

Limitations
1. The BUSH1D element nonlinear forces are defined with table look ups and equations. Equations are only available
if the default option ADAPT on the TSTEPNL entry is used, equations are not available for the options AUTO
and TSTEP.
2. The table look ups are all single precision. In nonlinear, round-off errors may accumulate due to single precision
table look ups.
3. For linear dynamic solution sequences, the damping forces are not included in the element force output.
4. The “LOG” option on the TABLED1 is not supported with the BUSH1D.

Connector Elements

Introduction
The Connector elements (CWELD, CFAST, CSEAM, and CINTC) can be used to address and solve many of the modeling
issues associated with the other structural connector and fastener elements such as the bars, the rigid elements and MPC’s.
See Connector Elements in the MSC Nastran Linear Static Analysis User’s Guide for the basic definition of these elements.
The connector family of elements enables you to connect surfaces with differing mesh densities and using a subset of spot
weld elements you can connect more than one element per surface.

CWELD Element
The CWELD /CFAST elements have been changed so that there is now a consistent formulation between linear and SOL
400 nonlinear analysis. Additionally, the CSEAM element is now supported in SOL 400 and has a consistent formulation
between linear and SOL 400 nonlinear analysis.

PARAM, NEWWELD, YES (DEFAULT)

 The CWELD/CFAST elements provide the same element output formats in both linear and SOL 400 nonlinear
solutions.
 In the CFAST/CSEAM/CWELD analysis, the auxiliary points generated are in the solution set.
 The CWELD/CFAST algorithm has been improved to find the Best Possible Projection with zero projection
tolerance improvements.
 The improved CWELD with options "PARTPAT" and "ELPAT", and CFAST elements do not move GA and
GB if both are supplied by the user, thus maintaining user mesh integrity.
 The 3x3 mesh limitation has been removed for the CWELD with options "PARTPAT" and "ELPAT" and the
CFAST elements.
 There are no required changes in the user element input description of the CFAST/CSEAM/CWELD elements.
 The CFAST/CWELD/CBUSH provides nonlinear force output for SOL 400 "ANALYSIS=NLTRAN".
 MPC Force output is available for the connector element constraints.

Main Index
 Chapter 3: Structural Elements 207
Connector Elements

 Besides global search algorithm control there is now local connector element connectivity control via the new
CONCTL bulk data entry.
 A brief summary of connector projection results is output in the F06 file for each connector type.
 A new "SWLDPRM, CSVOUT, UNITNUM" entry will produce a comma separated file useful for reports.
 The CSEAM and CWELD (not by default) can now contribute mass to the structure.

Main Index
 208 MSC Nastran Reference Guide
Connector Elements

Additional Information
 In SOL 400, for "ANALYSIS=NLSTAT" or "ANALYSIS=NLTRAN", the generated RBE3 constraints become
Lagrange elements and may undergo large rotation. For "ANALYSIS=NLTRAN" with initial conditions (IC=n)
in case control that cause large initial stresses in the structure at time t=0, the case control entry RIGID needs to
have the value "RIGID=LINEAR" to insure convergence.
 For user desiring to post process the CFAST/CSEAM/CWELD connectors with their own methods, the following
is useful:
a. The GEOM2 table contains, after module MODGM2, a record ELCORR that correlates the
CFAST/CSEAM/CWELD and its associated RBE3 elements. Also, this module will, for linear analysis and for
nonlinear SOL 400 analysis run with RIGID=LINEAR, place the internal generated RBE3 into the GEOM4
table.
b. In SOL 400 with "RIGID=LAGRANGE" (Default), internally generated RBE3 elements go into the GEOM2
(as do all other user specified rigid elements) not the GEOM4 table.
c. The CWELD/CFAST/CBUSH force output for "ANALYSIS=NLTRAN" in SOL 400 is OP2 file output on
OEFNLXX data block. If SCR=POST is run, then this force data is also written to the data base file OEFNL3
and op2 file OEFNL.op2 is also written.
 The DISPLACEMENT (CONNECTOR=) Case Control Command works in the same fashion for both the old
connector formulation and the new connector formulation.
 Reminder: For the CFAST with option "ELEM" and the CWELD with option "ELPAT," the shell elements
connected on each patch must have same property identification number of PSHELL entry.
 Reminder: If parameter OSWPPT is used to specify the offset for internally generated grid IDs, its value should be
greater than the maximum identification number of GRID entries to avoid conflict IDs.

Formulation Changes
In the new consistent formulation for the CWELD/CFAST elements for linear analysis, RBE3 elements are written
internally, and the auxiliary points are in the solution set and both are identifiable by the SWLDPRM, PRTSW entry.
 The auxiliary grids generated start with GRID ID 101000001. There are always four auxiliary grids for patch A
and four auxiliary grids for patch B.
 The RBE3 elements generated start with 100001002. An RBE3 is generated for each auxiliary point for each patch
A and B tying each patch grid to that auxiliary point. There is a RBE3 generated for GA tying GA to its patch
auxiliary points and a RBE3 generated for GB tying GB to its patch auxiliary points.
 Both linear and nonlinear output is consistent.
The CWELD can connect congruent as well as non congruent meshes. The connectivity options can handle spot welds
which connect more than one element per sheet and handle connections other than surface patches. The connector itself
is considered to be rigid in user defined degrees of freedom.
The CWELD using the ELEMID and GRIDID format, (see Figure 3-57) is limited to connect one element per shell sheet
in patch-to-patch type connections. The connectivity could be unsymmetric so that results became unsymmetric for a

Main Index
 Chapter 3: Structural Elements 209
Connector Elements

model with symmetric mesh, loads, and boundary conditions. The two patch-to-patch type connections, formats ELPAT
and PARTPAT,preserve symmetry and can connect more than one element per sheet, see Figure 3-58. In addition, the two
element connectivity options could miss elements if the mesh is so fine that the spot weld area spans over more than one
element, see Figure 3-59. The situation of meshes with element sizes smaller than spot weld sizes may occur in automotive
models for crash analysis.

Figure 3-57 Connected Elements for Format ELEMID

Main Index
 210 MSC Nastran Reference Guide
Connector Elements

Figure 3-58 Connected Elements for Format ELPAT

Figure 3-59 Spot Weld Covering a Patch of Elements

Starting in Nastran V2016.1 the CWELD has been enhanced by default: The original CWELD can be obtained with use
of PARAM, OLDWELD, YES.
CFAST and the CWELD with options "PARTPAT" and "ELPAT" with an improved formulation has removed the
restriction that a connector patch cannot span more than three elements. It will now span over a patch of as many elements
that the value of diameter D of the patch encloses and for which projections can be found.
In the following Figure 3-60 (The example for this figure can be found at /tpl/connectr14/cei_103.dat), all element grids
contained in the green circle region of say patch A will be used in a RBE3 connection in addition to the RBE3 connections
generated for the four auxiliary points. The green circle passes through the four auxiliary points of the patch (The nine digit
grid IDs.). The user-specified diameter D on the PWELD and
PFAST entries determine the locations of the four auxiliary points. (The green circle diameter is approximately 1.253D.).
The element grids shown outside the green circle belong only to the respective auxiliary points.
For higher-order shell elements CQUAD8 or CTRIA6 with no missing midside nodes, the RBE3 relationships use only
the midside nodes. If one or more midside node is missing (NEVER RECOMMENDED), then the corner nodes are used.
The diameter D on the PFAST/PWELD entry is used to determine the projection location of these auxiliary points as well
as the stiffness properties of the patch to patch connection.
A single RBE3 then connects the four auxiliary points and the shell grids within the green circle to the connector grid
GA=4000065.
Reminder: The CWELD ELEMID option still only connects by design two elements. The diameter is only used to
compute the beam stiffness.

Main Index
 Chapter 3: Structural Elements 211
Connector Elements

Figure 3-60 (a) Spot Weld Covering a Patch of Elements

The SWLDPRM, PRTSW entry will list the additional grids located within the green circle under FMESH SHELL A or
B GRIDS where FMESH is the f06 file listing title of the additional shell grids connected in the RBE3 relationships for
Finer MESH.
Table 3-15 shows the grids associated with auxiliary grid 101000023 of Figure 3-60 for its RBE3 generation. The WTi's are
weight factors based on patch shape functions. Grids G1, G2, G3 are selected for RBE3 EID
100001026 because they are the shell grids of the triangular element that contains the projected auxiliary point.

Table 3-15 Auxiliary Grid 101000023 and Associated RBE3 100001026

RBE3 EID GAH3 REFC Weight Ci Gi

100001026 101000023 123 WT1 123 G1=4007883
WT2 123 G2=4007884
WT3 123 G3=4007902

shows the grids associated with grid GA=4000065 of Figure 2-1 through RBE3=100001022. G5 through Gn are the grids
contained within the green circle. Grids G2=4007884, 4007869, 4007815, and 4007830 are NOT included in any of the
G5-Gn entries because they are included in their associated auxiliary point RBE3 elements 100001026 because they are the
shell grids of the triangular element that contains the projected auxiliary point.

Main Index
 212 MSC Nastran Reference Guide
Connector Elements

The CWELD element, using the ELEMID and GRIDID format is restricted to handling connectivities of shell elements

Table 3-16 All GHAi + Patch A Green Circle Grids for RGE3 100001022
RBE3 EID GRD_A REFC Weight Ci Gi
100001022 4000065 123456 WT1 123 GAH1=101000021
WT2 123 GAH2=101000022
WT3 123 GAH3=101000023
WT4 123 GAH4=101000024
1.0 123 G5= 4007885
... ... ...
1.0 123 Gn=last

or surface patches made up of grid point sequences. The point-to-patch type connectivity allows any element type, load or
boundary condition to be connected to the CWELD. The element can now be used to model bearings of large masses,
connectivities to springs, gaps, etc., see Figure 3-61.

Figure 3-61 Modeling Options using the Point-to-patch Type Connectivity

Main Index
 Chapter 3: Structural Elements 213
Connector Elements

Report Format

Theory
Finding CWELD End Points GA and GB

If the CWELD end points GA and GB are not defined, then they are created from the locator point GS as follows. For the
patch-to-patch type connectivity, the user supplied grid point GS is projected on to shell A and B according to Figure 3-62.
The end points GA and GB are created by a normal projection of GS to shell A and B if GS lies between shell A and B. A
two step projection is used if GS lies outside of shell A and B. First, the auxiliary points GA’ and GB’ are created by a normal
projection of GS to shell A and B. From the midpoint GC’ of section GA’,GB’ a second normal projection is done that
produces the final end points GA and GB. All the points described in this paragraph do not have degrees of freedom
attached.

Main Index
 214 MSC Nastran Reference Guide
Connector Elements

Figure 3-62 Projection of GS to Determine End Points GA and GB

Element Coordinate System

The element x-axis points from GA to GB.

xB – xA
e 1 = ---------------------- element x-axis
xB – xA

In case of zero length, the normal of shell A is taken. All vector components are in basic unless noted otherwise.
Find the smallest component j of e 1

j min  i 
e1 = e  .
i = 1 2 3  1 

In case of two equal components we take the one with the smaller i . The corresponding basic vector

 0 
 
b j , e.g., for j = 3, b 3 =  0 
 
 1 

provides a good directional choice for e 2 . In addition, the vector e 2 must be orthogonal to e 1 .

Main Index
 Chapter 3: Structural Elements 215
Connector Elements

e˜2
T
e1 bj
e˜2 = b j – ----------- e 1 e 2 = ---------- element y-axis
T
e1 e1 e˜2

and e 3 is just the cross product

e3 = e1  e2 element z-axis

The final transformation matrix is

T be = e 1 e 2 e 3

Figure 3-63 Element Coordinate System

Element Stiffness
The elastic part of the CWELD is a short beam from GA to GB with 2x6 degrees of freedom. The element is a shear flexible
Timoshenko type beam. The Young’s and shear modulus is taken from the material defined on the PWELD property entry.
The cross sectional properties are calculated from the user supplied spot weld diameter D defined on the PWELD property
entry,

E = Young’s modulus
E
G = shear modulus = --------------------
2 21 + 
A = area = D  4 4
D
I zz = I yy moment of inertia = ----------
64
4
D
----------
J =
32

Main Index
 216 MSC Nastran Reference Guide
Connector Elements

The length L of the beam element is from end points GA to GB. If the user defines the type as SPOT on the PWELD
property entry, then the effective element length L e is always

1
L e = ---  t A + t B 
2
regardless of the true length L. The user supplied values for the Young’s and shear modulus E and G are scaled by the ratio
of true length to effective length
L L
Ẽ = E ----- G̃ = G -----
Le Le

With the scaling, the spot weld stiffness is approximately constant for all elements. Extremely stiff elements from short
lengths L and extremely soft elements from long lengths L are avoided. For zero length L, the beam element degenerates to
6 uncoupled springs. If the user does not define SPOT as the type, the true length is used if it is inside the range of
L
LDMIN  ----  LDMAX
D
where LDMIN and LDMAX are user supplied parameters on the PWELD entry. By default LDMIN = 0.2 and
LDMAX = 5.0 . If the user does not want the true length adjusted at all, LDMIN = 0 . and LDMAX = 1.e+12
may be specified.
Constraints Connecting GA and GB to Shell Grids

The end points GA and GB are connected to the shell grids of shell A and B, respectively, using constraint equations. The
3 translations at grid GA are connected to the 3 translations of the shell grid points using the interpolation functions of the
corresponding shell surface. The 3 rotations at grid GA are connected to the 3 translations of the shell grid points using
Kirchhoff conditions.

 u   u 
   
 v  =
 
 NI  A A    v 
 w A  w I

A w
 x = ------- =  N I y  w I
y
A w
 y = – ------- = –  N I x  w I
x
A 1 v u 1
 z = ---  ----- – ------ = ---   N I x  v I –  N I y u I 
2  x y 2

Main Index
 Chapter 3: Structural Elements 217
Connector Elements

For the formats GRIDID and ELEMID on the CWELD entry, the grid points I are grid points of shell elements or grid
points of solid element faces. For the formats ELPAT and PARTPAT, the grid points I are auxiliary points GAHI and
GBHI, I=1,4, constructed according to Figure 3-64 and Figure 3-65. The auxiliary points are connected to shell element grids
with a second set of constraints

 u   u 
   
 v  =  G IK   v 
   
 w I K  w K

where G IK is a coefficient matrix derived from RBE3 type constraints. The grid points K are shell grid points.

Figure 3-64 Equivalent Square Cross Sectional Area for a Spot Weld of Diameter D

Main Index
 218 MSC Nastran Reference Guide
Connector Elements

Figure 3-65 Auxiliary Points GAHI and GBHI for Formats ELPAT and PARTPAT

With the formats ELPAT and PARTPAT, the element can connect from one to 3x3 elements on shell A and B, respectively,
see Figure 3-66.

Main Index
 Chapter 3: Structural Elements 219
Connector Elements

Figure 3-66 Connectivity for Formats ELPAT and PARTPAT

With the constraint equations above, the degrees of freedom of grid points GA, GB and the auxiliary points GAHI, GBHI
are condensed out on the element level. The final stiffness matrix of the CWELD consists of translational degrees of
freedom of grid points on shell A and B, respectively. There are no extra degrees of freedom generated except for the case
where the user requests explicit constraints by setting MSET=ON on the PWELD entry. The flag MSET=ON works only
for the formats ELEMID and GRIDID.

Connector Stiffness
Connector contribution to a structural model’s overall stiffness is sensitive to the model’s mesh size and the orientation of
the connector relative to the mesh. Thus, the discretization process itself may cause, for example, a model using a fine mesh
to be stiffer in torsion than a corresponding model using a coarse mesh. Also for production models that correlate well with
test, refining the mesh may cause an inherent overall loss of stiffness due to mesh refinement and hence loss of correlation.
To allow the user some control over this loss of stiffness, the new connectors (CWELD with ELPAT or PARTPAT or
CFAST) are provided with two options to provide additional connector stiffness. The two options may be used individually
or in combination.
The first stiffening technique is activated by "SWLDPRM, DRATIO, (1.0 ≤ value ≤ 10.0)” or “CONCTL, SETID, ,
DRATIO, (1.0 ≤ value ≤ 10.0)”. For this option the diameter, Dratio, is defined as Dratio = DRATIO * Dconnector. This

results in the diameter of the patch taking a value of Dpatch = Dratio. The default of DRATIO is a value=1.0 which
implies the diameter of the patch is computed in the standard fashion. For the patch to patch connection for the “beam”
properties of the CWELD, the area is still computed as Aconnector = π D2connector /4 as defined in the PWELD entry.

A disadvantage of this method is that as DRATIO is increased using the global command SWLDPRM, DRATIO, value;
some connector elements may begin to fail because they may no longer be able to find a patch projection.

Main Index
 220 MSC Nastran Reference Guide
Connector Elements

To overcome this, the “SWLDPRM, NREDIA, Integer_value” can be increased to a value as high as Integer_value = 8 to
allow failing welds to halve their patch diameters up to eight times.
If the “SWLDPRM, NREDIA, Integer_value” is not an approach the user wishes to pursue, then for these failing elements,
the bulk data entry CONCTL, SETID, ,DRATIO, value can be used to define a set for failing connectors and set a value
of DRATIO for these connectors that allows them to find a projection.
The second stiffening algorithm attempts, based on the diameter of the connector, to determine a measure of the mesh
discretization.
This feature is activated by "SWLDPRM, SKIN, 1" or "CONCTL, SETID,, SKIN, 1". The default is a 0 which implies
no stiffening. There is an associated stiffening factor "SWLDPRM, SCLSKIN" with value = 0.10 as default.
Depending on the complexity of the model and the overall mesh size and the number of connectors within the model and
the diameter of the connectors relative to the mesh, the default value tends to stiffen a structural model from about 0.4%
to about 4%. A value of SCLSKIN=10.0 stiffens coarser mesh models by about 10% to 11% and finer mesh models by
about 2% to 6%.
The contribution of the stiffening algorithm to the overall stiffness of the FEM model eventually reaches a limit. For
example, a very large value SCLSKIN=100 increases the stiffness of the models overall by only about 0.1% to 2% over the
stiffness obtained for SCLSKIN=10.
For a correlated structural model evaluated at a specific mesh size, with an aim to refine the mesh for some portion of this
model containing connectors, while leaving other portions containing connectors with an unmodified mesh, it is
recommended that the "SKIN, 1" and "SCLSKIN, real value" be entered on the CONCTL bulk data entry referring to the
connectors within the area of the refined mesh. Different refined mesh areas within the structural model can have different
values of SCLSKIN associated to the specific connectors in each refined region.
For post processing of the SKIN option, for the affected shell elements, an updated EPT table is available after module
MODGM2. It contains the PSKNSHL record that correlates the property data of the shells involved and a list of shell
elements for each patch modified.

Connector Control:
To provide the user with better connector control, the following new Bulk Data entry has been introduced:
CONCTL Parameter SWLDPRM override for CFAST, CSEAM, and CWELD Connector Elements.

This entry provides local connector search algorithm control to override SWLDPRM values.
Please see CONCTL (p. 1596) in the MSC Nastran Quick Reference Guide for the complete Bulk Data entry.

Enhanced Algorithm
For the new connector logic, the search algorithm has been enhanced based on user inputs in an attempt to achieve the best
possible connections. The new search tolerance starts with a zero projection tolerance. This may result in changes from the
previous connector results using the old CWELD/CFAST elements.
The list below gives a brief summary of the highlights of the improved CWELD/CFAST algorithm.

Main Index
 Chapter 3: Structural Elements 221
Connector Elements

 For the CFAST and the CWELD with options "PARTPAT" and "ELPAT", grids GA and GB internally keep the
user-specified IDs and the user-specified locations. This change was primarily introduced because many users
complained that the location of GA and GB represented their modeling procedures and desired mesh locations.
 For the CWELD with ELEMID/GRIDID option, grids GA and GB internally keep the user-specified IDs and
the user-specified locations, but in the case when GA and GB are associated with shell patches, a duplicate internal
grid is generated to avoid singularity of the generated RBE3.
CWELD, 5646, 22, , ELEMID, 3276, 3115
, 2191, 1941
CTRIA3, 2191, 8, 3272, 3276, 3271
Grid 3276 as input from standard mesh modeling procedures will automatically be placed in the independent
degree of freedom set, or may have been placed by the analyst in a SPC or MPC set at generation time. In either
case, the CWELD algorithm must create an internal constraint on this grid using a RBE3 element. This causes a
set conflict which is avoided by generating an internal grid.
 For two stacked connectors having a common patch with a common grid, the program checks duplicated GA/GB
and only a single RBE3 is generated for the common patch.
CWELD, 11, 100, 9001, PARTPAT, 3001, 3002
CWELD, 12, 100, 9002, PARTPAT, 3002, 3003
 If the user specifies both grids GA and GB, for the CFAST and the CWELD with options "PARTPAT" and
"ELPAT", the SWLDPRM, GSMOVE entry is nonfunctional. In the CFAST and the CWELD with options
"PARTPAT" and "ELPAT", if the user specifies both GA and GB they will not be moved. This may cause the
CFAST/CWELD search algorithm to fail for some welds that had passed under the old CWELD/CFAST search
algorithm. If this occurs, the user can do one of four things:
a. Determine a better location for GA and GB of the failing welds so that they may project.
b. Remove GA and GB from the CWELD/CFAST entry and replace with a GS entry allowing the
CWELD/CFAST algorithm to move and project and generate internal GA and GB locations.
c. Use "SWLDPRM, MOVGAB, 1" to generate internal GA and GB grids at the corrected locations for all
CWELD/CFAST. The locations of the original user specified GA and GB are unchanged.
d. Use the new CONCTL Bulk Data entry with "SWLDPRM, MOVGAB, 1" to allow local control of specific
welds to correct the locations of grids GA and GB. (See Connector Control section below for detail.)
CONCTL, 83, , MOVGAB, 1
Where:
SET3, 83, ELEM, 1345, 2678
 The maximum tolerance for SWLDPRM, PROJTOL has been relaxed.
a. Regardless of the value of SWLDPRM PROJTOL, the algorithm starts by assuming a zero projection tolerance
for the projections of GA/GB for the CWELD option "PARTPAT" or the CFAST option "PROP" and for
GAHi/GBHi for the CWELD options "PARTPAT" and "ELPAT" and any CFAST option.
b. The tolerance is increased by 0.02 until a projection is found or the PROJTOL value is reached.

Main Index
 222 MSC Nastran Reference Guide
Connector Elements

c. This can be turned off while computing the auxiliary grid projection onto EIDA/EIDB by setting PROJTOL=
- value where 0.0 ≤ value ≤ 1.0. In this case, the projection calculation starts at tolerance = |PROJTOL|. For the
rest of the projection search, the algorithm reverts back to (a) and (b) above.
 A brief "Connector Summary" of projection results is always output in the f06 file for each connector type: FST-
ELEM, WLD-PARTPAT, WLD-ELEMID, etc.

CONNECTOR SUMMARY
Elem type GOOD/BA MAX ANGLE B/N MAX TIMES GS MAX TIMES GS MAX TIMES
D SHELL
MOVED MOVED DIAMETER
NUMBER NORMALS (GSMOVE) AT (GSMOVE) AT
EID REDUCED
FOUND GAB/GH (GSPROJ) AT EID (NREDIA)

EID AT EID

FST-ELEM G 26 GAB 0.0 (20.00) 6000000 GH 0.0 (0.02) 0 ( 0) 6000000 1 (4) 6000018
6000000
B 0

 For linear connectors, MPCFORCE output is available. In nonlinear SOL 400, the RBE3 elements generated
become Lagrange elements if the default RIGID=LAGRANGE is used and are no longer in the MPC set; hence,
there will be NO MPCFORCE output for RIGID=LAGRANGE.
 In the improved CFAST and CWELD, GA and GB are not moved and internal coincident grids are not generated
at a new location; thus, two additional restrictions are required.
a. There can be no user-supplied constraints on GA and GB. A fatal message will be issued if there are any.
b. The CWELD length must be > 10-6. The point to patch option defined by ELEMID or GRIDID will,
however, still create a new GS internally to obtain a minimum required length; i.e., LDMIN ≤ length/D ≤
LDMAX. For the point-to-patch connection, GS is used as GB. The algorithm will use the new GS as GB but
keep the user-supplied GS unchanged. Since the point-to-patch is often used to "tack" two shell corners, the
default LDMIN may cause the connector to be unstable. To avoid this, it is recommended that the user set
LDMIN=1.E-6 on the appropriate PWELD entry.
 Warning: CWELD will not contribute to MASS by default even if its associated MATi entry has a nonzero
density. To react to a nonzero density "SWLDPARM, WMASS, 1" is required. If mass is computed, the PARAM,
COUPMASS effects the mass calculation.

Detailed Projection Algorithm for Best Possible Projection enhanced CWELD

 The Enhanced algorithm applies to the following projection calculations:
a. Find projections of GA/GB for the "PARTPAT" format of CWELD and "PROP" format of CFAST.

Main Index
 Chapter 3: Structural Elements 223
Connector Elements

b. For "ELPAT" format of the CWELD and the "ELEM" format of the CFAST, the user has specified the specific
shell elements and therefore no element search for GA/GB projections is required. Though no search is required,
GA/GB, however, must project onto the user specified EIDA/EIDB.
c. Find projections of the auxiliary grids GAHi/GBHi for the "PARTPAT" and "ELPAT" options on the
CWELD and any CFAST option respectively.
 The projection algorithm searches for possible projections from shell elements with shell grids that are closest to
GS. The closest grid may connect to several shell elements; hence, more than one shell element may get a
projection from GS for curved patches. The shell elements with projections from GS are collected and the
selection is based on the SHIDA/SHIDB pair with the smallest angle between their normal vectors.
a. The old CWELD/CFAST algorithm used the first shell elements found to get a projection and used
SWLDPRM, GMCHK, 1 and 2 to provide some control. These two options have no effect on the new
formulation.
b. Backward connections sometimes occur if the patch is near the boundary of a structure, and there is a "vertical"
flange associated with the patch elements. In this case, SWLDPRM, GMCHK, 3 may be used to prevent
backward projection. (See Figure 3-64a.)

 The minimum angle selected above must be ≤ SWLDPRM GSPROJ if GSPROJ ≥ 0.0
 If the user has not specified BOTH GA and GB and the algorithm cannot find a GSPROJ satisfied projection,
then for SWLDPRM GSMOVE entry, the point GS will be moved in an attempt to satisfy the projection
requirement.
 Reminder: If user has specified both GA and GB and CFAST and the CWELD with options "PARTPAT" and
"ELPAT" are used, then GSMOVE will be ignored and the connector will fail to connect if the user has taken the
default "SWLDPRM, NREDIA, 0" for NREDIA. Failed connectors issue USER FATAL MESSAGE 7635.
 If the GSMOVE specification limit is reached for the CFAST or the CWELD with options "PARTPAT" and
"ELPAT" and SWLDPRM NREDIA ≠ 0, then the diameter of the connector will be reduced by half to compute
new locations of auxiliary grids. If necessary this is repeated until the NREDIA specified value is reached.

Main Index
 224 MSC Nastran Reference Guide
Connector Elements

a. When the NREDIA ≠ 0 is initiated, the GS at its current location is used for GSMOVE ≥ 0.
b. When the NREDIA ≠ 0 is initiated, the GS at its original location is used for the new option GSMOVE < 0.

Outputs
Displacement Output of GA and GB are available for both the CWELD and CFAST Elements
The displacements of the projected grids GA and GB for CWELD elements with GRIDID, ELEMID, ELPAT or
PARTPAT format and CFAST elements are computed to display the relationship between these elements and their
connecting shell elements. The displacement output of GA and GB for CWELD element with GRIDID or ELEMID
format and MSET=OFF are calculated from the constraint equations described in the Basic Theory and Methods section.
If GA or GB is not specified for CFAST elements or CWELD elements with ELPAT or PARTPAT option, the program
will create a grid internally, with the grid ID number starting from OSWPPT+1 (OSWPPT is a parameter specified by
PARAM Bulk Data entry). The user may request a positive PRTSW parameter (1, 2, 11, or 12) in SWLDPRM Bulk Data
entry to view the grid ID of GA or GB in the diagnostic output.
The displacements of GA and GB are requested using the CONNECTOR keyword of the DISPLACEMENT Case
Control command. Outputs
The output is integrated with the displacements of the general grid points.

Basic Theory and Methods

For CWELD elements with GRIDID or ELEMID format, the displacements of the projected grids GA and GB in basic
coordinate are computed by the following equations:

dA =  AI uI
I

dB =  BJ uJ
J

where d A and d B are displacements at GA and GB. A I and B J are constraint matrices. u I and u J are displacements of
the connected shell grids.
For CWELD elements with ELPAT or PARTPAT format, the displacements of the auxiliary points in basic coordinate are
calculated first.

uI =  GIK uK
K

uJ =  GJL uL
L

Main Index
 Chapter 3: Structural Elements 225
Connector Elements

where u I and u J are displacements of the auxiliary points. G IK and G JL are RBE3 type constraint matrices. u K and
u L are displacements of the connected shell grids.
Then the displacements of the projected grids GA and GB in basic coordinate are computed by the same constraint
equations used for GRIDID and ELEMID options.

dA =  AI uI
I

dB =  BJ uJ
J

where d A and d B are displacements at GA and GB. A I and B J are constraint matrices. u I and u J are displacements of
the connected auxiliary points.

Example
This example demonstrates the various displacement output requests and their results for a small model with two CWELD
elements.
The input file follows:
nastran mesh
SOL 101
TIME 60
CEND
TITLE= two elements, identical location for GA, GB, GS
OLOAD= ALL
FORCE = ALL
SUBCASE 1
SUBTITLE= shear the weld
SPC= 1
LOAD= 1
DISP= ALL
SUBCASE 2
SUBTITLE= in plane twist
set 21 = 1002,1003,2011,thru,2014
spc= 1
LOAD= 2
DISP(CONN=ALL)=21
SUBCASE 3
SUBTITLE= upper bending
set 32 = 4
set 33 = 1012,1013,2001,thru,2004
spc= 1
LOAD= 3
DISP(CONN=32)=33
BEGIN BULK
$
$ Grids of lower shell

Main Index
 226 MSC Nastran Reference Guide
Connector Elements

grid, 1001, , 0., 0., 0.

grid, 1002, , 20., 0., 0.
:
grid, 1013, , 20., 10., 5.
grid, 1014, , 0., 10., 5.
$ Grids of upper shell
grid, 2001, , 0., 0., 0.0
grid, 2002, , 20., 0., 0.0
:
grid, 2013, , 20., 10., 6.0
grid, 2014, , 0., 10., 6.0
$ spot weld grid
grid, 3001, , 10.0, 5.0, 0.0
grid, 3011, , 10.0, 5.0, 10.0
$ quad4s
cquad4, 4001, 10, 1001, 1002, 1003, 1004
cquad4, 5001, 10, 2001, 2002, 2003, 2004
cquad4, 4011, 10, 1011, 1012, 1013, 1014
cquad4, 5011, 10, 2011, 2012, 2013, 2014
:
$ property and material
pshell, 10, 10, 1.0, 10
mat1, 10, 2.e+5 , , 0.3, 0.785e-8
$
$ spot welds
$
cweld, 4, 4, 3001, elemid, , , , , +cw4
+cw4, 4001, 5001
cweld, 5, 4, 3011, elemid, , , , , +cw5
+cw5, 4011, 5011
pweld, 4, 10, 5.0
$
enddata

The displacement results are shown as follows:

Main Index
 Chapter 3: Structural Elements 227
Connector Elements

0 SUBCASE 1

D I S P L A C E M E N T V E C T O R

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1001 G 0.0 0.0 0.0 0.0 0.0 0.0
1002 G 4.887498E-03 5.624979E-04 0.0 0.0 0.0 1.406245E-05
1003 G 4.887498E-03 -5.624979E-04 0.0 0.0 0.0 -1.406245E-05
1004 G 0.0 0.0 0.0 0.0 0.0 0.0
1011 G 0.0 0.0 0.0 0.0 0.0 0.0
1012 G 4.887498E-03 5.624979E-04 -7.685000E-01 -6.750000E-03 5.864999E-02 1.406245E-05
1013 G 4.887498E-03 -5.624979E-04 -7.685000E-01 6.750000E-03 5.864999E-02 -1.406245E-05
1014 G 0.0 0.0 0.0 0.0 0.0 0.0
2001 G 6.058339E-04 3.750000E-04 0.0 0.0 0.0 7.733938E-19
2002 G 5.605834E-03 3.750000E-04 0.0 0.0 0.0 1.084202E-1
2003 G 5.605834E-03 -3.750000E-04 0.0 0.0 0.0 1.191827E-18
2004 G 6.058339E-04 -3.750000E-04 0.0 0.0 0.0 7.623363E-19
2011 G 3.907157E-02 3.750000E-04 8.148733E-04 -2.562901E-16 3.850649E-02 2.836059E-17
2012 G 4.407157E-02 3.750000E-04 -7.693148E-01 -5.232793E-15 3.850649E-02 2.851452E-17
2013 G 4.407157E-02 -3.750000E-04 -7.693148E-01 -5.356609E-15 3.850649E-02 2.956961E-17
2014 G 3.907157E-02 -3.750000E-04 8.148733E-04 -2.636780E-16 3.850649E-02 2.803595E-17
3001 G 0.0 0.0 0.0 0.0 0.0 0.0
3011 G 0.0 0.0 0.0 0.0 0.0 0.0
101000001 G 2.443749E-03 8.917563E-18 0.0 0.0 0.0 8.402567E-19
101000002 G 3.105834E-03 8.944668E-18 0.0 0.0 0.0 9.774760E-19
101000003 G 2.443749E-03 2.517788E-16 -3.842500E-01 -2.775558E-15 3.842500E-02 2.303930E-17
101000004 G 4.157158E-02 3.029749E-15 -3.842500E-01 -2.789435E-15 3.850649E-02 2.786400E-17

0 SUBCASE 2
D I S P L A C E M E N T V E C T O R

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1002 G 1.499616E-02 3.324066E-02 0.0 0.0 0.0 2.330633E-03
1003 G -1.499616E-02 3.324066E-02 0.0 0.0 0.0 2.330633E-03
2011 G 6.107051E-03 -8.968769E-03 -1.542189E-16 3.303767E-17 1.479144E-16 1.890160E-03
2012 G 2.110705E-02 4.220943E-02 -3.132043E-15 5.117434E-17 1.498367E-16 3.390160E-03
2013 G -2.110705E-02 4.220943E-02 -2.615099E-15 5.202362E-17 1.383761E-16 3.390160E-03
2014 G -6.107051E-03 -8.968769E-03 1.645747E-16 3.118436E-17 1.398878E-16 1.890160E-03
101000001 G 6.505213E-18 1.662033E-02 0.0 0.0 0.0 1.580825E-03
101000002 G 6.722053E-18 1.662033E-02 0.0 0.0 0.0 2.640160E-03
101000003 G 3.035766E-18 1.662033E-02 -1.434168E-15 4.168935E-17 1.434168E-16 1.580825E-03
101000004 G 1.461505E-16 1.662033E-02 -1.434197E-15 4.178690E-17 1.439374E-16 2.640160E-03

0 SUBCASE 3
D I S P L A C E M E N T V E C T O R

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1012 G 3.460202E-16 9.473882E-16 -1.537000E+00 -1.350000E-02 1.173000E-01 6.294140E-17
1013 G -4.478185E-16 9.592263E-16 -1.537000E+00 1.350000E-02 1.173000E-01 6.130147E-17
2001 G 0.0 0.0 8.021409E-04 -1.965454E-14 7.693148E-02 0.0
2002 G 0.0 0.0 -1.537827E+00 6.397660E-15 7.693148E-02 0.0
2003 G 0.0 0.0 -1.537827E+00 6.758049E-15 7.693148E-02 0.0
2004 G 0.0 0.0 8.021409E-04 -2.021495E-14 7.693148E-02 0.0
101000001 G 0.0 0.0 -7.685000E-01 -6.591949E-15 7.685000E-02 0.0
101000002 G 0.0 0.0 -7.685127E-01 -6.664110E-15 7.693148E-02 0.0

Force output is also produced. For internal processing, the following element types are created depending on the user
specified formats.

Main Index
 228 MSC Nastran Reference Guide
Connector Elements

Element Name Element Number User Specification

CWELD 200 ALIGN
ELEMID, GRIDID with MSET=ON
CWELDC 117 ELEMID, GRIDID with MSET=OFF
CWELDP 118 ELPAT, PARTPAT

The output record of the CWELD forces is identical to the CBAR output record. The output is in the element coordinate
system, see the explanations above how the element coordinate system is constructed. The forces are output at the end
points GA and GB, see Figure 3-67.

Figure 3-67 CWELD Element Forces

CWELD stress computations are made using above forces and standard beam stress type calculations.
Guidelines and Limitations

Finding the connectivity is not straight forward for a complex topology. The user has an option to do a check out run first.
In addition, the user may choose to let the program move GS in case a projection can not be found. Furthermore, the user
may choose to reduce the diameter so that the spot weld can be placed near edges and corners. All parameters for the
connectivity search are lumped on the SWLDPRM Bulk Data entry. It is recommended to use the defaults first. If there
are problems, it is recommended that the check out option be used next and request diagnostic prints.
The CWELD element is not designed to provide accurate stresses at the spot weld location. The stresses and spot weld forces
are mesh dependent. The mesh dependency is more pronounced when the mesh size difference between shell A and B is
large, for example, when the element connects a coarse mesh on shell A with a fine mesh on shell B.
 The old CFAST/CSEAM/CWELD formulation and the new CSEAM/CWELD formulation only support
geometric nonlinear in SOL 400. Connectors are not supported in SOL106 or SOL129 as nonlinear elements.
 The CSEAM/CWELD elements do not currently support nonlinear material in SOL 400.
 The current CFAST/CSEAM/CWELD do not support thermal loading.
The MSET option on the PWELD entry is only available for the formats ELEMID and GRIDID. The default is
MSET=OFF, that means the constraints are eliminated on the element level and there are no external m-set constraints
generated. For the formats ELPAT and PARTPAT, all constraints are eliminated on the element level, there are no external
constraints generated.

Example
A symmetric hat profile with symmetric mesh, loads, and boundary conditions is used to demonstrate the difference
between the ELEMID and the ELPAT formats, see Figure 3-68. For the ELEMID format, it may happen that the
connectivity is defined in an unsymmetric manner, if the spot weld locator point GS is on a corner of an element, see

Main Index
 Chapter 3: Structural Elements 229
Connector Elements

mx

fx

m zB
xe
fz
GB

fy

m yA m yB
Plane 2

fy
GA ze fx Axial force

Plane 1 fy Shear force, plane 1

fz fz Shear force, plane 2

ye
m zA
fx m x Torque

m yA Bending moment end A, plane 2

zb
m yB Bending moment end B, plane 2
yb mx
m zA Bending moment end A, plane 1

xb m zB Bending moment end B, plane 1

Basic

Figure 3-69. Then the format ELEMID produces unsymmetric results. With the ELPAT format, even with a user defined
unsymmetric connectivity, the internal connectivity is symmetric and the results are symmetric, see the following print out.

Main Index
 230 MSC Nastran Reference Guide
Connector Elements

Figure 3-68 Hat Profile

Figure 3-69 Unsymmetric Connectivity for Format ELEMID

The following is the echo of the input for the formats ELEMID and ELPAT:
$ Weld definition with format ELEMID
PWELD 10 1 3.
CWELD 761 10 10001 ELEMID
259 22
CWELD 762 10 10002 ELEMID
299 62
CWELD 763 10 10003 ELEMID
639 402
CWELD 764 10 10004 ELEMID

Main Index
 Chapter 3: Structural Elements 231
Connector Elements

679 442

$ Weld definition with format ELPAT

PWELD 10 1 3.
CWELD 761 10 10001 ELPAT
259 22
CWELD 762 10 10002 ELPAT
299 62
CWELD 763 10 10003 ELPAT
639 402
CWELD 764 10 10004 ELPAT
679 442
Below is the print out for the element forces. The forces are not symmetric for the format ELEMID, see the first part with
the title CWELDC. The forces come out symmetric for the format ELPAT, see the second part with the title CWELDP.

1 HAT SECTION WELDED BEAM, NON-SYM ELEMID OPTION MAY 16, 2003 MSC Nastran 5/15/03 PAGE 16
MZ
0 SUBCASE 3

F O R C E S I N W E L D E L E M E N T S ( C W E L D C )

ELEMENT MOMENT
BEND- END-A BEND- MOMENT END-B - SHEAR - AXIAL
ID PLANE 1 (MZ) PLANE 2 (MY) PLANE 1 (MZ) PLANE 2 (MY) PLANE 1 (FY) PLANE 2 (FZ) FORCE
FX TORQUE MX
761 -1.422846E-02 1.642372E-01 -8.496036E-03 -3.994170E-01 -5.732428E-03 5.636542E-01 -5.534064E-04 4.985318E-02
762 -1.146957E-02 1.216688E-01 -1.872424E-02 -4.326421E-01 7.254668E-03 5.543109E-01 8.758141E-05 4.411347E-02
763 -2.167158E-02 -1.589983E-01 -1.698017E-02 4.059372E-01 -4.691408E-03 -5.649355E-01 5.010240E-04 4.672464E-02
764 -2.205327E-02 -1.208619E-01 -3.047966E-02 4.344850E-01 8.426391E-03 -5.553468E-01 -1.337890E-04 4.600000E-02

1 HAT SECTION WELDED BEAM, NON-SYM ELPAT OPTION MAY 16, 2003 MSC Nastran 5/15/03 PAGE 16
MZ
0 SUBCASE 3

F O R C E S I N W E L D E L E M E N T S ( C W E L D P )

ELEMENT MOMENT
BEND- END-A BEND- MOMENT END-B - SHEAR - AXIAL
ID PLANE 1 (MZ) PLANE 2 (MY) PLANE 1 (MZ) PLANE 2 (MY) PLANE 1 (FY) PLANE 2 (FZ) FORCE
FX TORQUE MX
761 -9.314521E-04 1.277454E-01 4.733397E-03 -4.993157E-01 -5.664849E-03 6.270611E-01 -2.275192E-04 5.223577E-02
762 9.314521E-04 1.277454E-01 -4.733397E-03 -4.993157E-01 5.664849E-03 6.270611E-01 2.275192E-04 5.223577E-02
763 -9.314521E-04 -1.277454E-01 4.733397E-03 4.993157E-01 -5.664849E-03 -6.270611E-01 2.275192E-04 5.223577E-02
764 9.314521E-04 -1.277454E-01 -4.733397E-03 4.993157E-01 5.664849E-03 -6.270611E-01 -2.275192E-04 5.223577E-02

Weld stiffening example

The figures below represent a "box" structure of overall physical dimensions of 100cm x 100cm x 100cm. One model is a
coarse mesh model and the other a fine mesh model. The vertical center section has the same mesh in both models. Both
Models have the same number of CWELD elements (Cluster of blue dots in fringe plots) each of D=5.0 cm at the same
spatial locations.
For the coarse mesh the outer horizontal top and bottom panels are 50 cm x 50 cm. The inner horizontal top and bottom
panels are 16 cm x 16 cm. The center vertical partition mesh is 20 cm x 20 cm. The outer vertical side panels have mesh
50 cm x 50 cm. The inner vertical side panels have a mesh of 20 cm x 20 cm. The mesh of the horizontal offset panel on
the upper right side is 10 cm x 10 cm. For the fine mesh (shown in Figure 3-71c) the inner horizontal top and bottom
panels remain a mesh of 16 cm x 16 cm. The outer horizontal top and bottom panels are meshed using CQUAD8 elements
at 4 cm x 4 cm. The center vertical partition mesh is 20 cm x 20 cm. The outer vertical side panels have mesh 3.4 cm x 3.4
cm. The inner vertical side panels remain at 20 cm x 20 cm. The mesh of the horizontal offset panel on the upper right side

Main Index
 232 MSC Nastran Reference Guide
Connector Elements

is 10 cm x 10 cm. The mesh of the horizontal offset panel on the upper right side is ~2.6 cm x ~2.6 cm. As shown in Figure
3-62a all the elements contained within the green outlined circle are included in a typical patch of this panel.

Figure 3-70 (a) Coarse and fine mesh model (b)shows the detail of the fine mesh of the horizontal offset panel.

Figure 3-71 Fine mesh horizontal offset panel

As the model mesh is refined there is a well-known tendency for welded models to exhibit a loss of stiffness due to model
discretization. This is shown if the next two figures. The coarse mesh model has a maximum displacement of 5.795 cm
where the center vertical partition meets the bottom plate. The fine mesh model has a maximum displacement at that
location of 13.75 cm at the same edge location.

Main Index
 Chapter 3: Structural Elements 233
Connector Elements

Figure 3-72 Coarse mesh model displacement

Figure 3-73 Fine mesh model displacement

For design purposes it is desired that the displacement at that location be brought to within 2% of the coarse mesh model.
To achieve this the SWLDPRM entry is modified to use a combination of two stiffening techniques available with the
CWELD element. This is achieved by modifying the entry SWLDPRM to include the DRATIO and SCLSKIN
parameters. The DRATIO factor of 5.95 increases the radius of influence of the CWELDs Dpatch = 5.95 Dweld = 29.75
cm. For patch to patch connection, the user supplied PWELD value of Dweld=5.0 is still used to compute the "beam area
and moment of area properties.
"SWLDPRM, PRTSW, 2, NREDIA, 4, DRATIO, 5.95, SCLSKIN, 0.255"
The SKIN parameter initiates a process which measures the discretizing of each CWELD relative to its patch and computes
an estimate of the local increase in patch bending resistance which is then applied to the shell element local to the patch.
For this model, the associated SWLDPRM parameter SCLSKIN is set to a value of 0.255 and is used to measure the
discretization of each patch and compute a "WELD" bending moment of inertia to be applied for the specific patch.

Main Index
 234 MSC Nastran Reference Guide
Connector Elements

The figure below shows the displacement pattern for this case. It is observed that the displacement of the fine mesh model
has a maximum displacement of 5.898 cm where the center vertical partition meets the bottom plate which is within the
required 2%.

Figure 3-74 Fine mesh model with DRATIO, 5.95 and SKIN, 1 with SCLSKIN, 0.255

For these models, the von Mises stress, in the area of the vertical partition stayed about the same. In the coarse mesh model
this was the area of maximum stress. In both fine mesh models the area of maximum stress occurred near the support point
of the upper plate furthest from the viewer.

CFAST Element
The CFAST element extends existing weld element capabilities by adding a flexible, user-defined connection between either
two surface patches or two shell elements. The PFAST entry gives the option of specifying longitudinal and rotational
stiffness, a lumped mass, and damping along a defined orientation.
When used with the new CFDIAGP and CFRANDEL parameters, you can elect to randomly remove a percentage of
elements and look at potential failures for the connection. See CFAST Element in the MSC Nastran Linear Static User’s Guide.
The CFAST element has the ability to define the properties of the connector itself. Introducing connector flexibility, mass,
and damping extends the uses of the CWELD. The remark section of the bulk data entry PFAST contains many details about
this element and the bulk data entry SWLDPRM shows how to modify the elements parameters.

CSEAM Element
The CWELD /CFAST elements have been changed so that there is now a consistent formulation between linear and SOL
400 nonlinear analysis. The details are described above in the Section CWELD Element.
The CFAST element which has been extended to include frequency dependency and material nonlinearity. The PFASTT
entry extends to the CFAST element nonlinear load deflection capabilities and limit load fusing ability of the current
CBUSH element in SOL400.

Main Index
 Chapter 3: Structural Elements 235
Connector Elements

PFASTT entry
The CFAST element with property entry PFAST has been extended to have a related PFASTT entry which requires the
same PID as the PFAST. The options specified are similar to the existing options on the PBUSHT entry with the exception
that the CFAST allows only a single GE value and no B values. In nonlinear analysis the PFASTT with “KN” options
support various material nonlinear features similar to the PBUSHT including the “FUSE” option.

Limitations and Guidelines

When using the “FUSE” option in conjunction with geometric nonlinear (LGDISP=1) the fuse stiffness retention “FSR”
may need adjustment as fusing to quickly may cause instant divergence while setting to large a FSR value may cause a
different convergence path than expected.
The CSEAM element connects surface patches. The seam weld is defined by selecting two surface patches by their property
IDs and specifying the width and the thickness of the seam. Extensive error checking and user diagnostics include detection
of missing connections for CSEAMs, defined across corners and cutouts. See CSEAM Element in the MSC Nastran Linear
Static User’s Guide.

Mathematical Model to Construct the CSEAM Auxiliary Points

The four auxiliary vertex points of the cross section at start point GS are constructed by the following equations (see
s s
Figure 3-75), where t 1 and t 2 are tangent vectors of the element coordinate system at start point GS, W is the width of the
s s
seam, and T is the thickness of the seam. For a continuous seam, t 1 and t 2 vectors are adjusted to a common face for the
two consecutive seam elements. Therefore, the four auxiliary grids at the start point S of the ith seam should be coincident
with the four auxiliary grids at the end point E of the (i-1)th seam, if both elements have same width and thickness.
WS TS
x SA1 = x S – ----- t 1 – --- t 2
2 2
WS TS
x SA2 = x S + ----- t 1 – --- t 2
2 2
WS TS
x SB1 = x S – ----- t 1 + --- t 2
2 2
WS TS
x SB2 = x S + ----- t 1 + --- t 2
2 2
S
The tangent vector t 2 is calculated based on the coordinates of the projection points from GS to patch A and patch B,
xSAand xSB. If these two patches are not too close, then

Main Index
 236 MSC Nastran Reference Guide
Connector Elements

S x SB – x SA
t 2 = ---------------------------
s SB – x SA

S
Otherwise, t 2 must be determined from the shell normal vectors of patch A and patch B, i.e. nSA and nSB.

If nSA.nSB > 0, then

S n SB + n SA
t 2 = -----------------------------
n SB + n SA
If nSA.nSB < 0, then

S n SB – n SA
t 2 = ----------------------------
n SB – n SA

S S S
For t 1 , if n SA  t 2  0.9 and n SB  t 2  0.9 then,

S S
S n SA –  t 2  n A t 2
t1 = --------------------------------------------
S S
n SA –  t 2  n A t 2

S e 1  n SA
t 1 = -------------------------
e 1  n SA
where,
x SC – x EC
e 1 = -----------------------------
x SC – x EC

x SA + x SB
x SC = -----------------------
2
x EA + x EB
x EC = ------------------------
2

Figure 3-75 Seam Weld Cross Section at Start Point S

The four auxiliary vertex points of the cross section at end point GE are calculated in the same way as that for the start point
GS. These eight auxiliary points form an auxiliary HEXA element with the following vertex points.

Main Index
 Chapter 3: Structural Elements 237
Connector Elements

s
t2
GSB1 GSB2

T/2

GS s
T/2 t1

GSA1 GSA2

W/2 W/2

1 2 3 4 5 6 7 8
GSA1 GSA2 GSB2 GSB1 GEA1 GEA2 GEB2 GEB1

CINTC Element
The CINTC Element is used to connect dissimilar meshes along the edges of finite element mesh subdomains. These
subdomains have boundaries usually associated with either two-dimensional shell elements or one dimensional beam
elements. This feature is particularly useful for global-local modeling, where a local fine mesh is required in a particular
region due to high stress gradients. A set of MPC (Multipoint Constraint) equations are internally generated with the
interface boundary grids to enforce the compatibility of displacements and rotations across the interface. See CINTC Element
in the MSC Nastran Linear Static User’s Guide.

Applications of Line Interface Element

Line Interface Element

A line interface element is a computational entity which consists of two or more interface boundaries along a virtual
common interface curve. An interface boundary coincides with an edge of a finite element mesh subdomain. Primarily, an
interface boundary is defined by a set of sequentially listed grid points, as specified by Bulk Data entry GMBNDC. On
each boundary, there are at least two grids for a linear interpolation scheme and three for a quadratic one.
There is a dependency (or primary-secondary) relationship among the interface boundaries. This dependency relation is
used to internally generate MPC equations between the independent and dependent degrees of freedom. At each grid point
on a boundary, there are six degrees of freedom, three translational and three rotational. An interface boundary is called the
dependent if all the degrees of freedom related to its grids are put in the m-set. An independent boundary is the one that

Main Index
 238 MSC Nastran Reference Guide
Connector Elements

all the degrees of freedom related to its grids are assigned to the n-set. There is only one independent boundary in an
interface element.
In terms of element characteristics, an interface element is similar to a rigid or constraint element, such as RBE1 and
RSPLINE. Unlike rigid elements, however, its processing is not selectable by Case Control command RIGID. In other
words, only the linear elimination method is applied to this element. The degrees of freedom associated to the grids on a
dependent interface boundary may not be simultaneously assigned dependent by other rigid, constraint or connection
elements, a multipoint constraint or another interface element. On the other hand, the degrees of freedom with the grids
on an independent interface boundary can be made dependent by another element or a multipoint constraint. This
recursive dependent-independent relationship should be carefully examined to avoid potential conflicts when an interface
element interacts with a multipoint constraint, a rigid, constraint or connection element, or another interface element. In
addition, the dependent set (m-set) specified by the dependent interface boundary may not be specified on other entries
that define mutually exclusive set (See “Degree-of-Freedom Sets (p. 1099) in the MSC Nastran Quick Reference Guide.
AUTOMSET can be used with the interface element. It is designed to handle potentially ill-conditioned matrix  R mg 
by reassigning dependent set (m-set) from the existing MPC equations, based on its rank deficiency. It needs further
verifications to see how well the interface element works with AUTOMSET.

Line Interface Boundary

An interface element has two or more boundaries. Each boundary is defined primarily by a set of grid points listed
sequentially from the initial grid through the final one, as specified by Bulk Data entry GMBNDC with Field
ENTITY=”GRID”. All the boundaries in a line interface element must have their initial grids at the starting end and final
grids at the terminal end of the interface.
One of the interface boundaries is specified as the independent by Bulk Data entry CINTC. If there are three or more
boundaries in an interface element, the boundaries are internally grouped in pairs by the dependent-independent
relationship. There are N-1 pairs of boundaries where N is the total number of boundaries with the interface element. MSC
Nastran generates the MPC equations between the dependent and independent grids on the pair of boundaries.
A boundary can be either open, i.e., the initial and final grids are different grids, or closed, i.e., the initial and final grids
are the same grid, as shown in Figure 3-76. For a closed boundary, we have GRIDI=GRIDF in GMBNDC. There is a
requirement for the minimum number of grids on a boundary. There must be at least two grids for the linear and three for
the quadratic interpolation scheme.

GRIDI
GRIDI GRIDF
GRIDF GRIDI=GRID
GRIDI=GRID

Open Boundary Closed Boundary

Closed Boundary
Figure 3-76 Open and Closed Boundaries

Main Index
 Chapter 3: Structural Elements 239
Connector Elements

Both dependent and independent open boundaries may share either initial or final grids or both of them. By assigning the
same end grid to a pair of boundaries, the you have more flexibility in applying the interface element when one interface
element connects or intersects with another interface element. This rule, however, may not apply to the closed boundaries.
A pair of interface boundaries can either be matching or non-matching, as shown in Figure 3-77. The non-matching
boundaries have at least one dangling end grid. The interface line integration is taken along the common arc-length shared
by both boundaries. Redundant dangling grids, as shown in Figure 3-78, are excluded from the processing of interface
element.

M atching Boundaries

Non-m atching Boundaries and Integration Path

Figure 3-77 Matching and Non-matching Boundaries

Redundant Dangling Grids

Figure 3-78 Redundant Dangling Grids on Boundary

For an interface element with closed boundaries, special caution must be taken when selecting the initial grids. Since the
initial grid can be any one on a closed boundary, it is required that the two initial grids on the dependent-independent
closed boundaries be as closed as possible, as shown in Figure 3-79. In other words, a selection of an initial grid being a
redundant grid in terms of the other initial grid is not supported. When one of the boundaries is open, this rule does not
apply.

Main Index
 240 MSC Nastran Reference Guide
Connector Elements

GRIDIA GRIDIA

GRIDIB GRIDIB

Acceptable Selection of Initial Unsupported Selection of Initial

Grids GRIDIA and GRIDIB Grids GRIDIA and GRIDIB
Figure 3-79 Selection of Initial Grids on Closed Boundaries

When the interface consists of curved boundaries, as shown in Figure 3-80, the interface boundary grids should be aligned
along the actual line of structural geometry. In this release we have not implemented a capability that enables MSC Nastran
to check the integrity of interface boundary geometries. The existence of awry grids in the boundary grid list will lead to
some interface malfunction without notice.

Main Index
 Chapter 3: Structural Elements 241
Element Geometry Checks

Actual structure geometry

Finite element model on

boundary A
Finite element model on
boundary B

Figure 3-80 Curved Boundaries

Element Geometry Checks

Mathematical models are usually composed of several different types of finite elements depending upon the simulation
being performed and the type of structure being investigated. For each of these finite element formulations, thorough
evaluations of the geometry presented are performed prior to and during element matrix generation operations. These
geometry tests are typically used to determine whether (1) the element geometry is adequate for finite element matrix
generation at all, and if so, (2) how close the geometry could be to producing poor formulations. Taking a beam element
for example, a fatal geometry condition would be locations of the two end points that result in an element with zero length.
An example of the second type of geometry check for a beam would be an excessive ratio of the length of the beam with
offset vectors to the length of the beam without them. In the first case, the geometry check is fatal because the finite element
cannot be formulated from the given geometry and the simulation cannot be done. In the second case, the geometry check
is informational and the simulation analysis can proceed. It is the user’s responsibility to inspect these informational
messages and determine if the results are acceptable.
Several different basic geometry tests are performed; for example, duplicate grid point location tests, mid-side node location
tests, warped element tests, and aspect ratio tests. One or more of these types of tests is performed using the geometry
supplied for each finite element. Obviously, not all tests apply to every element type. Each one of the geometry tests is
accompanied by a message that can generate up to four lines of informational output for an element if the test tolerance is
violated. For models that employ a small number of elements, this amount of output is not overwhelming. However, the
size and complexity of mathematical simulation models have continued to grow over time. It is no longer surprising to see
models that generate hundreds of thousands of lines of output from these tests alone. In an effort to reduce the amount of
output generated by geometry checks, an alternative output for reporting the results of the element geometry tests is
available. The results for all tests for a single element are reported on a single line together with a visual indication of which

Main Index
 242 MSC Nastran Reference Guide
Element Geometry Checks

tests have exceeded their tolerances and the severity of the failure. Default values for all test tolerances are defined such that
geometry tests result in the same type of pass/fail decision as previous releases. In a few cases, additional tests are performed.
At the same time, the GEOMCHECK Executive statement is available to give users more control over the test tolerances,
their severity levels and the number of messages output.

GEOMCHECK Statement
The GEOMCHECK statement provides users with a way to override the default values supplied for all tests. (Refer to the
Executive Statement, GEOMCHECK (p. 136) in the MSC Nastran Quick Reference Guide.) Except for a very few isolated
instances, bad geometry that does not allow the formulation of finite element matrix data is always fatal and is not under
user control. In these cases, the informational message formats have not been modified. Every test that fails will produce a
message for every element. For the cases where geometry does not prevent finite element matrix generation, additional tests
that evaluate various geometry parameters are performed and the results reported to the user. For these tests, the user can
control both the test tolerance and the severity level of test failure. Using the CBAR element as an example, there is a test
performed that evaluates the ratio of the CBAR’s length with offset effects to the length without offset effects. The tolerance
for this test is 15%, meaning that if the length with offset is different from the length without the offset by more than fifteen
percent, an informational message is issued and element processing continues. The user can change the tolerance if desired.
The severity of the test failure can also be changed from informational to fatal if desired. In that case, any CBAR element
that exceeds the offset ratio test tolerance will cause the job to stop after element matrix generation has been attempted for
all elements.
The GEOMCHECK statement can only be used to modify geometry test activities for the CQUAD4, CQUADR, CTRIA3,
CTRIAR, CBAR, CBEAM, CHEXA, CPENTA, and CTETRA finite elements at this time. As stated previously, only
informational tests can be affected using the GEOMCHECK statement. Results of geometry generated by MSC Nastran
that prevent the generation of a finite element matrix cannot be modified using the GEOMCHECK statement.
Keywords are available that allow control over the number of messages to be generated (MSGLIMIT) as well as the severity
of the message (MSGTYPE) and its associated effect on the job. Only a severity of FATAL will cause the job to abort after
the element matrix generation module has executed. In this case, the identifier associated with failed tests is FAIL and the
run will also terminate.

Geometry Test Descriptions

Several different types of geometry tests are performed. For each finite element type, the geometry must be tested to ensure
that it is adequate to allow generation of element matrices. No attempt is made here to describe those tests. Rather, tests
that optionally evaluate certain characteristics of the geometry are described. These tests fall into several categories
depending upon the element type. Thus, not all tests are applicable to every element type. The discussion that follows
summarizes the optional geometry tests that are applicable to various element types.

CBAR and CBEAM

The only optional test available for these element types is the offset test. The length of the element offset is compared to
the original length of the element. If the ratio of these lengths is greater than the tolerance, an informational message is
issued identifying the element and its length with and without the effects of offset.

Main Index
 Chapter 3: Structural Elements 243
Element Geometry Checks

CQUAD4 and CQUADR

Four optional tests are performed on the quadrilateral shaped elements. These tests are:
1. Interior Angle test: This test evaluates the interior angles measured at each of the four corner grid points. If any one
of the four angles exceeds minimum or maximum tolerance levels, an informational message is produced.
2. Taper test: Taper ratio for the quadrilateral element is defined to be the ratio of the area of the triangle formed at
each corner grid point to one half the area of the quadrilateral. The largest of the four ratios is compared against the
tolerance value. It may be noted that as the ratio approaches 1.0, the shape approaches a rectangle.
3. Skew test: This test evaluates the distortion or “parallelogram shape effect” by measuring the angle between lines
that join the midpoints of opposite sides of the element. If the angle exceeds the tolerance, an informational message
is produced.
4. Warp test: This test evaluates how far out of plane the four corner grid points are by measuring the distance of each
point from a “mean” plane passing through the locations of the four points. The corner points are alternately H
units above and H units below this mean plane. If the lengths of the diagonals of the element are denoted by D1
and D2, the warping coefficient is obtained from the equation WC = H / 2(D1+D2). If this value exceeds the
tolerance, an informational message is produced.

CTRIA3 and CTRIAR

Only one optional test is performed on the triangular shaped planar elements. It is the same as the interior angle test
discussed previously under the CQUAD4 and CQUADR test description.

CHEXA, CPENTA and CTETRA

Five optional tests are performed on the solid element family. Not all tests are applicable to every element type. These tests
are:
1. Aspect Ratio test: This test evaluates the ratio of the longest length (edge or height) to the shortest length (edge or
height) encountered in the element. If the ratio exceeds the tolerance, an informational message is produced.
2. Edge Point Length Ratio test: This test evaluates the location of the “mid-side” nodes if any are present. The node
should be located on the line connecting the two adjacent corner grid points at approximately the mid-way point.
This determination is made in two ways. First, the distance of the node from the two corner nodes is measured and
if the ratio of the distance from one node to the other exceeds the tolerance, an informational message is issued.
Next, the angles between the lines connecting the mid-side node to its two adjacent corner nodes, and the line
connecting the corner nodes themselves, is evaluated. If either of the angles exceeds 30 degrees, an informational
message is issued.
3. Integration Point Jacobian Determinant test: This test evaluates the determinant of the Jacobian at each integration
point. If it is zero, or changes sign from integration point to integration point, an informational message is issued.
Note that detection of a zero value should always be fatal because element matrices calculated from such geometry
rarely produce satisfactory analysis results.

Main Index
 244 MSC Nastran Reference Guide
Element Geometry Checks

4. Warped Face test: This test is applicable to CHEXA and CPENTA elements. It evaluates the warping coefficient of
the quadrilateral faces of such elements in a manner similar to that discussed previously under the CQUAD4 element
warping test description. If the warping coefficient exceeds the tolerance, an informational message is issued.
5. Grid Point Jacobian Determinant test: This test is applicable to CTETRA elements only. The test is the same as that
done for the integration point test discussed previously except it uses the location of the corner grid points to
perform the test. If any determinant is zero or changes sign, an informational message is produced.
Sample Output:

The following output from a small example provides a sample of the new message formats that are produced for geometry
tests that can be controlled using the GEOMCHECK statement. For this output, most of the default test tolerances were
modified so that the effect could be observed in the tolerance information lines of the messages that were produced. Also
note that a small table is generated at the end of all messages that summarizes the number of tests that actually exceeded the
tolerance value for each element type as well as a list of the elements that produced the worst violations. The summary table
itself can be produced by using the keyword SUMMARY on the GEOMCHECK statement.

Main Index
 Chapter 3: Structural Elements 245
Element Geometry Checks

*** USER INFORMATION MESSAGE 7555 (GMTSTD)

FINITE ELEMENT GEOMETRY CHECK RESULTS EXCEED TOLERANCE LEVELS FOR THE FOLLOWING ELEMENTS.
User Action: Use the GEOMCHECK (Executive Control Statement) keyword=value to change tolerance values if desired.
A MINIMUM OF 100 OFFSET LENGTH RATIO TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.

TOLERANCE LIMIT IS: BAR_/BEAM_OFF(SET) = .15 (xxxx = LIMIT VIOLATED)

ELEMENT TYPE ID LENGTH W/O OFFSET LENGTH W/ OFFSET OFFSET LENGTH RATIO
BAR 7101 1.00000E+00 3.66367E+00 2.66 xxxx

*** USER INFORMATION MESSAGE 7555 (GMTSTD)

FINITE ELEMENT GEOMETRY CHECK RESULTS EXCEED TOLERANCE LEVELS FOR THE FOLLOWING ELEMENTS.
A MINIMUM OF 4 OFFSET LENGTH RATIO TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY WARN.

TOLERANCE LIMIT IS: BAR_/BEAM_OFF(SET) = -.02 (WARN = LIMIT VIOLATED)

ELEMENT TYPE ID LENGTH W/O OFFSET LENGTH W/ OFFSET OFFSET LENGTH RATIO
BEAM 8101 1.00000E+00 1.10372E+01 10.04 WARN
BEAM 8102 1.00000E+00 1.00000E+00 .00 WARN

*** USER INFORMATION MESSAGE 7555 (EHEXGD)

FINITE ELEMENT GEOMETRY CHECK RESULTS EXCEED TOLERANCE LEVELS FOR THE FOLLOWING ELEMENTS.
A MINIMUM OF 100 EDGE LENGTH ASPECT RATIO (AR) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 EDGE NODE POINT LENGTH RATIO (ER)TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
EDGE ANGLE VALUES ARE INDICATED BY AN * AFTER THE VALUE. THE TOLERANCE LIMIT IS 30 DEGREES.
ALL NEG/ZERO DET(JACOBIAN) VALUE (DJ)TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY FAIL.
A MINIMUM OF 100 WARPED FACE COEFFICIENT TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.

TOLERANCE LIMITS ARE: HEX_AR = 300.00, HEX_EPLR = .40, HEX_DETJ = .01, HEX_WARP = .65 (xxxx = LIMIT VIOLATED)
LONGEST SHORTEST ASPECT EDGE POINT MIN. JACOBIAN FACE WARP
ELEMENT TYPE ID EDGE EDGE RATIO LENGTH RATIO DETERMINANT COEFFICIENT
HEXA 8602 4.00 1.00 4.00 N/A- ONLY 8 NODE .92 .52 xxxx
HEXA 8620 4.00 1.00 4.00 .60xxxx .50 1.00

*** USER INFORMATION MESSAGE 7555 (GMTSTD)

FINITE ELEMENT GEOMETRY CHECK RESULTS EXCEED TOLERANCE LEVELS FOR THE FOLLOWING ELEMENTS.
A MINIMUM OF 100 SKEW ANGLE (SA) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 MIN INT. ANGLE (IA) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 MAX INT. ANGLE (IA) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 WARPING FACTOR (WF) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 TAPER RATIO (TR) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.

TOLERANCE LIMITS ARE: SA = 30.00, IA(MIN) = 30.00, IA(MAX) = 150.00, WF = .05, TR = .50 (xxxx = LIMIT VIOLATED)
ELEMENT TYPE ID SKEW ANGLE MIN INT. ANGLE MAX INT. ANGLE WARPING FACTOR TAPER RATIO
QUAD4 105 29.74 xxxx 29.05 xxxx 150.95 xxxx .00 .05
QUAD4 106 28.39 xxxx 27.76 xxxx 152.24 xxxx .00 .05
QUAD4 107 27.15 xxxx 26.57 xxxx 153.43 xxxx .00 .05
QUAD4 108 26.00 xxxx 25.46 xxxx 154.54 xxxx .00 .05
QUAD4 109 24.94 xxxx 24.44 xxxx 155.56 xxxx .00 .05
QUAD4 2105 29.74 xxxx 29.05 xxxx 150.95 xxxx .00 .05
QUAD4 2106 28.39 xxxx 27.76 xxxx 152.24 xxxx .00 .05
QUAD4 2107 27.15 xxxx 26.57 xxxx 153.43 xxxx .00 .05
QUAD4 2108 26.00 xxxx 25.46 xxxx 154.54 xxxx .00 .05
QUAD4 2109 24.94 xxxx 24.44 xxxx 155.56 xxxx .00 .05

*** USER INFORMATION MESSAGE 7555 (GMTSTD)

FINITE ELEMENT GEOMETRY CHECK RESULTS EXCEED TOLERANCE LEVELS FOR THE FOLLOWING ELEMENTS.
A MINIMUM OF 100 SKEW ANGLE (SA) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 MIN INT. ANGLE (IA) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 MAX INT. ANGLE (IA) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 WARPING FACTOR (WF) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.
A MINIMUM OF 100 TAPER RATIO (TR) TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY xxxx.

Main Index
 246 MSC Nastran Reference Guide
Element Geometry Checks

TOLERANCE LIMITS ARE: SA = 30.00, IA(MIN) = 30.00, IA(MAX) = 150.00, WF = .05, TR = .50 (xxxx = LIMIT VIOLATED)
ELEMENT TYPE ID SKEW ANGLE MIN INT. ANGLE MAX INT. ANGLE WARPING FACTOR TAPER RATIO
QUADR 1105 29.74 xxxx 29.05 xxxx 150.95 xxxx .00 .05
QUADR 1106 28.39 xxxx 27.76 xxxx 152.24 xxxx .00 .05
QUADR 1107 27.15 xxxx 26.57 xxxx 153.43 xxxx .00 .05
QUADR 1108 26.00 xxxx 25.46 xxxx 154.54 xxxx .00 .05
QUADR 1109 24.94 xxxx 24.44 xxxx 155.56 xxxx .00 .05

*** USER INFORMATION MESSAGE 7555 (ETETGD)

FINITE ELEMENT GEOMETRY CHECK RESULTS EXCEED TOLERANCE LEVELS FOR THE FOLLOWING ELEMENTS.
A MINIMUM OF 4 EDGE LENGTH TO HEIGHT RATIO (AR)TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY WARN.
A MINIMUM OF 4 EDGE NODE POINT LENGTH RATIO (ER)TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY WARN.
EDGE ANGLE VALUES ARE INDICATED BY AN * AFTER THE VALUE. THE TOLERANCE LIMIT IS 30 DEGREES.
A MINIMUM OF 4 NEG/ZERO DET(JACOBIAN) VALUE (DJ)TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY WARN.
A MINIMUM OF 4 NEG DET(JAC) AT VERTEX NODE (DG)TOLERANCE LIMIT VIOLATIONS WILL BE IDENTIFIED AND INDICATED BY WARN.

TOLERANCE LIMITS ARE: TET_AR = 351.00, TET_EPLR = .45, TET_DETJ = -.01, TET_DETG = -.02 (xxxx = LIMIT VIOLATED)
LONGEST SHORTEST ASPECT EDGE POINT MIN. JACOBIAN MIN. DET(JAC)
ELEMENT TYPE ID EDGE HEIGHT RATIO LENGTH RATIO DETERMINANT AT VERTEX
TETRA 6601 1.41 .58 2.45 .25 WARN .14 -.20 WARN

E L E M E N T G E O M E T R Y T E S T R E S U L T S S U M M A R Y
TOTAL NUMBER OF TIMES TOLERANCES WERE EXCEEDED
ASPECT/ MINIMUM MAXIMUM SURFACE/FACE EDGE POINT JACOBIAN
ELEMENT TYPE SKEW ANGLE TAPER RATIO INTER. ANGLE INTER. ANGLE WARP FACTOR OFFSET RATIO LENGTH RATIO DETERMINANT
BAR N/A N/A N/A N/A N/A 0 N/A N/A
BEAM N/A N/A N/A N/A N/A 0 N/A N/A
HEXA N/A 0 N/A N/A 1 N/A 1 0
PENTA N/A 0 N/A N/A 0 N/A 1 0
QUAD4 10 0 10 10 0 N/A N/A N/A
QUADR 5 0 5 5 0 N/A N/A N/A
TETRA N/A 0 N/A N/A N/A N/A 1 1
TRIA3 0 N/A N/A 0 N/A N/A N/A N/A
TRIAR 0 N/A N/A 0 N/A N/A N/A N/A

N/A IN THE ABOVE TABLE INDICATES TESTS THAT ARE NOT APPLICABLE TO THE ELEMENT TYPE AND WERE NOT PERFORMED.
FOR ALL ELEMENTS WHERE GEOMETRY TEST RESULTS HAVE EXCEEDED TOLERANCES,

HEXA ELEMENT ID 8620 PRODUCED SMALLEST EDGE COS(ANGLE) OF .60 (TOLERANCE = .87).
HEXA ELEMENT ID 8602 PRODUCED SMALLEST FACE WARP FACTOR OF .52 (TOLERANCE = .65).
PENTA ELEMENT ID 3915 PRODUCED SMALLEST EDGE COS(ANGLE) OF .78 (TOLERANCE = .87).
QUAD4 ELEMENT ID 109 PRODUCED SMALLEST SKEW ANGLE OF 24.94 (TOLERANCE = 30.00).
QUAD4 ELEMENT ID 109 PRODUCED SMALLEST INTERIOR ANGLE OF 24.44 (TOLERANCE = 30.00).
QUAD4 ELEMENT ID 109 PRODUCED LARGEST INTERIOR ANGLE OF 155.56 (TOLERANCE = 150.00).
QUADR ELEMENT ID 1109 PRODUCED SMALLEST SKEW ANGLE OF 24.94 (TOLERANCE = 30.00).
QUADR ELEMENT ID 1109 PRODUCED SMALLEST INTERIOR ANGLE OF 24.44 (TOLERANCE = 30.00).
QUADR ELEMENT ID 1109 PRODUCED LARGEST INTERIOR ANGLE OF 155.56 (TOLERANCE = 150.00).
TETRA ELEMENT ID 6601 PRODUCED LARGEST EDGE POINT LR OF .25 (TOLERANCE = .45).
TETRA ELEMENT ID 6601 PRODUCED SMALLEST VERTEX DET(JAC) -UWM 6828- OF -.20 (TOLERANCE = -.02).

Introduction
A long-standing key feature of MSC Nastran is the ability of the PSHELL(, PCOMP), PBAR(, PBARL), PBEAM(,
PBEAML, PBCOMP), PROD(, CONROD), PBEND, PSHEAR, PTUBE, PCONEAX, and PRAC2D property entries
to include nonstructural mass. The five Bulk Data entries -- NSM, NSM1, NSML, NSML1, and NSMADD -- distribute
nonstructural mass by element lists or specific property lists associated with the above fourteen property entries.

Theory
NSML and NSML1 Entries

Main Index
 Chapter 3: Structural Elements 247
Element Geometry Checks

The NSML and NSML1 entries compute a nonstructural mass coefficient value for “area” elements identified by an element
list or property list or both by the relationship:
Lumped_non_structural_mass_value
NSM_value = ----------------------------------------------------------------------------------------
 elements AREA

The NSML and NSML1 entries compute a nonstructural mass coefficient value for “line” elements identified by an element
list or property list or both by the relationship:
Lumped_non_structural_mass_value
NSM_value = ----------------------------------------------------------------------------------------
 elements LENGTH

In the above two expressions, AREA corresponds to the area of each individual area element (CQUAD4 for example) and
LENGTH corresponds to the length of each individual length element (CBAR for example).
The NSML and NSML1 entries are then converted internally to NSM and NSM1 entries.

Inputs
Case Control Commands

NSM Case Control command allows for the selection of different NSM sets for residual and superelements.
Bulk Data Entries

NSM and its alternate form NSM1 allows the user to allocate an NSM_value to selected sets of elements.
NSML and its alternate form NSML1 allows the user to allocate and smear a lumped nonstructural mass value to selected
sets of elements.
NSMADD allows the user to form combinations of NSM, NSM1, NSML, and NSML1 sets and sum their results to
selected sets of elements.

Guidelines and Limitations

The following requirements are associated with the new capability:
 The NSML and NSML1 entries cannot mix “area” and “line” elements on the same entry.
 Area elements are: CQUAD4, CQUAD8, CQUADR, CTRIA3, CTRIA6, CTRIAR, CSHEAR, and CRAC2D. (The
CCONEAX is a stand alone element.)
 Line elements are: CBAR, CBEAM, CBEND, CTUBE, CROD, and CONROD.
 The CCONEAX is excluded from the NSML and NSML1 entries.

Example
The following example partial input demonstrates the form of the new entries and the expected result:

Main Index
 248 MSC Nastran Reference Guide
Element Geometry Checks

Partial Input File

SOL108 $
CEND
DISPL = ALL
.
.
.
ELSUM = ALL
NSM = 1222
BEGIN BULK
$
.
.
.
$
$ 2 3 4 5 6 7 8 9 0
NSML 9 PBEAM 211 .026 311 .028 411 .030
NSML 9 PBEND 21 .4869 31 .4869
NSML 10 PRAC2D 1 .132
NSML1 10 PBCOMP .126 21 THRU 41
NSML1 11 PBEAML .138 212 312 412
NSML 11 PROD 12 .17
NSML 12 CONROD 12 .19
NSML 12 PSHELL 51 4.5
NSML1 13 PSHELL 132. ALL
NSML1 13 PSHELL 107.8 50 THRU 52
NSML1 14 PSHELL 7.5 57
NSML1 14 ELEMENT .594 20 30 40 201 202
301 302 401 402
NSML1 15 PSHEAR .264 25 THRU 45 BY 10
NSML 15 PBAR 22 .22
NSML1 15 PBARL .52 ALL
NSML 16 ELEMENT 101 .077
NSML 16 ELEMENT 27 1.021 13 .67
NSML 17 ELEMENT 14 .41
NSML 17 PTUBE 14 .43
NSML1 18 ELEMENT 16.2 41 81234
NSML1 18 PSHEAR 1.476 ALL
NSML 18 PSHEAR 1 1.476
NSML1 19 ELEMENT .112 201 402
NSML1 19 ELEMENT 4.2 3123 6134 134
$
NSMADD 1222 9 11 12 14 13 15
16 1718 19 2010 12
$
$ FINAL PROPERTY
$ DISTRIBUTED VALUE
$ ELM NSM ID PID PROPERTY NSM
$ ---- --- -- --- ------- ----
$ CBAR .061 25 22 - PBAR .013
$ CBAR .039 26 26 - PBARL .013
$ CBEAM .065 20 21 - PBCOMP .011
$ .065
$ CBEAM .065 30 31 - PBCOMP .011
$ .065
$ CBEAM .065 40 41 - PBCOMP .011

Main Index
 Chapter 3: Structural Elements 249
Element Geometry Checks

$ .065
$ CBEAM .085 201 211 - PBEAM .011,.012,.014
$ .086 is1
$ .088 is2 all others zero
$ .088 end B
$ CBEAM .058 301 311 - PBEAM .011,.012,.014
$ .059 is1
$ .061 is2 all others zero
$ .061 end B
$ CBEAM .059 401 411 - PBEAM .011,.012,.014
$ .060 is1
$ .062 is2 all others zero
$ .062 end B
$ CBEAM .0671 202 212 - PBEAML .0111,.0112,.0113,.0114,.0115
$ .0672 is1
$ .0673 is2
$ .0674 is3 all others zero
$ .0675 is4
$ .0675 end B
$ CBEAM .0671 302 312 - PBEAML .0111,.0112,.0113,.0114,.0115
$ .0672 is1
$ .0673 is2
$ .0674 is3 all others zero
$ .0675 is4
$ .0675 end B
$ CBEAM .0951 402 412 - PBEAML .0111,.0112,.0113,.0114,.0115
$ .0952 is1
$ .0953 is2
$ .0954 is3 all others zero
$ .0955 is4
$ .0955 end B
$ CBEND .109 27 21 - PBEND .013
$ CBEND .044 37 31 - PBEND .013
$ CBEND .013 47 41 - PBEND .013
$ CONROD .032 12 1 - CONROD .013
$ CQUAD4 .257 41 57 - PCOMP .013
$ CQUAD4 .178 41234 50 - PSHELL .013
$ CQUAD4 .178 4123456 52 - PSHELL .013
$ CQUAD4 .193 4123456751 - PSHELL .013
$ CQUAD8 .178 81 50 - PSHELL .013
$ CQUAD8 .259 81234 50 - PSHELL .013
$ CQUAD8 .178 8123456 52 - PSHELL .013
$ CQUAD8 .193 8123456751 - PSHELL .013
$ CQUADR .178 1 50 - PSHELL .013
$ CQUADR .178 1234 50 - PSHELL .013
$ CQUADR .178 123456 52 - PSHELL .013
$ CQUADR .193 1234567 51 - PSHELL .013
$ CROD .097 13 12 - PROD .013
$ CSHEAR .158 25 1 - PSHEAR .013
$ CSHEAR .158 35 1 - PSHEAR .013
$ CSHEAR .158 45 1 - PSHEAR .013
$ CTRIA3 .206 3123 50 - PSHELL .013
$ CTRIA3 .178 3134 50 - PSHELL .013
$ CTRIA6 .178 6123 50 - PSHELL .013
$ CTRIA6 .206 6134 50 - PSHELL .013
$ CTRIAR .178 123 50 - PSHELL .013
$ CTRIAR .206 134 50 - PSHELL .013

Main Index
 250 MSC Nastran Reference Guide
Element Geometry Checks

$ CTUBE .097 14 14 - PTUBE .013

$
$
.
.
.
$
ENDDATA

Main Index
 Chapter 3: Structural Elements 251
Element Geometry Checks

Typical Element mass summary is given as:

E L E M E N T P R O P E R T Y S U M M A R Y
ELEMENT TYPE = BAR
ID MID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
25 1 1.000000E+01 2.000000E-02 2.000000E-01 0.000000E+00 6.100000E-01 6.100000E-01 6.100000E-01
26 1 1.000000E+01 3.141593E-02 3.141593E-01 0.000000E+00 3.900000E-01 3.900000E-01 3.900000E-01
SUBTOTAL MASS = 0.000000E+00 1.000000E+00 1.000000E+00 1.000000E+00
ELEMENT TYPE = BEAM
ID MID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
20 1 2.000000E+00 1.200800E-02 2.401600E-02 0.000000E+00 1.300000E-01 1.300000E-01 1.300000E-01
30 1 2.000000E+00 1.200800E-02 2.401600E-02 0.000000E+00 1.300000E-01 1.300000E-01 1.300000E-01
40 1 2.000000E+00 1.200800E-02 2.401600E-02 0.000000E+00 1.300000E-01 1.300000E-01 1.300000E-01
201 1 2.000000E+00 2.000000E-02 4.000000E-02 0.000000E+00 1.700000E-01 1.700000E-01 1.700000E-01
202 1 2.000000E+00 1.256637E+00 2.513275E+00 0.000000E+00 1.342000E-01 1.342000E-01 1.342000E-01
301 1 2.000000E+00 2.000000E-02 4.000000E-02 0.000000E+00 1.160000E-01 1.160000E-01 1.160000E-01
302 1 2.000000E+00 1.256637E+00 2.513275E+00 0.000000E+00 1.342000E-01 1.342000E-01 1.342000E-01
401 1 2.000000E+00 2.000000E-02 4.000000E-02 0.000000E+00 1.180000E-01 1.180000E-01 1.180000E-01
402 1 2.000000E+00 1.256637E+00 2.513275E+00 0.000000E+00 1.902000E-01 1.902000E-01 1.902000E-01
SUBTOTAL MASS = 0.000000E+00 1.252600E+00 1.252600E+00 1.252600E+00
ELEMENT TYPE = BEND
ID MID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
27 1 1.570796E+01 1.000000E+00 1.570796E+01 0.000000E+00 1.712104E+00 1.712104E+00 1.712104E+00
37 1 1.570796E+01 1.000000E+00 1.570796E+01 0.000000E+00 6.911035E-01 6.911035E-01 6.911035E-01
47 1 1.570796E+01 1.000000E+00 1.570796E+01 0.000000E+00 2.042035E-01 2.042035E-01 2.042035E-01
SUBTOTAL MASS = 0.000000E+00 2.607411E+00 2.607411E+00 2.607411E+00
ELEMENT TYPE = HEXA
ID MID VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
1001 1003 8.000000E+03 6.400000E+01 0.000000E+00 6.400000E+01 6.400000E+01
SUBTOTAL MASS = 6.400000E+01 0.000000E+00 6.400000E+01 6.400000E+01
ELEMENT TYPE = CONROD
ID MID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
12 51 1.000000E+01 1.100000E+00 1.100000E+01 6.600000E-01 3.200000E-01 9.800000E-01 9.800000E-01
SUBTOTAL MASS = 6.600000E-01 3.200000E-01 9.800000E-01 9.800000E-01
ELEMENT TYPE = QUAD4
ID MID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
41 100000057 2.000000E-01 1.000000E+02 2.000000E+01 5.040089E+01 2.570000E+01 7.610089E+01 7.610089E+01
41234 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.780000E+01 2.380000E+01 2.380000E+01
4123456 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.780000E+01 2.380000E+01 2.380000E+01
41234567 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.930000E+01 2.530000E+01 2.530000E+01
SUBTOTAL MASS = 6.840089E+01 8.060000E+01 1.490009E+02 1.490009E+02
ELEMENT TYPE = QUAD8
ID MID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
81 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.780000E+01 2.380000E+01 2.380000E+01
81234 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 2.590000E+01 3.190000E+01 3.190000E+01
8123456 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.780000E+01 2.380000E+01 2.380000E+01
81234567 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.930000E+01 2.530000E+01 2.530000E+01
SUBTOTAL MASS = 2.400000E+01 8.080000E+01 1.048000E+02 1.048000E+02
ELEMENT TYPE = QUADR
ID MID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
1 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.780000E+01 2.380000E+01 2.380000E+01
1234 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.780000E+01 2.380000E+01 2.380000E+01
123456 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.780000E+01 2.380000E+01 2.380000E+01
1234567 51 1.000000E+00 1.000000E+02 1.000000E+02 6.000000E+00 1.930000E+01 2.530000E+01 2.530000E+01
SUBTOTAL MASS = 2.400000E+01 7.270000E+01 9.670000E+01 9.670000E+01
ELEMENT TYPE = ROD
ID MID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
13 51 1.000000E+01 1.100000E+00 1.100000E+01 6.600000E-01 9.700000E-01 1.630000E+00 1.630000E+00
SUBTOTAL MASS = 6.600000E-01 9.700000E-01 1.630000E+00 1.630000E+00
ELEMENT TYPE = SHEAR
ID MID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
25 1 1.000000E+00 4.000000E+00 4.000000E+00 0.000000E+00 3.860000E-01 3.860000E-01 3.860000E-01
35 1 1.000000E+00 4.000000E+00 4.000000E+00 0.000000E+00 3.860000E-01 3.860000E-01 3.860000E-01
45 1 1.000000E+00 4.000000E+00 4.000000E+00 0.000000E+00 3.860000E-01 3.860000E-01 3.860000E-01
125 1 1.000000E+00 4.000000E+00 4.000000E+00 0.000000E+00 7.900000E-01 7.900000E-01 7.900000E-01
135 1 1.000000E+00 4.000000E+00 4.000000E+00 0.000000E+00 7.900000E-01 7.900000E-01 7.900000E-01
145 1 1.000000E+00 4.000000E+00 4.000000E+00 0.000000E+00 7.900000E-01 7.900000E-01 7.900000E-01
SUBTOTAL MASS = 0.000000E+00 3.528000E+00 3.528000E+00 3.528000E+00
ELEMENT TYPE = TRIA3
ID MID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
3123 51 1.000000E+00 5.000000E+01 5.000000E+01 3.000000E+00 1.030000E+01 1.330000E+01 1.330000E+01
3134 51 1.000000E+00 5.000000E+01 5.000000E+01 3.000000E+00 8.900000E+00 1.190000E+01 1.190000E+01
SUBTOTAL MASS = 6.000000E+00 1.920000E+01 2.520000E+01 2.520000E+01
ELEMENT TYPE = TRIA6
ID MID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
6123 51 1.000000E+00 5.000000E+01 5.000000E+01 3.000000E+00 8.900000E+00 1.190000E+01 1.190000E+01
6134 51 1.000000E+00 5.000000E+01 5.000000E+01 3.000000E+00 1.030000E+01 1.330000E+01 1.330000E+01
SUBTOTAL MASS = 6.000000E+00 1.920000E+01 2.520000E+01 2.520000E+01
ELEMENT TYPE = TRIAR
ID MID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
123 51 1.000000E+00 5.000000E+01 5.000000E+01 3.000000E+00 8.900000E+00 1.190000E+01 1.190000E+01
134 51 1.000000E+00 5.000000E+01 5.000000E+01 3.000000E+00 1.030000E+01 1.330000E+01 1.330000E+01
SUBTOTAL MASS = 6.000000E+00 1.920000E+01 2.520000E+01 2.520000E+01
ELEMENT TYPE = TUBE
ID MID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
14 51 1.000000E+01 3.141593E-01 3.141593E+00 1.884956E-01 9.700000E-01 1.158496E+00 1.158496E+00
SUBTOTAL MASS = 1.884956E-01 9.700000E-01 1.158496E+00 1.158496E+00
TOTAL MASS = 1.999094E+02 3.023480E+02 5.022575E+02 5.022575E+02

Main Index
 252 MSC Nastran Reference Guide
ELSUM Case Control Command

ELSUM Case Control Command

Introduction
The ELSUM Case Control command controls the generation of a printed table of informational properties for the various
element types present in the input data file. The information produced includes element measures (e.g. length, thickness,
area, volume) and mass property data. The information is grouped according to the finite element type. Several new options
have been added to the ELSUM command giving users more control over the amount of output produced and another
view of the data grouped by element property.

Method and Theory

For each element present in the model, physical metrics appropriate to the element type are computed. These metrics may
include length, thickness, area, and volume. Mass properties are computed for supported element types. For the CONM1
element, the mass is taken to be the average of the M11, M22 and M33 values on the CONM1 Bulk Data entry. In
addition, references to material properties and element properties are gathered for each element. For bar, beam and shell
elements, an attempt is made to determine whether the property reference is to a basic property entry, or has been derived
from one of the available alternate input entries (PBARL, PBEAML, PBCOMP, PCOMP). If it is determined that the
property is derived, a character (L for PBARL or PBEAML, P for PBCOMP or PCOMP) is appended to the property
identification number on output. For the case of the element type grouping, each element type is summarized in turn. For
each element type, information for each element is displayed. Subtotals are generated for mass property information. For
the case of the property type grouping, each property type referenced by an element type is summarized in turn. All of the
property ids referenced within the type are summarized in ascending property id order. Within this summary, information
for each element referencing the property id is displayed. Various subtotals are accumulated and displayed.

Inputs
The ELSUM Case Control command has been enhanced to give the user new controls over the amount and ordering of
the element summary information produced. The general format of the command is:

 ALL 
 
ELSUM   EID, PID, BOTH, PIDSUM, EIDSUM   =  n 
 
 NONE 
See ELSUM (Case) in the MSC Nastran Quick Reference Guide for a complete description of the ELSUM command. The EID
and PID keywords are used to select the particular grouping desired for the element summary output. The EID keyword
requests grouping by element type while the PID keyword requests grouping by element property type. The BOTH
keyword requests both groupings. The PIDSUM keyword requests that only mass property totals be output for the PID
grouping. The EIDSUM keyword requests that only mass property totals be output for the EID grouping.

Main Index
 Chapter 3: Structural Elements 253
ELSUM Case Control Command

Outputs
The ELSUM Case Control command causes printed output to be generated. Figure 3-81 is an example of the EID grouping
output. Figure 3-82 is an example of the PID grouping output. If PIDSUM or EIDSUM were used, then only the subtotals
and totals information would be present in the figures.

E L E M E N T P R O P E R T Y S U M M A R Y (BY ELEMENT TYPE / ID)

ELEMENT TYPE = BAR

ELEM ID PROP ID MATL ID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
3401 3401 1 1.00000E+00 1.00000E+00 1.00000E+00 5.00000E-02 5.00000E-01 5.50000E-01 5.50000E-01
3403 3403 1 3.00000E+00 1.00000E+00 3.00000E+00 1.50000E-01 1.50000E+00 1.65000E+00 1.65000E+00
3410 3403 1 2.00000E+00 1.00000E+00 2.00000E+00 1.00000E-01 1.00000E+00 1.10000E+00 1.10000E+00
17101 17103 L 1 1.00000E+00 1.01034E+00 1.01034E+00 5.05168E-02 0.00000E+00 5.05168E-02 5.05168E-02
17103 17103 L 1 3.00000E+00 1.01034E+00 3.03101E+00 1.51550E-01 0.00000E+00 1.51550E-01 1.51550E-01
17110 17103 L 1 2.00000E+00 1.01034E+00 2.02067E+00 1.01034E-01 0.00000E+00 1.01034E-01 1.01034E-01
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL BAR 6.03101E-01 3.00000E+00 3.60310E+00 3.60310E+00

.
.
.

ELEMENT TYPE = BEAM

ELEM ID PROP ID MATL ID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
200 200 1 1.00000E+00 1.00000E+02 1.00000E+02 5.00000E+00 0.00000E+00 5.00000E+00 5.00000E+00
17000 17000 L 1 1.00000E+00 9.99987E+01 9.99987E+01 4.99994E+00 0.00000E+00 4.99994E+00 4.99994E+00
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL BEAM 9.99994E+00 0.00000E+00 9.99994E+00 9.99994E+00

.
.
.

ELEMENT TYPE = PENTA

ELEM ID PROP ID MATL ID VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
6801 6801 2 4.00000E-02 2.00000E-03 0.00000E+00 2.00000E-03 2.00000E-03
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL PENTA 2.00000E-03 0.00000E+00 2.00000E-03 2.00000E-03

.
.
.

ELEMENT TYPE = QUAD4

ELEM ID PROP ID MATL ID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
3301 3301 1 1.00000E+00 1.00000E+00 1.00000E+00 5.00000E-02 5.00000E-01 5.50000E-01 5.50000E-01
3313 3313 1 6.12009E-01 6.36864E+00 3.89766E+00 1.94883E-01 1.91715E+00 2.11203E+00 2.11203E+00
3321 3321 P 3321 1.00000E+00 1.00000E+02 1.00000E+02 1.00000E+02 0.00000E+00 1.00000E+02 1.00000E+02
17503 17503 1 1.00000E+00 1.00000E+00 1.00000E+00 5.00000E-02 0.00000E+00 5.00000E-02 5.00000E-02
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL QUAD4 1.00295E+02 2.41715E+00 1.02712E+02 1.02712E+02

.
.
.

ELEMENT TYPE = ROD

ELEM ID PROP ID MATL ID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
101 101 1 1.00000E+00 1.00000E+00 1.00000E+00 5.00000E-02 5.00000E-01 5.50000E-01 5.50000E-01
107 101 1 1.00000E+00 1.00000E+00 1.00000E+00 5.00000E-02 5.00000E-01 5.50000E-01 5.50000E-01
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL ROD 1.00000E-01 1.00000E+00 1.10000E+00 1.10000E+00

.
.
.

ELEMENT TYPE = TETRA

ELEM ID PROP ID MATL ID VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
3901 3901 1 5.77350E-01 2.88675E-02 0.00000E+00 2.88675E-02 2.88675E-02
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL TETRA 2.88675E-02 0.00000E+00 2.88675E-02 2.88675E-02

.
.
.

TOTAL MASS FOR ALL SUPPORTED ELEMENT TYPES 1.38647E+02 1.53344E+01 1.53981E+02 1.53981E+02

Figure 3-81 Sample Element Summary by Element Type Output

Main Index
 254 MSC Nastran Reference Guide
ELSUM Case Control Command

E L E M E N T P R O P E R T Y S U M M A R Y (BY PROPERTY TYPE / ID)

PROPERTY TYPE = PBAR, ID = 3401

***************************************************************************************
ELEM ID ELEM TYPE MATL ID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
3401 BAR 1 1.00000E+00 1.00000E+00 1.00000E+00 5.00000E-02 5.00000E-01 5.50000E-01 5.50000E-01
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL BAR ELEMENTS FOR PBAR, ID = 3401 5.00000E-02 5.00000E-01 5.50000E-01 5.50000E-01

.
.
.

PROPERTY TYPE = PBEAM, ID = 200 ***************************************************************************************

ELEM ID ELEM TYPE MATL ID LENGTH AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
200 BEAM 1 1.00000E+00 1.00000E+02 1.00000E+02 5.00000E+00 0.00000E+00 5.00000E+00 5.00000E+00
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL BEAM ELEMENTS FOR PBEAM, ID = 200 5.00000E+00 0.00000E+00 5.00000E+00 5.00000E+00

.
.
.

PROPERTY TYPE = PSOLID, ID = 3901 ***************************************************************************************

ELEM ID ELEM TYPE MATL ID VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
3901 TETRA 1 5.77350E-01 2.88675E-02 0.00000E+00 2.88675E-02 2.88675E-02
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL SOLID ELEMENTS FOR PSOLID, ID = 3901 2.88675E-02 0.00000E+00 2.88675E-02 2.88675E-02

.
.
.

PROPERTY TYPE = (NONE) ***************************************************************************************

ELEM ID ELEM TYPE MATL ID STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
2901 CONM1 0 1.00000E+00 0.00000E+00 1.00000E+00 1.00000E+00
2904 CONM1 0 2.20000E+00 0.00000E+00 2.20000E+00 2.20000E+00
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL CONM1 ELEMENTS 3.20000E+00 0.00000E+00 3.20000E+00 3.20000E+00

.
.
.

PROPERTY TYPE = PSHELL, ID = 3301 ***************************************************************************************

ELEM ID ELEM TYPE MATL ID THICKNESS AREA VOLUME STRUCT.MASS NON-STR.MASS TOTAL MASS TM*WTMASS
3301 QUAD4 1 1.00000E+00 1.00000E+00 1.00000E+00 5.00000E-02 5.00000E-01 5.50000E-01 5.50000E-01
----------- ----------- ----------- -----------
SUBTOTAL MASS FOR ALL SHELL ELEMENTS FOR PSHELL, ID = 3301 5.00000E-02 5.00000E-01 5.50000E-01 5.50000E-01

.
.
.

=========== =========== ===========

===========
TOTAL MASS FOR ALL SUPPORTED ELEMENT TYPES 1.38647E+02 1.53344E+01 1.53981E+02 1.53981E+02
=========== =========== =========== ===========

Figure 3-82 Sample Element Summary by Element Property Type Output

Guidelines and Limitations

 The ELSUM command keywords can be used in any combination. However, if both EID and EIDSUM are
present, EIDSUM takes precedence over EID and only the mass property totals summary will be generated for the
grouping by element type. This is also true for the PID and PIDSUM keywords. PIDSUM takes precedence over
PID and only the mass property totals will be generated for the grouping by element property type.

Main Index
 Chapter 3: Structural Elements 255
ELSUM Case Control Command

 Only the following element types produce mass property information: CBAR, CBEAM, CBEND, CHEXA,
CMASSi, CONM1, CONM2, CONROD, CPENTA, CQUAD4, CQUAD8, CQUADR, CRAC2D, CRAC3D,
CROD, CSHEAR, CTETRA, CTRIA3, CTRIA6, CTRIAR, CTRIAX6 and CTUBE.
 The ELSUM command is ignored in heat transfer solution sequences.

Example
No new example is presented for this enhanced ELSUM Case Control command since any existing input data file available
can be used to demonstrate the new features. One simply inserts an ELSUM command with the desired keywords into the
Case Control Section of the input file. Element summary information will be generated and placed on the standard .f06
output file.

Main Index
 Chapter 4: Material Properties
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

4 Material Properties

 Isotropic Material 258


Orthotropic Material 261

Anisotropic Material 262

Nonuniform Material and Composite Properties for Shells 264

Hyperelastic Material 296

Main Index
 Chapter 4: Material Properties 257

The material properties used in structural analysis are described in this section. Material properties for heat transfer are
described in the MSC Nastran Thermal Analysis User’s Guide. The material definitions describe the stiffness or constitutive
relationships, strength, density, and thermal expansion characteristics of the material. Material Bulk Data entries can be
classified by type as isotropic (MAT1), orthotropic (MAT3, MAT8), or anisotropic (MAT2, MAT9).
The materials input data types MAT1, MAT2, MAT8 and MAT9 entries can be used to describe temperature-dependent
material when used in conjunction with MATT1, MATT2, MATT8 and MATT9 entries, respectively. In addition, the
MATS1 can be used with the MAT1entry to describe nonlinear material behavior.
Also included in this section is a description of composite materials (PBCOMP, PCOMP) which are used in the
formulation of equivalent material constitutive relations. The method used to model composites is also used in the analysis
of materials with nonhomogeneous cross sections. Composite material properties are used in the analysis of the composite
(nonuniform) beams and shells.
The advanced materials used in nonlinear solution, SOL 400, are described in the MSC Nastran Nonlinear User’s Guide,
Chapter 10: Materials. The advanced materials used in explicit nonlinear solution, SOL 700, are described in the MSC
Nastran Explicit Nonlinear (SOL 700) User’s Guide, Materials (Ch. 7).
Table 4-1 summarizes the relationships between elements and material input data types.

Table 4-1 Correspondence Between Element and Material Type

Hyper- Stress Temperature
Element Material Type elastic Dependent Dependent Composite
Type MAT1 MAT2 MAT3 MAT8 MAT9 MATHP MATS1 MATTi
CBAR X X
CBEAM X X X X
CBEAM3 X X X
CBEND X X
CCONEAX X X X
CFAST X
CRAC2D X X
CRAC3D X X
CHEXA X X X X X
CPENTA X X X X X
CQUAD X
CQUAD4 X X X X X X X
CQUAD8 X X X X X X
CQUADR X X X X
CQUADX X

Main Index
 258 MSC Nastran Reference Guide
Isotropic Material

Table 4-1 Correspondence Between Element and Material Type (continued)

Hyper- Stress Temperature
Element Material Type elastic Dependent Dependent Composite
Type MAT1 MAT2 MAT3 MAT8 MAT9 MATHP MATS1 MATTi
CROD X X X
CSEAM X
CSHEAR X X
CTETRA X X X X X
CTRIA3 X X X X X X X
CTRIA6 X X X X X X
CTRIAR X X X X
CTRIAX X
CTRIAX6 X X
CTUBE X X X
CWELD X

Isotropic Material
The MAT1 entry is used to describe the isotropic material properties for general, plain stress, plain strain, and axisymmetric
isotropic material behavior. This section describes the MAT1 Bulk Data entries used to define isotropic materials and how
that definition relates to the constitutive relations. See Table 4-1 for a summary of elements that use MAT1 entries.
The general three-dimensional isotropic material stress-strain constitutive relationship is:

Main Index
 Chapter 4: Material Properties 259
Isotropic Material

E  1 –   - -------------------------------------
E E
------------------------------------- - -------------------------------------
- 0 0 0
 1 +    1 – 2   1 +    1 – 2   1 +    1 – 2 
 E  1 –   - -------------------------------------
E
 x ------------------------------------- - 0 0 0
   1 +    1 – 2   1 +    1 – 2 
  y
  E1 –  -
------------------------------------- 0 0 0
 z   1 +    1 – 2 
  =
  xy  E -
------------------- 0 0
  2  1 + 

 yz 
  symmetric E -

 zx  ------------------- 0
21 + 
E -
-------------------
21 + 

 A
 x   
   A 
 y   
   
  z  A 
   –  T – T ref   
   xy  0 
   
   yz  0 
   
   zx  0 

where    ,    , and  A  are the stress, strain, and thermal expansions coefficients, respectively. E ,  , G , and
 T – T ref  are Young’s modulus, Poisson’s ratio, shear modulus, and the change in element temperature used to calculate
initial element thermal expansions.
For MAT1, G = E  21 +  is the relationship used to determine the isotropic elastic material constants when only two
of the three are input on the MAT1 entry. When all three are given, G will not be required to satisfy this relationship (see
Table 4-2 for the use of the elastic constants with respect to the finite elements usage).
Isotropic material constitutive relationship for plane stress behavior is:

Main Index
 260 MSC Nastran Reference Guide
Isotropic Material

E E
-------------- -------------- 0
2 2
 x  1– 1–   x  A 
  E
E --------------    
 y  = -------------- 0    y  –  T – T ref   A  
  2 2    
1– 1–
  xy     xy   0 
E -
0 0 -------------------
21 + 
Isotropic material constitutive relationship for plane strain behavior is:

E1 –  - E
------------------------------------- -------------------------------------
- 0

 x  1 +    1 – 2   1 +    1 – 2   x  A 
  E E1 –  -    
  y  = -------------------------------------- ------------------------------------- 0    y  –  T – T ref   1 +    A  
   1 +    1 – 2   1 +    1 – 2     
  xy  E -    xy   0 
0 0 -------------------
21 + 
When MAT1 is used for transverse shear

  zx G 0   zx
  =  

 zy 0 G   zy

Table 4-2 MAT1 – Material Property Usage Versus Element Types

Element Entry E NU G
CROD Extension and Not Used Torsion Transverse
CBEAM Bending Shear
CBAR
CQUADi Membrane and Bending Transverse Shear
CTRIAi
CCONEAX

Main Index
 Chapter 4: Material Properties 261
Orthotropic Material

Table 4-2 MAT1 – Material Property Usage Versus Element Types

Element Entry E NU G
CSHEAR Not Used Shear
CRAC2D All Terms Not Used
CHEXA All Terms Not Used
CPENTA
CTETRA
CRAC3D
CTRIAX6 Radial, Axial, All Coupled Ratios Shear
Circumferential

Orthotropic Material
This section describes the MAT3 and MAT8 Bulk Data entries which define orthotropic materials and their corresponding
constitutive relations. The MAT3 and MAT8 entries are used to describe the orthotropic material properties for two-
dimensional and three-dimensional element orthotropic material behavior. See Table 4-1 for a summary of elements that use
MAT3 and MAT8 entries.
The MAT3 entry is used to define orthotropic material matrix for axisymmetric solid elements in the stress-strain
relationship as given by:

1- –----------
-----
 x –  zx
- ---------- 0
Ex E Ez
 x  –  x 1 –  z  x   Ax 
  ----------- ------ ----------- 0     
   Ex E Ez    A 
  =   +  T – T ref    
 z  –  xz –  z 1  z   Az 
  ---------- ----------- ----- 0    
  zx  Ex E Ez   zx  0 
1
0 0 0 --------
G zx

where    ,    , and  A  are the stress, strain, and thermal expansions coefficients, respectively. E ,  , G , and
 T – T ref  are Young’s modulus, Poisson’s ratio, shear modulus, and the change in element temperature used to calculate
initial element thermal expansions.
Where  x  E x =  z  E  ,  zx  E z =  xz  E x , and  z  E  =  z  E z must be satisfied to preserve symmetry.
Only the  x ,  xz , and  z terms are used in the input.

Main Index
 262 MSC Nastran Reference Guide
Anisotropic Material

The MAT8 entry is used to define orthotropic material matrix for shell elements in the stress-strain relationship as given by:

1- –----------
-----
 21
- 0 0 0
E1 E2
 1  –  12 1  1   A1 
  ----------- ------ 0 0 0    
 2 E1 E2  2   A2 
  1    
  12  = 0 0 --------- 0 0   12  +  T – T ref   0 
  G 12    
  1z    1z  0
  1    
  2z  0 0 0 --------- 0  2z  0
G 1z
1
0 0 0 0 ---------
G 2z

where    ,    , and  A  are the stress, strain, and thermal expansions coefficients, respectively. E ,  , G , and
 T – T ref  are Young’s modulus, Poisson’s ratio, shear modulus, and the change in element temperature used to calculate
initial element thermal expansions.
Where  21  E 2 =  12  E 1 must be satisfied to preserve symmetry. Only the  12 term is used in the input.

When either G 1z or G 2z is input as zero, it implies that the corresponding shear modulus is infinite. When this is the
case, the transverse shear flexibility effects are neglected in the shell formulation.

Anisotropic Material
This section describes the MAT2 and MAT9 entries that define anisotropic materials and their corresponding constitutive
relations. The MAT2 and MAT9 entries are used to describe the anisotropic material properties for two-dimensional and
three-dimensional element anisotropic material behavior. See Table 4-1 for a summary of elements that use MAT2 and
MAT9 entries.
The MAT2 entry used to define two-dimensional anisotropic constitutive stress-strain relationship using the anisotropic
material matrix, [G]. The stress-strain relationship is given by:

 1  G 11 G 12 G 13    1   A1  
     
  2  = G 12 G 22 G 23    2  –  T – T ref   A 2  
     
  12  G 13 G 23 G 33    12   A 12  

Main Index
 Chapter 4: Material Properties 263
Anisotropic Material

where    ,    , and  A  are the stress, strain, and thermal expansions coefficients, respectively. E ,  , G, and
 T – T ref  are Young’s modulus, Poisson’s ratio, shear modulus, and the change in element temperature used to calculate
initial element thermal expansions.
The MAT2 entry is also used to define two-dimensional orthotropic constitutive stress-strain relation using the anisotropic
material matrix,  G  . The stress-strain relationship is given by:

 1  G 11 G 12 0   1   A1  
     
  2  = G 12 G 22 0    2  –  T – T ref   A 2  
     
  12  0 0 G 33    12   0 

where the elastic coefficients are given by engineering constants used to describe orthotropic material properties as:
E1
G 11 = ------------------------------
 1 –  12  21 

E2
G 22 = ------------------------------
 1 –  12  21 

 12 E 2  21 E 1
G 12 = ------------------------------ = ------------------------------
 1 –  12  21   1 –  12  21 

When MAT2 is used for transverse shear

  zx G 11 G 12   zx
  =  
  zy G 12 G 22   zy

G 33 must be set to 0.
The MAT9 entry is used to define general three-dimensional anisotropic constitutive stress-strain relationship using the
anisotropic material matrix,  G  , the stress-strain constitutive relationship is:

Main Index
 264 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

 G 11 G 12 G 13 G 14 G 15 G 16  A
 x  x   1 
     A 
 y G 22 G 23 G 24 G 25 G 26  y   2 
     
 z  G 33 G 34 G 35 G 36   z   A3  
  =    –  T – T ref    
  xy  G 44 G 45 G 46    xy   A4  
     
  yz  symmetric G 55 G 56    yz   A5  
     
  zx  G 66    zx   A6  

where    ,  , and  A  are the stress, strain and thermal expansions coefficients, respectively, and E ,  , G , and
 T – T ref  are Young’s modulus, Poisson’s ratio, shear modulus, and the change in element temperature used to calculate
initial element thermal expansions.
The MAT9 entry is also used to define three-dimensional orthotropic constitutive stress-strain relation using the anisotropic
material matrix  G  . The stress-strain relationship is given by:

 G 11 G 12 G 13 0 0 0 
 x  x   A1  
      
 y G 22 G 23 0 0 0  y   A2  
     
 z  G 33 0 0 0 
 z  A 
  =    –  T – T ref   3  
  xy  G 44 0 0    xy   0 
     
  yz  symmetric G 55 0    yz   0 
     
  zx  G 66    zx   0 

Nonuniform Material and Composite Properties for Shells

The CQUAD4, CQUAD8, CTRIA3, and CTRIA6 surface elements can be used to model laminated composites in finite
element models. Laminated composites may be conceptually viewed as a “stack” of laminae with different orientations of
the principal material directions in the individual lamina. An exploded view of three cross-ply laminated plates is illustrated
in Figure 4-1. The n-laminae (n = 1,2,3,4) of each of the three configurations are normal to the z-axis of the indicated
coordinate system and the 1- and 2-axes appended to the individual lamina denote principal material axis directions. The
directions of the principal material axes of each lamina alternate as implied by the use of the word “cross-ply” to describe
the configuration. The xy-plane of the coordinate axes is defined in the geometric middle plane of the laminae.
An entire stack of laminae may be modeled with a single plate or shell element because the geometry of the plate/shell is
thin and often the material properties of the “stack” are completely reflected in the matrices of elastic moduli for the
element. These matrices are automatically calculated from user-supplied definitions of the thickness, the material
properties, and the relative orientation of these properties for the individual lamina. Once these matrices of elastic moduli

Main Index
 Chapter 4: Material Properties 265
Nonuniform Material and Composite Properties for Shells

are calculated, the analysis proceeds in a standard manner as described in Chapter 7: Solution Sequences. This capability
for the automatic representation of laminated composites is available in linear static analysis, real and complex eigenvalue
analysis, buckling analysis, and dynamic analysis.
In static analysis and in static analysis with cyclic symmetry, the analyst has the option to request that the program evaluate
stresses or strains and appropriate failure indices in individual laminae. Interlaminar shear stresses and strains and a failure
index for bonding may also be requested.
If any of the plies have material nonlinear behavior including progressing failures, SOL 400 should be used. See the MSC
Nastran Nonlinear User’s Guide.

x
2 1
x
1 2
1 2
y 2 1
1
y 2
Two-Ply

Three-Ply

2
1 x
1
2
2
y 1
1
2

Four-Ply
Figure 4-1 Exploded View of Three Cross-Ply Laminated Plates

Layered Composite Material Analysis

Two Bulk Data entries, MAT8, PCOMP, and PCOMPG are used in the modeling of composite structures. The MAT8
Bulk Data entry (defined in Nonuniform Material and Composite Properties for Shells, 264) defines the material property for an

Main Index
 266 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

orthotropic material. The material properties of the individual lamina along the longitudinal direction (also called the fiber
direction or 1-direction) and the transverse direction (also called the matrix direction or 2-direction) are supplied using this
entry. The PCOMP Bulk Data entry defines the properties of the n-ply composite material laminate. The thickness,
orientation, and the material identification number of each of the individual lamina is provided using this entry. PCOMPG
is similar to PCOMP, but you can define unique ply layer IDs. For this section, they can be considered and only PCOMP
will be identified.
From the data supplied by the PCOMP and MAT8 Bulk Data entries, the program automatically calculates the membrane,
bending, transverse shear, and coupled membrane-bending material properties of the laminate as a whole. These calculated
properties are output in the form of PSHELL and MAT2 entries when a sorted echo of the Bulk Data Section is requested.

The coefficients of thermal expansion derived for membrane-bending coupling, which appear in the A1, A2, and A12 fields
of MAT2 entry and correspond to the MID4 Field on the PSHELL, require special interpretation. They are given by:
  MAT2  =  G ij    ACTUAL  (4-1)
MAT2

In order to obtain the actual values of A1, A2, and A12   MAT2  , the user must solve equation (4-1).

If the user specifies ECHO = PUNCH in the Case Control Section, the generated PSHELL and MAT2 entries are routed
to the PUNCH file, where they are written after the sorted bulk data echo. Since the MIDi fields use more than eight digits,
a large field format is used for the PSHELL and MAT2 entries. The PUNCH file thus contains the sorted Bulk Data echo
and the generated PSHELL and MAT2 entry images. After the MAT8 and PCOMP entry images are deleted, the entry
images contained in the PUNCH file may be included in the Bulk Data to be submitted in a subsequent run. The existence
of composite elements is recognized by a large value (more than eight digits) on the MIDi fields of the PSHELL entry, which
may be generated by PCOMP entries. If stress output is desired, ensure that PARAM,NOCOMPS is set to -1 for any
subsequent runs because composite stress recovery is not possible without the PCOMP entries.
Certain postprocessing features are available for Solution Sequences 101, 103, 105, 106, 114, 115, 153, 200, 400, and 600
which include all the statics, normal modes, steady nonlinear heat transfer, and buckling analyses. PARAM,NOCOMPS
may be used to modify these post processing options in superelement solution sequences. These features are (1) stresses in
individual laminae including interlaminar shear stresses and (2) a failure index for each lamina.
It is to be noted that two-dimensional plate theory used in the CQUAD4, CQUAD8, CTRIA3, or CTRIA6 elements does
not allow for the exact calculation of interlaminar stresses. An approximate procedure is therefore used to determine the
interlaminar shear stresses. The approximation consists of neglecting the effects of twisting moments in the computation
of interlaminar shear stresses. For the majority of composite material structural analyses, this approximation is expected to
yield satisfactory results for all elements in the model except those at the edges of the structure.
Voluminous output will result if the stresses for all elements for all laminae are requested. To help the analyst to review this
information, provisions are made to calculate a failure index for each ply. This failure index is obtained by considering the
failure criteria for unidirectional fiber composites as in the commonly used failure theories. The user has the option to
choose any one of the four commonly used failure criteria for composites. These are: (1) Hill's Theory, (2) Hoffman's
Theory, (3) Tsai-Wu Theory, and (4) Maximum Strain. The user must supply allowable stresses/strains for the laminae if a
failure index is to be calculated. If more modern failure criteria are desired (such as, Puck Hashin or Digimat), SOL 400
should be used. See the MSC Nastran Nonlinear User’s Guide.

Main Index
 Chapter 4: Material Properties 267
Nonuniform Material and Composite Properties for Shells

Classical Lamination Theory

Classical lamination theory makes the following assumption regarding the behavior of the laminae:
 The laminae are perfectly bonded together.
 The bonds are infinitesimally thin and no lamina can slip relative to another.
 Linear variation of strain through the laminate thickness is assumed.
Deformation in the X-Y plan of the plate at any point C at a distance z in the normal direction to plate middle surface is
U = U 0 + z y

V = V 0 + z x

where U , V , and W are the displacements along the X, Y, and Z directions in the element coordinate system, and  x ,
 y are the rotations.
The strain-displacement-middle surface strain and curvatures relationship is given by:
U 
 ---------0   -------y- 
  x   x   0 
  x      x   x 
   V 0    x   0  
  y  =  ---------  + z  – --------  =   y  – z   y 
   y   y     
  xy   U  V         0    xy 
 ---------0 + --------0-   -------y- – -------x-   xy 
 y x   y x 
0
where the  ’s and  ’s are the middle surface strains and curvatures, respectively.
The stress resultants for an N-layer laminate are obtained by integration of the stresses in each lamina through the laminate
thickness as:

 Nx  --t-
 x  N  x 
    zk  
2
 N y  =  t   y dz =     y  dz
  - ---   zk – 1  
 N xy  2   k=1   xy k
xy

 Mx  --t-
 x  N  x 
    zk  
2
 M y  = –  t   y z dz = –     y  z dz
  - ---   zk – 1  
 M xy  2   k=1   xy k
xy

The stress resultant to strain relationship is:

Main Index
 268 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

  0  
 Nx  N   x  x  
   zk  0  zk   
 Ny  =   G k     y  dz –    y  z dz 
   zk – 1   zk – 1   
 N xy  k=1   0    xy k 
  xy k 

M   0  
 x N   x  x  
   zk  0  zk   2 
 My  =   G k  –    y  z dz +    y  z dz 
   zk – 1   zk – 1   
 M xy  k=1   0    xy k 
  xy k 

where  G  k is the material matrix transformed from the laminate coordinate system into the lamina coordinate system.

These relations can be written in the following form used to describe composite elements:

F 0
A B  
  =  
M  B D 

where:

A = N

  G k  zk – zk – 1 
k=1
B = N
1--- 2 2
2   G k  zk – zk – 1 
k=1
D = N
1--- 3 3
3   G k  zk – zk – 1 
k=1

are named in composite element literature as the membrane, membrane-coupling, and bending matrices, respectively. In
the shell element formulation in MSC Nastran, these relationships take the following form:

Main Index
 Chapter 4: Material Properties 269
Nonuniform Material and Composite Properties for Shells

2
TG 1 T G4 0
F  0 
  3  
 M  = T2 G T-
----- G 0 
  4 12 2  
Q  
0 0 Ts G3

where:

 A  = TG 1
2
 B  = –T G4
D = 3
T
------ G 2
12
Q =
 Qx 
  = transverse shear resultants
 Qy 
 =
 x 
  = transverse shear strains
 y 
T = nominal plate thickness
Ts = effective transverse shear material thickness
G 3 = effective transverse shear material matrix

MSC Nastran allows G 1 , G 2 , G 4 , T , G 3 , and T s to be input directly in PSHELL or to have the composite equivalent
material matrices calculated internally from the PCOMP data.
Equivalent thermal properties are determined as follows:

  T 1  –1 G 
 0  3  A B 0  0 
  T   
    =  -----
- 2  = B D 0  G  
   12   
   0   2  0 0 Ts G3  G  0 
 T 3 
The following coefficients are used to determine equivalent thermal properties:

Main Index
 270 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

N
 G   =
0   G k   k  zk – zk – 1 
k=1

N
1 3 3
 G   = ---   G k   k  zk – zk – 1 
3
k=1

N
1 2 2
 G   = ---   G k   k  zk – zk – 1 
0 2
k=1
These coefficients are used to calculate the equivalent thermal properties as follows:
–1
    =  A   G  
0 0

and
–1
  x  =  D   G x 

where     and   x  are the membrane and bending equivalent thermal properties.
0

Note that     is not directly calculated, but is determined from  G   in MSC Nastran when the PCOMP input
0 0

is used when the MID4 field on the PSHELL is > 400,000,000. Also note that  G x  cannot be input directly using
0

PSHELL and     can be input only if  B  is invertible (which is generally not true).
0

The thermal expansion relationships in the shell element formulation take the following form:
  = T 1
0

3
T
  = ------  2
12
2
  = T  3
0

where  1 ,  2 , and  3 are the thermal expansion inputs on the materials referenced by the MID1, MID2, and MID4
fields on the PSHELL entry. If PCOMP is used, these relationships are automatically calculated.

Main Index
 Chapter 4: Material Properties 271
Nonuniform Material and Composite Properties for Shells

The terms G 1 , G 2 , and G 4 are defined by the following integrals:

1
G 1 = ---   G e  dz
T
1 2
G 2 = ---  z  G e  dz
I
1
G 4 = ------   – z   G e  dz
2
T
The limits on the integration are from the bottom surface to the top surface of the laminated composite. The matrix of
material moduli,  G e  , has the following form for isotropic materials:

E E
-------------- -------------- 0
2 2
1– 1–
E E
 Ge l = -------------- -------------- 0
2 2
1– 1–
E -
0 0 -------------------
21 + 

For orthotropic materials, the matrix,  G e  , is written as follows:

E1 1 E2
-------------------- -------------------- 0
1 – 1 2 1 – 1 2
2 E1 E2
 G e  0 = -------------------- -------------------- 0
1 – 1 2 1 – 1 2

0 0 G 12

Here,  1 E 2 =  2 E 1 in order to satisfy the requirement that the matrix of elastic moduli be symmetric. In general, the
analyst may supply element properties with respect to a particular orientation which does not necessarily correspond to the
principal material axes. In this case, the analyst must also supply the value of the angle,  or material coordinate system
that orients the element material axis relative to the side G1-G2 of the element. The material elastic modulus matrix is then
transformed by the program into the element modulus matrix through the relation

Main Index
 272 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

t
 Ge  =  U   Gm   U 

where:

2 2
cos  sin  cos  sin 
U = sin 
2
cos 
2
– cos  sin 
2 2
– 2 cos  sin  2 cos  sin  cos  – sin 
The finite element model for a structure composed of composite materials requires the evaluation of the matrix of elastic
moduli for each plate element of the model. The characteristics of the composite media are totally contained in these
matrices.

x
2 1
x
1 2
1 2
y 2 1
1
y 2
Two-Ply

Three-Ply

2
1 x
1
2
2
y 1
1
2

Four-Ply
Figure 4-2 Exploded View of Three Cross-Ply Laminated Plates

Main Index
 Chapter 4: Material Properties 273
Nonuniform Material and Composite Properties for Shells

To illustrate evaluation of these matrices, consider the cross-ply laminates of Figure 4-2. In this portion of the discussion, the
three configurations shown in Figure 4-2 are each assumed to be represented by a single quadrilateral plate element and the
coordinate axes shown are coincident with the element coordinate axes. Then, if it is assumed that each lamina of the n-ply
laminates is of thickness T  n , where T is the total thickness of each of three configurations, the matrices of elastic moduli
may be evaluated from the following relations:

 – T--- + T--- T 2T T 
– --- + ------ ---
 2 n 2 n 2 
1  
 G 1  = ---    G e  1 dz +   Ge 2 dz +  +   G e  n dz 
T 
 – T--- T T T  n – 1 T 
– --- + --- – --- + --------------------
 2 2 n 2 n 

 – T--- + T--- T 2T T 
– --- + ------ ---
 2 n 2 n 2 
1---  2 2 2 
 G 2  =    G e  1 z dz +   Ge 2 z dz +  +   G e  n z dz 
I 
 – T--- T T T  n – 1 T 
– --- + --- – --- + --------------------
 2 2 n 2 n 

 – T--- + T--- T 2T
– --- + ------
 2 n 2 n
1 
 G 4  = ------    G e  1  – z  dz +   Ge 2  –z  dz
2
T  T T T
 – --- – --- + ---
 2 2 n

T
--- 
2 

++   Ge n  –z  dz 
T n – 1 T-
--- + ------------------- 
2 n 
These relations reflect the assumption that the xy-plane of the element coordinate system is coincident with the geometric
middle plane of the laminate. The xy-plane of the element coordinate system is defined in the mean plane of the element
so that any offset between the mean plane of the connected grid points and the geometric middle plane of the laminate
would be reflected in the integration limits of the preceding relations.
3 2
Note that I = T  12 in the evaluation of  G 2  , i.e., the value  12I   T , will be assigned the default value of 1.0 on
the plate element property entry.

Main Index
 274 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

Transverse Shear Theory

The matrix of elastic moduli for transverse shear,  G 3  is defined as a two-by-two matrix of the form
m

G 11 G 12
 G3 m =
G 21 G 22

and the corresponding matrix transformed into an element coordinate system is given by
T
 G3 e =  W   G3 m  W 

where

W = cos  sin 

– sin  cos 
The mean value of the transverse shear modulus G for the laminated composite is defined in terms of the transverse shear
strain energy, U , through the depth
2 2
1V 1 z
U = --- -------- = ---  ----------------- dz (4-2)
2 GT 2 Gz
A unique mean value of the transverse shear strain is assumed to exist for both the x and y components of the element
coordinate system, but for ease of discussion, only the evaluation of an uncoupled x component of the shear moduli will be
illustrated here. From equation (4-2), the mean value of transverse shear modulus may be written in the following form
N zi 2
1- T   zx  z  
----- = --------   ---------------------- dz (4-3)
Gx 2
Vx i = 1 z  Gx i
i–1

where G is an “average” transverse shear coefficient used by the element code and  G x  is the local shear coefficient for
i
layer i . To evaluate equation (4-3), it is necessary to obtain an expression for   zx  z   . This can be accomplished by
assuming that the x- and y-components of stress are decoupled from one another. This assumption allows the desired
equation to be deduced through an examination of a beam unit cross-sectional width.

Main Index
 Chapter 4: Material Properties 275
Nonuniform Material and Composite Properties for Shells

z
x  x + d x
Vx  xz
M
 zx

 xz

The equilibrium conditions in the horizontal direction and for total moment are
 xz  x
---------- + -------- = 0 (4-4)
z x
M x
V x + ---------- = 0 (4-5)
x

Now, if the location of the neutral surface is denoted by z x and  is the radius of curvature of the beam, the axial stress
E x may be expressed in the form

Ex  zx – z 
 x + ------------------------ = 0
 EI  x

Relation 20 may be differentiated with respect to x combined with equations (4-4) and (4-5). In a region of constant E x the
result may be integrated to yield the following expression

Vx 2
z
 xz = C i + -------------  z x z – ----- E xi zi – 1  z  zi (4-6)
 EI  x  2

Equation (4-7) is particularly convenient to use in the analysis of n-ply laminates because sufficient conditions exist to
determine the constants Ci (i = 1,2,...,n) and the “directional bending center” z x . For example, consider the following
laminated configuration:

Main Index
 276 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

i=3

i=2 x

i=1

At the bottom surface ( i = 1 , z = z 0 , and  xz = 0 )

2
–Vx  z0 
C 1 = -------------  z x z 0 – ------- E x1
 EI  x  2

and for the first ply at the interface between plies i = 1 and i = 2 ( z = z 1 )

Vx 1 2 2
  xz  1 = + ------------- z x  z 1 – z 0  ---  z 1 – z 0  E x1
 EI  x 2

At this interface between plies i = 1 and i = 2 ,

2
Vx  z1 
  xz  2 = C 2 + -------------  z x z 1 – ------- E x2
 EI  x  2

and as   xz  =   xz  at z = z 1 ,
2 1

V x E x2 1 2
C 2 =   xz  1 – --------------- z x z 1 – --- z 1 (4-7)
 EI  x 2

Then, in the ply z 1  z  z 2 the shear is

V x E x2 1 2 2
 xz  z  =   xz  1 --------------- z x  z – z 1  – ---  z – z 1 
 EI  x 2

In general, for any ply, Z i – 1  z  z i , the shear is

Main Index
 Chapter 4: Material Properties 277
Nonuniform Material and Composite Properties for Shells

V x E xi 1 2 2
 xz  z  =   xz  i – 1 -------------- z x  z – z i – 1  – ---  z –  z i – 1   (4-8)
 EI  x 2

At any ply interface, z i , the shear is therefore

i
Vx 1
  xz  i = -------------  Exj Tj z – ---  z j + z j – 1 
 EI  x 2
j=1

where T j = z j –  z j – 1  .

Note that the shear at the top face,   xz  , is zero and therefore
n

n n
Vx
  xz  n = ---------- z x  Exj Tj –  Exj Tj  zj + zj – 1   2 = 0 (4-9)
 EI 
j=1 j=1

Equation (4-9) proves that if z x is the bending center, the shear at the top surface must be zero.

Equation (4-8) could be substituted into (4-4) and integrated. A better form of equation (4-9), for this purpose is
V x E xi 1 2 2
  xz  z   i = -------------- f xi + z x  z – z i – 1  ---  z –  z i – 1   (4-10)
 EI  x 2

where
i–1
1 1
f xi = -------  Exj Tj z x – ---  z j + z j – 1  (4-11)
E xi 2
j=1
Substituting equation (4-10) into equation (4-6) and after a considerable effort of integrating the results, we obtain
N
1-
----- T 1
= -------------  -------- R xi (4-12)
Gx 2 G xi
 EI  x i = 1

where

Main Index
 278 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

2  1 2
R xi =  E xi  T i  f xi +  z x – z i – 1 T i – --- T i f xi
 3 

1 1  2 1 2 1 1 2 2
+  ---  z x – 2 z i – 1  – --- T i z x T i +  --- z i – 1 + --- z i – 1 T i + ------ T i T i
3 4  3 4 20 

This expression for the inverse shear modulus for the x-direction may be generalized to provide for the calculation of each
term in the two-by-two matrix of shear moduli.

n –1
T i –1
 G kl  = ------------------   G kl  R ki (4-13)
2
 EI  kk i = 1

where:

k = 1,2
l = 1,2

i –1
Note that if no shear is given,  G  = 0 , and also that in equation (4-8)

where  G 2  is calculated in the same manner as  G 2  except that Poisson’s Ratio is set to zero.

The moduli for individual plies are provided through user input because, in general, G 12  G 21 will be used for the
coupling terms. Finally,

G 11  G 12  avg
 G3  =
 G 12  avg G 22

3
As an example, consider a single layer element. For this case let z i – 1 = – T  2 , z = 0 , f o = 0 , and EI = ET  12 .
Evaluating equation (4-13), we obtain
2 5
2 5 1 1 1 E T
R i = E T ------ – --- + ------ = ------------
12 8 20 120
and

Main Index
 Chapter 4: Material Properties 279
Nonuniform Material and Composite Properties for Shells

1-   12  2 T  E 2 T 5  6
--- =  -----------------   ------------------ = ----------
G  E T   120 G 1
2 6 5G 1

which provides the correct factor for a nonuniform shear distribution in a plate.

Data Recovery
As the material properties of the laminated composite plate are completely contained in the matrices of elastic moduli,
standard data recovery methods (see Nonuniform Material and Composite Properties for Shells, 264) may be used to calculate stresses
in individual laminae and the forces sustained by the laminate.
An additional aid to the analyst is provided through optional output of a failure index for individual laminae. Failure indices
assume a value of one on the periphery of a failure surface in stress space. If the failure index is less than one, the lamina
stress is interior to the periphery of the failure surface and the lamina is assumed safe. Conversely, if the failure is greater
than one, the lamina stress is exterior to the periphery of the failure surface and the lamina is assumed to have failed. These
failure indices represent a phenomenological failure criterion in that only an occurrence of a failure is indicated and not the
mode of failure.
In the present context, concern is with the analytical definition of a failure surface in stress or strain space for use with
laminae under biaxial loading. Four commonly used definitions of the failure surface are provided: Hill's Theory,
Hoffman's Theory, Tsai-Wu Theory, and maximum strain theory.
The discussion of data recovery activities concludes with a description of the approximate procedure provided to obtain
interlaminar shear stresses.

Failure Indices
In the analysis of isotropic materials, strength is independent of the orientation of the body under load and one may
compare the largest computed principal stress with an allowable stress to establish the integrity of the structure. Laminated
composites, on the other hand, are orthotropic materials and may exhibit unequal properties in tension and compression.
Thus, the strength of these orthotropic laminae is a function of body orientation relative to the imposed stresses.
As the evaluation of the matrices of material moduli for laminated composites provides sufficient information to determine
the actual stress field sustained by the material, the determination of structural integrity will depend on the definition of
an allowable stress field. The basic ingredient of this definition is the establishment of a set of allowable stresses or strengths
in the principal material directions.
Xt = Allowable tensile stress in the principal x (or 1) direction of the material.
X c = Allowable compressive stress in the principal x (or 1) direction of the material.
Yt = Allowable tensile stress in the principal y (or 2) direction of the material.
Y c = Allowable compressive stress in the principal y (or 2) direction of the material.
S = Allowable shear stress in the principal material system.

Main Index
 280 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

Expressions for the three available failure theories are shown in Table 4-3.

Table 4-3 Failure Theories Provided in MSC Nastran

Theory Failure Index Remarks
Hill  12  1  2  22  12 2 Orthotropic materials with
------ – ------------ + ------ + -------- = FI equal strengths in tension
x2 x2 y2 s2 and compression, i.e.,
x t = x c and y t = y c .
Hoffman  12  22  12 2 1 2 Orthotropic materials
 ---- 1
1- ----- 1 1
– -  1 +  ---- – -----  2 + --------- + --------- + --------- – ------------ = FI under a general state of
X X  y y  xx yy s2 xx plane stress with unequal
t c t c t c t c t c
tensile and compressive
strengths.
Tsai-Wu  12  22  12 2 Orthotropic materials
1 1
 ---- 1
1- ----
- – ------  1 +  --- – -  2 + --------- + --------- + --------- + 2F 12  1  2 = FI under a general state of
X X  y y  xt xc yt yc s 2 plane stress with unequal
t c t c
tensile and compressive
strengths.
MAX STRAIN
  1   2   12 
 ----- ,  ----- ,  ----------
 X t  Y t S

The following remarks apply to Table 4-3:

1. The failure index of Hill represents an analytical expression for an ellipse in the  1  2 plane. This original
expression for failure was subsequentially modified to that shown on the left-hand side of the expression in Table 4-3
by Tsai and hence this failure theory is sometimes known as the Tsai-Hill Theory.
2. The Tsai-Wu tensor polynomial theory for orthotropic materials is specialized to the expression given in Table 4-3
for orthotropic laminae under a condition of plane stress. The coefficient F 12 is experimentally determined from
test specimens under biaxial loading. This inconvenience along with the constraint that F 12 satisfy a stability
criterion of the form
1 -  --------
 -------- 1 - 2 0
– F 12
x x  y y 
t c t c

Main Index
 Chapter 4: Material Properties 281
Nonuniform Material and Composite Properties for Shells

creates some complication in the use of this theory. Narayanaswami and Adelman (see Reference 1) have suggested
that F 12 be set to zero and the use of Hoffman's Theory or the Tsai-Wu theory with F 12 = 0 are preferred
alternatives. The analyst may provide a value of f 12 to be used in the Tsai-Wu failure index if he wishes.

Hill’s Theory (Failure Theory ID = “HILL”)

 12  1  2  22  12 2
Failure Index = ------ – ------------ + ------ + --------
X2 X2 Y2 S2
where:

X =is allowable stress in 1-direction

Y =is allowable stress in 2-direction
S =is allowable stress in shear

and X = X t if  1 is positive or X = X c if  1 is negative and similarly for Y and  2 . For the interaction term

 1  2  X 2 , X = X t if  1  2 is positive or X = X c if  1  2 is negative.
Note that Hill’s theory is used for materials that have the same strength in tension and compression.
A plot of the above equation obtained by setting the failure index to 1 on the  1 -  2 plane yields an ellipse and is the
anisotropic yield criterion of Hill (modified later by Tsai, and hence also sometimes known as the Tsai-Hill theory).
Therefore, if the failure index so calculated is less than 1, the ply stresses are inside the yield ellipse and the ply is said to be
“safe”; conversely, if the failure index is greater than 1, the ply stresses are outside the yield ellipse and the ply has failed.
Hoffman’s Theory (Failure Theory ID = “HOFF”)

The Hoffman’s theory for an orthotropic lamina in a general state of plane stress is given by

1-
1- ----- 1 ----1-  12  22  12 2 1 2
 ---- – 1 +  ---- – 2 + ----------
- + ----------
- + - – ------------ = 1
--------
X X  Y Y  Xt Xc Yt Yc S 2 Xt Xc
t c t c

The failure index is obtained by evaluating the left-hand side of the above equation.
Note that this theory takes into account the difference in tensile and compressive allowable stresses by using linear terms in
the equation.

Tensor Polynomial Theory of Tsai-Wu (Failure Theory ID = “TSAI”)

The theory of strength for anisotropic materials proposed by Tsai and Wu specialized to the case of an orthotropic lamina
in a general state of plane stress is

Main Index
 282 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

F 1  1 + F 2  2 + F 11  12 + F 22  22 + 2F 12  1  2 + F 66  12
2 = 1

where:
1 1
F1 = ----- – ------
Xt Xc

1 1
F2 = ---- – -----
Yt Yc

1
F 11 = + -----------
Xt Xc

1
F 22 = + -----------
Yt Yc

1
F 66 = + -----
S2

and F 12 is to be evaluated experimentally.

The magnitude of F 12 is, however, constrained by the following inequality called a “stability criterion”:

2 0
F 11 F 22 – F 12

The necessity of satisfying the stability criterion, together with the requirement that F 12 be determined experimentally
from a combined stress state, poses difficulties in the use of this theory. Narayanaswami and Adelman (see Reference 1) have
suggested that F 12 be set to zero and that use of Hoffman's theory or Tensor Polynomial theory with F 12 = 0 is a
preferred alternative to the experimental determination of F 12 . If the user has a value for use with F 12 in the theory, the
same may be input in the MAT8 Bulk Data entry; otherwise, F 12 in the above theory will be set to 0.0. The left-hand side
of the above equation will be evaluated as the failure index by this theory.
The failure index of bonding material will be calculated as the maximum interlaminar shear stress divided by the allowable
bonding stress.
The failure indices for all the plies will be written into OEFIT (Output Element Failure Index Table) and output if stresses
are requested. The failure index for the element is the largest value of the failure indices for all plies of the element.

Main Index
 Chapter 4: Material Properties 283
Nonuniform Material and Composite Properties for Shells

Maximum Strain Theory (Failure Theory ID = “STRN”)

The midplane strains and curvatures are available in the element coordinate system. From these, the stresses and strains in
each individual lamina along the fiber direction and transverse direction can be calculated quite easily. The STRAIN Case
Control command is used to request the output of lamina strains.
The maximum strain criteria has no strain interaction terms. The strain allowables specified on the MAT8 entry for each
lamina include

X t , X c Allowable strains in tension and compression, respectively, in the longitudinal direction.

Y t , Y c Allowable strains in tension and compression, respectively, in the transverse direction.
S Allowable strain for inplane shear.

The Failure Index is calculated using

1 
----- (or -----1- if  1 compressive)
Xt Xc
2 
----- (or ----2- if  2 compressive)
Xt Yc

and
 12
-------
S

1  2  12 
i.e., the Failure Index = MAX  ----- ,  ----- ,  ---------- .
 X  Y  S 
In addition, the user needs to be informed which mode of Failure Index is critical; i.e., longitudinal (1), transverse (2), or
shear (12). Thus in the output for Failure Indices, the mnemonic 1, 2 or 12 is printed alongside the FP value to indicate
the critical direction. There is no change in the way the Failure Index is calculated for interlaminar shear stresses.
It is quite possible that for the maximum strain theory, the user may wish to specify lamina stress allowables instead of strain
allowables on the MAT8. This can be done by leaving the STRN field blank on the MAT8.
For this case, the failure indices are calculated using
 1  E 11  1  E 11
------------------ (or -----------------
- if  1 compressive)
Xt Xc
 2  E 22  2  E 22
------------------ (or -----------------
- if  2 compressive)
Yc Yc

Main Index
 284 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

and
 12  G 12
-----------------------
-
S

1 11  E 2  E
22 12  G 12
i.e., the Failure Index = MAX  ------------------ ,  ------------------ ,  ------------------------
 X   Y   S 

Strength Ratios
For laminated composites, Strength Ratio (SR) is a direct failure indicator compared to Failure Index (FI) which indicates
only if failure had occurred. Generally Strength Ratio is defined as:
Strength Ratio (SR) = Allowable Stress / Calculated Stress
For example a SR = 1.2 indicates that the applied loads can be increased by 20% before failure occurs. A FI = 0.8 indicates
that failure has not occurred and does not indicate 20% safety margin. Therefore the SR is a much more practical design
indicator for both analysis and strength-criteria based design.
This new result quantity is available for all solutions sequences that support composites data recovery.
A new parameter -- ‘PARAM,SRCOMPS,YES/NO’ requests Strength Ratio (SR) output. SR Output requires specification
of a failure theory and allowable stress/strain values. Note that SRs will be computed for plies with an output request
(PCOMP-SOUTi field).
For Maximum Stress/Strain and Transverse Stress theories, the Strength Ratio (SR) is simply the inverse of the FI. For the
quadratic failure theories such as Hill, Hoffman and Tsai-Wu the SR is calculated by solving the quadratic equation with
the FI set to unity and replacing the applied stress with the SR  applied stress.
Below the SR expression for each Failure Criteria is defined.

HILL Failure Criteria

The HILL failure criteria is defined as:
2 2 2
  1   2   1  2   12
FI =  ------ +  ------ –  ------------ +  --------
 X 2  Y 2  X 2   S 2 

where:

X = XT if  1 = Tensile
X = XC if  1 = Compressive
Y = YT if  2 = Tensiles

Main Index
 Chapter 4: Material Properties 285
Nonuniform Material and Composite Properties for Shells

Y = YC if  2 = Compressive
1 = Ply longitudinal stress
2 = Ply transverse stress
 12 = Ply shear stress

For the product term:

 1  2
---------------
2
X
where:

X = X T if  1 and  2 are of the same sign.

X = X C if  1 and  2 are of different signs.

The FI expression can be redefined in terms of a strength ratio. The roots of the resulting quadratic expression are the reserve
factors, the first one for the stresses as applied and the second for when the sign of all the components of stress are reversed.
The FI expression is re-arranged with the applied stress replaced by a ratio SR times the applied stress and equated to a FI
of unity, the SR can now be determined.
2
2 2 2 2 2 2  SR  1  2 
1.0 =   SR  1   1  X  +   SR  2    1  Y   – ----------------------------
2
X 
2 2
 SR  12 
+ ------------------------
2
Y 

2 2 2 2 2 2 2 2
   1  X  +   2  Y  –   1  2  X  +   12  Y  SR – 1 = 0
2
aSR + bSR + c = 0
from the general solution of a quadratic equation,

2 2
– b + b – 4ac- – b – b – 4ac
SR 1 = --------------------------------------
and SR 2 = --------------------------------------
2a 2a
For the HILL criteria,
2 2 2 2 2 2 2
a =   1  X  +   2  Y  –   1  2   X +   12  Y 

Main Index
 286 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

b = 0.0

c = –1

2
– b + b – 4ac
SR 1 = ---------------------------------------
2a

– 4  a  1.0 2 a
SR 1 = ----------------------------- = --------------
2a 2a
1
SR 1 = -------
a
1
SR 2 = – -------
a
where
2 2 2 2 2 2 2
a =   1  X  +   2  Y  –   1  2  X  +   12  Y  = FI

Therefore for the HILL quadratic failure criteria the SR can be directly derived from the FI:
1
SR = ---------
FI
HOFFMAN Failure Criteria

Unlike the Hill criteria the Hoffman failure criteria takes into account the difference in tensile and compressive strengths
in the longitudinal and transverse directions of the ply.
2 2 2 2
FI =  1   X T  X C   1 +  1   Y T  Y C   2 +  1   X T  X C     1  2  +   12  S 
+   1  X T  –  1  X C   1 +   1  Y T  –  1  Y C   2

defining:

F1 =  1  XT  –  1  XC 
F2 =  1  YT  –  1  YC 
F 11 = 1   X T  X C 
F 22 = 1   Y T  Y C 

Main Index
 Chapter 4: Material Properties 287
Nonuniform Material and Composite Properties for Shells

2
F 66 = 1   S 
2 2 2
FI = F 11  1 + F 22  2 – F 11   1  2  + F 66  12 + F 1  1 + F 2  2

Now introduce applied stress as a ratio R times applied stress equated to a FI of unity.
2 2 2 2 2 2 2
1.0 = F 11 SR  1 + F 22 SR  2 – F 11 SR   1  2  + F 66 SR  12 + F 1 SR 1 + F 2 SR 2

Arrange in quadratic form

2 2 2 2
 F 11  1 + F 22  2 – F 11   1  2  + F 66  12 R +  F 1  1 + F 2  2 R – 1 = 0

2 2
– b + b – 4ac – b – b – 4ac
SR 1 = --------------------------------------- and SR 2 = --------------------------------------
2a 2a
where:
2 2 2
a =  F 11  1 + F 22  2 – F 11   1  2  + F 66  12 
b =  F1 1 + F2 2 
c = – 1.0

This is now a quadratic that does not simplify in the same way as the Hill criteria so there is no direct relationship between
FI and SR. The roots of the equation have to be determined to obtain the SR.
TSAI-Wu Failure Criteria

The Tsai-Wu failure criterion is very similar to the Hoffman criteria but has a different value for the coefficient associated
with the interaction of the direct stresses. In the Hoffman criterion it is the same as the coefficient associated with the direct
longitudinal stresses.
For the Tsai-Wu criterion:
2 2 2
FI = F 11  1 + F 22  2 + 2F 12  1  2 + F 66  12 + F 1  1 + F 2  2

where:

F1 =  1  XT  –  1  XC 
F2 =  1  YT  –  1  YC 
F 11 = 1   X T  X C 
F 22 = 1   Y T  Y C 

Main Index
 288 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

2
F 66 = 1   S 
F 12 = See (p. 281) Tensor Polynomial Theory of Tsai-Wu for a discussion of
F12 and for the default settings.

Now introduce applied stress as a ratio R times applied stress equated to a FI of unity.
2 2 2 2 2 2
1.0 = F 11 SR  1 + F 22 SR 2 + 2F 12 SR  1  2 + F 66 SR  12 + F 1 SR 1 + F 2 SR 2

Arrange in quadratic form

2 2 2 2
 F 11  1 + F 22  2 + 2F 12  1  2 + F 66  12 SR +  F 1  1 + F 2  2 SR – 1 = 0

2 2
– b + b – 4ac – b – b – 4ac
SR 1 = --------------------------------------- and SR 2 = --------------------------------------
2a 2a
where:
2 2 2
a =  F 11  1 + F 22  2 + 2F 12  1  2 + F 66  12 
b =  F1 1 + F2 2 
c = – 1.0

This is a quadratic equation that does not simplify in the same way as the Hill criteria so there is no direct relationship
between FI and SR. The roots of the equation have to be determined to obtain the SR.
Maximum Stress (Strain)

Failure Index FI = Calculated Stress / Allowable Stress

By substituting the Allowable stress by RF*Allowable Stress and setting FI to unity
SR = 1   FI 
Transverse Shear Stress

For this case the Strength Ratio is simply the inverse of the failure index.

“LAM” Options
The PCOMP Bulk Data entry is used to provide input data for the properties of an n-ply composite material laminate. As
such, it is widely used in aerospace applications where a composite structure is built up from a number of discrete sheets of
material, such as a carbon-epoxy fiber.
1. The SMEAR options to the LAM field smears the ply effects while ignoring stacking sequence.

Main Index
 Chapter 4: Material Properties 289
Nonuniform Material and Composite Properties for Shells

2. The SMCORE option to the LAM field enables the modeling of a skin/core combination where the ply effects are
smeared in the skin, the core inertial terms are treated exactly and the core stiffness terms are ignored.
MSC Nastran develops mass and stiffness data from PCOMP input in a two-step process. First, the PCOMP input data
are considered together with the material data referenced by MIDi entries to produce PSHELL/MAT2 combinations that
will lead to the required stiffness results and then this spawned data are used in the actual stiffness and mass calculations.
The spawned PSHELL has four MIDis, identifying the MAT2s to be used for membrane, bending, transverse shear and
membrane-bending coupling. The SMEAR and SMCORE options use these MIDis in the following manner:

SMEAR
The spawned PSHELL has MID1=MID2 with MID3, MID4, 12I/T**3, and the TS/T fields set as blanks. This results in
a bending term given as
2
T A
 D  = --------------
12

SMCORE
Computation of the membrane and bending stiffness matrices is performed using the following derivation. Note that
membrane-bending coupling is ignored.
Definitions:

t face = T 1 + T 2 +  + T N – 1 (total thickness of both face sheets)

t core = T N (total core thickness)

Membrane Stiffness Matrix:

 A  is computed using method utilized by LAM=BLANK assuming that core stiffness (layer N) is zero.
Face Sheet Properties:
–1
–  A  12
 xy = -----------------
–1
 A  11

–1
–  A  12
 yx = -----------------
–1
 A  22

A 11  1.0 –  xy  yx 
E x = --------------------------------------------
t face

Main Index
 290 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

A 22  1.0 –  xy  yx 
E y = --------------------------------------------
t face

A 33
G xy = ----------
t face

Moment of inertia:

t face 2
t face  t core + ---------- 3
t face
 2 
I xx = I yy = -----------------------------------------------
- + ----------
4 48
3
Note, this collapses to t face  12 if t core is zero.

Bending Stiffness Matrix:

E x I xx
D 11 = -----------------------------------
 1.0 –  xy  yx 

E y I yy
D 22 = --------------------------------
 1. –  xy  yx 

D 11  xy E y
D 12 = D 21 = ------------------------
Ex

D 33 = G xy I xx

D 13 = D 31 = D 23 = D 32 = 0

Input

The format of the PCOMP entry is unchanged, but the description of the LAM field is expanded and an additional Remark
is added.

Field Contents
LAM Laminate Options (Character or Blank, default=Blank). See Remarks 3. and 4.
“Blank” All plies must be specified and all stiffness terms are developed

Main Index
 Chapter 4: Material Properties 291
Nonuniform Material and Composite Properties for Shells

Field Contents
“SYM” Only plies on one side of the element centerline are specified. The plies are numbered
starting with 1 for the bottom layer. If an odd number of plies are desired, the center ply
thickness (T1 or TN) should be half the actual thickness.
“MEM” All plies must be specified, but only membrane terms (MID1 on the derived PSHELL
entry) are computed.
“BEND” All plies must be specified, but only bending terms (MID2 on the derived PSHELL entry)
are computed.
“SMEAR” All plies must be specified, stacking sequence is ignored MID1=MID2 on the derived
PSHELL entry and MID3, MID4 and TS/T and 12I/T**3 terms are set to zero).
“SMCORE” All plies must be specified, with the last ply specifying core properties and the previous
plies specifying face sheet properties. The stiffness matrix is computed by placing half the
face sheet thicknesses above the core and the other half below with the result that the
laminate is symmetric about the mid-plane of the core. Stacking sequence is ignored in
calculating the face sheet stiffness.

Remarks
3. The SYM option for the LAM option computes the complete stiffness properties while specifying half the plies. The
MEM, BEND, SMEAR and SMCORE options provide special purpose stiffness calculations. SMEAR ignores
stacking sequence and is intended for cases where this sequence is not yet known, stiffness properties are smeared.
SMCORE allows simplified modeling of a sandwich panel with equal face sheets and a central core.
4. Element output for the SMEAR and SMCORE options is produced using the PARAM NOCOMPS –1
methodology that suppresses ply stress/strain results and prints results for the equivalent homogeneous element.

Output
Standard element response data are recovered for elements that employ the new LAM options with the key provision that
the SMEAR and SMCORE options results in element results that correspond to the equivalent homogenous element and
not the ply-by-ply results. This is equivalent to using PARAM NOCOMPS –1 and only the homogeneous results are
produced regardless of the user input specification of the NOCOMPS parameter.

Guidelines for Design Optimization

Special procedures must be employed when using the PCOMP with the SMEAR and SMCORE options in an automated
design task in SOL 200. The FSD option in SOL 200 does not support the SMEAR or SMCORE options since the resizing
of a ply layer in FSD is driven by a response in that layer. The SMEAR/SMCORE options do not produce layer responses,
therefore, FSD resizing is precluded. For the standard mathematical programming of SOL 200, composite stress, strain and
force results are not available for the PCOMP with SMEAR or SMCORE. Instead, the homogeneous element results must
be invoked. This is done by using the standard STRESS, STRAIN, FORCE response types (as opposed to the CSTRESS,
CSTRAIN, CFAILURE response types) and the PTYPE can be set to either PCOMP or PSHELL. The PID for the
PSHELL is identical to that of the PCOMP. The rules for the design of properties are not changed by these new options

Main Index
 292 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

in that the user can invoke a lamina thickness or orientation on the PCOMP entry from a DVPREL1 or DVPREL2.
Invoking the thickness of the equivalent PSHELL that is derived from a PCOMP is not allowed and produces a fatal error.

Temperature-Dependent Properties
Theory
Temperature Distribution

The temperature through the shell thickness is assumed to be linear:

t t
T      = T     +   T , – ---    ---
2 2

where T     is the membrane temperature defined by the weighted average of the grid temperatures from the TEMP
and/or TEMPD Bulk Data entries, or the element temperature from the TEMPP1 Bulk Data entry; and T is the
temperature gradient from the TEMPP1 Bulk Data entry.
Currently the smeared laminate properties are calculated using the reference temperature from the PCOMP Bulk Data
entry, which has no gradient defined through the thickness. This implies that all the plies are at the same temperature, which
is not a valid assumption in certain cases. For a more accurate representation of thermal strains and temperature-dependent
properties, use SOL 400.
The temperature at the midplane of the i-th ply will be calculated as:

 i–1 
1
T i = T + z 0 +  t j + --- t i  T

 2 
 j=1 

where T is the reference plane temperature; z 0 is the bottom surface offset from the reference plane; and t i are the ply
thicknesses defined on the PCOMP Bulk Data entry. This temperature will be used to calculate the temperature-dependent
material properties of each ply.
Coefficient of Thermal Expansion

Secant Approach

The coefficient of thermal expansion (in one dimension) is defined as the normalized change in length for a given change
in temperature:
1 u
 = --- --------
L T
such that the thermal strain is given by:

Main Index
 Chapter 4: Material Properties 293
Nonuniform Material and Composite Properties for Shells

 = T
which is the secant approach. This can be computed in several ways, depending on the case control and material
dependence.
For linear analysis using neither TEMP(MAT) nor TEMP(INIT):
 =   T load – T ref 

For linear analysis using TEMP(INIT) with no MATTi:

 =   T load – T init 

For linear analysis using TEMP(MAT) with MATTi:

 =  mat  T load – T ref 

For linear analysis using TEMP(INIT) with MATTi:

 =  init  T load – T init 

For nonlinear analysis using TEMP(INIT) with MATTi (TEMP(INIT) is required and TEMP(MAT) is not allowed), the
expression is more complex, in an attempt to better approximate the temperature-dependent properties:
 =  load  T load – T ref  –  init  T init – T ref 

Integral Approach

This secant approximation is adequate for small changes in temperature; however, for temperature-dependent properties, a
derivative approximation is more accurate:
1 u
  T  = --- ------
L T
such that the thermal strain is given by:
T load
 = T init
  T  dT

which is the integral approach.

Temperature dependence for each material property, such as the coefficient of thermal expansion, may be defined as a linear
function or a power series, and may include offsets and scale factors. These are input on the TABLEMi Bulk Data entries,
and may be integrated analytically for each interval of the table.

Main Index
 294 MSC Nastran Reference Guide
Nonuniform Material and Composite Properties for Shells

Input
For the temperature-dependent composites, two parameters have been added. The COMPMATT parameter controls
whether the smeared laminate properties are updated with the temperature-dependent material properties of the individual
plies. It has a default value of NO and must be changed to YES to invoke the more accurate integral approach. The
EPSILONT parameter controls whether the thermal strain is computed using an integral or a secant method. It has a
default value of SECANT and must be changed to INTEGRAL to invoke the enhancement.
The MATT8 Bulk Data entry is used for temperature-dependent shell orthotropic material properties, and thermal strains
may be input directly by using a negative TABLEMi ID for the coefficients of thermal expansion on the MATTi Bulk Data
entry. All of the other bulk data entries and case control remain the same.

Output
For the strains, the thermal strains will be calculated using the smeared coefficients of thermal expansion at the current load
temperature, in addition to using the appropriate secant or integral method. For the stresses and forces, the smeared material
properties will be calculated at the proper temperature. For the ply strains and stresses, the thermal strain will be calculated
using the coefficients of thermal expansion at the current load temperature and appropriate method for the specific ply, and
the ply properties will be calculated at the proper temperature, rather than using the reference temperature on the PCOMP
Bulk Data entry.
In addition, if the parameter COMPMATT is set, the composite elements will be included in the nonlinear stress/strain
output.

Guidelines and Limitations

Composite capabilities are currently available with the QUAD4, QUAD8, TRIA3, and TRIA6 h-elements. Therefore, only
the QUAD4 and TRIA3 element types will be supported. (Note that although smeared laminate properties may be applied
to other two-dimensional elements as well, ply-by-ply postprocessing is not performed on those elements.)
These capabilities are not implemented for SOL 101, Linear Statics, nor SOL 129, Nonlinear Transient Response; use SOL
400.

Examples
(TDCRN_*.DAT)
In order to illustrate the new enhancements, a very simple problem will be solved for each permutation of the parameters.
The problem consists of a simply-supported cantilever beam, ten units long by one unit wide, undergoing a uniform free
expansion from T init = 100 to T load = 200 with T ref = 50 , as shown in the following figure:

Main Index
 Chapter 4: Material Properties 295
Nonuniform Material and Composite Properties for Shells

x 1

10
Two independent versions of the model will comprise each test file: one will reference a PSHELL, and the other will
reference a PCOMP with a single ply of the same thickness and orientation as the shell. The material is isotropic, although
a MAT8/MATT8 will be used to demonstrate the new MATT8.
The coefficient of thermal expansion is temperature-dependent, as shown in the following table:

Temperature (T) CTE ()

T ref 50 0.0005

T init 100 0.001

T load 200 0.005

Most of the models will be run in SOL 106, Nonlinear Statics, though one will be run in SOL 101, Linear Statics, for
comparison purposes. (Note again that the enhancements are not implemented in SOL 101.)
The six example problems are listed in the following table, along with the thermal strains computed using the equations
given above. For the final problem, the thermal strains, marked with an asterisk, are input directly on the TABLEM4 entry
as the quadratic integral of the linear property variation of the coefficient of thermal expansion.

Parameters Thermal Strain (t)

Solution
Sequence COMPMATT EPSILONT PSHELL PCOMP
TDCRN_L 101 n/a n/a 0.10 0.05
TDCRN_NS 106 no secant 0.70 0.05
TDCRN_YS 106 yes secant 0.70 0.70
TDCRN_NI 106 no integral 0.30 0.05
TDCRN_YI 106 yes integral 0.30 0.30
TDCRN_YIE 106 yes integral 0.30* 0.30*

Main Index
 296 MSC Nastran Reference Guide
Hyperelastic Material

Results from the six example problems are listed in the following table, along with the axial displacement at the final load
step.

Parameters Axial Displacement (ux)

Solution
Sequence COMPMATT EPSILONT PSHELL PCOMP
TDCRN_L 101 n/a n/a 1.0 0.5
TDCRN_NS 106 no secant 7.0 0.5
TDCRN_YS 106 yes secant 7.0 7.0
TDCRN_NI 106 no integral 3.0 0.5
TDCRN_YI 106 yes integral 3.0 3.0
TDCRN_YIE 106 yes integral 3.0* 3.0*

The axial displacements consist of the thermal strains multiplied by the length, as expected; the strains are those given above;
and the stresses are numerically zero.

Hyperelastic Material
This section describes the MATHP Bulk Data entry, which defines hyperelastic analysis of rubber-like material used in SOL
106 and SOL 129. For more complex material modeling of elastomers, use SOL 400.
The generalized Mooney-Rivlin strain energy function may be expressed as follows:

NA ND
i j 2i
U  J, I 1 , I 2  =  Aij  I 1 – 3   I 2 – 3  +  Di  J – 1 – AV  T – T0  
i+j = 1 i=1
A00 = 0

where I 1 and I 2 are the first and second distortional strain invariants, respectively; J = det F is the determinant of
the deformation gradient; and 2D1 = K and 2  A10 + A01  = G at small strains, in which K is the bulk modulus
and G is the shear modulus. The model reduces to a Mooney-Rivlin material NA = 1 and to a Neo-Hookean material
if NA = 1 and A01 = 0.0 . For Neo-Hookean or Mooney-Rivlin materials no continuation entry is required. T is the
current temperature and T 0 is the initial temperature.

Conventional Mooney-Rivlin and Neo-Hookean materials are fully incompressible. Full incompressibility is not presently
available but may be simulated with a large enough value of D1. A value of D1 higher than 10 3   A10 + A01  is,
however, not recommended.

Main Index
 Chapter 4: Material Properties 297
Hyperelastic Material

Aij (material constants related to distortional deformation) and Di (material constants related to volumetric deformation)
are obtained from least squares fitting of experimental data. One or more of four experiments (TAB1 to TAB4) may be used
to obtain Aij. Di may be obtained from pure volumetric compression data (TABD). If all TAB1 through TAB4 are blank,
Aij must be specified by the user. Parameter estimation, specified through any of the TABLES1 entries, supersedes the
manual input of the parameters.
If ND = 10 and a nonzero value of D1 is provided or is obtained from experimental data in TABD, then the parameter
estimation of the material constants Aij takes compressibility into account in the cases of simple tension/compression,
equibiaxial tension, and general biaxial deformation. Otherwise, full incompressibility is assumed in estimating the material
constants.

Reference
1. Narayanaswami, R., and H. M. Adelman, “Evaluation of the Tensor Polynomial and Hoffman Strength Theories
for Composite Materials,” Journal of Composite Materials, Vol. II, 1977, p. 366.

Main Index
 Chapter 5: Applied Loads
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

5 Applied Loads

 Static Loads 299


Point Loads 299

Distributed Loads 299

Inertia Loads 300

Thermal and Enforced Element Deformation Loads 300

Loads Due to Enforced Motion 304

Frequency-Dependent Loads 306

Time-Dependent Loads 307

Simplified Static Loading Data in Dynamic Analysis 308

Combining of Loads 310

Summary of Nonlinear Static Loads 311

Summary of Nonlinear Transient Loads 316

Main Index
 Chapter 5: Applied Loads 299
Static Loads

Static Loads
In MSC Nastran, static loads are applied to geometric and scalar points in a variety of ways, including:
 Loads applied directly to grid points
 Pressure loads on surfaces
 Distributed and concentrated loads on CBAR, CBEAM, and CBEND elements
 Gravity loads
 Centrifugal loads due to steady rotation
 Tangential loads due to angular acceleration
 Loads resulting from thermal expansion
 Loads resulting from enforced deformations of a structural element
 Loads resulting from enforced displacements at a grid point
Additional information on static loads is given in Section 3.6 of The NASTRAN Theoretical Manual. Any number of load
sets can be defined in the Bulk Data Section. However, only those sets selected in the Case Control Section, as described in
Case Control Commands, will be used in the problem solution. The manner of selecting each type of load is specified on the
associated Bulk Data entry description in the Bulk Data Entries in the MSC Nastran Quick Reference Guide.

Point Loads
The FORCE entry is used to define a static load applied to a geometric grid point in terms of components defined by a
local coordinate system. The orientation of the load components depends on the type of local coordinate system used to
define the load. The directions of the load components are the same as those indicated on Figure 2-2 of Grid Point and Coordinate
System Definition for displacement components. The FORCE1 entry is used if the direction is determined by a vector
connecting two grid points, and a FORCE2 entry is used if the direction is specified by the cross product of two such
vectors. The MOMENT, MOMENT1 and MOMENT2 entries are used in a similar fashion to define the application of
a concentrated moment at a geometric grid point. The SLOAD entry is used to define a load at a scalar point. In this case,
only the magnitude is specified since only one component of motion exists at a scalar point.
The FORCEAX and MOMAX entries are used to define the loading of specified harmonics on rings of conical shell
elements. FORCE and MOMENT entries may be used to apply concentrated loads or moments to conical shell elements,
providing that such points have been defined with a POINTAX entry.

Distributed Loads
The PLOAD1 entry is used to define either distributed or concentrated loads on CBAR and CBEAM elements. This entry
is only used to define loads at intermediate points on CBAR and CBEAM elements. Applied loads at end points are defined
with either FORCE or MOMENT entries. The CBARAO entry is used to define intermediate points on CBAR elements
for which stress and/or force output is desired. The PLOAD1 entry is also used to define only distributed loads for the
CBEND element. The distributed load is always applied along the entire length of the element and does not define loads

Main Index
 300 MSC Nastran Reference Guide
Inertia Loads

at intermediate points on the CBEND. The PLOAD1 entry cannot be used to define concentrated loads at intermediate
stations on the CBEND element.
Pressure loads on three-node triangular and four-node quadrilateral elements are defined with a PLOAD2 entry. The
positive direction of the loading is determined by the order of the grid points on the element connection entry, using the
right hand rule. The magnitude and direction of the load is automatically computed from the value of the pressure and the
coordinates of the connected grid points. The load is applied to the connected grid points. The PLOAD entry is used in a
similar fashion to define the loading of any three or four grid points regardless of whether they are connected with two-
dimensional elements. The PLOAD4 entry may be used with CQUAD4, CQUAD8, CQUADR, CTRIA3, CTRIA6, and
CTRIAR elements. Since the surface hyperelastic elements CQUAD4, CQUAD8, CQUAD, CTRIA3, and CTRIA6 are
plane strain elements, no pressure load may be applied on these elements.
The PRESAX entry is used to define pressure loading on the conical shell element. The PLOADX1 entry is used to define
pressure loading on the CTRIAX6 element and on the hyperelastic CTRIAX and CQUADX elements.
Pressure loads on the CHEXA, CPENTA, and CTETRA elements are defined with the PLOAD4 entry. The magnitude
and direction of the equivalent grid point forces are automatically computed using the isoparametric shape functions of the
element to which the load has been applied.

Inertia Loads
The GRAV entry is used to specify a gravity load (or any other type of acceleration) by providing the components of the
gravity vector in any defined coordinate system. The gravity load is obtained from the gravity vector and the mass matrix.
The gravitational acceleration is not calculated at scalar points. The user is required to introduce gravity loads at scalar
points directly.
The RFORCE entry is used to define a static loading condition due to angular velocity and/or acceleration. These loads are
specified by the designation of a grid point that lies on the axis of rotation and by the components of rotational velocity or
angular acceleration in any defined coordinate system. There are two methods to compute these loads. In the first method,
the mass matrix is regarded as pertaining to a set of distinct rigid bodies connected to grid points. Deviations from this
viewpoint, such as the use of scalar points or the use of mass coupling between grid points, can result in errors. The second
method allows mass coupling, but does not allow offsets. The second method is the only method used for angular
acceleration.

Thermal and Enforced Element Deformation Loads

Temperatures may be specified for selected elements. The temperatures for CROD, CBAR, CBEAM, CBEND,
CONROD, or TUBE elements are specified on the TEMPRB Bulk Data entry. This entry specifies the average temperature
on both ends and, in the case of the CBAR, CBEAM and CBEND elements, is used to define temperature gradients over
the cross section. Temperatures for two-dimensional plate and membrane elements are specified on a TEMPP1, TEMPP2,
or TEMPP3 Bulk Data entry. The user-defined average temperature over the volume is used to produce in-plane loads and
stresses. Thermal gradients over the depth of the bending elements, or the resulting moments, may be used to produce
bending loads and stresses.

Main Index
 Chapter 5: Applied Loads 301
Thermal and Enforced Element Deformation Loads

If no thermal element data is given for an element, the temperatures of the connected grid points given on the TEMP,
TEMPD, or TEMPAX entries are simply averaged to produce an average temperature for the element. The thermal
expansion coefficients are defined on the material definition entries. Regardless of the type of thermal data, if the material
coefficients for an element are temperature dependent by use of the MATTi entry, they are always calculated from the
“average” temperature of the element. The presence of a thermal field does not imply the application of a thermal load. A
thermal load will not be applied unless the user makes a specific request in the Case Control Section.
Enforced axial deformations can be applied to CROD, CBAR, and CBEAM elements. Enforced deformations, for example,
are useful in the simulation of misfit or misalignment in engineering structures. As in the case of thermal expansion, the
equivalent loads are calculated by separate subroutines for each type of structural element and are applied to the connected
grid points. The magnitude of the axial deformation is specified on a DEFORM entry.

Temperature Loads
If temperature loads are present in one or more of the subcases, the thermal strain must be taken into account in the
calculation of element stresses and forces. That is, a body subjected to a temperature field will deform in a stress free manner
in the absence of constraints. However, if the deformation is resisted through boundary conditions and/or the presence of
applied loads, the total actual strain, at a point must be reduced by the thermal strain,   T – T o  , in order to obtain the
stress at the point. For example, the stress-strain relation for a rod that undergoes thermal expansion and also sustains an
axial force is written in the form

 –   T – T o  = ---
E
Element stresses and forces are calculated on an element-by-element basis from a knowledge of the displacement vector and
the temperature field; therefore, in a SUBCOM or a SYMCOM, the user must supply a definition of the temperature field
whenever element stresses and forces are required.
Consider the earlier example, with the modification that a temperature load is specified in SUBCASE 2.

Main Index
 302 MSC Nastran Reference Guide
Thermal and Enforced Element Deformation Loads

.
.
.

SUBCASE 1
LOAD=2
SUBCASE 2
TEMP(LOAD)=4
SUBCASE 10
LOAD=10
SUBCOM 20
SUBSEQ=.5,1.,1.6
TEMP(LOAD)=4
.
.
.

In this example, the multiplier for SUBCASE 2 is unity so that the temperature distribution required to calculate element
forces and stresses for the combined loading case is precisely TEMP(LOAD)=4. If, on the other hand, the SUBSEQ entry
is modified as indicated below, an additional TEMP(LOAD) must be defined.

.
.
.

SUBCASE 1
LOAD=1
SUBCASE 2
TEMP(LOAD)=4
SUBCASE 10
LOAD=10
SUBCOM 20
SUBSEQ=.5,.75,1.6
TEMP(LOAD)=40

Here, TEMP(LOAD)=40 must reflect a temperature distribution of the form

 T 40 – T o  = .75  T 4 – T o 

T 40 = T o + .75  T 4 – T o 

This new TEMP(LOAD) must be supplied by the user through standard Bulk Data entries such as TEMPD, TEMP,
TEMPP1, and TEMPP3.

Main Index
 Chapter 5: Applied Loads 303
Thermal and Enforced Element Deformation Loads

In general, with n temperature loading conditions in the subcases immediately preceding a SUBCOM, the temperature
distribution that must be defined under the SUBCOM if element force or stress output is desired is of the form
n
T SUBCOM = T o +  ai  Ti – To 
i=1

where the coefficients a i are defined on the SUBSEQ entry under the SUBCOM in question. The reference temperature
T o , defined on MATi Bulk Data entries, can conveniently be defined as zero in order to simplify the calculation of the
temperature distribution for the combined loading condition of the SUBCOM.
The same approach is used with combined loading condition defined through a SYMCOM Case Control entry. However,
the use of SYM subcases and SYMCOM entries implies the use of the concepts of reflective symmetry, in which case it is
strongly recommended that the DIH option of cyclic symmetry be used (Cyclic Symmetry). The appropriate boundary
conditions are generated internally in MSC Nastran, and the user simply defines the actual temperature distribution on
each portion of the structure without regard to decomposing the temperatures and/or mechanical loads into symmetric and
antisymmetric components.

Element Deformation
The DEFORM Bulk Data entry enables the user to define enforced axial deformations for the CBAR, CBEAM, CROD,
and CTUBE elements. The nature of this enforced deformation is analogous to a turnbuckle that may be adjusted to
provide a desired axial deformation in an element. The problem is analogous to the temperature problem discussed in the
previous section and, for example, the appropriate stress-strain relation for a rod element of length that undergoes an
enforced axial deformation,  is given by the following equation:

 
 – -- = ---
l E
Consider the final example of the previous section.

Main Index
 304 MSC Nastran Reference Guide
Loads Due to Enforced Motion

.
.
.

SUBCASE 1
LOAD=1
SUBCASE 2
DEFORM=5
SUBCASE 10
LOAD=10
SUBCOM 20
SUBSEQ=.5,.75,1.6
DEFORM=50
.
.
.

Here DEFORM = 50 must reflect a set of enforced deformations of the form

 50 = .75 5

For n DEFORM loading conditions in the subcases that immediately precede a SUBCOM, the DEFORM loading
condition that must be defined under the SUBCOM, if element force and stress output is desired, is of the form
n
 SUBCOM =  ai i
i=1

Loads Due to Enforced Motion

Statics
Enforced zero displacements may be specified on GRID, SPC, or SPC1 entries. Zero displacements that result in nonzero
forces of constraint are usually specified on SPC or SPC1 entries. If GRID entries are used, the constraints become part of
the structural model and modifications cannot be made at the subcase level.
Nonzero enforced displacements may be specified on SPC or SPCD entries. The SPC entry specifies both the component
to be constrained and the magnitude of the enforced displacement. The SPCD entry specifies only the magnitude of the
enforced displacement. When an SPCD entry is used, the component to be constrained must be specified on either an SPC
or SPC1 entry. The use of the SPCD entry avoids the decomposition of the stiffness matrix when changes are only made
in the magnitudes of the enforced displacements.
The equivalent loads resulting from enforced displacements of grid points are calculated by the program and added to the
other applied loads. The magnitudes of the enforced displacements are specified on SPC entries (SPCAX in the case of

Main Index
 Chapter 5: Applied Loads 305
Loads Due to Enforced Motion

conical shell problems) in the global coordinate system. The application of the load is automatic when the user selects the
associated SPC set in the Case Control Section.
The SPCD is selected by a load request in the Case Control Section.

Dynamics
Enforced motion in earlier versions of MSC Nastran used a nonzero value in the TYPE field (field 5) on the TLOAD1 and
TLOAD2 entries to specify whether an enforced displacement, velocity, or acceleration were to be applied using the large
mass method. Although these integer values are still allowed, the field has now been generalized to permit character input,
and a similar TYPE field has been included in field 8 of the RLOAD1 and RLOAD2 entries to support enforced motion
in frequency response analysis. The new TYPE field for these dynamic loading/enforced motion entries now reads as
follows:

Value Excitation Function

0, or LOAD Applied load (Default)
1, or DISP Enforced displacement using large mass or SPC/SPCD data
2, or VELO Enforced velocity using large mass or SPC/SPCD data
3, or ACCE Enforced acceleration using large mass or SPC/SPCD data

The character fields may be shortened to as little as a single character on input, if desired.
In order to allow maximum input flexibility and to support existing input data files that had used the large mass method,
the decision regarding whether to use the older, large mass method, or the newer direct enforced motion will be made based
on a search for SPC/SPC1/SPCD data as follows: If the TYPE field on the TLOAD1/2 entries indicates an enforced
motion, the EXCITEID field will first be assumed to point to SPC-type data. If not present, MSC Nastran will then assume
the excitation is enforced motion using large masses and will then look for DAREA and various static loading data, just as
in the case of applied loads. RLOAD1 and RLOAD2 entries will only look for SPC-type data in the case of enforced
motion. For all entries (TLOAD1/2, RLOAD1/2), if the TYPE field indicates an applied load, the program will search only
for static loading data.
With this feature, enforced motion data is supplied via the SPC/SPC1/SPCD Bulk Data entries. (Note that components
specified in SPCD data must be also be referenced on SPC or SPC1 entries.) Enforced motion could be defined solely with
SPC data (without the use of SPCD’s), but such usage is discouraged in favor of SPC/SPC1 data in connection with
SPCD’s. In this way, SPCD entries are selected via the EXCITEID field in the dynamic load data, while the SPC/SPC1
sets are used to identify the constrained displacement set in the Case Control Section. With SPC data alone, enforced
motion is applied solely via the Case Control Section. The entries referenced by the EXCITEID field on the TLOAD1/2
or RLOAD1/2 entries do not have to exist. This method of specifying enforced motion is less elegant and somewhat more
difficult to interpret.

Procedure for Using the Enforced Motion Feature

In summary, the procedure for applying enforced motion in a dynamic analysis:

Main Index
 306 MSC Nastran Reference Guide
Frequency-Dependent Loads

1. Specifies the appropriate type of enforced motion to be applied via the TYPE field in RLOAD1/RLOAD2 (field 8)
or TLOAD1/TLOAD2 (field 5) Bulk Data entry, as appropriate.
2. Defines the desired enforced motion using the SPCD Bulk Data entry. The set IDs of these SPCD data must match
the IDs appearing in the EXCITEID fields of the corresponding dynamic load data in step 1.
3. Ensures that the components referenced in the SPCD Bulk Data entry above are also specified in the SPC1 Bulk
Data entry and this SPC1 is also selected in the Case Control Section.

Frequency-Dependent Loads
A discussion of frequency response calculations is given in Frequency Response Analysis (p. 123) in the MSC Nastran Dynamic
Analysis User’s Guide. The DLOAD entry is used to define linear combinations of frequency dependent loads that are
defined on RLOAD1 or RLOAD2 entries. The RLOAD1 entry defines a frequency dependent load of the form
i   – 2f 
 P  f   = A  C  f  + iD  f  e

where A is defined on a DAREA entry, C  f  and D  f  are defined on TABLEDi entries,  is defined on a DPHASE
entry,  is defined on a DELAY entry, and f is defined on a FREQ, FREQ1 or FREQ2 entry. The RLOAD2 entry defines
a frequency-dependent load of the form
i    f  +  – 2f 
 P  f   = AB  f e

where A is defined on a DAREA entry, B  f  and   f  are defined on TABLEDi entries,  is defined on a DPHASE
entry, and  is defined on a DELAY entry. The coefficients on the DAREA, DELAY, and DPHASE entries may be different
for each loaded degree of freedom. The loads are applied to the specified components in the global coordinate system.
Static loading data may be used to supply DAREA values. Using a LOADSET Case Control request with LSEQ Bulk Data
entries, the user can build DAREA load vectors from any valid static load set. This option is available in the frequency
response solutions and in superelement analysis.
A discussion of random response calculations is given in Section 12.2 of The NASTRAN Theoretical Manual. The RANDPS
entry defines load set power spectral density factors for use in random analysis of the form
S jk  f  =  X + iY G  f 

where G  f  is defined on a TABRNDi entry. The subscripts j and k define the subcase numbers of the load definitions.
If the applied loads are independent, only the diagonal terms ( j = k ) need be defined. The RANDT1 entry is used to
specify the time lag constants for use in the computation of the autocorrelation functions.

Main Index
 Chapter 5: Applied Loads 307
Time-Dependent Loads

Time-Dependent Loads
A discussion of transient response calculations is given inTransient Response Analysis (p. 201) in the MSC Nastran Dynamic
Analysis User’s Guide. The DLOAD entry is used to define linear combinations of time-dependent loads that are defined
on TLOAD1 and TLOAD2 entries. The TLOAD1 entry defines a time-dependent load of the form
 P  t   =  AF  t –   
where A is defined on a DAREA entry,  is defined on a DELAY entry, and F  t –   is defined on a TABLEDi entry.
The TLOAD2 entry defines a time-dependent load of the form


 0 , t   T 1 +   or t   T 2 +  
Pt = 
 At̃ B e Ct̃ cos  2Ft̃ + P  ,  T 1 +    t   T 2 +  


where T 1 and T 2 are time constants, t̃ = t – T 1 –  , A and  are the same as on the TLOAD1 entry, B is the growth
coefficient, C is the exponential coefficient, F is the frequency, and P is the phase angle.
The coefficients on the DAREA and DELAY entries may be different for each loaded degree of freedom. The loads are
applied to the specified components in the global coordinate system.
Two methods are available to automatically create DAREA values from static load data. Simple static loads that supply load
values directly to grid points (such as FORCE, MOMENT, SLOAD, and PLOAD Bulk Data entries) will add to the
DAREA factors when the load identification matches the requested set.
More complicated loads may be generated using the LOADSET Case Control command. The requested LSEQ Bulk Data
entry matches DAREA sets to static loading sets. Any type of static load (e.g., LOAD, GRAV, RFORCE, and SPCD Bulk
Data entries) may be generated for either superelement or residual only (non-superelement) models.
Nonlinear effects are treated as an additional applied load vector, for which the components are functions of the degrees of
freedom. This additional load vector is added to the right side of the equations of motion and treated along with the applied
load vector during numerical integration. It is required that the points to which the nonlinear loads are applied and the
degrees of freedom on which they depend be members of the solution set; i.e., that they cannot be degrees of freedom
eliminated by constraints. It is further required that, if a modal formulation is used, the points referenced by the nonlinear
loads be members of the set of extra scalar points introduced for dynamic analysis (see Section 11.2 of The NASTRAN
Theoretical Manual).
There are four different types of nonlinear transient forcing functions: arbitrary function generators, multiplies, positive
power functions, and negative power functions. For a discussion of nonlinear forcing functions, see Section 11.2 of The
NASTRAN Theoretical Manual.
The NOLIN1 entry defines a nonlinear load of the form
Pi  t  = Si T  uj 

Main Index
 308 MSC Nastran Reference Guide
Simplified Static Loading Data in Dynamic Analysis

where P i is the load applied to u i , S i is a scale factor, T  u j  is a tabulated function defined with a TABLEDi entry, and
u j is any permissible displacement component or velocity component.
The NOLIN2 entry defines a nonlinear load of the form
Pi  t  = Si uj uk

where u j and u k are any permissible pair of displacement components and/or velocity components. They may be the
same.
The NOLIN3 entry defines a nonlinear load of the form

 –S  u  A , u  0
 i j j
Pi  t  = 
 0 , uj  0


where A is an exponent and u j is any permissible displacement component or velocity component.

The NOLIN4 entry defines a nonlinear load of the form

 –S  u  A , u  0
 i j j
Pi  t  = 
 0 , uj  0


where A is an exponent and u j is any permissible displacement component or velocity component.

Simplified Static Loading Data in Dynamic Analysis

Prior to MSC Nastran 2001, static loading data in dynamic loads had to be included through the use of the
LOADSET/LSEQ combination. The LOADSET command, which is selected in the Case Control Section, points to a
LSEQ Bulk Data entry. The LSEQ Bulk Data entry, in turn, acts as a “bridge” between the DAREA ID and the static load
ID/thermal load ID. If the DAREA ID of the selected dynamic load (RLOAD1/RLOAD2/TLOAD1/
TLOAD2/ACSRCE) matches the DAREA ID on the selected LSEQ entry, then all static loads and thermal loads
referenced in the selected LSEQ Bulk Data entry will be selected automatically. Note that the DAREA entry does even have
to exist in this case.
It is no longer necessary for the user to explicitly specify LOADSET/LSEQ combination in order to employ static loading
data in dynamic analysis. Instead, when the user selects a dynamic load, all static loads and thermal loads that have the same
ID as the DAREA ID on the dynamic load entry are automatically selected. The usage of this feature is illustrated below.

Main Index
 Chapter 5: Applied Loads 309
Simplified Static Loading Data in Dynamic Analysis

Pre-2001 Usage
.
CEND
.
$ THE FOLLOWING REQUEST SELECTS LSEQ ENTRY 1000
LOADSET = 1000
$ THE FOLLOWING REQUEST SELECTS RLOAD1 ENTRY 150
DLOAD = 150
.
BEGIN BULK
$ THE FOLLOWING ENTRY, WHICH ACTS AS A "BRIDGE"
$ BETWEEN EXCITEID ID 100 AND STATIC LOAD ID 200,
$ CAUSES THE SELECTION OF PLOAD4,200
LSEQ,1000,100,200
PLOAD4,200,…
RLOAD1,150,100,…
.

Current Usage
.
CEND
.
$ THE FOLLOWING REQUEST SELECTS RLOAD1 ENTRY 150
DLOAD = 150
.
BEGIN BULK
$ THE FOLLOWING PLOAD4 IS AUTOMATICALLY SELECTED
$ BECAUSE ITS ID OF 100 MATCHES THE EXCITEID ID OF THE
$ SELECTED RLOAD1 ENTRY
PLOAD4,100,…
RLOAD1,150,100,…
$ THE ABOVE USAGE IS EQUIVALENT TO THE USER
$ SELECTING AN LSEQ BULK DATA OF THE FORM:-
$ LSEQ,SET_ID,100,100,100
.

The DAREA Bulk Data entry is used to specify point loads in dynamic analysis. In the case of grid points, these loads are
implicitly assumed to be in the displacement (or local) coordinate systems of those points. Until now, the usage of these
data in superelement analysis had been restricted to the residual structure. Prior to MSC Nastran 2001, to specify DAREA-
type loads for upstream superelements in dynamic analysis, the user was forced to specify the LOADSET/LSEQ
combination in conjunction with the appropriate FORCE, FORCE1, FORCE2, MOMENT, MOMENT1, MOMENT2,
or SLOAD Bulk Data entry. An enhancement incorporated into MSC Nastran avoids this problem by automatically
converting all DAREA Bulk Data entries for grid and scalar points into the appropriate equivalent
FORCE/MOMENT/SLOAD Bulk Data entries.
Two additional, important advantages of the above enhancement follow.

Main Index
 310 MSC Nastran Reference Guide
Combining of Loads

 When performing dynamic analysis using the modal approach, it is very often desirable to employ residual vectors
to improve the quality of the solution. The user must specify static loads at those points that are dynamically
excited. However, with the above automatic conversion feature, it is not necessary to explicitly specify static loads
for the purpose of residual vector calculations. Such loads are automatically generated by the MSC Nastran.
 Prior to MSC Nastran 2001, the usage of the DAREA Bulk Data entry had been restricted to dynamic analysis.
The above automatic conversion feature permits the DAREA entry to be used in static analysis as well. This is
particularly advantageous when the user wishes to apply loads at grid points in the displacement (or local)
coordinate systems of those grid points.
When the above conversion is made, a message is issued at the end of the Preface module informing the user of the
conversion. This message also outputs an image not only of each converted DAREA Bulk Data entry, but also of the
corresponding FORCE/MOMENT/SLOAD Bulk Data entry into which it has been converted.

Combining of Loads
The LOADCYH Bulk Data entry is used to define the harmonic coefficients of a load in cyclic symmetry analysis. The
LOADCYN entry is used to define a physical load in cyclic symmetry analysis, and the LOADCYT entry specifies a tabular
load input for axisymmetric cyclic symmetry problems.
The LOAD entry defines a static loading condition that is a linear combination of load sets consisting of loads applied
directly to grid points, pressure loads, gravity loads, and centrifugal forces. This entry must be used if gravity loads are to
be used in combination with loads applied directly to grid points, pressure loads, or centrifugal forces. The application of
the combined loading condition is requested in the Case Control Section by selecting the set number of the LOAD
combination.
It should be noted that the equivalent loads (thermal, enforced deformation, and enforced displacement) must have unique
set identification numbers and be separately selected in Case Control. For any particular solution, the total static load will
be the sum of the applied loads (grid point loading, pressure loading, gravity loading, and centrifugal forces) and the
equivalent loads.
The LSEQ entry defines static load sets that may be referenced by dynamic load entries RLOAD1, RLOAD2, TLOAD1,
and TLOAD2. The LSEQ entry may be used to select static loading conditions in the superelement static solution
sequences and is used to correlate static loading conditions with dynamic loads in dynamic analysis.
The loads actually applied to points may be requested by use of the OLOAD entry.

Temperature Loads and Enforced Deformations in Subcase Combinations (SUBCOMS)

The use of the Case Control Section entries SUBCOM and SYMCOM in static analysis provides the user with a relatively
straightforward method to combine the results of two or more immediately preceding subcases or SYM cases respectively.
For example, the following portion of a Case Control Section illustrates the use of the SUBCOM.

Main Index
 Chapter 5: Applied Loads 311
Summary of Nonlinear Static Loads

.
.
.

SUBCASE 1
LOAD=1
TEMP(LOAD)=7
SUBCASE 2
LOAD=2
SUBCASE 10
LOAD=10
DEFORM=11
SUBCOM 20
SUBSEQ=.5,1.,1.6
.
.
.

Here, SUBCOM 20 provides results for a combined loading condition of the form
0.5   LOAD = 1 + TEMP(LOAD) = 7  + 1.0   LOAD = 2  + 1.6   LOAD = 10 + DEFORM = 11 

In MSC Nastran, the results for this combined loading condition are efficiently obtained by forming the linear combination
of the displacement vectors obtained in the previous subcases. Thus,
 U  SUBCOM 20 = 0.5   U  SUBCASE 1 + 1.0   U  SUBCASE 2 + 1.6   U  SUBCASE 10

and the element forces and stresses are calculated in the usual manner from the displacement vector for the combined
loading condition. This technique is quite straightforward when the several loading conditions are composed of forces and
moments. However, if any of the previous subcases contain thermal loads and/or enforced element deformations, the user
must exercise some care in order to obtain correct element stresses and forces for the combined loading condition defined
through the SUBCOM.

Summary of Nonlinear Static Loads

Loads Overview
The static loads in nonlinear analysis consist of a subset of the total MSC Nastran static load set. Most of the relevant loads
data applicable to the linear static analysis are also applicable to nonlinear static analysis except for the Bulk Data entry
DEFORM. The enforced displacements can be applied as loads using the Bulk Data entry SPCD, that defines an enforced
motion of a grid in a specific fixed direction. Each grid point with an enforced displacement must also appear on an SPC
or SPC1 entry.
The loads are selected in each subcase by the LOAD Case Control command. The load for a subcase is subdivided into the
number of increments specified for the subcase. The solution strategy in nonlinear is to apply the loads in an incremental

Main Index
 312 MSC Nastran Reference Guide
Summary of Nonlinear Static Loads

fashion until the desired load level is reached. The algorithms “remember” the loads from one subcase to the next. If the
load reaches the desired level in a subcase and if the load description is left out in the subsequent subcase inadvertently, the
nonlinear solution algorithm will begin in an incremental fashion to remove the load. The loads described below make up
the valid subset applicable for nonlinear static analysis.

Concentrated Loads
The concentrated loads consist of two categories: those that are stationary in direction throughout the analysis and those
that follow the grid motion. The first category consists of the following:
FORCE
MOMENT
SLOAD
The second category consists of the following:
FORCE1
FORCE2
MOMENT1
MOMENT2
The three FORCE entries differ only in the manner in which the direction of the force is specified. FORCE uses the
components of a vector. FORCE1 uses two grid points, not necessarily the same as the loaded grid points. FORCE2 defines
the direction of the force as a vector that is the vector product of two other vectors. The distinctions between the three
MOMENT entries are similar. The SLOAD entry is used to apply loads to scalar points.

Distributed Loads
The distributed loads consist of two categories: those that are stationary in direction throughout the analysis and those that
follow the element motion. The first category consists of the PLOAD1.
The second category consists of the following:
PLOAD
PLOAD2
PLOAD4
Loads may be applied to the interiors of finite elements. The PLOAD1 entry defines concentrated and linearly distributed
forces and moments to the CBEAM element. However, the stress calculation for nonlinear CBEAM element ignores the
distributed loads, but lump them onto the nodal points. The PLOAD entry is exceptional in that it references three or four
arbitrarily located grids rather than an element. The PLOAD2 entry defines a uniform static pressure load applied to the
CQUAD4 or CTRIA3 elements. The PLOAD4 defines a pressure load on surfaces of CHEXA, CPENTA, CTETRA,
CTRIA3, and CQUAD4 elements that need not be normal to the surface. Since the surface hyperelastic elements
CQUAD4, CQUAD8, CQUAD, CTRIA3, and CTRIA6 are plane strain elements, no pressure load may be applied on
these elements.

Main Index
 Chapter 5: Applied Loads 313
Summary of Nonlinear Static Loads

Mass Related Loads (GRAV and RFORCE Entries)

The mass related loads are gravity and centrifugal loads, which require specification of mass data. The gravity loads that can
be specified on the GRAV Bulk Data entry are stationary in direction throughout the analysis. The GRAV entry is used to
define the direction and magnitude of the gravity vector in some identified coordinate system. The components of the
gravity vector are multiplied by the mass matrix to obtain the components of the gravity force at each grid point.
Centrifugal loads are specified on the RFORCE Bulk Data entry. RFORCE defines the components of a spin vector that
is used internally to compute centrifugal forces. Each component of the spin vector is multiplied by a scale factor.

Thermal Loads
Thermal loads are selected in each subcase by the Case Control command TEMP(LOAD). Thermal loading that is a
follower loading is accomplished through the following entries:
TEMP
TEMPD
TEMPP1
TEMPP3
TEMPRB
The TEMP and TEMPD provide for grid point temperature specification which are then interpolated to points within
elements. For the nonlinear CQUAD4 and CTRIA3 elements the TEMPP1 provides for an average temperature and
thermal gradient in the thickness direction and the TEMPP3 provides for temperature at points through the thickness.
TEMPP1 and TEMPP3 are not available for the plane strain hyperelastic elements CQUAD4, CQUAD8, CQUAD,
CTRIA3, and CTRIA6. The TEMPRB provides for average temperature at ends A and B for the CROD and CBEAM
element. Average temperature specified directly for an element will take precedence over the temperature interpolated from
the elements connected grid points.
The Case Control command TEMP(INIT) is required above subcase level to specify an initial temperature distribution.
The use of TEMP(MAT) or TEMP(BOTH) will cause a fatal error. Subcases that do not contain a TEMP(LOAD) will
default to TEMP(INIT), which implies no thermal load. Thermal loads are often used to provide initial strains.

Load Combinations
Various load specifications can be combined using the LOAD, CLOAD, and LSEQ entries.
The LOAD provides for the linear combination of static load sets. The static loads (LOAD, FORCE, etc.) applied to the
upstream superelements cannot be referenced by the LOAD Case Control command. The CLOAD entry is designed to
apply static loads to upstream superelements by combining loads defined in LSEQ entries. The (CLOAD, LSEQ)
combination allows the nonlinear algorithms to apply in an incremental fashion on the upstream superelement loads to the
boundary of the residual structure.
The CLOAD defines a static load combination for superelement loads acting on residual boundary and the LSEQ defines
a static load combination for superelement upstream loads. The LSEQ assigns load vectors to the superelements and
numerically labels them via the DAREA field entry value. The LID and TID field entries point to the appropriate load and
temperature distribution entries. The CLOAD defines a static load as a linear combination of previously calculated

Main Index
 314 MSC Nastran Reference Guide
Summary of Nonlinear Static Loads

superelement loads defined by the LSEQ. The IDVi fields correspond directly to the numeric labels defined via the DAREA
fields of the LSEQ entries.
The LSEQ is selected by a LOADSET Case Control command above any residual subcase. Only one LOADSET may
appear in Case Control. The hierarchy of the loads data is shown schematically in Figure 5-1. An example for the Case
Control set-up follows:

.
.
.

SEALL=ALL
LOADSET = 1000 $ Selects LSEQ 1000 for upstream loads
SUPER = ALL $ Identify SEs to process
.
.
.

DISP = ALL
.
.
.

.
.
.

.
.
.

$ NONLINEAR SOLUTIONS FOR RESIDUAL SUPERELEMENTS

SUBCASE 10
CLOAD = 1001 $ Refers to CLOAD Bulk Data
NLPARM = 12 $ Iteration control
.
.
.

SUBCASE 20
CLOAD = 1002
LOAD = 10 $ Residual SE load
NLPARM = 22
.
.
.

Main Index
 Chapter 5: Applied Loads 315
Summary of Nonlinear Static Loads

SE 0 Upstream SE

LOAD CLOAD LOADSET

Bulk CLOAD

defined
on LSEQ
DAREA1 DAREA2
IDV1 IDV2
correlated
to CLOAD

LSEQ

Static Loads

Figure 5-1 Hierarchy of Load Data

Main Index
 316 MSC Nastran Reference Guide
Summary of Nonlinear Transient Loads

Summary of Nonlinear Transient Loads

Transient Loads And Initial Conditions
The methods employed to define loads in nonlinear transient analysis are similar to those used in the linear solutions. A
single degree of freedom or a set of grid points may be loaded with force pattern that varies with time. Functions may be
tabular such as an earthquake or a booster liftoff, or they may be simple analytic functions such as a sine wave. Simple static
load sets generated in both upstream superelements and the default residual may be used to create the dynamic loads. They
may be scaled and combined with other loads to simulate complex loading problems.
For special problems involving simple scalar nonlinearities, the traditional nonlinear loads (NOLINi option) may be used
in both types of transient solutions. These are useful for simulating nonlinear damping mechanisms and rotational coupling
not provided by the CGAP elements.
Transient loads define the loadings as functions of time and the location. They can be a load applied at a particular degree
of freedom, pressure over the surface area, or the body force simulating an acceleration. The time history is provided by the
TLOADi Bulk Data entries and the static loads may be converted to dynamic loads using LSEQ entry. The transient
dynamic loads are selected in the Case Control Section by the following commands:

DLOAD Selects TLOADi or DLOAD Bulk Data entries. Each stage of the transient solution may be defined with
a separate SUBCASE, each with its own DLOAD request. The DLOAD entry is restricted to act only in
the time defined in its subcase.
LOADSET Selects LSEQ entries. Must be placed above the solution subcases and/or in the Superelement Subcases.
NOLINEAR Selects NOLINi entries. May not be changed between solution subcases.
NLLOAD Required for NOLINi output request.

Bulk Data entries associated with the transient loads in SOL 129 are briefly described below:

TLOAD1, TLOAD2 Required data defining the coupling between load factors (DAREA entry) and functions of time
(TABLEDi, DELAY, or coefficients). Specifies the time functions, F i  t  , for dynamic load i .
TLOAD1 defines the load history in TABLEDi format while TLOAD2 defines the load history
by an analytic expression. Note that each TLOADi entry must have a unique set identification
number.
DAREA Defines load scale factors, A ij , for each degree of freedom, u j requested in set i . The DAREA
factors may also be derived from static load entries with the LSEQ Bulk Data entry.
LSEQ Generates transient load history for static loads and controls the assembly of static load vectors
to be used in dynamic analysis. Each set is equivalent to a subcase in Case Control. It will cause
a static load vector A j to be generated and will label it with the DAREA identification. This is
useful for modeling problems with distributed transient loading such as GRAV loads or
pressurized areas.

Main Index
 Chapter 5: Applied Loads 317
Summary of Nonlinear Transient Loads

DLOAD Combines different TLOADi functions into a single set and scales each by a factor, C ki as
shown in (5-8)(5-8). Performs the same function as the static LOAD Bulk Data input. Unlike
statics, this is the only method to combine loads in dynamic analysis.
TABLEDi, i=1,2,3,4 Used with TLOAD1 entries to specify general time varying tabular functions. Since
extrapolation is used beyond the range of data the user is cautioned to be careful with the end
points.
DELAY Specifies the delay time for applying the forcing function defined in TLOADi to each GRID
point. The DELAY entry is associated with the degree of freedom and not the forcing function
in each TLOADi. This option is useful for defining loads that travel across a structure such as an
oblique wave along a dam or a bump in the road exciting the wheels of a traveling vehicle. This
saves the analyst the effort of defining multiple TLOADi entries for simple time lags.
NOLINi, i=1,2,3,4 Define specific nonlinear scalar loads as functions of velocity and/or displacement. Basic analog
nonlinear transfer functions are defined. However, these functions do not generate tangent
matrix terms and the Newton iteration method may experience difficulties in convergence. Line
search and BFGS solution methods are recommended for these models.
NOLIN1 Nonlinear transient load as a tabular function.
NOLIN2 Nonlinear transient load as products of two variables.
NOLIN3 Nonlinear transient load as a positive variable raised to a power.
NOLIN4 Nonlinear transient load as a negative variable raised to a power.
NLRGAP The NLRGAP is an idealization of the contact between any circular shaft enclosed by a circular
housing (Figure 5-2). In this idealization, grid point A represents the outer surface of the shaft
while grid point B represents the inner surface of the housing. Grid points A and B are coincident
and must have a parallel displacement coordinate system. A rudimentary friction capability is
included to account for the friction-induced torque and friction-induced lateral loads that occur
with a spinning shaft. However, friction can be neglected, in which case the solution is consistent
with that for a nonrotating shaft in a frictionless housing.

Main Index
 318 MSC Nastran Reference Guide
Summary of Nonlinear Transient Loads

2
N B6

N A6

K Clearance (TABG)

N B1 N A1
1
Grid A K
(RADIUS)

Grid B
N A2

N B2
Figure 5-2 Radial Gap Orientations and Nonlinear Load Sign Conventions
The equations underlying the NLRGAP measure the relative radial displacement between the
shaft grid and the housing grid in the displacement coordinate system of grid points A and B.
When the relative radial displacement is greater than the clearance (i.e. when contact has
occurred), contact forces as well as frictional forces are automatically applied to the grids. The
contact and frictional forces are dependent on the contact stiffness and the coefficient of friction.

NLRGAP Underlying Equations

Definitions

K(t,u) Contact stiffness, function of time or displacement

G(t) Radial clearance, function of time
t Coefficient of friction, function of time. Positive value implies shaft rotation from axis 1
towards axis 2.
R Shaft radius for friction induced torque calculation

Main Index
 Chapter 5: Applied Loads 319
Summary of Nonlinear Transient Loads

Equations

The NLRGAP is internally composed of the following equations. It is helpful to visualize Grid A as belonging to the shaft
and Grid B as belonging to the housing. Consider each grid to have DOF 1 and 2 in the plane of action (e.g., DOF 1 = X
and DOF 2 = Y for XY orientation).
The element relative displacement in directions 1 and 2 are:
U 1 = U A1 – U B1

U 2 = U A2 – U B2

The relative radial displacement is:

r = U 12 + U 22

The force in the gap when it is open  r  G  t   is zero. When the gap is closed  r  G  t   , the penetration is defined
as:
P = r – G  t 
The nonlinear gap forces are:

Gt
S = K  t,u   1 – ----------- (5-1)
 r 

N A1 = – U 1 S + U 2 S  t  (5-2)

N B1 = U 1 S – U 2 S  t  (5-3)

N A2 = – U 2 S – U 1 S  t  (5-4)

N B2 = U 2 S + U 1 S  t  (5-5)

If the shaft radius R is input, then the following friction induced torque loads are generated.
N A6 = – RrS  t  (5-6)

N B6 = RrS  t  (5-7)

Benefit
Allows you to model contact between two coaxial cylinders, for example shaft and housing.

Main Index
 320 MSC Nastran Reference Guide
Summary of Nonlinear Transient Loads

Input
The Bulk Data entry NLRGAP is used to model the radial gap.
The contact stiffness can be input either as a function of time or as a function of penetration. In both cases, the stiffness
curves are input on TABLEDi Bulk Data entries which are referenced by the NLRGAP entry. Inputting stiffness as a
function of time allows the user to model special situations where structural changes over time cause the contact stiffness
to vary over time. Inputting stiffness as a function of penetration allows the user to input a nonlinear stiffness curve which
might be required, for example, if the shaft or housing were covered by some nonlinear-elastic material.
As with stiffness, both the coefficient of friction and the clearance can be functions of time. The time-varying coefficient
of friction might be used where the quality of lubrication varied with time, or if experimental data showed that the surface
characteristics of the shaft and housing varied with time. The clearance can be input as a function of time to model
situations where an accelerating, spinning rotor expands due to centripetal loads.
The NLRGAP entry can also approximate the torque on the shaft and housing due to friction. If this effect is desired, the
RADIUS field of the NLRGAP must be specified. If not specified, then friction effects will still be included in the lateral
loads as shown in equations (5-2) through (5-5). The torque loads, equations (5-6) and (5-7), will not be computed nor applied
if RADIUS is input as 0.0 or left blank.
As with the NOLINs, the NLRGAP is selected by the NONLINEAR Case Control command.

Output
The output is the nonlinear loads applied to the NLRGAP grids to simulate the contact. The loads can be recovered with
the NLLOAD Case Control command.

Guidelines
 There should be mass on both grids of the NLRGAP. Otherwise erratic results and divergence can occur. A small
amount of damping (e.g. CDAMPi) between the two grids may improve stability for some problems.
 The two grids listed on the NLRGAP should be coincident and have parallel displacement coordinate systems.
 The friction capability is somewhat arbitrary in sign. Referring to Figure 5-2, a positive coefficient of friction is
consistent with shaft rotation from axis 1 towards axis 2 (counter-clockwise). If the shaft is rotating in the opposite
direction, then a negative coefficient of friction must be input in order for the friction induced loads to have the
proper signs.
 In SOL 129, setting ADJUST = 0 on the TSTEPNL entry will turn off the adaptive time step. This is useful if
large contact stiffness is causing an excessively small time step.
 As with the NOLINs, the forces applied to the NLRGAP grids to simulate the radial contact are based upon the
relative displacements of the grids from the previous time step. Therefore, the larger the contact stiffness, the
smaller the necessary time steps. The smaller the contact stiffness, the larger the penetration. Use the smallest
contact stiffness possible to model the contact to the accuracy desired. The time step may need to be decreased by
10 times or more compared to the same model without NLRGAPs and NOLINs.

Main Index
 Chapter 5: Applied Loads 321
Summary of Nonlinear Transient Loads

 The NLRGAP is based upon small displacement theory. Therefore, for most accurate results, especially with
respect to friction induced torque, the clearance should be small relative to the housing diameter. The friction
induced torque applied to grids A and B will be equal and opposite in direction.

Limitations
 The NLRGAP can only be positioned in the XY, YZ or ZX plane of the displacement coordinates of the
referenced grids.
 The DOF in the plane of the NLRGAP must be in the d-set. This means that these DOF cannot be constrained
by SPCs or be dependent DOF on MPCs and RBE type elements.
 The NLRGAP can only be used in SOLution sequences 109 and 129.
A diagram of input data linkages is shown in Figure 5-3. The general form of the transient load, P kj , for each degree of
freedom, u j , is:

P kj = C ki  A j  F i  t  (5-8)

where the factors are defined by various Bulk Data entries as shown in Figure 5-4.

Main Index
 322 MSC Nastran Reference Guide
Summary of Nonlinear Transient Loads

Case Control Section

DLOAD (NONLINEAR) LOADSET

Bulk Data Section

(NOLINi
DLOAD
NLRGAP)

TLOAD1 TLOAD2

DAREA DAREA

LSEQ

LOAD

FORCES MOMENT PLOAD

Figure 5-3 Dynamic Loads for Transient Analysis

Main Index
 Chapter 5: Applied Loads 323
Summary of Nonlinear Transient Loads

Dynamic Load Entries

DLOAD
 •S 

 •S1   •S2   •S3 

TLOAD1 TLOAD1 TLOAD2

TABLED1 TABLED2  •f  t  
 •f  t    •f  t  

DAREA DAREA DAREA

FORCE1 LOAD PLOAD2

MOMENT PLOAD4

Static Load Entries

Note that each load is scaled by S  Si  f  t  at each time step in dynamic loads.

Figure 5-4 Example of Combining Loads

Main Index
 Chapter 6: Constraint and Set Notation
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

6 Constraint and Set Notation

 Degree of Freedom Sets 325


Multipoint Constraints 328

Single Point Constraints 330

Rigid Body Supports 331

Sets for Dynamic Reduction 332

Sets for Aerodynamics 332

Rules of Sets for Undefined Degrees of Freedom 333

Output Selection Via Set Specification 334

Main Index
 Chapter 6: Constraint and Set Notation 325
Degree of Freedom Sets

Degree of Freedom Sets

Structural matrices are initially assembled in terms of all structural points, which excludes only the extra points introduced
for dynamic analysis. These matrices are generated with six degrees of freedom for each geometric grid point and a single
degree of freedom for each scalar point. These degrees of freedom are partitioned into sets, based on user inputs. The sets
are used to successively eliminate variables during the solution process, as described in General Operations, 361.

Degree-of-Freedom Set Definitions

Each degree of freedom is a member of one mutually exclusive set. Set names have the following definitions:

Set Name Definition

mp degrees of freedom eliminated by multipoint constraints.

mr degrees of freedom eliminated by multipoint constraints created by the rigid elements using the
LGELIM method on the Case Control command RIGID.
sb* degrees of freedom eliminated by single-point constraints that are included in boundary
condition changes and by the AUTOSPC feature.
sg* degrees of freedom eliminated by single-point constraints that are specified on the PS field on
GRID Bulk Data entries.
o degrees of freedom omitted by structural matrix partitioning.
q Generalized degrees of freedom assigned to component modes and residual vectors.
r reference degrees of freedom used to determine free body motion.
c degrees of freedom that are free during component mode synthesis or dynamic reduction.
b degrees of freedom fixed during component mode analysis or dynamic reduction.

lm Lagrange multiplier degrees of freedom created by the rigid elements using the LAGR method
on the Case Control command, RIGID.
e extra degrees of freedom introduced in dynamic analysis.
sa Permanently constrained aerodynamic degrees of freedom.
k Aerodynamic mesh point set for forces and displacements on the aero mesh.
j Aerodynamic mesh collocation point set (exact physical interpretation is dependent on the
aerodynamic theory).

*Strictly speaking, sb and sg are not exclusive with respect to one another. Degrees of freedom may exist in both sets
simultaneously. Since these sets are not used explicitly in the solution sequences, this need not concern the user.
However, those who use these sets in their own DMAPs should avoid redundant specifications when using these sets for
partitioning or merging operations. That is, a degree of freedom should not be specified on both a PS field of a GRID

Main Index
 326 MSC Nastran Reference Guide
Degree of Freedom Sets

entry (sg set) and on a selected SPC entry (sb set). Redundant specifications will cause UFM 2120 in the VEC module
and behavior listed in MSC Nastran DMAP Programmer’s Guide i for the UPARTN module. These sets are exclusive,
however, from the other mutually exclusive sets.
Each degree of freedom is also a member of one or more combined sets called “supersets.” Supersets have the following
definitions:

sets supersets
mp
m
mr
sb
sg s
o
g
q p
ne
r n
a f
c d fe
t
b l
lm
e
k
ks
sa
js
j

Set Name Meaning (+ indicates union of two sets)

s = sb + sg all degrees of freedom eliminated by single point constraints
l = b + c + lm the structural degrees of freedom remaining after the reference degrees of freedom
are removed (degrees of freedom left over)
t = l+r the total set of physical boundary degrees of freedom for superelements
a = t+q the set assembled in superelement analysis
d = a+e the set used in dynamic analysis by the direct method
f = a+o unconstrained (free) structural degrees of freedom
fe = f + e free structural degrees of freedom plus extra degrees of freedom
n = f+s all structural degrees of freedom not constrained by multipoint constraints
ne = n + e all structural degrees of freedom not constrained by multipoint constraints plus extra
degrees of freedom
m = mp + mr all degrees of freedom eliminated by multiple constraints
g = n+m all structural (grid) degrees of freedom including scalar degrees of freedom

Main Index
 Chapter 6: Constraint and Set Notation 327
Degree of Freedom Sets

Set Name Meaning (+ indicates union of two sets)

p = g+e all physical degrees of freedom
ks = k + sa the union of k and the re-used s-set (6 degrees of freedom per grid)
js = j + sa the union of j and the re-used s-set (6 degrees of freedom per grid)
fr = o + l statically independent set minus the statically determinate supports
 fr = f – q – r 
v = o+c+r the set free to vibrate in dynamic reduction and component mode synthesis

The a-set and o-set are created in the following ways:

1. If only OMITi entries are present, then the o-set consists of degrees of freedom listed explicitly on OMITi entries.
The remaining f-set degrees of freedom are placed in the b-set, which is a subset of the a-set.
2. If ASETi or QSETi entries are present, then the a-set consists of all degrees of freedom listed on ASETi entries and
any entries listing its subsets, such as QSETi, SUPORTi, CSETi, and BSETi entries. Any OMITi entries are
redundant. The remaining f-set degrees of freedom are placed in the o-set.
3. If there are no ASETi, QSETi, or OMITi entries present but there are SUPORTi, BSETi, or CSETi entries present,
then the entire f-set is placed in the a-set and the o-set is not created.
4. There must be at least one explicit ASETi, QSETi, or OMITi entry for the o-set to exist, even if the ASETi, QSETi,
or OMITi entry is redundant.
In dynamic analysis, additional vector sets are obtained by a modal transformation derived from real eigenvalue analysis of
the a-set. These sets are as follows:

 o = rigid body (zero frequency) modal degrees of freedom

 f = finite frequency modal degrees of freedom

 i =  o +  f , the set of all modal degrees of freedom

One vector set is defined that combines physical and modal degrees of freedom:

u h =  i + u e , the set of all modal degrees of freedom

The membership of each degree of freedom can be printed by use of the Bulk Data entries PARAM,USETPRT and
PARAM,USETSEL.

Degree-of-Freedom Set Bulk Data Entries

Degrees of freedom are placed in sets as specified by the user on the following Bulk Data entries:

Main Index
 328 MSC Nastran Reference Guide
Multipoint Constraints

Name Bulk Data Entry Name

m MPC, MPCADD, MPCAX, POlNTAX, RBAR, RBE1, RBE2, RBE3, RROD, RSPLINE,
RTRPLT, GMBC, GMSPC*
sb SPC, SPC1, SPCADD, SPCAX, FLSYM, GMSPC*, BNDGRID,
(PARAM,AUTOSPC,YES)
sg GRID, GRIDB, GRDSET (PS field)
o OMIT, OMIT1, OMlTAX, GRID (SEID field), SESET
q QSET, QSET1
r SUPORT, SUPORT1, SUPAX
c CSET, CSET1
b BSET, BSET1
e EPOINT
sa CAEROi
k CAEROi
a ASET, ASET1, Superelement exterior degrees of freedom, CSUPEXT
*Placed in set only if constraints are not specified in the basic coordinate system.

In superelement analysis, the appropriate entry names are preceded by the letters SE, and have a field reserved for the
superelement identification number. This identification is used because a boundary (exterior) grid point may be in one
mutually exclusive set in one superelement and in a different set in the adjoining superelement. The SE-type entries are
internally translated to the following types of entry for the referenced superelement:

Entry Type Equivalent Type

SEQSETi QSETi
SESUP SUPORT
SECSETi CSETi
SEBSETi BSETi

Multipoint Constraints
Each multipoint constraint is described by a single equation that specifies a linear relationship for two or more degrees of
freedom. In static analysis, multiple sets of multipoint constraints can be provided in the Bulk Data Section, with selections
made at execution time by using the subcase structure in Case Control as explained in Case Control Section, 14. Multipoint

Main Index
 Chapter 6: Constraint and Set Notation 329
Multipoint Constraints

constraints are also discussed in Executive Control Section, 14 of this manual and Section 5.4 of The NASTRAN Theoretical
Manual.
Multipoint constraints are defined on MPC, MPCADD, and MPCAX entries. MPC is the basic entry for defining
multipoint constraints. The first component specified on the entry is the dependent degree of freedom, i.e., that degree of
freedom that is removed from the equations of motion. Dependent degrees of freedom may appear as independent terms
in other equations of the set; however, they may appear as dependent terms in only a single equation. The MPCADD entry
defines a union of multipoint constraints. The MPCAX defines multipoint constraints in conical shell problems. Some uses
of multipoint constraints are:
 To describe rigid elements and mechanisms such as levers, pulleys, and gear trains. In this application, the degrees
of freedom associated with the rigid element that are in excess of those needed to describe rigid body motion are
eliminated with multipoint constraint equations. Treating very stiff members as rigid elements eliminates the ill-
conditioning associated with their treatment as ordinary elastic elements.
 To enforce zero motion in directions other than those corresponding with components of the global coordinate
system. In this case, the multipoint constraint will involve only the degrees of freedom at a single grid point. The
constraint equation relates the displacement in the direction of zero motion to the displacement components in
the global system at the grid point.
 To be used with scalar elements to generate nonstandard structural elements and other special effects.
 To describe parts of a structure by local vibration modes. This application is treated in Section 14.1 of The
NASTRAN Theoretical Manual. The general idea is that the matrix of local eigenvectors represents a set of
constraints relating physical coordinates to modal coordinates.
In general, the user must provide the coefficients in the multipoint constraint equations. However, several rigid elements
have been introduced which will generate the MPC equations for some applications:

RROD A pin-ended rod which is rigid in extension.

RBAR A rigid bending element with six degrees of freedom at each end.
RTRPLT A rigid triangular plate with six degrees of freedom at each vertex.
RBE1 A rigid body connected to an arbitrary number of grid points. The independent and
dependent degrees of freedom can be arbitrarily selected by the user.
RBE2 A rigid body connected to an arbitrary number of grid points. The independent degrees
of freedom are the six components of motion at a single grid point. The dependent
degrees of freedom are specified at an arbitrary number of grid points.
RBE3 Defines the motion at a reference grid point as the weighted average of the motions at a
set of other grid points.
RSPLlNE Defines multipoint constraints for the interpolation of displacements at grid points.
RSSCON Defines multipoint constraints to model clamped connections of shell-to-solid elements.

The rigid elements will always meet equilibrium and continuity requirements, whereas this is a user responsibility for MPC
equations. (See Rigid Elements and Multipoint Constraints (R-type, MPC), 159 for further information on rigid elements.)

Main Index
 330 MSC Nastran Reference Guide
Single Point Constraints

Multipoint forces of constraint may be output with MPCFORCE Case Control command in SOLution Sequences 101
through 200.

Single Point Constraints

A single point constraint (SPC) applies a fixed value to a translational or rotational component at a geometric grid point or
to a scalar point. Common uses of single point constraints are to specify the boundary conditions of a structural model by
fixing the appropriate degrees of freedom and to eliminate unwanted degrees of freedom with zero stiffness. Multiple sets
of single point constraints can be provided in the Bulk Data Section, with selections made at execution time by using the
subcase structure in the Case Control Section as explained in Case Control Section, 14. This procedure is particularly useful in
the solution of problems having one or more planes of symmetry.
The elements connected to a grid point may not provide resistance to motion in certain directions, causing the stiffness
matrix to be singular. Single point constraints are used to remove these degrees of freedom from the stiffness matrix. A
typical example is a planar structure composed of membrane and extensional elements. The translations normal to the plane
and all three rotational degrees of freedom must be constrained since the corresponding stiffness matrix terms are all zero.
If a grid point has a direction of zero stiffness, the single point constraint needs not be exactly in that direction, but only
needs to have a component in that direction. This allows the use of single point constraints for the removal of such
singularities regardless of the orientation of the global coordinate system. Although the displacements will depend on the
direction of the constraint, the internal forces will be unaffected.
One of the tasks performed by the Grid Point Singularity Processor (GPSP) (see Constraint and Mechanism Problem Identification
in SubDMAP SEKR, 370) is to examine the stiffness matrix for singularities at the grid point level. Singularities remaining at
this level, following the application of the multipoint and single point constraints, are listed in the Grid Point Singularity
Table (GPST), which is automatically printed. The GPST lists all singular degrees of freedom, in the global coordinate
system, and the ratio of stiffness between the softest and stiffest degree of freedom for the grid point. The user may request
that single point constraints be generated for all identified singularities by use of the Bulk Data entry
PARAM,AUTOSPC,YES.
Single point constraints are defined on SPC, SPC1, SPCADD and SPCAX entries. The SPC entry is the most general way
of specifying single point constraints. The SPC1 entry is a less general entry that is more convenient when several grid points
have the same components constrained to a zero displacement. The SPCADD entry defines a union of single point
constraint sets specified with SPC or SPC1 entries. The SPCAX entry is used only for specifying single point constraints
in problems using conical shell elements.
Single point constraints can also be defined on the GRID entry. In this case, however, the points are constrained for all
subcases. The default value for enforced displacement on points constrained on GRID entries is zero. The default value can
be overridden at the subcase level with SPC entries.
The printed output for single point forces of constraints can be requested in Case Control. All nonzero forces are printed,
whether they originate from SPCi entries, the PS field on GRID entries, or by PARAM,AUTOSPC,YES.

Main Index
 Chapter 6: Constraint and Set Notation 331
Rigid Body Supports

Rigid Body Supports

In the following discussion, a free body is defined as a structure that is capable of motion without internal stress; i.e., it has
one or more rigid body degrees of freedom. The stiffness matrix for a free body is singular with the defect equal to the
number of stress-free, or rigid body modes. A solid three-dimensional body has up to six rigid body modes. Linkages and
mechanisms can have a greater number. In order to permit the analysis of mechanisms, no restriction is placed in the
program on the number of stress-free modes.
Free-body supports are defined with a SUPORT or SUPORT1 entry. In the case of problems using conical shell elements,
the SUPAX entry is used. Free-body supports must be defined in the global coordinate system. The SUPORT1 entry must
be selected by the SUPORT1 Case Control command.
In static analysis by the displacement method, the rigid body modes must be restrained in order to remove the singularity
of the stiffness matrix. The required constraints may be supplied with single point constraints, multipoint constraints, or
free body supports. If free body supports are used, the rigid body characteristics will be calculated and a check will be made
on the sufficiency of the supports. Such a check is obtained by calculating the rigid body error ratio and the strain energy
as defined in the Rigid Body Matrix Generator operation in Static Solutions in SubDMAP SEKRRS, 376. This error ratio and the
strain energy are automatically printed following the execution of the Rigid Body Matrix Generator. The error ratio and
the strain energy should be zero, but may be nonzero for any of the following reasons:
 Round-off error accumulation.
 Insufficient free body supports have been provided.
 Redundant free body supports have been provided.
The redundancy of the supports may be caused by improper use of the free body supports themselves or by the presence of
single point or multipoint constraints that constrain the rigid body motions.
Static analysis with inertia relief is necessarily made on a model having at least one rigid body motion. Such rigid body
motion must be constrained by the use of free body supports. These supported degrees of freedom define a reference system,
and the elastic displacements are calculated relative to the motion of the support points. The element stresses and forces
will be independent of any valid set of supports.
Rigid body vibration modes are calculated by a separate procedure provided that a set of free body supports is supplied by
the user. This is done to improve efficiency and, in some cases, reliability. The determinant method, for example, has
difficulty extracting zero frequency roots of high multiplicity, whereas the alternate procedure of extracting rigid body
modes is both efficient and reliable. If the user does not specify free body supports (or he specifies an insufficient number
of them), the (remaining) rigid body modes will be calculated by the method selected for the finite frequency modes,
provided zero frequency is included in the range of interest. If the user does not provide free body supports, and if zero
frequency is not included in the range of interest, the rigid body modes will not be calculated.
Free body supports must be specified if the mode acceleration method of solution improvement is used for dynamic
problems having rigid body degrees of freedom (see Rigid-body Modes (p. 155) in the ). This solution improvement technique
involves a static solution, and although the dynamic solution can be made on a free body, the static solution cannot be
performed without removing the singularities in the stiffness matrix associated with the rigid body motions.

Main Index
 332 MSC Nastran Reference Guide
Sets for Dynamic Reduction

Sets for Dynamic Reduction

There are several methods for reducing the size of models in dynamic analysis. The method described here uses shapes
derived as generalized functions by static analysis. The same entries may be used in static analysis with the same
nomenclature being used internally as in superelement analysis.
The statically independent degrees of freedom remaining after constraint elimination are partitioned into two sets. The
partitions are defined by listing the degrees of freedom for one of the partitions on the ASETi entries. These degrees of
freedom are referred to as the “analysis set.” The remaining degrees of freedom are referred to as the “omitted set.” The
OMITi entry can be used to place degrees of freedom in the omitted set with the remaining degrees of freedom being placed
in the analysis set. This is easier if the analysis set is large. In the case of problems using conical shell elements, the OMITAX
entry is used.
One of the more important applications of partitioning is Guyan reduction, described in Dynamic Reduction (p. 690) in the
. This technique is a means for reducing the number of degrees of freedom used in dynamic analysis with minimum loss of
accuracy. Its basis is that many fewer grid points are needed to describe the inertia of a structure than are needed to describe
its elasticity with comparable accuracy. The error in the approximation is small provided that the set of displacements used
for dynamic analysis is judiciously chosen. Its members should be uniformly dispersed throughout the structure, and all
large mass items should be connected to grid points that are members of the analysis set.
The user is cautioned to consider the fact that the matrix operations associated with this partitioning procedure tend to
create nonzero terms and to fill what were previously very sparse matrices. The partitioning option is most effectively used
if the members of the omitted set are either a very large fraction or a very small fraction of the total set. In most of the
applications, the omitted set is a large fraction of the total and the matrices used for analysis, while small, are usually full.
If the analysis set is not a small fraction of the total, a solution using the large, but sparser, matrices may be more efficient.
The a-set is further partitioned into the q-, r-, c-, and b-sets. The q-set is used to store the coefficients for the generalized
coordinates determined by dynamic reduction and/or component mode calculations. The r-set is again used to determine
rigid body modes. The c-set contains coordinates considered free to move while computing the generalized coordinate
functions. The b-set contains coordinates fixed during this process.
The set used to describe the generalized functions for dynamic reduction and component mode synthesis is the v-set, the
union of the o-, r-, and c-sets.
The d-set (“dynamic”) for direct formulations is formed from the union of the structural degrees of freedom in the a-set
and the extra points of the e-set. The d-set can be used to input unsymmetric terms into the equations of motion using
DMIG and TF Bulk Data entries. In the modal formulations, the h-set is composed of the union of the ci (modal) set and
the e-set. Again, the DMIG and TF Bulk Data entries are a means to introduce unsymmetric matrix coefficients.

Sets for Aerodynamics

Aerodynamic calculations are made in a Cartesian aerodynamic coordinate system. By convention, the flow is in the positive
x direction. The basic coordinate system was not chosen, since it would place an undesirable restriction upon the description
of the structural model. Any MSC Nastran Cartesian system may be specified, and flow will be defined in the direction of
its positive x-axis. All element and aerodynamic grid point data, computed initially in the basic coordinate system, will be
converted to the aerodynamic coordinate system. The global (displacement) coordinate system of the aerodynamic grid

Main Index
 Chapter 6: Constraint and Set Notation 333
Rules of Sets for Undefined Degrees of Freedom

points will have its T1 direction in the flow direction. T3 is normal to the element for boxes, and parallel to the aerodynamic
T3 in the case of bodies. Coordinate system data are generated for the aerodynamic grid points.
The grid points are physically located at the centers of the boxes and body elements. Permanent constraints are generated
for the unused degrees of freedom. A second set of grid points, used only for undeformed plotting, is located at the element
corners. All six degrees of freedom associated with each grid point in this second set are permanently constrained. Grid point
numbers are generated based upon the element identification number. For any panel, the external grid point numbers for
the boxes start with the panel identification number and increase consecutively.
Aerodynamic degrees of freedom, along with the extra points, are added after the structural matrices and modes have been
determined. This introduces the following displacement sets:

k Aerodynamic box and body degrees of freedom

sa Permanently constrained degrees of freedom associated with aerodynamic grid points
ks = k + sa the union of k and the re-used s-set (6 degrees of freedom per grid)

js = j + sa the union of j and the re-used s-set (6 degrees of freedom per grid)

Rules of Sets for Undefined Degrees of Freedom

Most of the Bulk Data entries used to list set membership specify mutually exclusive sets. This means that degrees of
freedom listed on more than one such entry will result in a fatal error. For example, a degree of freedom may not be in the
m-set on two RBAR entries, or on a selected SPC entry and a SUPORT entry.
The a-set, by contrast, defines a combined set. Degrees-of-freedom listed on Bulk Data entries that define its subsets, the
q-, r-, c-, and b-set, may also be listed on ASETi Bulk Data entries, although this is a redundant specification and is not
necessary. However, automatic restarts are more reliable if ASETi entries are present for all degrees of freedom listed on
SUPORT entries.
Bulk Data entries used to define mutually exclusive sets automatically place them in their specified set and any combined
set to which they belong. Since the ASETi entries define a combined set, the subsets of the a-set, if not defined, must be
placed in a mutually exclusive set by an arbitrary convention. The conventions described below are used because they allow
the user to specify these sets with the minimum input for most cases. In superelement analysis, exterior (“boundary”)
degrees of freedom are automatically determined and placed in the a-set of the superelement. Exterior degrees of freedom
for superelements are therefore equivalent to degrees of freedom listed on ASETi entries in the discussion below.
The f-set is the union of the mutually exclusive o-, q-, r-, c-, and b-sets and includes the a-set. The o-set exists under two
conditions only:
1. If OMITi entries are present, the o-set consists of degrees of freedom listed explicitly on OMITi entries. The
remaining f-set degrees of freedom are placed in the b-set. They are then also members of the a-set.

Main Index
 334 MSC Nastran Reference Guide
Output Selection Via Set Specification

2. If ASETi or QSETi entries are present, the a-set consists of all degrees of freedom listed on ASETi entries and any
entries listing its subsets, such as QSETi, SUPORTi, CSETi, and BSETi entries. The remaining f-set degrees of
freedom are placed in the o-set. Note, however, there must be at least one explicit ASETi or QSETi entry type
present for the o-set to exist, even if it is a redundant specification. If there are no ASETi or QSETi entries present,
and SUPORT, BSETi, or CSETi are present, the entire f-set is placed in the a-set; that is, the o-set is not formed.
If both OMITi and ASETi entries are present, the OMITi entries are ignored, so that condition 2, above, applies. Similarly,
if ASETi entries are used in conjunction with entries defining its subsets, undefined degrees of freedom in the a-set are
placed in mutually exclusive sets as shown in the following table (0 = no entry types present; 1 = entry type present):

Case
Set Defined by Entry Type 1 2 3 4
c 0 0 1 1
b 0 1 0 1
Undefined a-set placed in b c b b

Output Selection Via Set Specification

The membership of grid points and their components and scalar points in the various sets is stored in the USET table as
digits in a binary number. Some modules use input parameters that are translations of these binary numbers as a shorthand
notation. Three formats are used:

Set name
(DEFUSET, USET and USET1 Bulk Data entries, MATGPR, UPARTN, VEC, UMERGE, UMERGE1, and PARAML
(OP = USET) Modules)
The data block naming convention in the solution sequences for matrices related to the sets is based on the set names. For
example, the system stiffness matrix  K gg  has the name KGG. Some sets are used that do not appear in the USET table.
The h-set, for example, is the set used for modal formulations. Its size is not known until after the USET table is formed,
so it does not appear in the table. The j-set, a set synonymous with the g-set, is used in superelement analysis. There are
several other sets used that also do not appear in the USET table.
Some names may be used for other purposes. The name H, for example, may be input to the MATGPR module when one
wants to label rows or columns of matrices sequentially, rather than by their external sequence numbers. This feature may
be used to number vectors in load matrices or in other instances not related to modal analysis, even though “H” is defined
for modal analysis.

Main Index
 Chapter 7: Solution Sequences
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

7 Solution Sequences

 Introduction 336

Output Description 339

Restart Procedures 360

General Operations 361

Summary of Solution Sequence Operations 398

Main Index
 336 MSC Nastran Reference Guide
Introduction

Introduction
MSC Nastran is composed of a large number of building blocks called modules. A module is a collection of FORTRAN
subroutines designed to perform a specific task—processing model geometry, assembling matrices, applying constraints,
solving matrices, calculating output quantities, conversing with the database, printing the solution, and so on. The modules
are controlled by an internal language called the Direct Matrix Abstraction Program (DMAP). Each type of analysis
available in MSC Nastran is called a solution sequence, and each solution sequence is a prepackaged collection of hundreds
or thousands of DMAP commands. Once a solution sequence is chosen, its particular set of DMAP commands sends
instructions to the modules that are needed to perform the requested solution. All of this happens automatically with no
effort on your part beyond choosing a solution sequence.
The solution sequences contain complete DMAP sequences required to perform certain types of analysis as well as
superelement generation (phase I) and data recovery (phase III). The solution sequences are listed in Table 7-1. An automatic
restart capability is employed in these sequences. The automatic restart is not only driven by modifications to the Case
Control and Bulk Data Sections but also by modifications in upstream superelements. A full description of automatic
restarts is given in Restart Procedures, 360.
A solution is a feature selected by a SOL Executive Control command. It is a collection of DMAP statements that performs
a specific type of analysis. For example, static analysis is requested by the command SOL 101. There are two types of
solutions at present.
The major type is called a Solution Sequence. They are called with numbers ranging from 100 to 200, 400, 600 and 700.
There were formerly Structured Solution Sequences and Unstructured Solution Sequences, but the latter class has been
removed, so the name Solution Sequence is now used, although you may find references to Structured Solutions in older
documentation. All new technology is introduced through Solution Sequences. They also have automatic restart capability.
This is accomplished with the NASTRAN Data Definition Language feature, a file that lists the names and attributes of
data types saved for restart. The evidence that a Solution Sequence has restart capability can be found by the existence of
TYPE statements in its DMAP compile listings. It uses the SubDMAP feature, which allow breaking a large collection of
DMAP statements into more manageable data structures similar to subroutines in other programming languages.

Main Index
 Chapter 7: Solution Sequences 337
Introduction

Solution Sequence
SESTATIC

DMAP Introduction SubDMAP PHASE1DR Matrix Operation Functional Module

DMAP (Accessible)

SubDMAP PHASE1A S S

Solution Sequence Operations

SubDMAP . . . SubDMAP SEMG

Functional Module General Operations

EMG
(Not Accessible)
Source

S S S

Figure 7-1 Hierarchy in the Structural Solution Sequences

Main Index
 338 MSC Nastran Reference Guide
Introduction

Table 7-1 Structured Solution Sequences

SOL SOL
Number Name Description References
101 SESTATIC Statics with options: page 376
Linear Heat Transfer page 424
Alternate Reduction
Inertia Relief
Design Sensitivity -- Statics
103 SEMODES Normal Modes with option: page 408
Design Sensitivity -- Modes
105 SEBUCKL Buckling with options: page 424
Static Analysis
Design Sensitivity -- Buckling
106 NLSTATIC Nonlinear Statics page 684
107 SEDCEIG Direct Complex Eigenvalues page 423
108 SEDFREQ Direct Frequency Response page 423
109 SEDTRAN Direct Transient Response page 424
110 SEMCEIG Modal Complex Eigenvalues page 423
111 SEMFREQ Modal Frequency Response page 423
112 SEMTRAN Modal Transient Response page 424
114 CYCSTATX Cyclic Statics with option: page 703
Alternate Reduction
115 CYCMODE Cyclic Normal Modes page 703
118 CYCFREQ Cyclic Direct Frequency Response page 703
129 NLTRAN Nonlinear Transient Response page 568
144 AESTAT Static Aeroelastic Response See the MSC Nastran Aeroelastic
Analysis User’s Guide
145 SEFLUTTR Aerodynamic Flutter See the MSC Nastran Aeroelastic
Analysis User’s Guide
146 SEAERO Aeroelastic Response See the MSC Nastran Aeroelastic
Analysis User’s Guide
153 NLSCSH Static Structural and/or Steady State Heat page 438
Transfer Analysis
159 NLTCSH Transient Structural and/or Transient Heat page 438
Transfer Analysis

Main Index
 Chapter 7: Solution Sequences 339
Output Description

Table 7-1 Structured Solution Sequences (continued)

SOL SOL
Number Name Description References
200 DESOPT Design Optimization See the MSC.Nastran Design
Sensitivity and Optimization
User’s Guide
400 NONLIN Nonlinear Static and Transient Analysis See the MSC Nastran Nonlinear
User’s Guide
600,ID SESTATIC MSC Nastran, Marc in MSC Nastran See the MSC Nastran Implicit
Nonlinear User’s Guide
700 NLTRAN Explicit nonlinear analysis See the MSC Nastran Explicit
Nonlinear (SOL 700) User’s Guide

The DMAP listings for any of the solutions sequences can be obtained by specifying a COMPILER LIST statement in the
Executive Control Section. A description of the data processing and matrix operations for the solution sequences is given
in Summary of Solution Sequence Operations, 398. The complete details for each of the functional modules are given in Chapter
4 of the MSC Nastran Programmer’s Manual. Those DMAP modules which are of general use (matrix operations and utility
modules) are described along with examples of their use in General Operations, 361.
If the user wishes to modify a DMAP sequence in some manner, he can use the alter feature described in MSC Nastran
DMAP Programmer’s Guide. Typical uses are to schedule an EXIT prior to completion in order to check intermediate
output, schedule the printing of a table or matrix for diagnostic purposes, and to delete or add a functional module to the
DMAP sequence. The user should be familiar with the rules for DMAP programming, as described in the Msc Nastran
DMAP Programmer’s Guide, to make alterations to a solution sequence. A library of DMAP alters is delivered with MSC
Nastran. These alters are used to request options not available in the solution sequences and for exceptional operations that
do not fit in with the normal flow of the solution sequences.

Output Description
Although most of the solution sequence output is optional and requested in the Case Control Section, some of the printer
output is automatic or under control of DIAG statements in the Executive Control Section. The printer output is designed
for 132 characters per line, with the lines per page controlled by the LINE command in the Case Control Section. The
default is LINE = 50 for 11-inch paper. Optional titles are printed at the top of each page from information in the Case
Control Section. These titles may be defined at the subcase level. The pages are automatically dated and numbered.
The output from data recovery and plot modules is all optional and its selection is controlled by commands in the Case
Control Section. The details of making selections in the Case Control Section are described in Case Control Section, 14 for
printer and punch output and in Plotting, 467 for plotter output. The available output for each solution sequence is described
in Summary of Solution Sequence Operations, 398. Detailed information on the force and stress output available for each element
type is given in Structural Elements, 40.

Main Index
 340 MSC Nastran Reference Guide
Output Description

A few printer output items are under the control of PARAM Bulk Data entries. The use of PARAM entries is described in
Use of Parameters, 31. The DIAG statement is used to control the printing of diagnostic output. The available output under
the DIAG statement is given in the description of the Executive Control statements in the Executive Control Statements
(p. 107) in the MSC Nastran Quick Reference Guide.
The first part of the output for an MSC Nastran run is prepared during the execution of the preface, prior to the beginning
of the solution sequence. The following output is either automatically or optionally provided during execution:
 MSC Nastran title page — automatic
 File Management Section echo — automatic if statements are present
 Executive Control Section echo — automatic
 Case Control Section echo — automatic
 Unsorted Bulk Data Section echo — optional, selected in the Case Control Section
 Sorted Bulk Data Section echo — automatic, unless suppressed in the Case Control Section

MSC Nastran Output

MSC Nastran generates the following output files.

File Contents
Execution Summary Computer model, operating system, MSC Nastran version, system creation date, problem
Table in *.f04 run date, MSC Nastran file assignments, and a summary of the MSC Nastran execution:
time of day, elapsed time, I/O seconds, CPU seconds, module execution messages, matrix
trailers (DIAG 8), value of REAL (DIAG 13), MPYAD, SDCOMP and FBS estimates
(DIAG 19), and statistical information on module expected/actual CPU item (DIAG 49).
Standard Output in *.f06 MSC Nastran results (i.e., displacements, stresses, etc.). See the MSC Nastran Linear Static
Analysis User’s Guide and the MSC Nastran Numerical Methods User’s Guide for more
information.

An example follows which shows the execution summary (.f04) and the standard output (.f06) for SOL 101.
Execution Summary Contents (see Listing 7-1)

The information in the Execution Summary file can provide the user with CPU time, I/O time, elapsed time and disk space
resource usage information on the execution of the problem. To aid in the interpretation of these data, an example is
presented on the following pages. Circled integers have been added for ease of reference.

1 The column headed DAY TIME indicates the time of day as the run progresses.
2 The column headed ELAPSED indicates the number of elapsed “wall clock” minutes:seconds at the beginning on
the job. This number may vary greatly depending on the system load in a time-sharing environment.

Main Index
 Chapter 7: Solution Sequences 341
Output Description

3 The column headed I/O SEC is an internally generated measure of I/O. This measure is a count of the number
of BUFFSIZE blocks transferred, divided by the value of SYSTEM(84) (the estimated rate of number of blocks
moved per second). This number does not include I/O for loading the executable file or I/O for FORTRAN files.
The column headed DEL I/O is the I/O seconds spent in the previous DMAP statement.
4 The column headed CPU SEC shows the amount of CPU seconds that elapsed when this line was printed. It is
recommended that the user examine these entries and look for large jumps in this number and try to understand
why each one occurred. For example, the cost for “formatting” the output (modules SDR2 and OFP) may be
greater than the cost of generating the numbers. In such situations, simple techniques such as reducing output
requests may save more money than other more well-known techniques such as sequencing, OMITs, better matrix
methods, etc. In short, the user should be aware of the expensive operations on each large run. The column headed
DEL CPU is the CPU time spent in the previous DMAP statement.
5 Primary database file assignments.
6 The column headed BUFFSIZE is the buffer size of each DBset. This buffer size may be defined by the INIT
command for each DBset (see the File Management Statements (p. 43) in the MSC Nastran Quick Reference Guide).
7 The column headed CLUSTER SIZE is the unit of allocation for GINO and executive blocks. This unit may be
defined on the INIT command (see the File Management Statements (p. 43) in the MSC Nastran Quick Reference
Guide).
8 The column headed TIME STAMP is the time the file was created. The format is YYMMDDHHMMSS.
9 The delivery database file assignments. These files contain the Solution Sequences and timing constants.
10 Indicates the FMS statement INIT MASTER (RAM) has been invoked (by default).
11 SUBDMAP name.
12 DMAP module name or subDMAP name.
13 The line REAL CORE IS XXX shows the value of the NASTRAN keyword REAL. This message is generated by
using the DIAG 13 command in the Executive Control Section of the MSC Nastran data file. It only appears after
module executions which use the value. The open core (i.e., working space) for other modules is set via HICORE
(see the ).
14 The number that precedes the module name on lines that terminate with BEGN or END is the DMAP statement
number. This number corresponds to the statement number in the DMAP listing for the solution sequence being
used. The DMAP sequence listing will be printed as a part of the output if DIAG 14 or COMPILER LIST is
inserted in the Executive Control Section.

Main Index
 342 MSC Nastran Reference Guide
Output Description

15 BEGN indicates the start of execution of an MSC Nastran module or subDMAP. The functions of these modules
are given in Section . Section 4 of the MSC Nastran Programmer’s Manual provides additional detailed information
on MSC Nastran functional modules.
END indicates the end of execution of an MSC Nastran module or subDMAP. The resources utilized by each
module can be estimated by computing the incremental CPU SECs and I/O SECs required to execute the
module. This line of information will not normally appear unless the module required more than the CPU
seconds defined in SYSTEM(20). SYSTEM is defined via the NASTRAN statement and has a default of 5. M is
the timing constant for the user’s computer type, as described in the . The user may force the output of this line
by including the keyword SYSTEM(20) = 0 on the NASTRAN statement.
16 The line “*8**” is a typical matrix trailer. This line is output for each MSC Nastran matrix that is generated if
DIAG 8 is included in the Executive Control Section. Some parameters may be redefined using DMAP as
explained in the MSC Nastran DMAP Programmer’s Guide. Other terms in the DIAG 8 output are: the number
of GINO blocks in the matrix (BLOCKS); the average number of terms per string (STRL); the number of strings
in a matrix data block (NBRSTR); the number of bank terms in a matrix data block on the average (BNDL); the
number of banks in a matrix data block (NBRBND); the average first nonzero row position (ROWL); the average
of full bank width (BNDAVG); the maximum full bank width of all columns of data blocks (BNDMAX); the
number of null columns in a data block (NULCOL).
17 User Information Message 4157 describes parameters for decomposition. The format of this message varies
slightly if either sparse or parallel methods are selected.
18 The time estimation and method selection data are printed for DCMP, FBS, and MPYAD modules if DIAG 19
is included in the Executive Control Section.
19 A detailed analysis of timing data will be printed in the .f04 file if DIAG 49 is set. The results of this analysis
appear at the end of the .f04 file in three separate sections. The first section contains information about each
module and submodule. The second section reports statistical information about the submodules (FBS, MPYAD,
and DCMP) grouped by specific methods. Instances where the CPU estimate does not pass the following criteria
are output:
Estimated CPU – Actual CPU-
--------------------------------------------------------------------------  0.05
Actual CPU
or
Estimated CPU – Actual CPU  5% of the total CPU of the job
The third section contains information about the modules listed in alphabetical order. (The numbers shown are
for a very small model with execution times less than the integer seconds used in these calculations.)
20 This table provides the CPU usage for each process created and executived by a parallel module.
21 The column headed BLOCKS USED contains the number of GINO and executive blocks used at the end of the
job. One GINO Block = BUFFSIZE-1 words.
22 The column headed HIWATER BLOCK contains the maximum number of GINO and executive blocks used
during the entire execution.

Main Index
 Chapter 7: Solution Sequences 343
Output Description

Standard Output Contents

Main Index
 344 MSC Nastran Reference Guide
Output Description

Listing 7-1 Example Execution Summary File (.f04)

MACHINE MODEL OPERATING SYSTEM MSC/NASTRAN BUILD DATE RUN DATE
DIGITAL EQUIPMENT VAX 6000-410 VMS V5.4-3 VERSION 68 MAR 3, 1994 MAR 11, 1994

=== M S C / N A S T R A N E X E C U T I O N S U M M A R Y ===

DAY TIME ELAPSED I/O SEC DEL_I/O CPU SEC DEL_CPU SUB_DMAP/DMAP_MODULE MESSAGES

10:01:52 0:00 0.0 0.0 0.0 0.0 DBINIT BGN

** CURRENT PROJECT ID = ' "BLANK" ' ** CURRENT VERSION ID = 1

S U M M A R Y O F F I L E A S S I G N M E N T F O R T H E P R I M A R Y D A T A B A S E ( DBSNO 1, V68 )
9
ASSIGNED PHYSICAL FILE NAME (/ORIGINAL) LOGICAL NAME DBSET STATUS BUFFSIZE CLUSTER SIZE TIME STAMP
--------------------------------------- ------------ ----- ------ -------- ------------ ------------
SCR_KNIGHT:[SCRATCH]N10013850.MASTER MASTER MASTER NEW 1409 1 940311100153
SCR_KNIGHT:[SCRATCH]N10013850.USROBJ USROBJ USROBJ NEW 1409 1 940311100154
SCR_KNIGHT:[SCRATCH]N10013850.USRSOU USRSOU USRSOU NEW 1409 1 940311100155
SCR_KNIGHT:[SCRATCH]N10013850.DBALL DBALL DBALL NEW 1409 1 940311100156
SCR_KNIGHT:[SCRATCH]N10013850.OBJSCR OBJSCR OBJSCR NEW 1409 1 940311100157
**** MEM FILE **** * N/A * SCRATCH
SCR_KNIGHT:[SCRATCH]N10013850.SCRATCH SCRATCH SCRATCH NEW 1409 1 940311100158
SCR_KNIGHT:[SCRATCH]N10013850.SCR300 SCR300 SCRATCH NEW 1409 1 940311100159
10:01:57 0:05 4.5 4.5 0.6 0.7 DBINIT END
10:01:57 0:05 4.7 0.2 0.7 0.0 XCSA BGN

S U M M A R Y O F F I L E A S S I G N M E N T F O R T H E D E L I V E R Y D A T A B A S E ( DBSNO 2, V68 )

ASSIGNED PHYSICAL FILE NAME (/ORIGINAL) LOGICAL NAME DBSET STATUS BUFFSIZE CLUSTER SIZE TIME STAMP
--------------------------------------- ------------ ----- ------ -------- ------------ ------------
nast:[nast68.DEL]SSS.MASTERA MASTERA MASTER OLD 1409 1 940303153920
/NAST:[NAST68.DEL.SOURCE.OUT]SSS.MASTERA
10
nast:[nast68.DEL]SSS.MSCOBJ
/NAST:[NAST68.DEL.SOURCE.OUT]SSS.MSCOBJ
MSCOBJ MSCOBJ OLD 1409 1 940303154017

nast:[nast68.DEL]SSS.MSCSOU MSCSOU MSCSOU OLD 1409 1 940303154032

/NAST:[NAST68.DEL.SOURCE.OUT]SSS.MSCSOU
10:02:57 1:05 115.3 110.7 7.1 6.4 XCSA END
10:02:57 1:05 115.3 0.0 7.1 0.0 CGPI BGN
10:02:57 1:05 115.3 0.0 7.1 0.1 CGPI END
10:02:57 1:05 115.3 0.0 7.1 0.0 LINKER BGN
10:03:09 1:17 133.6 18.3 9.1 2.0 LINKER END

** MASTER DIRECTORIES ARE LOADED IN MEMORY. 11 12

USER OPENCORE (HICORE) = 6074952 WORDS
EXECUTIVE SYSTEM WORK AREA = 76045 WORDS
MASTER(RAM) = 30000 WORDS
SCRATCH(MEM) AREA = 140900 WORDS ( 100 BUFFERS)
BUFFER POOL AREA (GINO/EXEC) = 52503 WORDS ( 37 BUFFERS)
TOTAL MSC/NASTRAN MEMORY LIMIT = 6374400 WORDS

DAY TIME ELAPSED I/O SEC DEL_I/O CPU SEC DEL_CPU SUB_DMAP/DMAP_MODULE MESSAGES

10:03:11 1:19 135.6 2.0 9.4 0.3 SESTATIC 19 IFPL BEGN

10:03:11 1:19 135.8 0.1# 9.4 0.0# SESTATIC 19 IFPL END
10:03:12 1:20 136.1 0.4 9.5 0.0 IFPL 22 IFP1 BEGN
10:03:12 1:20 136.1 0.0# 9.5 0.0# IFPL 22 IFP1 END
10:03:12 1:20 136.1 0.0 9.5 0.0 IFPL 24 XEQUIV BEGN
10:03:12 1:20 136.1 0.0# 9.5 0.0# IFPL 24 XEQUIV END
10:03:12 1:20 136.1 0.0 9.5 0.0 IFPL 25 XEQUIV BEGN
10:03:12 1:20 136.1 0.0# 9.5 0.0# IFPL 25 XEQUIV END
10:03:12
10:03:12
1:20
1:20
136.1
136.1
0.0
0.0#
9.5
9.5 13
0.0
0.0#
IFPL
IFPL
26
26
XEQUIV BEGN
XEQUIV END
10:03:12 1:20 136.1 0.0 9.5 0.0 IFPL 27 XEQUIV BEGN
10:03:12 1:20 136.1 0.0# 9.5 0.0# IFPL 27 XEQUIV END
10:03:12 1:20 136.1 0.0 9.6 0.0 IFPL 28 XEQUIV BEGN
10:03:12 1:20 136.1 0.0# 9.6 0.0# IFPL 28 XEQUIV END
10:03:12 1:20 136.1 0.0 9.6 0.0 IFPL 35 XSORT BEGN
REAL CORE IS 6074952
10:03:14 1:22 136.1 0.0# 10.8 1.3# IFPL 35 XSORT END
10:03:36 1:44 137.1 1.0 14.8 4.0 IFPL 61 IFP BEGN
10:03:37 1:45 137.1 0.0 14.9 0.1 IFP
10:03:39 1:47 137.1 0.0# 15.1 0.3# IFPL 61 IFP END
10:03:39 1:47 137.1 0.0 15.1 0.0 IFPL 62 MODEPT BEGN
10:03:39 1:47 137.1 0.0# 15.1 0.0# IFPL 62 MODEPT END

Main Index
 Chapter 7: Solution Sequences 345
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

10:03:39 1:47 137.1 0.0 15.2 0.0 IFPL 63 MODGM2 BEGN
10:03:40 1:48 137.1 0.0# 15.2 0.0# IFPL 63 MODGM2 END
10:03:40 1:48 137.1 0.0 15.2 0.0 IFPL 64 PVT BEGN
10:03:40 1:48 137.1 0.0# 15.3 0.0# IFPL 64 PVT END
10:03:40 1:48 137.6 0.5 15.3 0.1 IFPL 79 PROJVER BEGN
10:03:41 1:49 137.6 0.0# 15.3 0.0# IFPL 79 PROJVER END
10:03:41 1:49 137.6 0.0 15.3 0.0 IFPL 81 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 81 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 82 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 82 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 83 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 83 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 84 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 84 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 85 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 85 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 86 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 86 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 87 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 87 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 88 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 88 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 89 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 89 XEQUIV END
10:03:41 1:49 137.6 0.0 15.4 0.0 IFPL 90 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.4 0.0# IFPL 90 XEQUIV END
10:03:41 1:49 137.6 0.0 15.5 0.0 IFPL 92 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.5 0.0# IFPL 92 XEQUIV END
10:03:41 1:49 137.6 0.0 15.5 0.0 IFPL 93 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.5 0.0# IFPL 93 XEQUIV END
10:03:41 1:49 137.6 0.0 15.5 0.0 IFPL 94 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.5 0.0# IFPL 94 XEQUIV END
10:03:41 1:49 137.6 0.0 15.5 0.0 IFPL 96 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.5 0.0# IFPL 96 XEQUIV END
10:03:41 1:49 137.6 0.0 15.5 0.0 IFPL 97 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.5 0.0# IFPL 97 XEQUIV END
10:03:41 1:49 137.6 0.0 15.5 0.0 IFPL 98 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.6 0.0# IFPL 98 XEQUIV END
10:03:41 1:49 137.6 0.0 15.6 0.0 IFPL 99 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.6 0.0# IFPL 99 XEQUIV END
10:03:41 1:49 137.6 0.0 15.6 0.0 IFPL 100 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.6 0.0# IFPL 100 XEQUIV END
10:03:41 1:49 137.6 0.0 15.6 0.0 IFPL 101 XEQUIV BEGN
10:03:41 1:49 137.6 0.0# 15.6 0.0# IFPL 101 XEQUIV END
10:03:41 1:49 137.6 0.0 15.6 0.0 IFPL 115 PVT BEGN
10:03:41 1:49 137.6 0.0# 15.6 0.0# IFPL 115 PVT END
10:03:41 1:49 137.6 0.0 15.6 0.0 IFPL 128 DELETE BEGN
10:03:41 1:49 137.6 0.0# 15.6 0.0# IFPL 128 DELETE END
10:03:41 1:49 137.8 0.2 15.6 0.0 IFPL 154 PLTSET BEGN
10:03:41 1:49 137.8 0.0# 15.7 0.0# IFPL 154 PLTSET END
10:03:41 1:49 137.8 0.0 15.7 0.0 IFPL 157 GP0 BEGN
10:03:42 1:50 137.8 0.0# 15.7 0.0# IFPL 157 GP0 END
10:03:42 1:50 137.8 0.0 15.7 0.0 IFPL 176 SEQP BEGN
10:03:43 1:51 137.8 0.0# 15.9 0.1# IFPL 176 SEQP END
10:03:43 1:51 137.8 0.0 15.9 0.0 IFPL 182 GP1 BEGN
10:03:44 1:52 137.8 0.0# 16.1 0.2# IFPL 182 GP1 END
10:03:44 1:52 137.9 0.1 16.1 0.0 SESTATIC 36 PHASE0 BEGN
10:03:44
10:03:44
1:52
1:52
138.1
138.1
0.2#
0.0
16.1
16.1
0.0#
0.0
SESTATIC
PHASE0
36
24
14PHASE0
DTIIN
END
BEGN
15
10:03:44 1:52 138.1 0.0# 16.2 0.0# PHASE0 24 DTIIN END
10:03:44 1:52 138.1 0.0 16.2 0.0 PHASE0 25 GP1 BEGN
10:03:45 1:53 138.1 0.0# 16.4 0.2# PHASE0 25 GP1 END
10:03:46 1:54 138.1 0.0 16.4 0.0 PHASE0 28 SEP1 BEGN
10:03:46 1:54 138.1 0.0# 16.4 0.0# PHASE0 28 SEP1 END
10:03:46 1:54 138.1 0.0 16.4 0.0 PHASE0 37 XEQUIV BEGN
10:03:46 1:54 138.1 0.0# 16.4 0.0# PHASE0 37 XEQUIV END
10:03:46 1:54 138.1 0.0 16.5 0.0 PHASE0 40 PVT BEGN
10:03:48 1:56 138.1 0.0# 16.6 0.2# PHASE0 40 PVT END
10:03:48 1:56 138.1 0.0 16.6 0.0 PHASE0 46 TABPRT BEGN
10:03:48 1:56 138.1 0.0# 16.7 0.0# PHASE0 46 TABPRT END
10:03:48 1:56 138.1 0.0 16.7 0.0 PHASE0 47 SEP3 BEGN
10:03:48 1:56 138.1 0.0# 16.7 0.0# PHASE0 47 SEP3 END
10:03:48 1:56 138.2 0.1 16.7 0.1 PHASE0 88 PVT BEGN
10:03:49 1:57 138.2 0.0# 16.8 0.1# PHASE0 88 PVT END
10:03:49 1:57 138.9 0.7 16.8 0.0 PHASE0 96 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.8 0.0# PHASE0 96 XEQUIV END
10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 97 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.9 0.0# PHASE0 97 XEQUIV END
10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 98 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.9 0.0# PHASE0 98 XEQUIV END
10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 99 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.9 0.0# PHASE0 99 XEQUIV END
10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 100 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.9 0.0# PHASE0 100 XEQUIV END

Main Index
 346 MSC Nastran Reference Guide
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 101 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.9 0.0# PHASE0 101 XEQUIV END
10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 102 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.9 0.0# PHASE0 102 XEQUIV END
10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 103 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.9 0.0# PHASE0 103 XEQUIV END
10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 104 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 16.9 0.0# PHASE0 104 XEQUIV END
10:03:49 1:57 138.9 0.0 16.9 0.0 PHASE0 105 XEQUIV BEGN
10:03:49 1:57 138.9 0.0# 17.0 0.0# PHASE0 105 XEQUIV END
10:03:50 1:58 140.7 1.8 17.1 0.2 PHASE0 224 PROJVER BEGN
10:03:50 1:58 140.7 0.0# 17.1 0.0# PHASE0 224 PROJVER END
10:03:50 1:58 140.7 0.0 17.1 0.0 PHASE0 241 SETQ BEGN
10:03:50 1:58 140.8 0.1# 17.1 0.0# PHASE0 241 SETQ END
10:03:50 1:58 140.8 0.0 17.1 0.0 SETQ 10 XEQUIV BEGN
10:03:50 1:58 140.8 0.0# 17.1 0.0# SETQ 10 XEQUIV END
10:03:50 1:58 140.8 0.1 17.3 0.2 PHASE0 307 GP1 BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
GP1 307 VGFS 1 12 2 2 0 0.0000 1 0 1 0 0 0 0 0 1 *8**
10:03:51 1:59 140.8 0.0# 17.5 0.2# PHASE0 307 GP1 END
10:03:51 1:59 140.9 0.1 17.5 0.0 PHASE0 311 XEQUIV BEGN
10:03:51 1:59 140.9 0.0# 17.5 0.0# PHASE0 311 XEQUIV END
10:03:51 1:59 140.9 0.0 17.5 0.0 PHASE0 313 SEPREP2 BEGN
10:03:51 1:59 141.0 0.1# 17.5 0.0# PHASE0 313 SEPREP2 END
10:03:51 1:59 141.0 0.0 17.5 0.0 SEPREP2 6 GP2 BEGN
10:03:51 1:59 141.0 0.0# 17.5 0.0# SEPREP2 6 GP2 END
10:03:51 1:59 141.1 0.1 17.6 0.0 SEPREP2 12 XEQUIV BEGN
10:03:51 1:59 141.1 0.0# 17.6 0.0# SEPREP2 12 XEQUIV END
10:03:51 1:59 141.1 0.0 17.6 0.0 SEPREP2 13 XEQUIV BEGN
10:03:51 1:59 141.1 0.0# 17.6 0.0# SEPREP2 13 XEQUIV END
10:03:51 1:59 141.1 0.0 17.6 0.0 SEPREP2 14 XEQUIV BEGN
10:03:51 1:59 141.1 0.0# 17.6 0.0# SEPREP2 14 XEQUIV END
10:03:51 1:59 141.1 0.0 17.6 0.0 SEPREP2 15 XEQUIV BEGN
10:03:51 1:59 141.1 0.0# 17.6 0.0# SEPREP2 15 XEQUIV END
10:03:51 1:59 141.1 0.0 17.6 0.0 PHASE0 332 XEQUIV BEGN
10:03:51 1:59 141.1 0.0# 17.6 0.0# PHASE0 332 XEQUIV END
10:03:51 1:59 141.1 0.0 17.6 0.0 PHASE0 333 XEQUIV BEGN
10:03:51 1:59 141.1 0.0# 17.6 0.0# PHASE0 333 XEQUIV END
10:03:51 1:59 141.1 0.0 17.7 0.0 PHASE0 335 GP3 BEGN
10:03:51 1:59 141.1 0.0# 17.7 0.0# PHASE0 335 GP3 END
10:03:51 1:59 141.1 0.0 17.7 0.0 PHASE0 336 LCGEN BEGN
10:03:51 1:59 141.1 0.0# 17.7 0.0# PHASE0 336 LCGEN END
10:03:51 1:59 141.1 0.0 17.7 0.0 PHASE0 341 XEQUIV BEGN
10:03:51 1:59 141.1 0.0# 17.7 0.0# PHASE0 341 XEQUIV END
10:03:51 1:59 141.1 0.0 17.8 0.0 PHASE0 352 BCDR BEGN
10:03:51 1:59 141.1 0.0# 17.8 0.0# PHASE0 352 BCDR END
10:03:51 1:59 141.2 0.1 17.8 0.0 PHASE0 367 GP4 BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
GP4 367 YSB 1 6 2 2 0 0.0000 1 0 1 0 0 0 0 0 1 *8**
10:03:52 2:00 141.2 0.0# 17.8 0.0# PHASE0 367 GP4 END
10:03:52 2:00 141.2 0.0 17.9 0.0 PHASE0 373 XPURGE BEGN
10:03:52 2:00 141.2 0.0# 17.9 0.0# PHASE0 373 XPURGE END
10:03:52 2:00 141.2 0.0 17.9 0.0 PHASE0 378 UMERGE1 BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
UMERGE1 378 SCRATCH 1 12 2 1 6 0.5000 1 6 1 0 0 0 6 6 0 *8**
10:03:52 2:00 141.2 0.0# 17.9 0.0# PHASE0 378 UMERGE1 END
10:03:52 2:00 141.3 0.1 18.0 0.1 SESTATIC 39 SETQ BEGN
10:03:52 2:00 141.3 0.0# 18.0 0.0# SESTATIC 39 SETQ END
10:03:52 2:00 141.3 0.0 18.0 0.0 SETQ 10 XEQUIV BEGN
10:03:52 2:00 141.3 0.0# 18.0 0.0# SETQ 10 XEQUIV END
10:03:52 2:00 141.3 0.0 18.2 0.2 SESTATIC 40 DBSTATUSBEGN
10:03:52 2:00 141.3 0.0# 18.2 0.0# SESTATIC 40 DBSTATUSEND
10:03:52 2:00 141.3 0.0 18.2 0.0 SESTATIC 45 MATGEN BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
MATGEN 45 TEMPALL 2 2 6 2 2 0.5000 1 1 2 0 0 0 1 1 0 *8**
10:03:52 2:00 141.3 0.0# 18.2 0.0# SESTATIC 45 MATGEN END
10:03:52 2:00 141.3 0.0 18.2 0.0 SESTATIC 46 RESTART BEGN
TEMPALL HAS CHANGED
10:03:52 2:00 141.9 0.6# 18.3 0.0# SESTATIC 46 RESTART END
10:03:53 2:01 142.0 0.1 18.3 0.0 SESTATIC 47 DELETE BEGN
10:03:53 2:01 142.0 0.0# 18.3 0.0# SESTATIC 47 DELETE END
10:03:53 2:01 142.1 0.1 18.3 0.0 SESTATIC 49 PHASE1DRBEGN
10:03:53 2:01 142.2 0.1# 18.3 0.0# SESTATIC 49 PHASE1DREND
10:03:53 2:01 142.4 0.2 18.4 0.1 PHASE1DR 70 PVT BEGN
10:03:53 2:01 142.4 0.0# 18.4 0.1# PHASE1DR 70 PVT END
10:03:53 2:01 142.4 0.0 18.4 0.0 PHASE1DR 71 SETQ BEGN
10:03:53 2:01 142.4 0.0# 18.5 0.0# PHASE1DR 71 SETQ END
10:03:53 2:01 142.4 0.0 18.5 0.0 SETQ 10 XEQUIV BEGN
10:03:53 2:01 142.4 0.0# 18.5 0.0# SETQ 10 XEQUIV END
10:03:53 2:01 142.4 0.0 18.6 0.2 PHASE1DR 74 PHASE1A BEGN
10:03:53 2:01 142.6 0.2# 18.7 0.0# PHASE1DR 74 PHASE1A END
10:03:53 2:01 142.6 0.0 18.7 0.0 PHASE1A 24 TA1 BEGN
REAL CORE IS 6073541

Main Index
 Chapter 7: Solution Sequences 347
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
EMA 79 KJJZ 12 12 6 2 8 0.2777 1 0 40 0 0 0 9 11 0 *8**
10:03:54 2:02 142.9 0.0# 19.0 0.0# SEMG 79 EMA END
10:03:54 2:02 142.9 0.0 19.0 0.0 SEMG 95 XPURGE BEGN
10:03:54 2:02 142.9 0.0# 19.0 0.0# SEMG 95 XPURGE END
10:03:55 2:03 143.1 0.2 19.0 0.0 SEMG 98 MTRXIN BEGN
REAL CORE IS 6074952
10:03:55 2:03 143.1 0.0# 19.0 0.0# SEMG 98 MTRXIN END
10:03:55 2:03 143.1 0.0 19.0 0.0 SEMG 103 SEMG1 BEGN
10:03:55 2:03 143.1 0.1# 19.1 0.0# SEMG 103 SEMG1 END
10:03:55 2:03 143.1 0.0 19.1 0.0 SEMG1 8 XEQUIV BEGN
10:03:55 2:03 143.1 0.0# 19.1 0.0# SEMG1 8 XEQUIV END
10:03:55 2:03 143.1 0.0 19.1 0.0 SEMG1 15 XEQUIV BEGN
10:03:55 2:03 143.1 0.0# 19.1 0.0# SEMG1 15 XEQUIV END
10:03:55 2:03 143.1 0.0 19.1 0.0 SEMG1 17 XEQUIV BEGN
10:03:55 2:03 143.1 0.0# 19.1 0.0# SEMG1 17 XEQUIV END
10:03:55 2:03 143.1 0.0 19.1 0.0 SEMG1 19 XEQUIV BEGN
10:03:55 2:03 143.1 0.0# 19.1 0.0# SEMG1 19 XEQUIV END
10:03:55 2:03 143.1 0.0 19.1 0.0 PHASE1A 31 SESUM BEGN
10:03:55 2:03 143.1 0.0# 19.1 0.0# PHASE1A 31 SESUM END
10:03:55 2:03 143.1 0.0 19.2 0.0 SESUM 22 DELETE BEGN
10:03:55 2:03 143.1 0.0# 19.2 0.0# SESUM 22 DELETE END
10:03:55 2:03 143.4 0.3 19.2 0.0 SESUM 25 XPURGE BEGN
10:03:55 2:03 143.4 0.0# 19.2 0.0# SESUM 25 XPURGE END
10:03:55 2:03 143.6 0.2 19.2 0.0 PHASE1A 35 XEQUIV BEGN
10:03:55 2:03 143.6 0.0# 19.2 0.0# PHASE1A 35 XEQUIV END
10:03:55 2:03 143.6 0.0 19.2 0.0 PHASE1A 60 SELG BEGN
10:03:55 2:03 143.7 0.1# 19.3 0.0# PHASE1A 60 SELG END
16
10:03:55
*8** MODULE
2:03 143.7
DMAP MATRIX
0.0
COLS
19.3
ROWS F T
0.0
NZWDS
SELG
DENS BLOCKS
11 SSG1 BEGN
STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL

SSG1 11 SCRATCH 1 12 2 2 6 0.2500 1 3 1 0 0 0 3 3 0 *8**

SSG1 11 PJX 1 12 2 2 6 0.2500 1 3 1 0 0 0 3 3 0 *8**
10:03:56 2:04 143.7 0.0# 19.3 0.1# SELG 11 SSG1 END
10:03:56 2:04 143.7 0.0 19.4 0.0 SELG 22 XEQUIV BEGN
10:03:56 2:04 143.7 0.0# 19.4 0.0# SELG 22 XEQUIV END
10:03:56 2:04 143.7 0.0 19.4 0.0 SELG 60 XEQUIV BEGN
10:03:56 2:04 143.7 0.0# 19.4 0.0# SELG 60 XEQUIV END
10:03:56 2:04 143.7 0.0 19.4 0.0 SELG 62 VECPLOT BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
VECPLOT 62 PPJ 1 12 2 1 3 0.2500 1 3 1 0 0 0 3 3 0 *8**
10:03:56 2:04 143.7 0.0# 19.4 0.0# SELG 62 VECPLOT END
10:03:56 2:04 143.7 0.0 19.5 0.0 PHASE1A 61 SESUM BEGN
10:03:56 2:04 143.7 0.0# 19.5 0.0# PHASE1A 61 SESUM END
10:03:56 2:04 143.7 0.0 19.5 0.0 SESUM 22 DELETE BEGN
10:03:56 2:04 143.8 0.1# 19.5 0.0# SESUM 22 DELETE END
10:03:56 2:04 144.4 0.6 19.6 0.0 SESUM 25 XPURGE BEGN
10:03:56 2:04 144.4 0.0# 19.6 0.0# SESUM 25 XPURGE END
10:03:57 2:05 144.5 0.1 19.6 0.0 PHASE1A 65 SELA1 BEGN
10:03:57 2:05 144.6 0.1# 19.6 0.0# PHASE1A 65 SELA1 END
10:03:57 2:05 144.6 0.0 19.6 0.0 SELA1 9 XEQUIV BEGN
10:03:57 2:05 144.6 0.0# 19.6 0.0# SELA1 9 XEQUIV END
10:03:57 2:05 144.6 0.0 19.6 0.0 SELA1 11 XEQUIV BEGN
10:03:57 2:05 144.6 0.0# 19.6 0.0# SELA1 11 XEQUIV END
10:03:57 2:05 144.6 0.0 19.7 0.0 PHASE1DR 78 DBSTATUSBEGN
10:03:57 2:05 144.6 0.0# 19.7 0.0# PHASE1DR 78 DBSTATUSEND
10:03:57 2:05 144.6 0.0 19.7 0.0 PHASE1DR 80 XEQUIV BEGN
10:03:57 2:05 144.6 0.0# 19.7 0.0# PHASE1DR 80 XEQUIV END
10:03:57 2:05 144.6 0.0 19.7 0.0 PHASE1DR 98 BCDR BEGN
10:03:57 2:05 144.6 0.0# 19.7 0.0# PHASE1DR 98 BCDR END
10:03:57 2:05 144.6 0.0 19.7 0.0 PHASE1DR 100 PHASE1B BEGN
10:03:57 2:05 144.7 0.1# 19.8 0.0# PHASE1DR 100 PHASE1B END
10:03:57 2:05 144.7 0.0 19.8 0.0 PHASE1B 31 DBSTATUSBEGN
10:03:57 2:05 144.7 0.0# 19.8 0.0# PHASE1B 31 DBSTATUSEND
10:03:57 2:05 144.7 0.0 19.8 0.0 PHASE1B 32 SEKR BEGN
10:03:57 2:05 144.9 0.2# 19.8 0.0# PHASE1B 32 SEKR END
10:03:57 2:05 144.9 0.0 19.9 0.0 SEKR 13 XEQUIV BEGN
10:03:57 2:05 144.9 0.0# 19.9 0.0# SEKR 13 XEQUIV END
10:03:57 2:05 144.9 0.0 19.9 0.0 SEKR 14 XPURGE BEGN
10:03:57 2:05 144.9 0.0# 19.9 0.0# SEKR 14 XPURGE END
10:03:57 2:05 144.9 0.0 19.9 0.0 SEKR 29 GPSP BEGN
10:03:57 2:05 144.9 0.0# 20.0 0.0# SEKR 29 GPSP END
10:03:57 2:05 144.9 0.0 20.0 0.0 SEKR 30 PMLUSET BEGN
10:03:57 2:05 144.9 0.0# 20.0 0.0# SEKR 30 PMLUSET END
10:03:57 2:05 144.9 0.0 20.0 0.0 SEKR 45 XEQUIV BEGN
10:03:57 2:05 144.9 0.0# 20.0 0.0# SEKR 45 XEQUIV END
10:03:57 2:05 144.9 0.0 20.0 0.0 SEKR 46 XPURGE BEGN
10:03:57 2:05 144.9 0.0# 20.0 0.0# SEKR 46 XPURGE END
Main Index
10:03:57 2:05 144.9 0.0 20.0 0.0 SEKR 47 UPARTN BEGN
 348 MSC Nastran Reference Guide
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
UPARTN 47 SCRATCH 1 12 2 1 6 0.5000 1 6 1 0 0 0 6 6 0 *8**
UPARTN 47 KFF 6 6 6 2 4 0.2777 1 0 10 0 0 0 3 5 0 *8**
UPARTN 47 KSF 6 6 6 2 4 0.2777 1 0 10 0 0 0 3 5 0 *8**
UPARTN 47 KFS 6 6 6 2 4 0.2777 1 0 10 0 0 0 3 5 0 *8**
UPARTN 47 KSS 6 6 6 2 4 0.2777 1 0 10 0 0 0 3 5 0 *8**
10:03:58 2:06 144.9 0.0# 20.1 0.0# SEKR 47 UPARTN END
10:03:58 2:06 144.9 0.0 20.1 0.0 SEKR 48 XEQUIV BEGN
10:03:58 2:06 144.9 0.0# 20.1 0.0# SEKR 48 XEQUIV END
10:03:58 2:06 144.9 0.0 20.1 0.0 SEKR 49 XPURGE BEGN
10:03:58 2:06 144.9 0.0# 20.1 0.0# SEKR 49 XPURGE END
10:03:58 2:06 144.9 0.0 20.1 0.0 SEKR 54 XPURGE BEGN
10:03:58 2:06 144.9 0.0# 20.1 0.0# SEKR 54 XPURGE END
10:03:58 2:06 144.9 0.0 20.1 0.0 SEKR 110 SESUM BEGN
10:03:58 2:06 144.9 0.0# 20.1 0.0# SEKR 110 SESUM END
10:03:58 2:06 144.9 0.0 20.2 0.0 SESUM 22 DELETE BEGN
10:03:58 2:06 144.9 0.0# 20.2 0.0# SESUM 22 DELETE END
10:03:58 2:06 145.3 0.4 20.2 0.0 SESUM 25 XPURGE BEGN
10:03:58 2:06 145.3 0.0# 20.2 0.0# SESUM 25 XPURGE END
10:03:58 2:06 145.4 0.1 20.3 0.0 PHASE1B 33 SESUM BEGN
10:03:58 2:06 145.4 0.0# 20.3 0.0# PHASE1B 33 SESUM END
10:03:59 2:07 145.4 0.0 20.3 0.1 SESUM 22 DELETE BEGN
10:03:59 2:07 145.4 0.0# 20.3 0.0# SESUM 22 DELETE END
10:03:59 2:07 145.9 0.4 20.4 0.0 SESUM 25 XPURGE BEGN
10:03:59 2:07 145.9 0.0# 20.4 0.0# SESUM 25 XPURGE END
10:03:59 2:07 146.0 0.1 20.4 0.0 PHASE1B 36 XEQUIV BEGN
10:03:59 2:07 146.0 0.0# 20.4 0.0# PHASE1B 36 XEQUIV END
10:03:59 2:07 146.1 0.1 20.4 0.0 PHASE1B 41 PMLUSET BEGN
10:03:59 2:07 146.1 0.0# 20.4 0.0# PHASE1B 41 PMLUSET END
10:03:59 2:07 146.5 0.4 20.5 0.1 PHASE1B 76 SEGOA BEGN
10:03:59 2:07 146.6 0.1# 20.5 0.0# PHASE1B 76 SEGOA END
10:03:59 2:07 146.6 0.0 20.5 0.0 PHASE1B 77 DBSTATUSBEGN
10:03:59 2:07 146.6 0.0# 20.5 0.0# PHASE1B 77 DBSTATUSEND
10:03:59 2:07 146.6 0.0 20.5 0.0 PHASE1B 78 SELR BEGN
10:03:59 2:07 146.6 0.1# 20.6 0.0# PHASE1B 78 SELR END
10:03:59 2:07 146.6 0.0 20.6 0.0 SELR 16 XEQUIV BEGN
10:03:59 2:07 146.6 0.0# 20.6 0.0# SELR 16 XEQUIV END
10:03:59 2:07 146.6 0.0 20.6 0.0 SELR 37 XEQUIV BEGN
10:03:59 2:07 146.6 0.0# 20.6 0.0# SELR 37 XEQUIV END
10:03:59 2:07 146.6 0.0 20.6 0.0 SELR 38 XPURGE BEGN
10:03:59 2:07 146.6 0.0# 20.6 0.0# SELR 38 XPURGE END
10:03:59 2:07 146.6 0.0 20.6 0.0 SELR 40 XPURGE BEGN
10:03:59 2:07 146.6 0.0# 20.6 0.0# SELR 40 XPURGE END
10:03:59 2:07 146.6 0.0 20.6 0.0 SELR 41 UPARTN BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
UPARTN 41 SCRATCH 1 12 2 1 6 0.5000 1 6 1 0 0 0 6 6 0 *8**
10:03:59 2:07 146.6 0.0# 20.6 0.0# SELR 41 UPARTN END
10:03:59 2:07 146.6 0.0 20.6 0.0 SELR 52 XEQUIV BEGN
10:03:59 2:07 146.6 0.0# 20.6 0.0# SELR 52 XEQUIV END
10:03:59 2:07 146.6 0.0 20.7 0.0 SELR 64 SSG2 BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
SSG2 64 SCRATCH 1 12 2 1 6 0.5000 1 6 1 0 0 0 6 6 0 *8**
SSG2 64 SCRATCH 1 6 2 2 6 0.5000 1 3 1 0 0 0 3 3 0 *8**
SSG2 64 PSS 1 6 2 2 0 0.0000 1 0 1 0 0 0 0 0 1 *8**
REAL CORE IS 6074952
SSG2 64 PA 1 6 2 2 6 0.5000 1 3 1 0 0 0 3 3 0 *8**
10:04:00 2:08 146.6 0.0# 20.7 0.1# SELR 64 SSG2 END
10:04:00 2:08 146.6 0.0 20.7 0.0 SELR 65 XPURGE BEGN
10:04:00 2:08 146.6 0.0# 20.7 0.0# SELR 65 XPURGE END
10:04:00 2:08 146.6 0.0 20.7 0.0 SELR 75 XPURGE BEGN
10:04:00 2:08 146.6 0.0# 20.8 0.0# SELR 75 XPURGE END
10:04:00 2:08 146.6 0.0 20.8 0.0 PHASE1B 80 XEQUIV BEGN
10:04:00 2:08 146.6 0.0# 20.8 0.0# PHASE1B 80 XEQUIV END
10:04:00 2:08 146.6 0.0 20.8 0.0 PHASE1B 81 SESUM BEGN
10:04:00 2:08 146.6 0.0# 20.8 0.0# PHASE1B 81 SESUM END
10:04:00 2:08 146.6 0.0 20.8 0.0 SESUM 22 DELETE BEGN
10:04:00 2:08 146.6 0.0# 20.8 0.0# SESUM 22 DELETE END
10:04:00 2:08 147.1 0.4 20.8 0.0 SESUM 25 XPURGE BEGN
10:04:00 2:08 147.1 0.0# 20.9 0.0# SESUM 25 XPURGE END
10:04:00 2:08 147.5 0.4 21.0 0.1 SESTATIC 63 BCDR BEGN
10:04:00 2:08 147.5 0.0# 21.0 0.0# SESTATIC 63 BCDR END
10:04:00 2:08 147.5 0.0 21.0 0.0 SESTATIC 64 DBSTATUSBEGN
10:04:01 2:09 147.5 0.0# 21.0 0.0# SESTATIC 64 DBSTATUSEND
10:04:01 2:09 147.5 0.0 21.0 0.0 SESTATIC 71 PHASE1C BEGN
10:04:01 2:09 147.6 0.1# 21.0 0.0# SESTATIC 71 PHASE1C END
10:04:01 2:09 147.6 0.0 21.0 0.0 PHASE1C 9 DBSTATUSBEGN
10:04:01 2:09 147.6 0.0# 21.1 0.0# PHASE1C 9 DBSTATUSEND
10:04:01 2:09 147.6 0.0 21.1 0.0 PHASE1C 10 SEKRRS BEGN
10:04:01 2:09 147.6 0.1# 21.1 0.0# PHASE1C 10 SEKRRS END
10:04:01 2:09 147.6 0.0 21.1 0.0 SEKRRS 10 XEQUIV BEGN
10:04:01 2:09 147.6 0.0# 21.1 0.0# SEKRRS 10 XEQUIV END
10:04:01 2:09 147.6 0.0 21.1 0.0 SEKRRS 13 XEQUIV BEGN
10:04:01 2:09 147.6 0.0# 21.1 0.0# SEKRRS 13 XEQUIV END
10:04:01 2:09 147.6 0.0 21.1 0.0 SEKRRS 18 DCMP BEGN

Main Index
 Chapter 7: Solution Sequences 349
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

*** USER INFORMATION MESSAGE 4157 (DFMSA) ---PARAMETERS FOR SPARSE DECOMPOSITION OF DATA BLOCK KLL ( TYPE=RDP ) FOLLOW

17 MATRIX SIZE = 6 ROWS NUMBER OF NONZEROES = 8 TERMS

NUMBER OF ZERO COLUMNS = 0 NUMBER OF ZERO DIAGONAL TERMS = 0
CPU TIME ESTIMATE = 0 SEC I/O TIME ESTIMATE = 0 SEC
ESTIMATED MEMORY REQUIREMENT = 15 K WORDS MEMORY AVAILABLE = 6074 K WORDS
EST. INTEGER WORDS IN FACTOR = 1 K WORDS EST. NONZERO TERMS = 1 K TERMS
ESTIMATED MAXIMUM FRONT SIZE = 2 TERMS
*** USER INFORMATION MESSAGE 6439 (DFMSA) ---ACTUAL MEMORY AND DISK SPACE REQUIREMENTS FOR SPARSE SYM. DECOMPOSITION
SPARSE DECOMP MEMORY REQUIRED = 15 K WORDS MAXIMUM FRONT SIZE = 2 TERMS
INTEGER WORDS IN FACTOR = 1 K WORDS NONZERO TERMS IN FACTOR = 1 K TERMS
*8** MODULE DMAP MATRIX COLS ROWS F T IBLKS NBLKS NUMFRT FRTMAX
DCMP 18 LLL 6 6 13 2 1 1 4 2 *8**
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
DCMP 18 SCRATCH 1 6 2 2 6 0.5000 1 3 1 0 0 0 6 6 0 *8**
DCMP 18 SCRATCH 1 6 2 2 6 0.5000 1 3 1 0 0 0 6 6 0 *8**
10:04:02 2:10 147.6 0.0# 21.2 0.1# SEKRRS 18 DCMP END
10:04:02 2:10 147.6 0.0 21.3 0.0 SEKRRS 26 XPURGE BEGN
10:04:02 2:10 147.6 0.0# 21.3 0.0# SEKRRS 26 XPURGE END
10:04:02 2:10 147.6 0.0 21.3 0.0 PHASE1C 11 SESUM BEGN
10:04:02 2:10 147.6 0.0# 21.3 0.0# PHASE1C 11 SESUM END
10:04:02 2:10 147.6 0.0 21.3 0.0 SESUM 22 DELETE BEGN
10:04:02 2:10 147.6 0.0# 21.3 0.0# SESUM 22 DELETE END
10:04:02 2:10 147.9 0.3 21.3 0.0 SESUM 25 XPURGE BEGN
10:04:02 2:10 147.9 0.0# 21.3 0.0# SESUM 25 XPURGE END
10:04:02 2:10 148.1 0.2 21.4 0.0 PHASE1C 12 DBSTATUSBEGN
10:04:02 2:10 148.1 0.0# 21.4 0.0# PHASE1C 12 DBSTATUSEND
10:04:02 2:10 148.1 0.0 21.4 0.0 PHASE1C 18 DBSTATUSBEGN
10:04:02 2:10 148.1 0.0# 21.4 0.0# PHASE1C 18 DBSTATUSEND
10:04:02 2:10 148.1 0.0 21.4 0.0 PHASE1C 21 SELRRS BEGN
10:04:02 2:10 148.1 0.1# 21.4 0.0# PHASE1C 21 SELRRS END
10:04:02 2:10 148.1 0.0 21.4 0.0 SELRRS 5 XEQUIV BEGN
10:04:02 2:10 148.1 0.0# 21.4 0.0# SELRRS 5 XEQUIV END
10:04:02 2:10 148.1 0.0 21.4 0.0 SELRRS 6 XPURGE BEGN
10:04:02 2:10 148.1 0.0# 21.4 0.0# SELRRS 6 XPURGE END
10:04:02 2:10 148.1 0.0 21.5 0.0 PHASE1C 22 SESUM BEGN
10:04:02 2:10 148.1 0.0# 21.5 0.0# PHASE1C 22 SESUM END
10:04:02 2:10 148.1 0.0 21.5 0.1 SESUM 22 DELETE BEGN
10:04:02 2:10 148.1 0.0# 21.5 0.0# SESUM 22 DELETE END
10:04:03 2:11 148.4 0.3 21.5 0.0 SESUM 25 XPURGE BEGN
10:04:03 2:11 148.4 0.0# 21.5 0.0# SESUM 25 XPURGE END
10:04:03 2:11 148.6 0.2 21.6 0.0 SESTATIC 72 DBSTATUSBEGN
10:04:03 2:11 148.6 0.0# 21.6 0.0# SESTATIC 72 DBSTATUSEND
10:04:03 2:11 148.6 0.0 21.6 0.0 SESTATIC 73 STATRS BEGN
10:04:03 2:11 148.6 0.1# 21.6 0.0# SESTATIC 73 STATRS END
10:04:03 2:11 148.6 0.0 21.6 0.0 STATRS 6 XPURGE BEGN
10:04:03 2:11 148.6 0.0# 21.6 0.0# STATRS 6 XPURGE END
10:04:03 2:11 148.6 0.0 21.6 0.0 STATRS 7 XEQUIV BEGN
10:04:03 2:11 148.6 0.0# 21.6 0.0# STATRS 7 XEQUIV END
10:04:03 2:11 148.6 0.0 21.6 0.0 STATRS 27 SSG3 BEGN
REAL CORE IS 6074952
REAL CORE IS 6074952
REAL CORE IS 6074952
REAL CORE IS 6074952

0*** DIAG 19 --FBS-- FACTOR TRAILER=( 6 6 13 2 2 2222), METHOD S2-PASSES= 1, CPU= 0.0, I/O= 0

R.H.S. TRAILER=( 1 6 2 2 6 5000)

*** USER INFORMATION MESSAGE 4153. FBS METHOD S2 TIME ESTIMATE TO FORM UL - CPU= 0, I/O= 0, TOTAL= 0, PASSES= 1
10:04:03 2:11 148.6 0.0 21.7 0.0 FBS BEGN
10:04:03 2:11 148.6 0.0# 21.7 0.0# FBS END
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
SSG3 27 UL 1 6 2 2 10 0.8333 1 2 2 0 0 0 6 6 0 *8**
REAL CORE IS 6074952
* M MATRIX KLL TRAILER=( 6 6 6 2 4 2777), METHOD 1 -PASSES= 1, CPU= 0.0, I/O= 0.3, TOTAL= 0.3
* P
* Y 18
MATRIX UL
MATRIX PLI
TRAILER=(
TRAILER=(
1
1
6 2 2
6 2 2
10
6
8333), METHOD 2 -PASSES=
5000), METHOD 3 -PASSES=
1, CPU=999999.0, I/O=
1, CPU=999999.0, I/O=
0.2, TOTAL=999999.0
0.3, TOTAL=999999.0
* A WORKING MEMORY= 6074952 SYSTEM( 66)= 0 METHOD 4 -PASSES= 1, CPU=999999.0, I/O= 0.8,
TOTAL=999999.0
* D TRANSPOSE FLAG= 0 SYSTEM(126)= 57 METHOD S -PASSES= 1, CPU= 0.0, I/O= 0.2, TOTAL= 0.2
METHOD S NT, NBR PASSES= 1, EST. CPU= 0.0, I/O= 0.2, TOTAL= 0.2
10:04:03 2:11 148.6 0.0 21.7 0.0 MPYAD BGN P=1
10:04:03 2:11 148.6 0.0# 21.7 0.0# MPYAD END

Main Index
 350 MSC Nastran Reference Guide
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
SSG3 27 RUL 1 6 2 2 6 0.5000 1 3 1 0 0 0 3 3 0 *8**
10:04:03 2:11 148.6 0.0# 21.7 0.1# STATRS 27 SSG3 END
10:04:03 2:11 148.6 0.0 21.8 0.0 SESTATIC 74 DELETE BEGN
10:04:03 2:11 148.6 0.0# 21.8 0.0# SESTATIC 74 DELETE END
10:04:03 2:11 148.6 0.0 21.8 0.0 SESTATIC 75 DBSTATUSBEGN
10:04:03 2:11 148.6 0.0# 21.8 0.0# SESTATIC 75 DBSTATUSEND
10:04:03 2:11 148.6 0.0 21.8 0.0 SESTATIC 94 SESUM BEGN
10:04:03 2:11 148.6 0.0# 21.8 0.0# SESTATIC 94 SESUM END
10:04:04 2:12 148.6 0.0 21.9 0.0 SESUM 22 DELETE BEGN
10:04:04 2:12 148.6 0.0# 21.9 0.0# SESUM 22 DELETE END
10:04:04 2:12 148.9 0.3 21.9 0.0 SESUM 25 XPURGE BEGN
10:04:04 2:12 148.9 0.0# 21.9 0.0# SESUM 25 XPURGE END
10:04:04 2:12 149.1 0.2 21.9 0.0 SESTATIC 98 XEQUIV BEGN
10:04:04 2:12 149.1 0.0# 21.9 0.0# SESTATIC 98 XEQUIV END
10:04:04 2:12 149.1 0.0 21.9 0.0 SESTATIC 103 SUPER3 BEGN
10:04:04 2:12 149.3 0.2# 22.0 0.1# SESTATIC 103 SUPER3 END
10:04:04 2:12 149.3 0.0 22.0 0.0 SUPER3 26 XEQUIV BEGN
10:04:04 2:12 149.3 0.0# 22.0 0.0# SUPER3 26 XEQUIV END
10:04:04 2:12 149.3 0.0 22.0 0.0 SUPER3 28 XEQUIV BEGN
10:04:04 2:12 149.3 0.0# 22.0 0.0# SUPER3 28 XEQUIV END
10:04:05 2:13 149.3 0.0 22.2 0.2 SUPER3 61 XEQUIV BEGN
10:04:05 2:13 149.3 0.0# 22.2 0.0# SUPER3 61 XEQUIV END
10:04:05 2:13 149.3 0.1 22.2 0.0 SUPER3 67 SEP4 BEGN
10:04:05 2:13 149.3 0.0# 22.2 0.0# SUPER3 67 SEP4 END
10:04:05 2:13 149.3 0.0 22.3 0.1 SUPER3 92 SEDRDR BEGN
10:04:05 2:13 149.3 0.0# 22.3 0.0# SUPER3 92 SEDRDR END
10:04:05 2:13 149.3 0.0 22.3 0.0 SUPER3 99 PVT BEGN
10:04:05 2:13 149.3 0.0# 22.4 0.1# SUPER3 99 PVT END
10:04:05 2:13 149.4 0.1 22.4 0.0 SUPER3 106 SEDR BEGN
10:04:05 2:13 149.4 0.0# 22.5 0.1# SUPER3 106 SEDR END
10:04:05 2:13 149.4 0.0 22.5 0.0 SUPER3 112 XEQUIV BEGN
10:04:05 2:13 149.4 0.0# 22.5 0.0# SUPER3 112 XEQUIV END
10:04:05 2:13 149.4 0.0 22.5 0.0 SUPER3 117 SETQ BEGN
10:04:05 2:13 149.4 0.0# 22.5 0.0# SUPER3 117 SETQ END
10:04:05 2:13 149.4 0.0 22.6 0.0 SETQ 10 XEQUIV BEGN
10:04:05 2:13 149.4 0.0# 22.6 0.0# SETQ 10 XEQUIV END
10:04:05 2:13 149.4 0.0 22.7 0.2 SUPER3 135 DBSTATUSBEGN
10:04:05 2:13 149.4 0.0# 22.7 0.0# SUPER3 135 DBSTATUSEND
10:04:05 2:13 149.4 0.0 22.7 0.0 SUPER3 140 LCGEN BEGN
10:04:05 2:13 149.4 0.0# 22.7 0.0# SUPER3 140 LCGEN END
10:04:05 2:13 149.4 0.1 22.8 0.1 SUPER3 160 XEQUIV BEGN
10:04:05 2:13 149.4 0.0# 22.8 0.0# SUPER3 160 XEQUIV END
10:04:05 2:13 149.4 0.0 22.8 0.0 SUPER3 168 SEDISP BEGN
10:04:05 2:13 149.6 0.2# 22.9 0.1# SUPER3 168 SEDISP END
10:04:05 2:13 149.6 0.0 22.9 0.0 SEDISP 46 XEQUIV BEGN
10:04:05 2:13 149.6 0.0# 22.9 0.0# SEDISP 46 XEQUIV END
10:04:05 2:13 149.6 0.0 22.9 0.0 SEDISP 57 XEQUIV BEGN
10:04:05 2:13 149.6 0.0# 23.0 0.0# SEDISP 57 XEQUIV END
10:04:06 2:14 149.6 0.0 23.0 0.0 SEDISP 65 BCDR BEGN
10:04:06 2:14 149.6 0.0# 23.0 0.0# SEDISP 65 BCDR END
10:04:06 2:14 149.6 0.0 23.0 0.0 SEDISP 68 XEQUIV BEGN
10:04:06 2:14 149.6 0.0# 23.0 0.0# SEDISP 68 XEQUIV END
10:04:06 2:14 149.7 0.1 23.0 0.0 SEDISP 79 SEGOA BEGN
10:04:06 2:14 149.8 0.1# 23.1 0.0# SEDISP 79 SEGOA END
10:04:06 2:14 149.8 0.0 23.1 0.0 SEDISP 105 XEQUIV BEGN
10:04:06 2:14 149.8 0.0# 23.1 0.0# SEDISP 105 XEQUIV END
10:04:06 2:14 149.8 0.0 23.1 0.0 SEDISP 106 XEQUIV BEGN
10:04:06 2:14 149.8 0.0# 23.1 0.0# SEDISP 106 XEQUIV END
10:04:06 2:14 149.8 0.0 23.1 0.0 SEDISP 107 XEQUIV BEGN
10:04:06 2:14 149.8 0.0# 23.1 0.0# SEDISP 107 XEQUIV END
10:04:06 2:14 149.8 0.0 23.1 0.0 SEDISP 108 XEQUIV BEGN
10:04:06 2:14 149.8 0.0# 23.1 0.0# SEDISP 108 XEQUIV END
10:04:06 2:14 149.8 0.0 23.1 0.0 SEDISP 110 SDR1 BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
SDR1 110 SCRATCH 1 12 2 1 6 0.5000 1 6 1 0 0 0 6 6 0 *8**
SDR1 110 SCRATCH 1 12 2 2 10 0.4166 1 2 2 0 0 0 6 6 0 *8**
SDR1 110 UGI 1 12 2 2 10 0.4166 1 2 2 0 0 0 6 6 0 *8**
10:04:06 2:14 149.8 0.0# 23.2 0.0# SEDISP 110 SDR1 END
10:04:06 2:14 149.8 0.0 23.2 0.0 SEDISP 135 COPY BEGN
10:04:06 2:14 149.8 0.0# 23.2 0.0# SEDISP 135 COPY END
10:04:06 2:14 149.8 0.0 23.2 0.0 SEDISP 139 COPY BEGN
10:04:06 2:14 149.8 0.0# 23.2 0.0# SEDISP 139 COPY END
10:04:06 2:14 149.8 0.0 23.2 0.0 SEDISP 141 COPY BEGN
10:04:06 2:14 149.8 0.0# 23.2 0.0# SEDISP 141 COPY END
10:04:06 2:14 149.8 0.0 23.2 0.0 SEDISP 213 SESUM BEGN
10:04:06 2:14 149.8 0.0# 23.2 0.0# SEDISP 213 SESUM END
10:04:06 2:14 149.8 0.0 23.3 0.0 SESUM 22 DELETE BEGN
10:04:06 2:14 149.8 0.0# 23.3 0.0# SESUM 22 DELETE END
10:04:06 2:14 150.1 0.3 23.3 0.0 SESUM 25 XPURGE BEGN
10:04:06 2:14 150.1 0.0# 23.3 0.0# SESUM 25 XPURGE END
10:04:06 2:14 150.2 0.1 23.3 0.0 SUPER3 172 XEQUIV BEGN
10:04:06 2:14 150.2 0.0# 23.3 0.0# SUPER3 172 XEQUIV END
10:04:06 2:14 150.3 0.1 23.4 0.0 SUPER3 181 XEQUIV BEGN
10:04:06 2:14 150.3 0.0# 23.4 0.0# SUPER3 181 XEQUIV END
10:04:06 2:14 150.3 0.0 23.4 0.0 SUPER3 190 XEQUIV BEGN
10:04:06 2:14 150.3 0.0# 23.4 0.0# SUPER3 190 XEQUIV END

Main Index
 Chapter 7: Solution Sequences 351
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

10:04:07 2:15 150.3 0.0 23.4 0.0 SUPER3 203 SEDRCVR BEGN
10:04:07 2:15 150.4 0.2# 23.5 0.1# SUPER3 203 SEDRCVR END
10:04:07 2:15 150.4 0.0 23.5 0.0 SEDRCVR 31 XEQUIV BEGN
10:04:07 2:15 150.4 0.0# 23.5 0.0# SEDRCVR 31 XEQUIV END
10:04:07 2:15 150.4 0.0 23.5 0.0 SEDRCVR 32 XEQUIV BEGN
10:04:07 2:15 150.4 0.0# 23.5 0.0# SEDRCVR 32 XEQUIV END
10:04:07 2:15 150.4 0.0 23.5 0.0 SEDRCVR 36 VECPLOT BEGN
10:04:07 2:15 150.4 0.0# 23.6 0.0# SEDRCVR 36 VECPLOT END
10:04:07 2:15 150.4 0.0 23.6 0.0 SEDRCVR 37 VECPLOT BEGN
10:04:07 2:15 150.4 0.0# 23.6 0.0# SEDRCVR 37 VECPLOT END
10:04:07 2:15 150.4 0.0 23.6 0.0 SEDRCVR 38 VECPLOT BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
VECPLOT 38 UGVB 1 12 2 1 5 0.4166 1 2 2 0 0 0 6 6 0 *8**
10:04:07 2:15 150.4 0.0# 23.6 0.0# SEDRCVR 38 VECPLOT END
10:04:07 2:15 150.4 0.0 23.6 0.0 SEDRCVR 39 VECPLOT BEGN
*8** MODULE DMAP MATRIX COLS ROWS F T NZWDS DENS BLOCKS STRL NBRSTR BNDL NBRBND ROW1 BNDAVG BNDMAX NULCOL
VECPLOT 39 PJB 1 12 2 1 3 0.2500 1 3 1 0 0 0 3 3 0 *8**
10:04:07 2:15 150.4 0.0# 23.6 0.0# SEDRCVR 39 VECPLOT END
10:04:07 2:15 150.4 0.0 23.7 0.0 SEDRCVR 48 SDR2 BEGN
10:04:08 2:16 150.4 0.0# 23.8 0.1# SEDRCVR 48 SDR2 END
10:04:08 2:16 150.5 0.1 23.8 0.0 SEDRCVR 49 SDR2 BEGN
10:04:08 2:16 150.5 0.0# 23.8 0.0# SEDRCVR 49 SDR2 END
10:04:08 2:16 150.5 0.0 23.8 0.0 SEDRCVR 58 XEQUIV BEGN
10:04:08 2:16 150.5 0.0# 23.8 0.0# SEDRCVR 58 XEQUIV END
10:04:08 2:16 150.5 0.0 23.8 0.0 SEDRCVR 59 XEQUIV BEGN
10:04:08 2:16 150.5 0.0# 23.8 0.0# SEDRCVR 59 XEQUIV END
10:04:08 2:16 150.5 0.0 23.8 0.0 SEDRCVR 60 XEQUIV BEGN
10:04:08 2:16 150.5 0.0# 23.8 0.0# SEDRCVR 60 XEQUIV END
10:04:08 2:16 150.5 0.0 23.8 0.0 SEDRCVR 61 XEQUIV BEGN
10:04:08 2:16 150.5 0.0# 23.8 0.0# SEDRCVR 61 XEQUIV END
10:04:08 2:16 150.6 0.1 23.9 0.1 SEDRCVR 125 SDRX BEGN
10:04:09 2:17 150.6 0.0# 23.9 0.0# SEDRCVR 125 SDRX END
10:04:09 2:17 150.6 0.0 23.9 0.0 SEDRCVR 129 XEQUIV BEGN
10:04:09 2:17 150.6 0.0# 23.9 0.0# SEDRCVR 129 XEQUIV END
10:04:09 2:17 150.6 0.0 23.9 0.0 SEDRCVR 130 XEQUIV BEGN
10:04:09 2:17 150.6 0.0# 23.9 0.0# SEDRCVR 130 XEQUIV END
10:04:09 2:17 150.6 0.0 24.0 0.0 SEDRCVR 132 XEQUIV BEGN
10:04:09 2:17 150.6 0.0# 24.0 0.0# SEDRCVR 132 XEQUIV END
10:04:09 2:17 150.6 0.0 24.0 0.0 SEDRCVR 133 XEQUIV BEGN
10:04:09 2:17 150.6 0.0# 24.0 0.0# SEDRCVR 133 XEQUIV END
10:04:09 2:17 150.7 0.1 24.0 0.0 SEDRCVR 205 OFP BEGN
10:04:10 2:18 150.7 0.0# 24.1 0.1# SEDRCVR 205 OFP END
10:04:10 2:18 150.7 0.0 24.1 0.0 SEDRCVR 206 OFP BEGN
10:04:10 2:18 150.7 0.0# 24.1 0.0# SEDRCVR 206 OFP END
10:04:10 2:18 150.7 0.0 24.1 0.0 SEDRCVR 208 OFP BEGN
10:04:10 2:18 150.7 0.0# 24.1 0.0# SEDRCVR 208 OFP END
10:04:11 2:19 150.7 0.0 24.1 0.0 SEDRCVR 211 SDR2 BEGN
10:04:11 2:19 150.7 0.0# 24.2 0.0# SEDRCVR 211 SDR2 END
10:04:11 2:19 150.7 0.0 24.2 0.0 SEDRCVR 212 SDRCOMP BEGN
10:04:11 2:19 150.7 0.0# 24.2 0.0# SEDRCVR 212 SDRCOMP END
10:04:11 2:19 150.7 0.0 24.2 0.0 SEDRCVR 213 STRSORT BEGN
10:04:11 2:19 150.7 0.0# 24.2 0.0# SEDRCVR 213 STRSORT END
10:04:13 2:21 150.8 0.1 24.2 0.0 SEDRCVR 214 STRSORT BEGN
10:04:13 2:21 150.8 0.0# 24.2 0.0# SEDRCVR 214 STRSORT END
10:04:13 2:21 150.8 0.0 24.3 0.0 SEDRCVR 215 OFP BEGN
10:04:13 2:21 150.8 0.0# 24.3 0.0# SEDRCVR 215 OFP END
10:04:13 2:21 150.8 0.0 24.3 0.0 SEDRCVR 222 XEQUIV BEGN
10:04:14 2:22 150.8 0.0# 24.3 0.0# SEDRCVR 222 XEQUIV END
10:04:14 2:22 150.8 0.0 24.3 0.0 SEDRCVR 224 OFP BEGN
10:04:14 2:22 150.8 0.0# 24.3 0.0# SEDRCVR 224 OFP END
10:04:15 2:23 150.9 0.1 24.5 0.2 SESTATIC 132 PRTSUM BEGN
10:04:15 2:23 150.9 0.0# 24.5 0.0# SESTATIC 132 PRTSUM END
10:04:15 2:23 150.9 0.0 24.5 0.0 PRTSUM 4 PROJVER BEGN
10:04:15 2:23 150.9 0.0# 24.5 0.0# PRTSUM 4 PROJVER END
10:04:15 2:23 150.9 0.0 24.5 0.0 PRTSUM 5 DBDICT BEGN
10:04:15 2:23 150.9 0.0# 24.6 0.1# PRTSUM 5 DBDICT END
10:04:15 2:23 150.9 0.0 24.7 0.0 PRTSUM 6 PRTPARM BEGN
10:04:15 2:23 150.9 0.0# 24.7 0.0# PRTSUM 6 PRTPARM END
10:04:15 2:23 150.9 0.0 24.7 0.0 SESTATIC 133 EXIT BEGN

Main Index
 352 MSC Nastran Reference Guide
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

1*** USER INFORMATION MESSAGE 1009 (F04SUM)
STATISTICAL SUMMARY OF ABOVE LOG FILE FOLLOWS.

STATISTICS FOR INDIVIDUAL EVENTS

LINE DMAP EVNT METH EST CPU ACT CPU DEL CPU % DIF <25% <50% >50% % TOT
65 35 XSOR 1. 5.
67 61 IFP 0. 1.
120 154 PLTS 0. 0.
124 176 SEQP 0. 1.
126 182 GP1 0. 1.
132 25 GP1 0. 1.
134 28 SEP1 0. 0.
140 46 TABP 0. 0.
142 47 SEP3 0. 0.
172 307 GP1 0. 1.
180 6 GP2 0. 0.
194 335 GP3 0. 0.
196 336 LCGE 0. 0.
202 367 GP4 0. 0.
208 378 UMER 0. 0.
218 45 MATG 0. 0.
237 24 TA1 0. 0.
241 16 ELTP 0. 0.
243 22 EMG 0. 0.
249 79 EMA 0. 0.
278 11 SSG1 0. 0.
287 62 VECP 0. 0.
319 29 GPSP 0. 0.
327 47 UPAR 0. 0.
371 41 UPAR 0. 0.
377 64 SSG2 0. 0.
410 18 DCMP 0. 0.
459 27 FBS S 2 0. 0. 0. 0. 1 0 0 0.
459 27 MPYA S NT 0. 0. 0. 0. 1 0 0 0.
459 27 SSG3 0. 0.
498 67 SEP4 0. 0.
504 106 SEDR 0. 0.
514 140 LCGE 0. 0.
538 110 SDR1 0. 0.
568 36 VECP 0. 0.
570 37 VECP 0. 0.
572 38 VECP 0. 0.
576 39 VECP 0. 0.
580 48 SDR2 0. 0.
582 49 SDR2 0. 0.
592 125 SDRX 0. 0.
602 205 OFP 0. 0.
604 206 OFP 0. 0.
606 208 OFP 0. 0.
608 211 SDR2 0. 0.
610 212 SDRC 0. 0.
612 213 STRS 0. 0.
614 214 STRS 0. 0.
616 215 OFP 0. 0.
620 224 OFP 0. 0.

STATISTICS FOR EVENTS BY METHOD

COUNT EST CPU ACT CPU DEL CPU <25% <50% >50% A.M. % S.D. % % TOT

MPYAD S NT 1 0. 0. 0. 1 0 0 0. 0. 0.
MPYAD N TOTALS 1 0. 0. 0. 0.
FBS S 2 1 0. 0. 0. 1 0 0 0. 0. 0.
FBS TOTALS 1 0. 0. 0. 0.

METHOD SUMMARY:
TOTAL OF EVENTS RECORDED = 0.
TOTAL CPU = 25.
PERCENT RECORDED = 0.

Main Index
 Chapter 7: Solution Sequences 353
Output Description

Listing 7-1 Example Execution Summary File (.f04) (continued)

STATISTICS FOR EVENTS BY MODULE

COUNT ACT CPU % TOT

ELTPRT 1 0. 0.
EMA 1 0. 0.
EMG 1 0. 0.
GPSP 1 0. 0.
GP1 3 1. 2.
GP2 1 0. 0.
GP3 1 0. 0.
GP4 1 0. 0.
IFP 1 0. 1.
LCGEN 2 0. 0.
MATGEN 1 0. 0.
OFP 5 0. 0.
PLTSET 1 0. 0.
SDRCOMP 1 0. 0.
SDRX 1 0. 0.
SDR1 1 0. 0.
SDR2 3 0. 1.
SEDR 1 0. 0.
SEP1 1 0. 0.
SEP3 1 0. 0.
SEP4 1 0. 0.
SEQP 1 0. 1.
SSG1 1 0. 0.
SSG2 1 0. 0.
SSG3 1 0. 0.
STRSORT 2 0. 0.
TABPRT 1 0. 0.
TA1 1 0. 0.
UMERGE1 1 0. 0.
UPARTN 2 0. 0.
VECPLOT 5 0. 0.
XSORT 1 1. 5.
DCMP 1 0. 0.

MODULE SUMMARY:
TOTAL OF EVENTS RECORDED = 4.
TOTAL CPU = 25.
PERCENT RECORDED = 15.

*** TOTAL MEMORY USAGE STATISTICS ***

+------SPARSE DECOMPOSITION MODULES ---------+

HIWATER SUB_DMAP DMAP

WORDS DAY_TIME NAME MODULE
314448 10:04:01 SEKRRS 18 DCMP

*** DATABASE USAGE STATISTICS ***

+------------------------- LOGICAL FILES -------------------------+ +----------------- PHYSICAL FILES ----------------+

DBSET BLOCKS BLOCKS % BLOCKSIZE BLOCKS PER PHYS FILE BLOCKS HIWATER FILE SIZE MAX %
ALLOCATED USED USED (WORDS) CLUSTER ALLOCATED BLOCK (WORDS) USED
MASTER 5000 185 3.70 1408 1 MASTER 5000 185 260480 3.70
USROBJ 5000 7 0.14 1408 1 USROBJ 5000 7 9856 0.14
20
USRSOU
DBALL
5000
25000
7
7
0.14
0.03
1408
1408
1
1
USRSOU
DBALL
5000
25000
7
7
9856
9856
0.14
0.03

OBJSCR 5000 124 2.48 1408 1 OBJSCR 5000 124 174592 2.48
SCRATCH 350100 23 0.01 1408 1 MEMFILE 100 58 81664 58.00
SCRATCH 175000 1 1408 0.00
21 SCR300 175000 1
22
1408 0.00

*** BUFFER POOL AND SCRATCH 300 USAGE STATISTICS ***

+----------------- BUFFER POOL -----------------+ +-------------------------- SCRATCH 300 --------------------------+

OPTION BLOCKS BLOCKS BLOCKS OPTION HIWATER SUB_DMAP DMAP OPN/CLS
SELECTED ALLOCATED REUSED RELEASED SELECTED BLOCK DAY_TIME NAME MODULE COUNTER
GINO,EXEC 37 910 67 2 0 10:01:52 PREFACE 0 PREFACE 0

The Standard Output contains the MSC Nastran analysis results. This file also contains information about the version of
MSC Nastran being run.

Main Index
 354 MSC Nastran Reference Guide
Output Description

Listing 7-2 MSC Nastran Standard Output

Warning: This computer program is protected by copyright law and international treaties.
Unauthorized use, reproduction or distribution of this computer program, or any portion of it, may
result in severe civil and criminal penalties.
Copyright (C) 2016 MSC Software Corporation and its licensors. All rights reserved.

* * * * * * * * * * * * * * * * * * * *
* * * * * * * * * * * * * * * * * * * *
* * * *
* * * *
* * MSC Software * *
* * CORP * *
* * * *
* * M S C N A S T R A N * *
* * * *
* * VERSION 2016.0.0-CLXXXXXX * *
* * * *
* * MAR 30, 2016 * *
* * * *
* * * *
* * x86-64 Based System * *
* * MODEL Intel x8664 * *
* * Linux 2.6.32-279.el6.x86_64 * *
* *Compiled for x8664 (SINGLE Mode) * *
* * * *
* * * *
* * * * * * * * * * * * * * * * * * * *
* * * * * * * * * * * * * * * * * * * *

Main Index
 Chapter 7: Solution Sequences 355
Output Description

Listing 7-2 MSC Nastran Standard Output (continued)

MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 2

N A S T R A N E X E C U T I V E C O N T R O L D E C K E C H O

ID MSC, AM761B $ JJG 5MAR91

DIAG 8,13,19,49 $
TIME 2
SOL 101 $ STRUCTURED STATICS
CEND

EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 3

C A S E C O N T R O L D E C K E C H O
CARD
COUNT
1 TITLE = EXAMPLE: ONE ELEMENT PROBLEM AM761B
2 LOAD = 11
3 DISPLACEMENT = ALL
4 EL FORCE = ALL
5 BEGIN BULK
INPUT BULK DATA CARD COUNT = 7

EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 4

S O R T E D B U L K D A T A E C H O
CARD
COUNT . 1 .. 2 .. 3 .. 4 .. 5 .. 6 .. 7 .. 8 .. 9 .. 10 .
1- CBEAM 1 1 10 20 0. 1. 0.
2- FORCE 11 20 100. 1. .8 1.
3- GRID 10 0. 0. 0. 123456
4- GRID 20 10. 0. 0.
5- MAT1 100 1.+7 .3
6- PBEAM 1 100 1. .08 .064 .1
ENDDATA
TOTAL COUNT= 7

EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 5

SEQUENCE PROCESSOR OUTPUT

THERE ARE 2 POINTS DIVIDED INTO 1 GROUP(S).

CONNECTION DATA
ELEMENT TYPE NUMBER ASSEMBLY TIME(SEC)

BEAM 1 0.02
-----------------------------------------------

TOTAL MATRIX ASSEMBLY TIME FOR 1 ELEMENTS IS 0.02 SECONDS.

ORIGINAL PERFORMANCE DATA
SUPER(GROUP) ID NO. GRIDS AV. CONNECTIVITY C-AVERAGE C-RMS C-MAXIMUM P-GROUPS P-AVERAGE DECOMP TIME(SEC)
(6.0 DOF/ GRID )
0 2 2.00 1.50 1.58 2 0 0.00 0.000

RESEQUENCED PERFORMANCE DATA

SUPER(GROUP) ID NO. GRIDS AV. CONNECTIVITY C-AVERAGE C-RMS C-MAXIMUM P-GROUPS P-AVERAGE DECOMP TIME(SEC) METHOD
(6.0 DOF/ GRID )
0 2 2.00 1.50 1.58 2 0 0.00 0.000 ACTIVE

EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 6

OLOAD
RESULTANT
T1 T2 T3 R1 R2 R3
1 1.0000000E+02 8.0000000E+01 1.0000000E+02 0.0000000E+00 -1.0000000E+03 8.0000000E+02
*** USER INFORMATION MESSAGE 5293 FOR DATA BLOCK KLL
LOAD SEQ. NO. EPSILON EXTERNAL WORK EPSILONS LARGER THAN 0.001 ARE FLAGGED WITH ASTERISKS
1 -4.6914561E-17 3.9638200E+00
EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 7

MAXIMUM DISPLACEMENTS
T1 T2 T3 R1 R2 R3
1 9.9999997E-05 3.3541333E-02 5.2343331E-02 0.0000000E+00 7.8124995E-03 4.9999999E-03
EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 8

MAXIMUM APPLIED LOADS

T1 T2 T3 R1 R2 R3
1 1.0000000E+02 8.0000000E+01 1.0000000E+02 0.0000000E+00 0.0000000E+00 0.0000000E+00

Main Index
 356 MSC Nastran Reference Guide
Output Description

Listing 7-2 MSC Nastran Standard Output (continued)

EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 9

D I S P L A C E M E N T V E C T O R

POINT ID. TYPE T1 T2 T3 R1 R2 R3

10 G 0.0 0.0 0.0 0.0 0.0 0.0
20 G 1.000000E-04 3.354133E-02 5.234333E-02 0.0 -7.812500E-03 5.000000E-03
EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 10

F O R C E S I N B E A M E L E M E N T S ( C B E A M )
STAT DIST/ - BENDING MOMENTS - - WEB SHEARS - AXIAL TOTAL WARPING
ELEMENT-ID GRID LENGTH PLANE 1 PLANE 2 PLANE 1 PLANE 2 FORCE TORQUE TORQUE
1
10 0.000 8.000000E+02 1.000000E+03 8.000000E+01 1.000000E+02 1.000000E+02 0.0 0.0
20 1.000 0.0 0.0 8.000000E+01 1.000000E+02 1.000000E+02 0.0 0.0
EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 11

EXAMPLE: ONE ELEMENT PROBLEM AM761B MARCH 11, 1994 MSC/NASTRAN 3/ 3/94 PAGE 12

* * * * D B D I C T P R I N T * * * * SUBDMAP = PRTSUM , DMAP STATEMENT NO. 5

* * * * A N A L Y S I S S U M M A R Y T A B L E * * * *
SEID PEID PROJ VERS APRCH SEMG SEMR SEKR SELG SELR MODES DYNRED SOLLIN SOLNL LOOPID DESIGN CYCLE SENSITIVITY
--------------------------------------------------------------------------------------------------------------------------
0 0 1 1 ' ' T F T T T F F T 0 F -1 0 F
0SEID = SUPERELEMENT ID.
PEID = PRIMARY SUPERELEMENT ID OF IMAGE SUPERELEMENT.
PROJ = PROJECT ID NUMBER.
VERS = VERSION ID.
APRCH = BLANK FOR STRUCTURAL ANALYSIS. HEAT FOR HEAT TRANSFER ANALYSIS.
SEMG = STIFFNESS AND MASS MATRIX GENERATION STEP.
SEMR = MASS MATRIX REDUCTION STEP (INCLUDES EIGENVALUE SOLUTION FOR MODES).
SEKR = STIFFNESS MATRIX REDUCION STEP.
SELG = LOAD MATRIX GENERATION STEP.
SELR = LOAD MATRIX REDUCTION STEP.
MODES = T (TRUE) IF NORMAL MODES OR BUCKLING MODES CALCULATED.
DYNRED = T (TRUE) MEANS GENERALIZED DYNAMIC AND/OR COMPONENT MODE REDUCTION PERFORMED.
SOLLIN = T (TRUE) IF LINEAR SOLUTION EXISTS IN DATABASE.
LOOPID = THE LAST LOOPID VALUE USED IN THE NONLINEAR ANALYSIS. USEFUL FOR RESTARTS.
SOLNL = T (TRUE) IF NONLINEAR SOLUTION EXISTS IN DATABASE.
DESIGN CYCLE = THE LAST DESIGN CYCLE (ONLY VALID IN OPTIMIZATION).
SENSITIVITY = SENSITIVITY MATRIX GENERATION FLAG.

* * * END OF JOB * * *

The Grid Point Singularity Table (GPST) is automatically output following the execution of the Grid Point Singularity
Processor (GPSP) if singularities remain in the stiffness matrix at the grid point level. The table lists singular degrees of
freedom in the global coordinate system. These singular degrees of freedom may be automatically constrained, at the user’s
option. Further information on this topic is given in Constraint and Mechanism Problem Identification in SubDMAP SEKR, 370.
Any of the matrices or tables that are prepared by the functional modules can be printed by using selected utility modules
described in the MSC Nastran DMAP Programmer’s Guide. These utility modules can be scheduled at any point in a
solution sequence by using the alter feature. They should be scheduled immediately after the functional module that
generates the table or matrix to be printed. However, the user is cautioned to check the calling sequence for the utility
module, in order to be certain that all required inputs have been generated prior to this point.
The following descriptions relate to automatic output that is associated with particular types of solution sequences.
Real Eigenvalue Analysis

Each eigenvalue is identified with a mode number determined by sorting the eigenvalues by their magnitude. The following
summary is also automatically printed for all methods:

Main Index
 Chapter 7: Solution Sequences 357
Output Description

 Mode number
 Extraction order (this is irrelevant under the Lanczos method)
 Eigenvalue
 Frequency (radians)
 Frequency (Hz)
 Generalized mass
 Generalized stiffness
For all methods except the Inverse Power Method (see below), if the orthogonality criterion is failed, then User Warning
Message 3034 is issued:
The following summary is also automatically printed when using the Lanczos method:
 Block size used
 Number of decompositions
 Number of roots found
 Number of (block) solves required
 One of the following termination messages:
 Required number of eigenvalues found
 Not all the roots found within the range
 All eigenvalues found within the range
 Insufficient time to find more eigenvalues
The following summary is automatically printed when using the Inverse Power Method:
 Number of eigenvalues extracted
 Number of starting points used
 Number of starting point moves
 Number of triangular decompositions
 Number of vector iterations
 The reason for termination is indicated by one of the following numbers:

Main Index
 358 MSC Nastran Reference Guide
Output Description

Reason Description
Number
1 Two consecutive singularities encountered while performing triangular decomposition
2 Four shift points while tracking a single root
3 All eigenvalues found in the frequency range specified
4 Three times the number of roots estimated in the frequency range have been extracted
5 All eigenvalues that exist in the problem have been found
6 The number of roots desired have been found
7 One or more eigenvalues have been found outside the frequency range specified
8 Insufficient time to find another root
9 Unable to converge

 Largest off-diagonal modal mass term identified by mode pair

 Number of off-diagonal mass terms failing orthogonality criterion
Complex Eigenvalue Analysis

Each complex eigenvalue is identified with a root number determined by sorting the complex eigenvalues by magnitude.
The following summary of the complex eigenvalues extracted is automatically printed for all methods:
 Root number
 Extraction order (this is irrelevant under the Lanczos method)
 Real and imaginary parts of the eigenvalue
 The coefficients f (frequency) and g (damping coefficient) in the following representation of the eigenvalue
 p =  + i  :
p = 2f  i – 1  2 g 
Therefore,
-
f = -----
2
– 2 C
g = ---------- = 2 --------
 C cr
When using the complex Lanczos method, User Information Messages 5444, 5445, and 5453 are also issued. The
following summary is also automatically printed when the Determinant Method is used:
 Number of eigenvalues extracted
 Number of passes through starting points.

Main Index
 Chapter 7: Solution Sequences 359
Output Description

 Number of criteria changes

 Number of starting point moves
 Number of triangular decompositions
 Number of failures to iterate to a root
 Number of predictions outside region
 The reason for termination is indicated by one of the following integers:

Reason Description
Number
1 The number of roots desired have been found
2 All predictions for eigenvalues are outside the regions specified
3 Insufficient time to find another root
4 Matrix is singular at the first three starting points

 Swept determinant functions for each starting point

The following summary is automatically printed for each region specified when the Inverse Power Method is used:
 Number of eigenvalues extracted
 Number of starting points used
 Number of starting point or shift point moves
 Number of triangular decompositions
 Number of vector iterations
 The reason for termination is indicated by one of the following integers:

Reason Description
Number
1 Two consecutive singularities encountered while performing triangular decomposition
2 Four starting point moves while tracking a single root
3 All eigenvalues found in the region specified
4 Three times the number of roots estimated in the region have been extracted
5 All eigenvalues that exist in the problem have been found
6 The number of roots desired have been found
7 One or more eigenvalues have been found outside the region specified
8 Insufficient time to find another root
9 Unable to converge

Main Index
 360 MSC Nastran Reference Guide
Restart Procedures

Superelement Analysis
In addition to the previously mentioned automatic output, the superelements solution sequences also print the following:
 Superelement mapping information that lists the membership of grid points and elements in the superelements,
the order in which the superelements will be processed, and timing and space estimates for each superelement
 A page header that identifies the superelement for which the printed information on the page pertains
 Matrices wherein mechanisms are identified
 Resultants of external loads
 Maximum applied loads
 Maximum displacements
 Maximum SPC forces
For SOL 101, if PARAM,INREL,-1 or PARAM,INREL,-2, the following data is printed automatically:
INTERMEDIATE MATRIX...QRR

This is the rigid body mass matrix of the total structure, including superelements. It is presented as a 6  6 matrix,
measured at the reference grid point, in its global coordinates. Masses on scalar points are ignored.
INTERMEDIATE MATRIX...QRL
This is the resultant of the applied loads, measured at the reference point. There is one column for each loading condition.
INTERMEDIATE MATRIX...URACCEL
This is the user-supplied rigid body acceleration, input on the DMIG,UACCEL entry. This output does not appear if the
entry is not present.
INTERMEDIATE MATRIX...URA
This is the rigid body acceleration matrix, computed from the applied loads.

Restart Procedures
Scheduled exits can be requested at any point in a structured solution sequence format by means of the SEMG, SEKR,
SELG, SELR, SEMR, and SEALL Case Control commands. If none of these commands are specified, then SEALL = ALL
is the default.
Unscheduled exits are usually caused by errors on input entries or errors in the structural model resulting from missing or
inconsistent input data. When such errors are detected, an unscheduled exit is performed accompanied with the output of
the applicable user error messages. Following the correction of the input data errors, a modified restart can be performed.
Unscheduled exits may also occur because of machine failure or insufficient time allowance. In these cases, an unmodified
restart is usually made. In some cases, where a portion of the problem has been completed, including the output for the
completed portion, a modified restart must be made following an unscheduled exit due to insufficient time allowance. The
situations are discussed under case control requirements in the sections dealing with the individual solution sequences.

Main Index
 Chapter 7: Solution Sequences 361
General Operations

The initial execution of any problem must be made with complete MSC Nastran input data, including all of the bulk data.
However, all or part of the bulk data may be assembled from alternate input sources.
For restart, the Bulk Data Section consists only of delete “/” entries (see Bulk Data Section, 31) and new entries that the user
wishes to add. The previous Bulk Data is read from the database. All other parts of the MSC Nastran Input Data (including
the File Management statements, the Executive Control statements, the Case Control commands, the BEGIN BULK
command and the ENDDATA entry) must be resubmitted even though no changes are made in the case control and no
new bulk data is added. When changing solution sequences, the solution number (SOL) must be changed to the number
or name of the new solution sequence.
Any changes in the Case Control Section associated with Bulk Data entry selection or subcase definition, or changes in the
Bulk Data Section.
Restarts always start at the beginning of a specified DMAP or SOL sequence and queries the database in two phases.

Phase 1: Conceptually this phase marks all appropriate existing database data blocks and parameters as existing for
the current restart run.
Phase 2: This phase then conceptually checks the current input against the version from which the restart is starting
from and deletes from the Phase 1 determined database data blocks and parameters any data blocks that
were modified because of input. This phase is performed by the RESTART module(s) contained within the
solution’s DMAP.

Restart only executes DMAP Modules for which some or all of the output does not exist on the current version database.
The S-type parameter is considered as an output data block for restart purposes, hence if it is not listed in the NDDL, then
the module that contains it will be re-executed. All DMAP modules are executed until the first RESTART module after
which output checking is performed. Forced execution of modules after this point may be manually controlled by
SYSTEM (109) flag.
If no such changes are made, the executive system performs an unmodified restart. If changes have been made only in the
output requests, the restart is considered unmodified. However, some modules may have to be re-executed in order to
prepare the output. Output requests that were satisfied, plots made, etc., should be removed if they are not to be output
again.
For modified restarts, a number of previously executed DMAP instructions may have to be re-executed, depending on the
nature of the modifications made by the user. The DMAP instructions that need to be executed in a modified restart are
automatically determined within the program by comparing all changes made in Case Control commands and Bulk Data
entries.

General Operations
This section describes the data processing and matrix operations that are performed by the functional modules in the
solution sequences (SOLs 101 through 200) unless otherwise indicated. Detailed descriptions of the functional modules
are given in Chapter 4 or the MSC Nastran Programmers Manual. Additional information regarding the use of DMAP
modules for matrix operations is given in the MSC Nastran DMAP Programmer’s Guide.

Main Index
 362 MSC Nastran Reference Guide
General Operations

The following sections will refer to the structural analysis matrices such as stiffness, damping, and mass, but the discussion
also applies to heat transfer analysis for conductance and capacitance. There is no analogy for mass in heat transfer analysis.

Geometry Processing in SubDMAP PHASE0

The total model is given its initial processing in PHASE0. Data is taken from the input files and stored in data structures
that feed more expensive processes in later SubDMAPS. Flags are set to control branching for special features. All
superelement inputs are processed. Tables are indexed for faster processing in post-processors. Parameters are read from Case
Control and Bulk Data files by module PVT. The shell element normals at each grid point are computed in module
TASNP2, and output in table GPSNT.
Tables for internal program use are generated by the geometry processors from the grid point entries, coordinate system
definition entries, sequence entries, connection entries, static load entries, temperature definition entries, and constraint
entries. The computer time for all geometry processing is usually only a small part of the time required for problem solution.
The basic geometry processing is performed by GP1 (Geometry Processor — Phase 1). A list of all grid points and scalar
points is assembled in internal order for use in relating internal and external identification numbers. Matrices for
transformation from basic to all defined global coordinate systems are computed, and all grid point locations are
transformed to the basic coordinate system.
The element connection tables (ECT) are generated by GP2 (Geometry Processor — Phase 2) from the connection entries.
The external grid point numbers are converted to the internal indices defined in GP1.
The temperature data and static loads are processed by GP3 (Geometry Processor — Phase 3). The temperature data is
collected by sets and the external grid point numbers and element numbers are converted to the internal indices defined in
GP1 and GP2. In the case of static analysis, the static loads are also collected by sets and the external identification numbers
are converted to internal indices.
The constraint data is processed by GP4 (Geometry Processor — Phase 4). The displacement set definition table (USET)
is generated from the constraint definition entries. USET contains one coded word for each degree of freedom in the model
to identify the set or sets of coordinates to which the degree of freedom belongs (see Constraint and Set Notation, 324 for the
definition of degree of freedom sets). The multipoint constraint matrix  R mg  is formed from the MPC and rigid element
Bulk Data entries. In static analysis, the enforced displacement matrix  Y s  is formed for use in the generation of
equivalent static loads.
The BNDSPC module moves SPCs on boundary grid points to downstream superelements.

Model Checkout
A checkout procedure is provided for the development of new models. It includes the bookkeeping and control blocks, and
part of the generation block of Figure 7-3 in Superelement Analysis, 425. It is requested by PARAM,CHECKOUT,YES.
SEP1 Module

Generates the SEMAP table, which lists superelement membership of grid points and elements. Provides estimates for
computation costs and storage requirements. This module will identify any structural partitioning errors.

Main Index
 Chapter 7: Solution Sequences 363
General Operations

There are two paths through the superelement capability. If the bulk data entries have a BEGIN SUPER and/or BEGIN
SUPER = [SEID] entries the data flow goes through the partitioned superelement capability modules, SEP1X and SEP2X.
If only a BEGIN BULK entry is present the grid point list capability modules SEP1 and SEP2 are used. The SEP1X path
also allows grid point list superelements, with the exception of fluid structure analysis, SOL 200 (Design Optimization),
and the RELEASE entry. Models using these features must use the grid point list path. More details are given in Superelement
Analysis, 425.
SEP2, SEP2CT Modules

Partitions out the data for any superelement selected by SE-type Case Control commands. These modules identify load and
rigid element partitioning errors, and errors in the Case Control Section.
Constraint Modules

The GP1 and GP4 modules process all selected superelements. They check for the presence of requested coordinate systems;
for illegal duplicate specifications of mutually exclusive sets, such as SPCi and ASETi entries referencing the same point;
and for correct specifications on SPCi, xSETi, and MPCi entries, and rigid elements.
The user set table may be printed by use of PARAM,USETPRT, when PARAM,CHECKOUT,YES is present. If rigid
elements or selected MPC entries exist, the constraint matrix  R mg  (see Multipoint Constraint Operations in SubDMAPs SEKR and
SEMR2 and SEMRB, 366) is given three tests:

o
1. A matrix of rigid body vectors  u gh  is calculated by the VECPLOT module. This matrix includes zeroes for the
rows corresponding to scalar points. The product
o
 E mh    R mg   u gh 
is calculated. Terms of  E mh  larger than PARAM,TINY are printed. If the constraint matrix contains internal
constraints, nonzero terms will exist in  E mh  . If an MPC equation includes scalar points, nonzero terms may
exist in the corresponding row of  E mh  although the equation may be in equilibrium.

2. The product
g T
 R mm    R mg   R mg 
is calculated. The product matrix is decomposed by the DCMP module. Diagonal terms of the factor matrix that
are larger than PARAM,MAXRATIO are printed. These terms indicate that the printed row is linearly dependent
on prior rows. Unless the constraint matrix is changed, it is likely that the model will encounter fatal errors due to
singularity in the MCE1 module, or will have a poorly-conditioned constraint elimination process.
3. The product
m T
 R mm    R mm   R mm 

Main Index
 364 MSC Nastran Reference Guide
General Operations

is calculated, where  R mm  is the partition of  R mg  . This product is also processed by the DCMP module. If a
row appears here that also appears in the messages from the prior step, it confirms that the constraint matrix is
inherently ill-conditioned. If a row appears here that did not appear in the prior step, it indicates that a change in
the m-set exists that can also make the  R mm  matrix well conditioned.

The DBC module stores data for use in postprocessors. It runs at many points in the Solution Sequences. In this SubDMAP
it stores basic structural geometric data. The AEROE SubDMAP generates the geometry of the aerodynamic model. The
Automated Component Mode Synthesis (ACMS) capability controls branches at many places in the Solution Sequences.
The RESTART module controls the modules that will be run on restart, based on changes in the input files. The run exits
after these checks are performed.

Matrix Assembly Operations in SubDMAP SEMG

The stiffness, mass, damping, differential stiffness, and heat conduction matrices are formed during the matrix assembly
operations in subDMAP SEMG.
Stiffness is generated from the following sources:
 Structural elements defined on connection entries (e.g., CBAR and CROD).
 General elements defined on GENEL entries.
 Scalar springs defined on CELASi entries.
Mass is generated from the following sources:
 A 6  6 matrix of mass coefficients at a grid point defined on a CONM1 entry.
 A concentrated mass element defined on a CONM2 entry in terms of its mass and moments of inertia about its
center of gravity.
 Structural mass for all elements, except plate elements without membrane stiffness, using the mass density on the
material definition entry.
 Nonstructural mass for all elements specifying a value on the property entry.
 Scalar masses defined on CMASSi entries.
Damping is generated from the following sources:
 Viscous rod elements defined on CVISC entries.
 Scalar viscous dampers defined on CDAMPi entries.
 Element structural damping by multiplying the stiffness matrix of an individual structural element by a damping
factor obtained from the material properties (MATi) entry for the element.
These matrices are formed from the connection and geometric tables prepared by the geometry processors and the material
and element property tables prepared in the Phase 0. The matrix assembly operations will usually represent a significant
portion of the total solution time. The computer time for matrix assembly operations is linear with the number and type
of elements. Unit times for stiffness matrix assembly are given in the . Assembly times for coupled mass matrices, differential

Main Index
 Chapter 7: Solution Sequences 365
General Operations

stiffness matrices, and heat conduction matrices are about the same as those for stiffness matrices. Assembly times for
lumped mass matrices and damping matrices are usually relatively small.
The matrix assembly process begins with the generation of various tables in TA1 (Table Assembler) from the geometric data,
connection data and property data. These tables are used in data recovery operations as well as in the matrix assembly
operations.
The matrix assembly operations are performed by EMG (Element Matrix Generator) and EMA (Element Matrix
Assembler) modules. All of the element matrices (KELM, KDICT, MELM, MDICT, BELM, and BDICT) are generated
x 4
by EMG. The stiffness matrix exclusive of general elements,  K gg  , as well as the structural damping matrix,  K gg  , the
mass matrix,  M gg  , and the viscous damping matrix,  B gg  , are assembled in separate executions of EMA.

The parameter COUPMASS is used to select either coupled mass or lumped mass generation in EMG. If the parameter
WTMASS is present, the mass matrices generated in EMG are multiplied term by term by the value in field 3 of the
PARAM entry. This operation allows the user to specify mass data in units of weight.
4
The structural damping matrix  K gg  is assembled from elements with structural damping. Structural damping is
specified in the GE field of the MATi entry.
NE
4
 K gg  =  GE e  K e  (7-1)
e=1
where:

NE = number of elements
GE e = element structural damping coefficient from MATi entries
 Ke  = element stiffness matrix

The direct matrix inputs requested by the K2GG, M2GG, and B2GG Case Control commands defined the stiffness matrix
2 2 2
 K gg  , mass matrix  M gg  , and damping matrix  B gg  , and are assembled by the MTRXIN module. The outputs from
EMA and MTRXIN are combined by the ADD module.
The stiffness contributions from the general elements (GENEL Bulk Data entry) are assembled in SMA3 (Structural Matrix
x
Assembler -- Phase 3). The final stiffness matrix  K gg  is formed by combining  K gg  prepared by the EMA module
with the contributions from general elements generated by SMA3.
Weight and balance information is calculated from the mass matrix  M gg  by GPWG (Grid Point Weight Generator).
The execution of GPWG is controlled by the parameter GRDPNT. The details of the output are described in GRDPNT
(p. 838) in the MSC Nastran Quick Reference Guide.

Main Index
 366 MSC Nastran Reference Guide
General Operations

Multipoint Constraint Operations in SubDMAPs SEKR and SEMR2 and SEMRB

The multipoint constraint operations apply constraints to the stiffness matrices using the linear relationships among the
displacements specified by the user on MPC, MPCAX, and rigid element Bulk Data entries.
The multipoint constraint equations are initially expressed in the form,
 Rg   ug  = 0 , (7-2)

where the coefficients are supplied by the user. The user also specifies the degree of freedom that is made dependent by each
equation of constraint, so that the  u g  matrix may immediately be partitioned into two subsets,

 un 
 ug  =   , (7-3)
 um 

where the set, u m , is the set of dependent degrees of freedom. The matrix of constraint coefficients is similarly partitioned

 Rg  =  Rn Rm  , (7-4)

so that equation (7-2) becomes

 Rn   un  +  Rm   um  = 0 . (7-5)

 R m  is a nonsingular matrix. We can, therefore, form the multipoint constraint matrix,

–1
 Gm  = – Rm   Rn  , (7-6)

so that equation (7-4) may be stated as

 um  =  Gm   un  . (7-7)

Prior to the imposition of constraints, the structural problem may be written as

 K gg   u g  =  P g  , (7-8)

or, partitioning in terms of the coordinate sets, u n and u m

K nn K nm  u n   Pn 
  =   (7-9)
K nm K mm  u m   Pm 
T

Bars over symbols are used to designate arrays that are replaced in the reduction process.

Main Index
 Chapter 7: Solution Sequences 367
General Operations

The addition of constraints to the structure requires that the forces of constraint be added to the equilibrium equations.
Thus, writing the equilibrium and constraint equations together in partitioned form,

T    
K nn K nm G m  u n   Pn 
   
=  P ,
K nm K mm – I  u m 
T (7-10)
 m 
Gm –I 0  qm   0 
   

where  q m  is the vector of constraint forces on  u m  . Straightforward elimination of u m and q m gives

T T T T
 K nn + K nm G m + G m K mm + G m K mm G m   u n  =  P n  +  G m   P m  , (7-11)

or
 K nn   u n  =  P n  , (7-12)

where
T T T
K nn = K nn + K nm G m + G m K nm + G m K mm G m , (7-13)

and
T
Pn = Pn + Gm Pm . (7-14)

The initial partition of K gg and the operations indicated by (7-6), (7-13) and (7-14) are performed by appropriate modules
of the program. The multipoint constraint matrix, G m , is used in structural matrix reduction ((7-13)), load vector
reduction, ((7-14)) and data recovery ((7-7)).
The multipoint constraint equations, R mg , formed in the GP4 module, are partitioned by the MCE1 module (Multipoint
Constraint Eliminator -- Phase 1) as follows:

 R mg  = R mm R mn (7-15)

MCE1 also solves the equation

 R mm   G mn  = –  R mn  (7-16)

Main Index
 368 MSC Nastran Reference Guide
General Operations

for the transformation matrix  G mn  . An alternate path is available for module MCE1 in SubDMAP XMCE1, requested
–1
by PARAM,MCE1,NEW. It contains a partitioned solution for R mm  R mn , where the part of R mm that is diagonal is
partitioned from the coupled part. It was developed at a time when the partitioned solution was faster than the single-pass
solution used in MCE1. The single pass solution was improved so that it is now faster than the partitioned solution. The
partitioned solution is maintained to provide an alternate solution when the single-pass solution is implausible, or does not
complete in a normal manner.
The MCE2 module (Multipoint Constraint Eliminator — Phase 2) partitions the stiffness matrix

K nn K nm
 K gg  = (7-17)
K mn K mm

and performs matrix reduction

T T
 K nn  =  G mn   K mm G mn + K mn  +  K mn G mn + K nn  (7-18)

4
The other structural matrices,  K gg  ,  B gg  and  M gg  , are transformed by formulas that are identical in form to
equation (7-13), in subDMAPs SEMR2 and SEMRB.
The MCE2 module is also used in subDMAP SEMR2 to obtain  M nn  .

The UPARTN module partitions  K gg  to  K mg  form which is equivalent to

 K mg  =  K mn K mm  (7-19)

and is used to obtain multipoint forces of constraint. See Data Recovery Operations in SubDMAP SEDISP, 391.

Single Point Constraint Operations in SubDMAPs SEKR, SEMR2 and SEMRB

The single point constraint operations apply constraints to the structural matrices using the single point constraints
specified by the user on SPCi Bulk Data entries, the PS field on GRID entries, or constrained by the AUTOSPC feature
described in Constraint and Mechanism Problem Identification in SubDMAP SEKR, 370.
Single point constraints are applied to the set, u s , in the form

 us  =  Ys  , (7-20)

where  Y s  is a vector of enforced deformations, any or all of whose elements may be zero. The set, u n , is partitioned
into u s and u f (the free or unconstrained set)

Main Index
 Chapter 7: Solution Sequences 369
General Operations

 uf 
 un  =   (7-21)
 us 

The stiffness matrix, K nn , is similarly partitioned

K ff K fs
 K nn  = (7-22)
T
K fs K ss

The complete structural equations including the single point forces of constraint, q s , may be written in partitioned matrix
form as

   
K ff K fs 0  u f   Pf 
   
K fs K ss – I  u s  =  P s 
T (7-23)
   
0 I 0  qs   Ys 
   
Straightforward elimination gives

 K ff   u f  =  P f  –  K fs   Y s  =  P f  (7-24)

The forces of constraint are recovered by means of the middle row of equation (7-23), i.e.,
T
 q s  = –  P s  +  K fs   u f  +  K ss   u s  (7-25)

The distinct partitions K fs and K ss , are needed in subsequent calculations of SPC forces in static analysis. For the other
4
structural matrices ( K nn , B nn , and M nn ) the (sf ) and (ss) partitions are saved, for dynamic SPC forces calculation.

The partially constrained stiffness matrix  K nn  is partitioned by UPARTN module as follows:

K ff K fs
 K nn  = (7-26)
K sf K ss

Similar partitioning operations are performed on  M nn  in subDMAP SEMR2 to obtain  M ff  ,  M sf  , and M ss . In

subDMAP SEMRB, the structural and viscous damping matrices

Main Index
 370 MSC Nastran Reference Guide
General Operations

4 4
 K sf  ,  K ss  ,  B sf  and  B ss  (7-27)

are obtained from the MATREDU module. All (sf ) and (ss) matrices are required for the recovery of single point constraint
reaction forces described in Data Recovery Operations in SubDMAP SEDISP, 391.

Constraint and Mechanism Problem Identification in SubDMAP SEKR

A singularity in a linear equation solution leads to a condition where a unique solution is not possible. For example, if
 K   u  =  F  , and  K  is zero, then if  F  is not zero, no value of  u  will satisfy the equation. However, if  F 
is zero, any value of  u  will satisfy the equation. The equation can always be satisfied by prescribing  u  , rather than
 F  . This is equivalent to constraining u with a single-point constraint.
There are two types of singularity:
1. The singularity can be identified by considering the stiffness terms of only one grid point. This is called a grid point
singularity.
2. The singularity requires consideration of stiffness terms of more than one grid point. This is called a mechanism
type of singularity.

Grid Point Singularity

Grid point singularities are automatically identified by the GPSP module. They may be automatically constrained, under
user control. First, multipoint constraints are eliminated, leading to the matrix  K nn  (see Multipoint Constraint Operations in
SubDMAPs SEKR and SEMR2 and SEMRB, 366). Singularities are identified by grid point and component number of the failed
direction. The ratio of stiffness at the degree of freedom to the largest stiffness at that grid point is listed. If automatic
constraint is selected (see AUTOSPC (p. 804) in the MSC Nastran Quick Reference Guide), the set membership of the singular
degrees of freedom is listed before the automatic constraint and after the automatic constraint. Both the mutually exclusive
set and the highest combined set identical to the mutually exclusive set are listed for each identified singularity.
The output will include a Grid Point Singularity Table with the following format:

G R I D P O I N T S I N G U L A R I T Y T A B L E

POINT TYPE FAILED STIFFNESS OLD NEW

ID DIRECTION RATIO USET USET
3801 G 1 0.00E+00 L S *
3801 G 2 0.00E+00 L S *
3801 G 3 0.00E+00 L S *
3801 G 4 0.00E+00 L S *
3801 G 5 0.00E+00 L S *
3801 G 6 0.00E+00 L S *

In this table,
1. G (under the heading TYPE) stands for grid point (S for scalar point is also output).

Main Index
 Chapter 7: Solution Sequences 371
General Operations

2. FAILED DIRECTION is the degree of freedom component in the displacement coordinate system and closest to
the weakest direction. See Grid Point and Coordinate System Definition, 35 for definitions of component numbers.
3. The STIFFNESS RATIO is the ratio of the stiffness in the weakest direction to the stiffness in the strongest
direction at the grid point (considering all possible directions rather than just the coordinate directions).
–8
Translational and rotational stiffnesses are considered separately. The default value for the failure criterion is 10 ,
which may be changed by the PARAM,EPZERO entry.
4. Note that, for the example shown, the failed degrees of freedom were originally in the u l set and were placed in the
u s set. This may not always be the case, for reasons noted earlier, or because the user has not requested the
AUTOSPC feature.
5. The user should carefully inspect the Grid Point Singularity Table because a singularity is very often the result of a
modeling error, which should be corrected.

User Options
The identification of singularities is an automatic feature; however, the user may exercise the following options through
PARAM Bulk Data entries described in Parameters (p. 793) in the MSC Nastran Quick Reference Guide.

Parameter
Name Description
AUTOSPC Controls automatic constraint of a singularity.
EPPRT Specifies the maximum value of a singularity to print.
EPZERO Specifies the maximum value of a singularity.
PRGPST Controls printout of singularity table.
SPCGEN Writes SPC entries to the PUNCH file based on the singularity table.

The singular degrees of freedom are shifted to other sets by the following logic if PARAM,AUTOSPC,YES is specified:

o-set Set Before Set After

Condition Constraint Constraint
If o-set exists o s
c b
If no o-set c s
b s

Other mutually exclusive sets not listed above will not be constrained. In addition, singular degrees of freedom on CYAX,
CYJOIN, and SPCOFFi entries are identified but not constrained. The most usual cause of grid point singularity is degrees

Main Index
 372 MSC Nastran Reference Guide
General Operations

of freedom not constrained by elements because they are not of interest to the modeler. An example is out-of-plane rotations
on a component made of plates lying in a plane. The automatic constraint provides the correct action for this type of
singularity. Another cause is elements left out by oversight. Here, one typically observes that many or all degrees of freedom
at a grid point are singular. The automatic constraint may mask a modeling error in this instance. For this reason, all
singularity messages should be carefully inspected during the modeling checkout phase to ensure that modeling errors are
not being masked.
When using the K2PP, M2PP, B2PP, or TFL Case Control command, PARAM,AUTOSPC,NO is recommended.

Mechanism-Type Singularities
A more complicated type of instability arises when two or more grid points participate in a singularity in a linear equation
solution. This is known as a mechanism type of instability. Again, for the simple example shown in Figure 7-2, the equations
of equilibrium can not be satisfied at all, or may have non unique solutions for special loading conditions. Note, however,
that the structure of this example will pass the grid point singularity checks.

u1 u2

Figure 7-2 A Mechanism Type of Instability

The static equilibrium equation for the example in Figure 7-2 is

u P
k –k  1  =  1  (7-28)
   
–k k  u2   P2 

If P 1 = – P 2 , many solutions are possible, where  u 1 – u 2  = P 1  k , but u 1 may have any value. If P 1 and P 2 are
arbitrary, the equation cannot be satisfied.
This condition is detected during the matrix decomposition phase of the linear equation solution process by the DCMP
module. The stiffness matrix is decomposed into a lower triangular matrix,  L  , and a diagonal matrix,  D  . As discussed
in User Interface (p. 78) in the a diagonal term of  D  (“ d i ”) will approach zero for every mode of instability of the
structure. (In this example there is one instability.) These instabilities are identified by dividing all d i ’s into their

Main Index
 Chapter 7: Solution Sequences 373
General Operations

corresponding stiffness term, then listing all ratios larger than a large number. See MAXRATIO (p. 899) in the MSC Nastran
Quick Reference Guide for a description of several other parameters used to control this operation.
T
Let  K  =  L   D   L  (7-29)

Then  D  =
k 0 , L = 1 0 (7-30)
0 k – k –1 1

– 10
If  k – k  = 0.0 , replace d i with 10 (7-31)

Ratio i = k ii  d i (7-32)

Identify Grid Component for All Ratios > PARAM,MAXRATIO.

A corrective action for this example would be to constrain one of the grid points or to add springs connecting this spring
to other parts of the structure (not shown). Since, in general, there may be many grid points active in a singularity, there is
little reason for picking one over another for connection. The usual cause of this type of instability is elements left out by
oversight. While this may be the user’s intention, it cannot be detected by the program. For these reasons, there is no
automatic method to constrain mechanisms. The discussion of MAXRATIO (p. 899) in the MSC Nastran Quick Reference
Guide describes the various actions and user options that are available in the several contexts where singularities may be
detected. Diagnosis techniques and corrective actions are described in the MSC Nastran Numerical Methods User’s Guide.

Static Condensation in SubDMAPs SEKR and SEMR2

Structural partitioning is used to refer structural matrices to the boundary points in substructuring operations and to
perform a static condensation in dynamic analyses.
At user option the set of free coordinates, u f , may be partitioned into two sets, u o and u a , such that the u o set is
eliminated first. Thus

 ua 
 uf  =   (7-33)
 ou 
The equilibrium equations after the elimination of constraints equation ((7-24)) may be written in partitioned form as

K aa K ao  u a   Pa 
  =   (7-34)
T
K ao K oo  u o   Po 

Rearrange the bottom half of equation (7-34):

Main Index
 374 MSC Nastran Reference Guide
General Operations

T
 K oo   u o  =  P o  –  K ao   u a  , (7-35)

and solve for  u o  :

–1 –1 T
 u o  =  K oo   P o  –  K oo   K ao   u a  (7-36)

Substitute for u o into the top half of equation (7-34):

–1 T –1
 K aa – K ao K oo K ao   u a  =  P a  –  K ao   K oo   P o  (7-37)

It is convenient to define the matrix

–1 T
 G o  = –  K oo   K ao  , (7-38)

so the equation (7-37) becomes

T
 K aa + K ao G o   u a  =  P a  +  G o   P o  , (7-39)

where advantage is taken of the symmetry of  K oo  .

Following the practice of condensation established in preceding subsections,

 K aa   u a  =  P a  , (7-40)

where

 K aa  =  K aa  +  K ao   G o  , (7-41)

T
 Pa  =  Pa  +  Go   Po  (7-42)

The  G o  matrix defined in (7-38) is obtained practically from the solution of

T
 K oo   G o  = –  K ao  , (7-43)

where
T
 K ao  (7-44)

is treated as a set of load vectors. Each such vector produces a column of  G o  .

Main Index
 Chapter 7: Solution Sequences 375
General Operations

Once  u a  is obtained the set of omitted coordinates,  u o  , is obtained as follows. Define the set

 o
 uo  (7-45)
 
as the solution of

 o
 K oo   u o  =  P o  (7-46)
 
Then, using equation (7-38)(7-46) in equation (7-35),

 o
 uo  =  uo  +  Go   ua  . (7-47)
 

In subDMAP SEKR, the constrained stiffness matrix,  K ff  , is partitioned by UPARTN as follows:

K aa K ao
 K ff  = (7-48)
K oa K oo

DCMP and FBS are used to solve the equation

t
 K oo   G oa  = –  K oa  (7-49)

t
for the static transformation matrix,  G oa  , and MPYAD performs the matrix reduction

t
 K aa  =  K aa  +  K ao   G oa  (7-50)

x
If PARAM,ALTRED,YES is specified then the  K ff  matrix is partitioned and reassembled to form  K ff  which has the
x
a-set degrees of freedom sequenced last.  K ff  is partially decomposed

x
 K ff    L fo ,  K aa  (7-51)

where  L fo  is a trapezoidal matrix and  K aa  is the boundary stiffness matrix.

The  L fo  matrix is partitioned

Main Index
 376 MSC Nastran Reference Guide
General Operations

L oo
 L fo  = (7-52)
L ao

T t
While  L ao  is of the same dimension as the  G oa  matrix, it is on the average half as dense, thereby reducing the
storage requirements.
In subDMAP SEMR2, UPARTN partitions the constrained mass matrix  M ff  as follows:

M aa M ao
 M ff  = (7-53)
M oa M oo

and successive executions of MPYAD and SMPYAD modules perform the matrix reduction

t T t T t
 M aa  =  G oa   M oo G oa + M oa  +  M oa G oa + M aa  (7-54)

4
In subDMAP SEMRB, the structural and viscous damping matrices  B ff  and  K ff  are condensed by the MATREDU
module.
The partitioned solution for statics is an exact solution. It merely changes the order of operations of the unpartitioned
solution. The static reduction for dynamics is exact for o-set degrees of freedom that are massless, but represents an
approximate relumping of o-set masses to the a-set. The user must select enough a-set degrees of freedom to adequately
define the eigenvectors of the system. The dynamic reduction technique discussed in Dynamic Reduction and Component Mode
Synthesis in SubDMAP SEMR3, 383 provides a way of improving the approximation by computing the normal modes for the
o-set degrees of freedom.

Static Solutions in SubDMAP SEKRRS

There are two major branches through this subDMAP, standard static analysis and static analysis with inertia relief effects.

Inertia Relief Not Selected

When inertia relief is not selected and there are no SUPORTi entries are present the a-set equations are solved directly
K aa  u a = P a (7-55)

K aa is decomposed into its factor matrices L aa and D a . These factors are passed on to SubDMAP STATRS for solution
of u a .

Main Index
 Chapter 7: Solution Sequences 377
General Operations

When inertia relief is not selected but there are SUPORTi entries are present the a-set equations are partitioned into the r-
set, defined on the SUPORTi entries, and the l-set, what is “left over” after the r-set is removed.

K ll K lr u l P1
 (7-56)
K rl K rr u r Pr

In this context, the r-set is treated as an alternate form of single-point constraints. The partitions of K involving r are
discarded. u r is set to zero. u l is found from the first row of this equation. Loads in P r go directly to ground, and are
output with the SPC forces, along with the product – K rl  u l .

Inertia Relief Selected

The basic assumption of inertia relief is that the structure is in a state of steady acceleration, and that all transients due to
dynamic effects are negligible. The solution is obtained in a moving reference frame from which all displacements are
measured. For inertia relief analysis inertial loads are computed automatically that exactly balance the applied loads. A
support system is applied to make the system numerically stable. The inertial loads assume rigid body accelerations when
computing the loads.
The derivations given here assume that the system mass matrix M gg has been reduced to M aa and is available, as
described elsewhere. In the actual DMAP used in MSC Nastran 2001 equivalent methods that never generate M aa are
used instead, in the interests of efficiency. M aa can be a larger size data block, and costly to compute. It can be shown that
the results provided by the two derivations are identical. For example, the rigid body mass matrix M rr , defined below as

M rr = D ar  M aa  D ar (7-57)

can just as easily be formed from the g-set equations

M rr = D gr  M gg  D gr (7-58)

Similar comments can be made about computing the inertial loads that balance the applied loads. The derivatives are
presented here in the form of a-set equations that are more readily understood, if less efficient in practice, than the g-set
equations.
There are at present two forms of inertia relief: automatic selection of a support system (“auto-support”); and manual
selection of the support system, using SUPORTi entries. Both react to PARAM,GRDPNT, the parameter used to locate a
reference point for grid point weight summaries. The more automatic method is described first.

Main Index
 378 MSC Nastran Reference Guide
General Operations

Auto-Support (PARAM,INREL,-2)
Vectors of rigid body motion D gr for six unit motions of a reference point are computed from grid point geometry in
module VECPLOT. The D ar component is partitioned from this matrix. D ar is used to compute the rigid body mass
M rr about the reference point.

MD ar = M aa  D ar (7-59)

M rr = D ar  MD ar (7-60)

M rr is printed out with the matrix name of QRR.

The acceleration of the reference point ar due to the applied loads is computed,

M rr  ar  D ar  P a (7-61)

The accelerations ar are printed with the matrix name of URA.

The resulting inertial loads P ai and total loads P a are

P ai = M aa  ar (7-62)

P a1 = P a – P ai (7-63)

A constraint equation is written that states that the average weighted motion of the system is zero,
MD ar  u ar = 0 (7-64)

This equation is appended to the a-set to define the 1-set equations in their augmented form,
K ll  u l = P l (7-65)

K aa MD ar u a P a1
= (7-66)
MD ar 0 rr – q r 0r

q r are the unknown forces required on the system to produce equilibrium. They should be computational zeros, and are
output with the SPC forces. The equation involving K ll is solved directly, with u a and q r partitioned from u l .

Main Index
 Chapter 7: Solution Sequences 379
General Operations

Manual Support (PARAM, INREL, -1 and SUPORTi)

The user prescribes a set of DOFs that constrain the structure in a statically-determinate manner on SUPORTi entries.
These entries define the r-set. There must be six or less DOFs listed, and they must constrain all rigid body modes that are
not constrained by SPCs or other modeling devices. The rigid body modes are computed from the statics equation, where
there are no loads applied, and u rr is the identity matrix. Given

K ll K lr D lr 0 lr
= (7-67)
K le K rr I rr 0 rr

D lr is found by solving the first row of this equation, then merged with I rr to form D ar ,

K ll  D lr  – K lr  I rr (7-68)

The rigid body mass, reference point accelerations, and inertial loads are formed with the same equations used for auto-
support. The loads are partitioned into the l- and r-sets, and the partitioned form of K ll is used for solution for
displacements,
K ll  u l = P ll (7-69)

u a is merged from u l and a null partitioned for u r .

A Comparison of Auto- and Manual-Support Solution

If one analyzes the same model with both methods, you should find that stresses, constraint forces, and other internal
quantities are identical except for truncation errors, but the displacements may differ appreciably. This is because they use
a different reference frame. For the many supports, the reference frame is set up so that the motion at the r-set DOFs are
identically zero. One can think of a reference frame drawn through these points, and all displacements measured relative to
it. This reference frame is moving through space at a constant acceleration rate, so the displacements are relative, rather than
inertial displacements.
For the auto-support option, one can visualize the reference frame as a set of axes connected to every point with mass with
a very complicated set of levers that allows the structure to “breathe”, as loads are applied to it, but keep the weighted average
motion to a zero value. If the model happens to have a grid point at the reference point, you will observe very small motion
there, computational zeros but not the binary zeros of the manual support method. This is because this motion is computed
for the auto-support method, but assumed to be zero for the manual method.
The choice of an r-set as defined by SUPORTi DOFs requires meeting both a geometric requirement and a stiffness
requirement. The geometric requirement is that the set should form a statically determinate tie to ground. This can be
difficult to do with complicated geometries. The stiffness requirement is that the DOFs should be stiff enough to provide
a stable solution. A method that has been used by some modelers is to pick points stiff enough that they could be used to
lift the structure without damaging it. Sometimes such points do not exist. The difficulties in picking a stable, reliable
support system was one of the main motivations for development of the more modern auto-support systems.

Main Index
 380 MSC Nastran Reference Guide
General Operations

PARAM, GRDPNT, GID is optional for both methods when superelements are not present, but required by the manual
method when superelements are present. The auto-method does not support superelement analysis. The default for
GRDPNT is GID = -1, which is a flag to use the basic origin as the reference point when computing rigid body modes.
Use of a GID near the c.g. of the structure produces more esthetically pleasing results. The rigid body accelerations
produced by the solution are then measured near the center of the structure, rather than at a point that may be outside of
the structure.

Static and Dynamic Load Generation

Applied Static Loads in SubDMAPs SELG, SELR, and SELRRS
The static load vectors  P g  are generated by SSG1 (Static Solution Generator — Phase 1) in subDMAP SELG from the
user input of applied loads, temperature fields, and enforced deformations.

The direct matrix input requested by the P2G Case Control command define the load matrix  P g2  which is assembled by
the MTRXIN module and is added to the output from SSG1.
If PARAM,INREL,-1 is specified, then
 The rigid body transformation matrix  D gr  is generated with the VECPLOT module. Its columns contain the
motion of all degrees of freedom due to unit motion of the corresponding reference point degree of freedom. (See
PARAM,INREL in Use of Parameters, 31 for a discussion of the reference point (“GIC”).)
 Loads due to unit rigid body accelerations are computed,

 P ji  = –  M jj   D gr 
 The loads above are appended to the applied loads  P j  ,

 Pj    Pj P ji 
The SELA module assembles  P j  into  P g  .

The constraints are applied to the static load vectors by SSG2 (Static Solution Generator — Phase 2) in subDMAP
SELR as follows:

 Pn 
 Pg  =   (7-70)
 Pm 

T
 P n  =  P n  +  G mn   P m  (7-71)

Main Index
 Chapter 7: Solution Sequences 381
General Operations

 Pf 
 Pn  =   (7-72)
 Ps 

 P f  =  P f  –  K fs   Y s  (7-73)

 Pa 
 Pf  =   (7-74)
 Po 

T
 P a  =  P a  +  G oa   P o  (7-75)

If PARAM,ALTRED,YES is specified, the diagonal extracted from  L oo    D oo 

and the reduced load matrix is computed from

 P a  =  P a  –  L ao   D oo   u ox  (7-76)

where  u ox  is obtained on a forward pass from the equation

 L oo D oo   u ox  =  P o  (7-77)

for PARAM,ALTRED,NO, the SSG3 module (Static Solution Generation — Phase 3) solves the equation

 K oo   u oo  =  P o 

for  u oo  , the displacements of the omitted coordinates.

SSG3 also calculates the residual vector, P o , and the residual vector error ratio,  o , for the omitted coordinates

 P o  =  K oo   u oo  –  P o  (7-78)

T
 u oo   P o 
o = -------------------------------
- (7-79)
T o
 Po   uo 

Except for round-off error, the error ratio  o should be zero. Large values of these error ratios usually indicate singularities
in the stiffness matrix. The residual load vector, RUOV, may be output by use of PARAM,IRES,1.
The quantity

Main Index
 382 MSC Nastran Reference Guide
General Operations

T
1  2  Po   uo  (7-80)

is calculated by the SSG3 module and printed under the heading “External Work.” This component of strain energy
includes effects of thermal loads, element deformations, and enforced displacements that may be subtracted later in the
solution process. For example, an enforced displacement that causes zero strain will result in external work.
If a r-set is present, subDMAP SELRRS calculates determinate forces of reaction
T
 q r  = –  D ar   P a  (7-81)

where:

 D ar  = D
er
I rr

Applied Dynamic Loads

Dynamic loads are produced in many subDMAPs. They are generally of the form:

Px =  POi  Fi  x  (7-82)

i

The PO i vectors are spatial functions that describe what DOFs are loaded and their scale factors. The F i functions are
temporal functions that define how the loads change with time in transient analysis x = t , or with excitation frequency
in frequency response analysis x = f . All spatial loads except those on extra points are calculated in the static analysis load
generator module, SSG1, to obtain the P g vector described in the prior paragraphs.

Enforced motions are produced similarly.

Yx =  YOi  Fi  x  (7-83)

i

The YO i vectors are again the spatial distribution, and the F i  x  temporal functions. The YO i vectors are produced by
the GP4 module to produce the Y g vector. It is the same vector described for static analysis in (7-73), except that it is merged
to g-set size with null terms from Ys.
The static load and enforced motion vectors are converted into dynamic loading tables in DPD module, then passed on to
the dynamic load generators, FRLG and TRLG. These modules produce the applied loads and enforced motions at the p-
set level, Pp  x  and Yp  x  , as a function of excitation frequency or time. For transient analysis velocity and acceleration
vectors are also provided, Yp·  x  and Yp
·· x  .

Main Index
 Chapter 7: Solution Sequences 383
General Operations

A reduction matrix Rpx is made in the FDRMGEN module, where the subscript “x” implies the d-set, for direct analysis,
and h, for modal analysis. It combines the operations shown in equations (7-70) through (7-75) previously. In subDMAP
CFORCE2 equation (7-73) is modified to add velocity and acceleration effects,

P f = P – K fs  Y s – B fs  Y· s – M fs  Y··s (7-84)
f

In frequency response the velocity and acceleration effects can be computed from Y s by multiplying it by i   once or
twice, where  is the excitation frequency.

P f is reduced to P d or P h by the same techniques used in static analysis.

For modal analysis with enforced motion it is essential that RESVEC be used. It causes static solutions to be formed from
K fs  Y s -type terms. When the structure is tied to ground in a statically-determinate manner, these solutions are linear
combinations of the rigid body modes that would exist if the ties to ground were removed. If the ties to ground are
redundant, a condition named multi-base input, the resulting shapes will have significant stresses associated with them.
They represent the effect of low-frequency excitation on the structure when RESVEC is not present, for the statically
determinate case the element stresses will be correct but the displacement-type data will be missing the effect of the static
shapes. For the multi-base input case, both stresses and displacement-type output will be incorrect.

Dynamic Reduction and Component Mode Synthesis in SubDMAP SEMR3

These capabilities use several subsets of the a-set not used by other solution techniques. The v-set matrix is formed in
subDMAP SEKR

K oo K or K oc
K vv = K ro K rr K rc (7-85)
K co K cr K cc

The v-set mass matrix is formed similarly in subDMAP SEMR3. The v-set degrees of freedom are free to vibrate during
dynamic reduction and component mode computation. They consist of the omitted terms (o-set), the free boundary points
used for rigid-body mode designation (r-set), and the other free boundary points (c-set). The  M v  matrix is formed in a
similar manner.
If generalized dynamic reduction is requested by the DYNRED Case Control command, the shifted stiffness matrix is
formed,
 A vv  =  K vv  +  s  M vv  (7-86)

Main Index
 384 MSC Nastran Reference Guide
General Operations

where  s is determined from data on the specified DYNRED Bulk Data entry. The approximate eigenvectors are found
from  A vv  and  M vv  using inverse iteration based on solving the equation

 A vv    v  i + 1 =  M vv    v  i (7-87)

for   v  , starting with random vectors for   v  . The   v  are stored, then made mass orthogonal with respect to
i+1 o i
each other to form the matrix   vy  , where N y the number of generalized functions used, may be determined
automatically. N y must be equal to or less than N q , the upper limit for the number of generalized degrees of freedom
specified by the user.
The generalized mass and stiffness matrices are computed,
T
 M yy  =   vy   M vv    vy  (7-88)

T
 K yy  =   vy   K vv    vy  (7-89)

If the METHOD command is specified for superelements, then the component modes are computed. The auto-omit
operations described in Real Eigenvalue Analysis in SubDMAPs SEMR3 and MODERS, 408 are carried out resulting in the reduced
eigensolution equation
 K yy – M yy    yz  = 0 (7-90)

where N z is the number of eigenvectors requested by the EIGR or EIGRL Bulk Data entries. The back-transformations
for the auto-omit-spc are performed, resulting in the expanded eigenvector matrix   yz  . This matrix is used to purify the
approximate eigenvectors found in dynamic reduction,
  vz  =   vy    yz  (7-91)

The uncoupled eigensolution components are assembled from   vz 

  oz 
 
  vz  =   rz  (7-92)
 
  cz 

Main Index
 Chapter 7: Solution Sequences 385
General Operations

  rz 
 
  az  =   cz  (7-93)
 
  bz 

where   bz  =  0 

  oz  and   az  are stored for use with the uncoupled solution obtained using PARAM,FIXEDB,-1, as described in Use
of Parameters, 31 and Data Recovery Operations in SubDMAP SEDISP, 391.
The residual vector capability is requested with the RESVEC Case Control command. Details on this capability are given
in Matrix Assembly Operations in SubDMAP SEMG, 364, where real eigensolutions are discussed. The RESVEC SubDMAP returns
auxiliary functions u o2 . They are appended to the eigenvectors, and are given all of the operations given to eigenvectors,

 oz u o2
H vz =  rz 0 r2 (7-94)
 cz 0 c2

The inertia relief mode shapes are calculated by statically applying the rigid-body inertial loads to the structure.

i
K oo K oa H o M oo M oa V o
= (7-95)
K ao K aa H i M ao M aa V a
a

i i
Setting  H a  to zero and solving for  H o  results in,

i –1
 H o  =  K oo    M oo   V o  +  M oa   V a   (7-96)

The  G oq  matrix is formed,

1
 G oq    G oz H oi 0 (7-97)

The  G oq  matrix is truncated or padded with null columns to provide Nq columns, where Nq is the q-set size.  G oq 
is expanded to a-size,

 G oq q 
0 ot    G oa (7-98)

The total transformation matrix is formed by adding dynamic and static components and is reassembled whenever needed,

Main Index
 386 MSC Nastran Reference Guide
General Operations

t  + Gq 
 G oa  =  G oa (7-99)
oa

The generalized stiffness coefficients are formed from the dynamic transformation,
T
 K qq  =  G oq   K oo   G oq  (7-100)

The physical and generalized degrees of freedom are not stiffness coupled.  K qq  is merged to N a size,

l  =
K qq 0
 K aa (7-101)
0 0

where superscript l refers to matrices formed from dynamic reduction.  K qq  is stored in the database. When the total
stiffness boundary matrix is needed, it is formed from its partitions,

m  =
K qq 0
 K aa (7-102)
0 K tt

or
m l 
 K aa  =  K aa  +  K aa (7-103)

The physical and generalized degrees of freedom are mass-coupled. The terms are formed from the equations,
T
 M qq  =  G oq   M oo   G oq 
T t 
 M qt  =  G oq   M ot + M oo G oa
(7-104)

l  =
M qq M qt
 M aa1
M tq 0

l  to form  M l  .
Virtual mass effects, if any, are added to  M aa1 aa
The total mass matrix is formed by:
m  = M  + Ml 
 M aa (7-105)
aa aa

For the rigid formats  M aa  is formed directly from its partitions,

Main Index
 Chapter 7: Solution Sequences 387
General Operations

M qq M qt
 M aa  = (7-106)
M tq M tt

The a-set matrices do not contain static effects in the t-set components, and dynamic effects in the q-set components.
m .
Virtual mass effects, if any, are added to  M aa  to form  M aa

4  are reduced to the a-set by the MATREDU module which performs the
The damping matrices  B gg  and  K gg
following operations:
1. Eliminate multipoint constraints

B nn B nm
 B gg  
B mn B mm

T T
 B nn  =  G mn   B mn G mn + B mn  +  B mn   G mn  +  B nn 
2. Eliminate single point constraints

B ff B fs
 B nn  
B sf B ss
3. Partition omitted degrees of freedom

B aa B ao
 B ff  
B oa B oo
4. Perform static condensation
T T
 B aa  =  G oa   B oo G oa + B oa  +  B oa   G oa  +  B aa 

Formulation of Dynamic Equations in SubDMAP GMA

The tables prepared by the geometry processors are modified by DPD (Dynamic Pool Distributor) to include the extra
points (e-points) introduced for dynamic analysis.

Main Index
 388 MSC Nastran Reference Guide
General Operations

m  ,  M m  ,  B  , and  K 4  are expanded to the d-set to form  K 1  ,  M 1  ,  B 1  , and  K 4  . The

 K aa aa aa aa dd dd dd dd
2  ,  M 2  , and  B 2  requested by the K2PP, M2PP, and B2PP Case Control commands are
direct input matrices  K pp pp pp
2  ,  M 2  , and  B 2  , respectively.
reduced to the d-set to form  K dd dd dd
The general dynamic equation used in the direct methods is
2
  M dd p +  B dd p +  K dd    U d  =  P d  (7-107)

p is the derivative operator

 Ua 
 Ud  =   (7-108)
 Ue 
where:

 U d  = the extra points

For direct frequency response and complex eigenvalue analysis the stiffness, damping, and mass matrices are generated as
follows:
1  + K 2  + iK 4 
 K dd  =  1 + ig   K dd (7-109)
dd dd

1  + B2 
 B dd  =  B dd (7-110)
dd

and
1  + M2 
 M dd  =  M dd (7-111)
dd

For direct transient response the stiffness, damping and mass matrices are generated as follows:
1  + K2 
 K dd  =  K dd (7-112)
dd

g
1  +  B 2x  + ------ 1
1  + ------ 4 
 B dd  =  B dd dd  K dd  K dd (7-113)
 3  4

and
1  + M2 
 M dd  =  M dd (7-114)
dd

where:

Main Index
 Chapter 7: Solution Sequences 389
General Operations

1
  K dd  is the reduced structural stiffness matrix plus the reduced direct input K2GG (symmetric).
2
  K dd  is the reduced direct input matrix K2PP plus the reduced transfer function input (symmetric oar
unsymmetric).
4
  K dd  is the reduced structural damping matrix obtained by multiplying the stiffness matrix  K e  of an
individual structural element by an element damping factor g e and combining the results for all structural
elements (symmetric).
1
  B dd  is the reduced viscous damping matrix plus the reduced direct input B2GG (symmetric).
2
  B dd  is the reduced direct input matrix B2PP plus the reduced transfer function input (symmetric or
unsymmetric).
1
  M dd  is the reduced mass matrix plus the reduced direct input M2GG (symmetric).
2
  M dd  is the reduced direct input matrix M2PP plus the reduced transfer function input (symmetric or
unsymmetric).
 g is the structural damping coefficient on the PARAM Bulk Data entry G. The frequencies  3 and  4 are
specified on the PARAM Bulk Data entries as W3 and W4, respectively.
For direct solutions  K dd  ,  M dd  , and  B dd  are examined to identify rows and columns which are null in all three
matrices. If the parameters ASING = 0 (the default) the singularities caused by the null rows and columns are removed as
follows:

 x  is formed from  K  by placing

For direct frequency and direct transient response, the stiffness matrix  K dd dd
unity on the diagonal for each null row and column.
 For direct complex eigenvalue analysis null rows and columns are discarded from  K dd  ,  M dd  , and  B dd  to
d  ,  M d  , and  B d  . This is performed in subDMAP DCEIGRS. See Complex Eigenvalue Analysis in
form  K xx xx xx
SubDMAP CEIGRS, 423.
If parameter ASING = -1 and null rows and/or columns exist, a fatal error will result.
In modal frequency response, and complex eigenvalue analysis, structural damping and viscous damping are included in
the dynamic matrices as follows:
2  =  K 2x  + ig  K 1  + i  K 4 
 K dd (7-115)
dd dd dd

Main Index
 390 MSC Nastran Reference Guide
General Operations

and
2  =  B 1  +  B 2x 
 B dd (7-116)
dd dd

In modal transient response, the structural damping and viscous damping are included as follows:
2  =  K 2x 
 K dd (7-117)
dd

and

 B dd
g
2  =  B 1  +  B 2x  + ------ 1  + 1.0
 K dd 4 
-------  K dd (7-118)
dd dd  
3 4

The general dynamic equation used in the modal method is:

2
  M hh p +  B hh p +  K hh    U h  =  P h  (7-119)

where p is the derivative operator

 i 
 Uh  =   (7-120)
 Ue 

 i are the modal coordinates.

The transformation from  U h  to  U d  is

 U d  =   dh   U h  (7-121)

where:
 ai 0
  dh  =
0 I ee

The dynamic matrices expressed in modal coordinates are assembled in GKAM (General K Assembler-Modal) as follows:
T 2  
 K hh  =  k  +   dh   K dd dh (7-122)

T 2  
 M hh  =  m  +   dh   M dd dh (7-123)

and
T 2  
 B hh  =  b  +   dh   B dd dh (7-124)

Main Index
 Chapter 7: Solution Sequences 391
General Operations

where, if PARAM,KDAMP = 1 (default)

m i = modal mass
b i = 2f i g  f i m i (7-125)
k i = 4 2 f i2 m i

If PARAM,KDAMP = -1 in complex eigenvalue analysis, modal frequency response, aerodynamic flutter, and aeroelastic
response the viscous modal damping will be incorporated into the complex stiffness matrix as follows:
m i = modal mass
bi = 0 (7-126)
k i =  1 + ig  f i  4 2 f i2 m i

Values of g  f i  are specified on the TABDMP1 Bulk Data entry, selected by the SDAMPING Case Control command. If
KDAMP = 1 (default), the matrices  b  and  k  are formulated as in modal transient response.
In SubDMAP CMPMODE damping of CMS(o-set) modes is calculated by similar techniques. This q-set damping is
expanded to g-set size, and placed in the B aa matrix for summing into downtown superelements.

Data Recovery Operations in SubDMAP SEDISP

The dependent displacements and the single-point forces of constraint are calculated in SDR1 (Stress Data Recovery —
Phase 1). The inputs to SDR1 are columns of the solution vectors corresponding to each eigenvalue, loading condition, or
output time step. In the case of transient analysis, there are three columns, corresponding to displacement, velocity and
acceleration, for each output time step.

Load Modification
If PARAM,INREL,-1 or PARAM,INREL,-2 are specified, then any component of the load vector is first multiplied by
 C rx  to produce a set of loads that are a linear combination of applied and inertial loads called  P li 

 P li    P l   C rx  (7-127)

Static Analysis
In the case of static analysis and PARAM,ALTRED,NO, SDR1 recovers dependent displacements

 ul 
    ua  (7-128)
 ur 

Main Index
 392 MSC Nastran Reference Guide
General Operations

 ua 
    uf  (7-129)
 uo 
where
t u  + u o
 u o  =  G oa (7-130)
a o

If PARAM,ALTRED,YES is specified, then the displacements  u o  are obtained from a backward pass from the equation

T
 L oo   u o  =  u oy  (7-131)

T
where  u oy  =  u ox  –  L ao   u a  (7-132)

Enforced displacements are combined with u f

 uf 
    un  (7-133)
 Ys 

 un 
    ug  (7-134)
 um 
where
 u m  =  G mn   u n  (7-135)

and recovers single point forces of constraint

T
 q s  =  K fs   u f  +  K ss   Y s  –  P s  (7-136)

and multipoint forces of constraint

I mm
q mg =   K mg   u g  –  P m   (7-137)
t
– G mn

In superelement static analysis, if PARAM,FIXEDB,-1 is present, then u a is set to zero, and data recovery is performed
using the modified equation

 u o  =  u oo  (7-138)

Main Index
 Chapter 7: Solution Sequences 393
General Operations

This is equivalent to obtaining a solution with all exterior (“boundary”) degrees of freedom fixed. Displacements are due
only to interior displacements due to interior loads, that is, the  u oo  component.

Eigenvalue Analysis
In the case of eigenvalue analysis SDR1 recovers dependent components of the eigenvectors
  o  =  G oa    a  (7-139)

 a 
 -----  =  f  (7-140)
 o 

 f 
 ----
-  =  n  (7-141)
 s 

  m  =  G mn    n  (7-142)

 n 
 ------  =  g  (7-143)
 m 
and single-point forces of constraint
T
 q s  =  K fs    f  (7-144)

If PARAM,DYNSPCF,NEW (default) then the single-point forces are computed as follows:

 q s  =   K sf  –  2  M sf     f  (7-145)

In superelement analysis, if PARAM,FIXEDB,-1 is present, the uncoupled eigenvectors are recovered from the modified
equations
  a  =   az  (7-146)

  o  =   oz  (7-147)

with   az  and   oz  from Superelement Analysis, 425.

The multipoint forces of constraint qm q are recovered by

Main Index
 394 MSC Nastran Reference Guide
General Operations

I mm 2
qm q =  –  M mg + K mg  (7-148)
T
– G mn

Dynamic Analysis
In the case of dynamic analysis with the direct formulation the extra points are partitioned out of  u d  .

 ua 
 ----
-  =  ud  (7-149)
 ue 
SDR1 recovers the dependent components
 u o  =  G oa   u a  (7-150)

 ua 
 ----
-  =  uf  (7-151)
 uo 
and recovers single point forces of constraint
T
 q s  =  K fs   u f  –  P s  (7-152)

If PARAM,DYNSPCF,NEW (default) then the forces are computed as follows.

In transient response analysis, the formulation is
xu
q s = M sf ·· + B x u· + K x u – P (7-153)
f sf f sf f s

where:

x = M + M 2pp
M sf sf sf

g
x = B x + B 2pp + ------ 1
B sf sf sf K + ------ K 4gg
 sf  sf 3 4

x = K + K 2pp
K sf sf sf

In frequency response and complex eigenvalue analysis, the formulation is

Main Index
 Chapter 7: Solution Sequences 395
General Operations

q s =  –  2 M sf
x + iB x + K x u – P
sf sf f s (7-154)

where:

x
M sf 2pp
= M sf + M sf

x
B sf x + B 2pp
= B sf sf

x
K sf 2pp + iK 4gg
=  1 + ig K sf + K sf sf

The multipoint forces of constraint qm g are recovered

I mm
qm g =  Xm – Pm  (7-155)
T
– G mn

where qm g is the reactions expanded to the g-set and X m and P m are the internal and external loads on the m-set degrees
of freedom, respectively. (See also Eq. 9 in Section 3.5.1 in The NASTRAN Theoretical Manual).
In transient response analysis,
x u
X m = M mg ·· + B x u· + K x u (7-156)
g mg g mg g

where:

x
M mg 2pp
= M mg + M mg

g
2pp + ------ K 1
x
B mg = B mg + B mg + ------ K 4gg
 mg  mg
3 4

x
K mg 2pp
= K mg + K mg

In frequency response and complex eigenvalue analysis,

2
x + iB x + K x u
X m =  –  M mg mg mg g (7-157)

where:

Main Index
 396 MSC Nastran Reference Guide
General Operations

x
M mg 2pp
= M mg + M mg

x
B mg 2pp
= B mg + B mg

x
K mg 2pp + iK 4gg
=  1 + ig K mg + K mg mg

For frequency response there is one complex vector for each loading condition. For complex eigenvalue analysis there is one
complex eigenvector for each complex root and there is no loading term  P s  associated with the determination of the
single point constraints. For transient analysis there is a displacement, velocity, and acceleration associated with each output
time step.
In the case of dynamic analysis with the modal formulation, the user may obtain improved element stresses by requesting
the mode acceleration data recovery technique.
The use of the mode acceleration technique requires the decomposition of the l -set stiffness matrix  K ll  . This
decomposition is not otherwise performed in dynamic analysis unless the model has rigid body modes. Also, the technique
requires a forward/backward substitution operation with a number of right-hand sides equal to the number of solution
vectors. A subset of the solution vectors can be obtained can be obtained by using the OTIME or OFREQUENCY entries
in the Case Control Section. The subset of solution vectors is prepared by the MODACC functional module.
In the case of dynamic analysis with the modal formulation, two methods of data recovery are available. The default
procedure is usually much more efficient for data recovery operations when the modal formulation is used for dynamic
analysis. In this procedure, the dependent components of the eigenvectors used in the modal formulation are first recovered
by SDR1 using the same sequence of operations described for problems in real eigenvalue analysis (after eliminating
unwanted modes). The complete eigenvectors are used as the input to SDR2 in order to determine the requested forces and
stresses in terms of modal coordinates.
In dynamic analysis, the user may restrict the output request to the response of selected points in the solution set. In this
case, the time-consuming operations in the recovery of the dependent degrees of freedom can be avoided. This solution set
processing is performed by VDR (Vector Data Recovery).
The modal data generated by SDR1 and SDR2 are formed into a data matrix by DDRMM (Dynamic Data Recovery
Matrix Method). The output quantities are determined by multiplying the data matrix by the modal solution vectors. The
computer time for this matrix multiplication is proportional to the number of output times, the number of modes used in
the modal formulation, and the number of output frequencies or output time steps. If both the number of output items
and the number of modes are small, the computer time required will not be significant when compared to the total problem
solution time. Although the DDRMM method of data recovery is more efficient, a complete  u g  is not formed. For this
reason the method does not permit the creation of deformed structure plots, grid point force balance or the use of the mode
acceleration technique.

Main Index
 Chapter 7: Solution Sequences 397
General Operations

The alternate procedure may be invoked using PARAM,DDRMM,-1. With this method the solution vectors are first
transformed from modal to physical coordinates
 u d  =   dh   u h  (7-158)

This operation is performed by MPYAD in subDMAP MODACC. Having determined the vectors  u d  in physical
coordinates, the data recovery operations can be completed using the same sequence of operations described for the direct
formulation of dynamics problems.
Most of the operations in the recovery of the dependent components do not require significant amounts of computer time.
However, in the case of the multiply-add operation associated with the recovery of the omitted coordinates
  G oa   u a ,  G oa    a   a significant amount of computer time will be required if there is a large number of vectors
and a large number of components in the a-set. This operation is likely to be especially costly for transient problems because
there is usually a large number of output time steps and there are three vectors for each time step.

Data Recovery Operations in SubDMAP SEDRCVR

The element forces and stresses are calculated in SDR2 (Stress Data Recovery — Phase 2). The matrix operations in this
phase of data recovery are similar to those performed in EMG for stiffness matrix generation. The computer time will be
proportional to the number of elements for which output is requested.
If static loads are applied to intermediate points on the CBAR element, the user may request the output of stresses and forces
at a selected set of intermediate points. The additional calculations required for output at intermediate points on CBAR
elements is performed in functional module SDRX.
The solutions and data recovery operations for transient analysis are initially sorted by the time step (SORT1). A transpose
operation is performed by SDR3 (Stress Data Recovery — Phase 3) in order to prepare the requested output sorted by
external point identification number or external element identification number (SORT2). In frequency response problems,
the user may request either SORT1 or SORT2 output. The SORT2 output will usually be desirable if the solution is made
for a large number of frequencies. In any event, the SORT2 output must be prepared if XY-plots are requested.
Computer time for the preparation of SORT2 output will be directly related to the number of items requested for output.
If a large number of items are requested for output, the transpose operation will require a significant amount of computer
time.
The user may request the output of a grid point force balance table for a selected set of points. This table lists the forces
acting at each selected point due to element forces, single point forces of constraints and applied loads. The grid point force
balance table is prepared by GPFDR (Grid Point Force Data Recovery).
The user may also request the output of an element strain energy table and/or an element kinetic energy table for a selected
set of elements. (See the Case Control commands, EDE (Case), EKE (Case), and ESE (Case) in the MSC Nastran Quick Reference
Guide.) The tables include the energy for each element, the percentage of the total energy for each element, and the energy
density for each element. The energy density is computed by dividing the element energy by the element volume. Some
elements do not have enough data input to allow calculation of element volume, such as CELASi elements. The density
output for these elements is set to zero. Simplified equations for other element volumes may be used. For example, offsets

Main Index
 398 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

are ignored for bar-like elements, and midside node locations are ignored for plate and solid elements. Element volumes are
calculated by the ELPRT module, and can be output by use of PARAM,EST,1. The strain energy table is prepared by
functional module GPFDR.

Summary of Solution Sequence Operations

This section describes the data processing and algorithms that are performed by special purpose functional modules,
sequences of functional modules (subDMAPs) and complete solution sequences.
Detailed descriptions of the functional modules are given in the MSC Nastran DMAP Programmer’s Guide and Chapter 4 of the
MSC Nastran Programmer’s Manual.

Static Solutions
There are several major paths for static solutions and some minor ones. For standard static analysis the stiffness matrix is
decomposed in SubDMAP SEKRRS, in preparation for a linear solution in SubDMAP STATRS. There is also a form of
static analysis called inertia relief analysis.
There are now several paths through inertia relief analysis. This revision of this manual will discuss inertia relief in SOL 101
and a little in SOL 111. There is still another method in SOL 144, Static Aeroelasticity. It is documented elsewhere.

General Formulation
Consider common 3-D unconstrained structure with six 0 Hz rigid body modes.
Rigid body mechanics loads balance (small motion):

 Fi –  mi ai = 0 and   M i + r i  o  F i  –   I i  i + r i  o  m i a i  = 0 (7-159)

In finite element matrix notation:

T T
 R   P  –  R   M   R   u··o  =  0  (7-160)

R is a (a-degree of freedom x 6) geometric rigid body matrix resulting from unit displacements in each basic direction with
respect to GRDPNT or (0,0,0). R provides summation and cross-product utilities for loads and motion at each degree of
freedom i . Rigid body accelerations a and  are represented by 6 x 1 u ·· at PARAM,GRDPNT. All F and M are
i i o i i
entered into load vector P ; m i and I i are entered into a-set mass matrix M . Solve for the rigid body accelerations:

T –1 T
 u··o  =   R   M   R    R   P  (7-161)

T
R MR is the total 6 x 6 a-set mass, nonsingular for normal 3-D models with appropriate mass properties.

Main Index
 Chapter 7: Solution Sequences 399
Summary of Solution Sequence Operations

Apply the balanced loads to the finite element structure in linear statics formulation. This form is employed by the
PARAM,INREL,-2 method:
 K   U  =  P  –  M   R   u··o  (Inrel = -2) (7-162)

In contrast, older method INREL = -1 and SOL 111 employ the following:
·· 
KU = P – MDU (Inrel = -1) (7-163)
r

 K   U  =  P  –  M    r   q··r  (SOL 111 free-free RESVEC’s) (7-164)

Each method uses a different representation for the rigid body matrix and accelerations. Stiffness matrix K is singular (i.e.,
rank l = a – 6 ), and each method likewise employs different techniques to solve for displacement shape U.

Inrel = -2
Inrel = -2 with auto-suport is designed for 3-D models with exactly 6 rigid body modes. The geometric R matrix is generated
about reference GRDPNT, used instead of suport entries. R is a linear combination of normal rigid body modes  r and
can be used to decouple rigid motion and provide a unique inertia relief solution, convergent to the SOL 111 f  0 result.

SOL 111 Basis

In MSC Nastran modal frequency response SOL 111, the elastic solution on a free-free structure converges statically as
follows (using elastic modes notation e):
–1 T
 U e  = lim f  0  U e  f   =   e   k ee    e   P  (7-165)

U e also satisfies equation (7-164). It is obtained in MSC Nastran SOL 111 doing the following:
 Set PARAM,LFREQ or EIGRL f1 to low frequency value (i.e., 0.1 Hz) to exclude use of rigid body modes.
 Remove stiffness damping or load phasing to get real-valued, true static result.
An important observation is that U e is orthogonal or “decoupled” from eigensolution rigid body modes  r . For “clean”
low-strain rigid modes r  6 , geometric rigid body vectors R can be expressed as a linear combination of  t . SOL 111
static U e is likewise decoupled from rigid motion R:

T T
 r   M   Ue  =  0    R   M   Ue  =  0  (7-166)

Main Index
 400 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

SOL 101 Implementation

SOL 101 can use geometric R at low computational cost without having to compute normal rigid modes  r . For INREL
= -2, impose the rigid body decoupling constraint observed in equation (7-166). Re-arrange equations (7-162), (7-161), and
(7-166):

 K   U  +  M   R   u··o  =  P 
T T
 R   M   R   u··o  =  R   P  (7-167)
T
R MU = 0
Add together to obtain:

 
K MR  U   P 
  =   (7-168)
T T ··
R M R MR  u o   RT P 
 
Equation (7-168) is the resultant INREL = -2 formulation. Augmented equation (7-168) matrix is nonsingular for models
with 6 rigid mode singularities in matrix K, because of the addition of 6 acceleration + 6 decoupling equations. For this
scenario the solution is unique and identical to the SOL 111 elastic convergence solution. If K has more than 6 singularities
- local, massless, or true mechanisms - equation (7-168) will become singular. Auto-support removes extra singularities in
the matrix reduction above, but the quality of the inertia relief solution could reduce. Check for high factor diagonal ratios
in the .f06 output. It is recommended that K have only the 6 rigid mode singularities for inertia relief with auto-suport to
work most effectively.

Inrel = -1
Suport DOF (r-set) are commonly used in MSC Nastran to define reference DOF and partition singular stiffness K to
calculate a relative displacement shape U . They are part of structure and solution  U r  , and must be independent DOF
··
in the a-set. In inertia relief, Suport DOF also define the rigid body accelerations of the structure with nonzero U r . INREL
= -1 uses a stiffness/suport-generated rigid body matrix D applying unit displacements at each r-DOF in the flexible body:

–1 D lr
– K ll K lr
D = = (7-169)
I rr I rr

INREL= -1, equation (7-163) becomes:

Main Index
 Chapter 7: Solution Sequences 401
Summary of Solution Sequence Operations

K ll K lr  Ul   P l  M ll M lr D lr ··
 =  –  Ur  (7-170)
K rl K rr  Ur   P r  M rl M rr I rr

Premultiply by D
T ··
and solve for acceleration U r similar to equation (7-161):

T –1
··  =   D   M   D    D   P  T
U r (7-171)

T
Note:  D   K  =  0  is assumed, and valid when the structure is not grounded, and support DOF are a well-
chosen statically determinate set, i.e., at 6 stiff DOF on the structure. D is comprised of strain-free rigid body
vectors.

Solve top portion of equation (7-163) for U i in terms of U r :

–1 ··  
 U l  =  D lr   U r  +  K ll    P l  –  M lr   U (7-172)
r

Setting U r = 0 yields following relative displacement solution U:

Ui –1
U = =  K ll   P l  , ·· 
 P l  =  P l  –   M ll D lr + M lr    U (7-173)
r
Ur 0
A significant property of INREL = -1 is that displacement U solution is not unique and differs with selection of suport
DOF. But deformed shape of U and corresponding internal loads and stresses should not vary with suport sets. For well-
chosen, statically determinate, stiff suport sets 1a and 1b:
1a 1b 1a 1b 1a 1b
 U    U  shape  U  = shape  U    U  =   U  (7-174)
On standard unconstrained 6 rigid-body-mode structure, ideal suport selections are:
 All six degrees of freedom at stiff independent grid at or near CG.
 Six degrees of freedom spread out to 3 distant stiff locations on the structure, in 3-2-1 statically determinate
balance (i.e., support 123 on one corner, 23 and 3 on other corners).

Internal Loads and Accuracy

Inertia relief theory is predicated upon strain-free rigid body vectors available in the finite element model to formulate the
free-free deformed solution:

Main Index
 402 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

  R     D     r    0  (7-175)

When this assumption is valid, the following convergence and accuracy can be obtained:

2 1  111 
  U  =   U  =   Ue  (7-176)
 
INREL = -1 and matrix D tend to be more susceptible to inaccuracies above, because of the vagaries of suport selection,
and the potential for “soft” suports. INREL = -2 tends to be more numerically robust for internal loads as well as
displacement, as long as 6 rigid body modes exist which remain distinct from the 7th mode. Ideally rigid body mode
frequencies should be on the order E-3 or E-4 less than first flexible mode. If difficult to achieve, both methods may be
required to increase confidence in the analyses.

General Discussion
All methods of inertia relief analysis take the standard static analysis equation
K aa  u a = P a (7-177)

where P a are the applied loads, and add a set of constraint equations R ra and automatically-computed inertial loads P i .
The constraints are required because K aa is for a free structure and cannot be solved reliably without some form of
constraint.
K aa  u a = P a – P i = P ai (7-178)

R ra  u a = 0 (7-179)

In theory, if P ai is a set of balanced loads (no resultant) the solution of u a is possible without the need for constraint
equations. In practice, P ai may have some unbalance due to numerical truncation effects. Even a small resultant load will
provide a displacement solution at infinity in theory, or with very large numbers in practice, usually. The same truncation
effects that unbalance the loads may provide enough “ground springs” in the stiffness matrix to allow a solution, but it will
be a poor quality solution. Everything will first move 1.E5 units in the same direction. Stresses will be determined by
subtracting numbers of this size to find element displacement perhaps on the order of 1.E-3. Ironically, if all goes well the
forces in the support system are miniscule, but it must be present to control stability when truncation errors are present.

Types of Inertia Relief

Inertia Relief Analysis is a quasi-dynamic analysis. It is assumed that the structure is not tied to ground and is in a state of
steady acceleration caused by the applied loads. The displacements measured relative to ground are constantly increasing
and cannot be determined. Displacements are therefore determined with respect to an accelerating reference frame that is
attached to the structure. The applied loads are balanced by inertial loads that are calculated automatically from the steady

Main Index
 Chapter 7: Solution Sequences 403
Summary of Solution Sequence Operations

acceleration acting on the mass properties of the model. Inertia relief analysis is requested with PARAM,INREL,[x], where
[x] may have the values -1 or -2.
The reference frame may be obtained by two different methods. For the manual support method (PARAM,INREL,-1) the
user selects 6 or fewer DOFs using SUPORTi entries. The reference frame is defined from these DOFs. For the auto
support method (PARAM,INREL,-2) 6 constraint equations are written that constrain the structure to zero motion, on
average, while leaving it free to “breath” and deflect due to the applied loads and inertial loads. Both methods find it
convenient to have “rigid body modes shapes” that represent a set of vectors that represent all possible strain-free motion of
the structure when not loaded. The technique used to obtain rigid body mode shapes differs between the two methods.

Rigid Body Modes Using Manual Support (PARAM,INREL,-1 and SUPORTi)

The user prescribes a set of DOFs that constrain the structure in a statically-determinate manner on SUPORTi entries.
These entries define the r-set. There must be six or less DOFs listed and they must constrain all rigid body modes that are
not constrained by SPCs or other modeling devices. The rigid body modes are computed from the static analysis equation.
There are no loads applied and I rr is the identity matrix. Given

K ll K lr D O lr
lr
= (7-180)
T I rr O rr
K K
lr rr

D lr is found by solving the first row of this equation, then merged with I rr to form D ar ,

K ll  D lr  – K lr  I rr (7-181)

D lr
 D ar  = (7-182)
I rr

Rigid Body Modes Using Auto-Support (PARAM,INREL,-2)

Vectors of rigid body motion D gr for six unit motions of a reference point are computed from grid point geometry in
module VECPLOT. The D ar component is partitioned from this matrix.

The Basic Equation of Motion for Inertia Relief

Inertia relief analysis starts with the a-set equation
K aa  u a  P ai (7-183)

and adds a constraint equation

Main Index
 404 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

T
R ar  u a = 0 (7-184)

The last equation implies a similar equation on a generalized constraint force variable vector, q r

R ar  q r = P ai – K aa  u a (7-185)

If P ai represents balanced loads, q r is zero. If there is a small unbalance due to numerical truncation effects, q r is then
small, but finite. There are two methods for defining R ar , as discussed below. Both have the property of defining some
part of u a to have zero motion, so that the other parts of u a are in effect measured relative to the constrained DOFs.

On of the solution methods partitions K aa into its r-set and l-set components. The first equation is partitioned, and the
a-set equations and the constraint equation may combine into one equation,

T
K ll K lr R rl ul P ail
T T  u r = P air (7-186)
K lr K rr R rr
R rl R rr O rr –qr Yr

The methods of computing P i are described below. Y r , the generalized displacement input, is set to zero.

Constraint Equations Using Manual Supports

In the constraint equation R rr = I rr , the identity matrix. R rl is null. The last row can then be solved for
u r = Y r = 0 . The second equation can then be used to eliminate q r ,

T
K lr  u l – I rr  q r = P air (7-187)

T
q r = K lr  u l – P air (7-188)

Because R rl is null, the first equation then becomes

K ll  u l = P ail (7-189)

This equation can be solved for u l reduced from u a directly. It will be shown later that P r terms are transferred to P l
terms in the inertial loads. u a is formed from a merge of u l and null terms for u r .

Main Index
 Chapter 7: Solution Sequences 405
Summary of Solution Sequence Operations

Constraint Equations Using Automatic Supports

A constraint equation is written that states that the average weighted motion of the system is zero,
T
R ra = D ar  M aa (7-190)

R ra  u ar = 0 (7-191)

R ra is generally a dense matrix, with nonzero terms at most DOFs with mass. It could be used to reduce the K aa matrix
with equations similar to those used for multi-point constraints, but that would produce a dense K ll matrix, much more
expensive to decompose and solve than the K ll used in the manual support method. A solution method is used which
increases solution costs only slightly.
The r-set l-set partitioning is not needed. The constraint equation is appended to the a-set equations to define the l-set
equations in their augmented form,
K ll  u l = P l (7-192)

T
K aa R ra u a P al
= (7-193)
R ra 0 rr – q r 0r

q r are the unknown forces required on the system to produce equilibrium. They should be computational zeros, and are
output with the SPC forces. The equation involving the augmented K ll is solved directly for u l . u a and q r are then
partitioned from u l . In superelement analysis the constraints are applied in the residual structure only.

Matrix Decomposition in SubDMAP SEKRRS

There are two major branches through this SubDMAP, standard static analysis and static analysis with inertia relief effects.

Inertia Relief Not Selected

When inertia relief is not selected and there are no SUPORTi entries present the a-set is decomposed directly in the DCMP
module,
T
K aa = L aa  D aa  L aa (7-194)

L aa is lower triangular, and D aa is diagonal. Both are stored in an unconventional (“packed”) format, and output in a
single data block named LLL. This name is used because the l-set and the a-set are synonymous when no SUPORTi entries
exist. The factor matrices L aa and D aa are passed on to SubDMAP STATRS for solution of u a .

Main Index
 406 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

When inertia relief is not selected but there are SUPORTi entries present, the a-set equations are partitioned into the r-set,
defined on the SUPORTi entries, and the l-set, what is “left over” after the r-set is removed.

K ll K lr u l P
 l (7-195)
K rl K rr u r Pr

In this context, the r-set is treated as an alternate form of single-point constraints. The partitions of K involving r are
discarded. u r is set to zero. u l is found from the first row of this equation. Loads in P r go directly to ground, and are
output with the SPC forces, along with the product – K rl  u l from the second equation. K ll is decomposed with the same
equations used for conventional static analysis.

Decomposition of K ll

For the manual support method K ll is defined above. Its factors, used to compute the rigid body modes are also used later
to compute the static solution. For the auto support method the factors of the expanded form of K ll are passed on.

Solution in SubDMAP STATRS

Computation of Inertial Loads

The basic assumption of inertia relief is that the structure is in a state of steady acceleration and that all transients due to
dynamic effects are negligible. The solution is obtained in an accelerating reference frame from which all displacements are
measured. Inertial loads are automatically computed that exactly balance the applied loads. A support system is applied to
make the system numerically stable. The inertial loads assume rigid body accelerations when computing the loads.
The derivations given here assume that the system mass matrix M gg has been reduced to M aa and is available, as described
elsewhere. Unit inertial loads are computed from M gg , then reduced by load reduction techniques, avoiding the need for
M aa . M aa can be a large size data block, and costly to compute. It can be shown that the results provided by the two
derivations are identical. For example, the rigid body mass matrix M rr , defined below as

T
M rr = D ar  M aa  D ar (7-196)

can just as easily by formed from the g-set equations

T
M rr = D gr  M gg  D gr (7-197)

D ar is used to compute the rigid body mass M rr about the reference point.

Main Index
 Chapter 7: Solution Sequences 407
Summary of Solution Sequence Operations

T
M rr = D ar  M aa  D ar (7-198)

M rr is printed out with the matrix name of QRR.

The acceleration of the reference point a r due to the applied loads is computed,

T
M rr  a r = D ar  P a (7-199)

The acceleration a r are printed with the matrix name of URA.

The resulting inertial loads P ai and total loads P al are

P i = M aa  a r (7-200)

P ai = P a – P i (7-201)

For both methods, the mass matrix is reduced from M gg size to M rr size with load vector operations that can be shown
to be identical to the M aa operations described here, except for the order of operations. The load and mass transformations
are based on rigid body modes produced from geometry.

A Comparison of Auto- and Manual-Support Solutions

When the same model is analyzed with both methods you should find that stresses, constraint forces, and other internal
quantities are identical except for truncation errors, but the displacements may differ appreciably. This is because they use
a different reference frame. For the manual supports, the reference frame is set up so that the motion at the r-set DOFs are
identically zero. One can think of a reference frame drawn through these points, and all displacements measured relative to
it. This reference frame is moving through space at a constant acceleration rate, so that displacements are relative rather
than inertial displacements. There are many sets of SUPORT i entries valid for one model. If you change SUPORTi DOFs
you may also expect to see changes in displacements, but not in internal loads or element stresses.
For the auto-support option, one can visualize the reference frame as a set of axes connected to every point with mass with
a very complicated set of levers that allow the structure to “breathe” as loads are applied to it, but keep the weighted average
motion to a zero value. If the model happens to have a grid point at the reference point, you will observe very small motion
there, computational zeros but not the binary zeros of the manual support method. This is because this motion is computed
for the auto-support method, but assumed to be zero for the manual method.
The choice of an r-set as defined by SUPORTi DOFs requires meeting both a geometric requirement and a stiffness
requirement. This geometric requirement is that the set should form a statically determinate tie to ground. This can be
difficult to do with complicated geometries. The stiffness requirement is that the DOFs should be stiff enough to provide
a stable solution. A method that has been used by some modelers is to pick points stiff points do not exist. The difficulties
in picking a stable, reliable support system was one of the main motivations for development of the more modern auto-

Main Index
 408 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

support system. SUPORTi entries are not allowed for the auto support method, and will cause a fatal error exit when
present.
PARAM,GRDPNT,GID is optional for both methods when superelements are not present. It is required when
superelements are present, and must be on a boundary point attached to all superelements. This same point should be used
for the SUPORTi DOFs for the manual support method. It defines the reference point used for generating the geometry-
based rigid body vectors. The default for GRDPNT is GID=-1, which is a flag to use the basic origin as the reference point
when computing rigid body modes. Use of a GID near the c.g. of the structure produces more esthetically pleasing results.
The rigid body accelerations produced by the solution are then measured near the center of the structure, rather than at a
point that may be outside the structure.

Applications of Inertia Relief

Although this capability could be used in an environment where the reference frame is accelerating steadily, such as a free-
falling elevator, the more common usage is when the modeler attempts to apply loads in equilibrium. He wishes the net
acceleration of the structure to be zero. A ship floating in the sea, with pressure loads balancing gravity loads, is an example.
Developing balanced load sets can be tedious on large models with complex geometry. The accelerations printed out in the
URA matrix can be used as a measure of load quality. If the accelerations are negligibly small, the load balance is adequate.
If the accelerations are large, some adjustment of the loads sets may be necessary. For good inertia relief analysis a good mass
distribution is needed. As the structural weight is usually only a small proportion of the total weight, this means that a
realistic payload mass distribution must be modeled also.
Another use of inertia relief analysis becoming more prevalent with the advent of auto support is in checkout of a new
model. Although six and only six rigid body modes are provided, they are provided in the form of a “flying waterbed”. To
take an extreme example, suppose that a structure is made up of two components bolted together, but the bolts were not
modeled, inadvertently. This model has twelve rigid body modes before the constraints are applied. Both components sit
on the same waterbed. This means that all singularities will be constrained, and the model will run to completion on the
first pass. If loads are applied to one component only, the constraints will cause the other component to move too, in a
strain-free manner. A plot of the motion of the model should reveal that there is no displacement continuity between the
two components. This should make the cause of the unexpected behavior apparent. As the goal of this analysis is model
checkout only the structure mass distribution is usually adequate for stable solution. The payload masses may be added later.

Real Eigenvalue Analysis in SubDMAPs SEMR3 and MODERS

The basic equation solved in these SubDMAPs is
K –   M   = 0 (7-202)

where K and M are real and symmetric, and M is positive semi-definite. This equation is defined for the a-set variables
when called from MODERS for overall structural modes of vibration, and for o-set variables when called from SEMR3
during CMS (Component Mode Synthesis) analysis. Many steps are taken in DMAP to better automate the removal of
DOFs that are ill-defined, and to augment the mode shapes with residual flexibility effects. The SubDMAPs that perform
these operations are described below.

Main Index
 Chapter 7: Solution Sequences 409
Summary of Solution Sequence Operations

Although a knowledge of DMAP is not necessary for dynamic analysis, such a study can increase the understanding of the
steps that are performed.

Massless Mechanism Identification in SubDMAP MMFIND

SubDMAP MMFIND identifies potential instabilities in the matrices used for real eigensolutions. The K and M matrices
presented to the READ module for real eigensolutions may contain massless mechanisms (MMs). A massless mechanism
has a constraint shape X such that
T
X  K  X = k (7-203)

T
X  M  X = m (7-204)

T
X  X  0.0 (7-205)

If  contains the eigenvectors of the system, then

T
X  = 0 (7-206)

T
X M = 0 (7-207)
The constraint shapes have nonzero terms only where the eigenvectors have zero terms. The massless mechanism shapes
can be thought of as “deficient” eigenvectors describing parts of the system that are unconnected (or very weakly connected)
to other parts of the structure that have both stiffness and mass that are active in the well-conditioned eigenvectors.  k
and  m may be binary zeroes, or computational zeroes due to truncation effects. An “eigenvalue”  associated with
massless mechanism shapes is
 = k  m (7-208)

 is indeterminate in the limit as the  i approach zero. It can be a very large number, a very small number, and either
negative or positive. It is likely to change a large amount when small changes are made to the model. This is the definition
of an unstable eigensolution. The assumption taken here is that the cause of this condition is either stiffness and/or mass
effects inadvertently left out of this part of the model, with the result that the load paths are ill-defined. This load path is
constrained, with a warning message to the user. That is, this mode of motion is later removed from the K and M matrices
prior to eigensolution by steps equivalent to single point constraint elimination.
The massless mechanisms are defined from the shifted stiffness matrix A = K + s  M , where s is a shift that the user may
change from its default value with user parameter PARAM, SHIFT1, (default=1.234 Hz.). The default is chosen to make
detection of rigid body modes as MMs unlikely. A is decomposed into its factors,
A = LDL (7-209)

Main Index
 410 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

The ratios of the diagonal terms of A over D are stored in a vector Ry. Terms of D approach a value of zero at the ends of
massless mechanisms. Terms of Ry larger in magnitude than user parameter MAXRATIO are reset to unity and are kept in
vector R1. It can be shown that each of these terms is at the end of a massless mechanism constraint shape. R1 is expanded
into a set of loading vectors P, with one unit term per column. P is used along with the factors to find the constraint shapes,
Au P (7-210)
The constraint shapes are normalized, and small numbers are discarded. Each column of u is then a potential constraint
mode. It could also be a rigid body mode. The rigid body modes are detected by computing the diagonal terms of the
generalized mass matrix for these shapes,
Mg = u  M  u (7-211)

Terms of Mg  i i  greater than PARAM, MECHFIL, default value = 1.e-6 are an indication that the corresponding
column is a rigid body mode. Columns of this type in u are discarded to produce u1, the matrix of constraint mode shapes.
The location of the end of the constraint shapes is passed back in R. The constraint mode shapes themselves are passed back
in MECHU. R is used later to remove rows and columns in K and M to constrain the massless mechanisms. The nonzero
terms in MECHU are used to indicate which DOFs will discard loads applied to them, because they will have zero
coefficients in the true eigenvectors of the system, leading to a null contribution to generalized force for all modes.
MECHLD contains P, and is intended for a feature not yet implemented.

Massless Mechanism Control in SubDMAP XREAD

At present, the auto-omit DMAP is duplicated at several places, followed by a call to XREAD. It is therefore an implicit
part of the XREAD operations now, and may be moved to XREAD in a later version to reduce the duplication of code.
SubDMAP XREAD modifies the input matrices used for eigensolution, when needed, then restores the modified solutions
back to the original size of the matrices. Solutions for both structural vibration and buckling analyses are provided. This
discussion describes vibration analysis first, with a concluding section on changes made for a buckling solution.

Auto-Omit Operation
All of the eigensolution methods require that each column of  K  plus  M  have some nonnull terms. Any degree of
freedom that has only null terms in both matrices is regarded as undefined and causes a fatal error. The GIV and HOU
methods of eigenvalue extraction also require that the mass matrix be positive-definite. The MGIV, MHOU, and the
Lanczos methods allow null columns in the mass matrix but are generally faster if the massless degrees of freedom are
removed by the static reduction method described in Static Solutions in SubDMAP SEKRRS, 376. However, the static reduction
tends to increase the number of active columns in the a-set matrices. The INV, SINV, and Lanczos methods, by contrast,
are generally slower if the static reduction is performed on massless degrees of freedom, because these methods take
advantage of sparse matrices, whereas the other methods do not.
Three options are provided to deal with the various null column possibilities. If PARAM,ASING,-1 is used, any degrees of
freedom with null columns in both the stiffness and mass matrices (“undefined degrees of freedom”) are identified and a
fatal error exit is taken. This is an option to use to identify modeling errors before performing an expensive eigenvalue
solution.

Main Index
 Chapter 7: Solution Sequences 411
Summary of Solution Sequence Operations

If ASING = 0 (the default) the method used depends on the eigenvalue extraction method. In the tridiagonal methods of
eigenvalue extraction, degrees of freedom with null columns in the mass matrix are identified and placed in the w-set.

M xx M xw
M = (7-212)
M wx M ww

where  M ww  and  M wx  are null by definition. Degrees of freedom in the r-set (SUPORT Bulk Data entry) must have
mass, or a fatal error will result. The stiffness matrix is similarly partitioned.

K xx K xw
K = (7-213)
K wx K ww

with until diagonal terms placed on null columns of  K ww  . The static condensation matrices are formed

 K ww   G wx  = –  K wx  (7-214)

and
T
 K xx  =  K xx  +  K wx   G wx  (7-215)

This method will prevent errors due to mass matrix singularly in the tridiagonal methods if the singularity is caused by null
columns in  M  . It will not detect singularities due to mechanisms (caused, for example, by a point mass offset from a grid
point). This type of singularity does not necessarily prevent obtaining eigensolutions, but will cause poor numerical
stability. The MGIV or MHOU methods do not suffer from numerical instability due to singular or nearly singular mas
matrices. They are therefore more reliable than the GIV HOU methods. In well conditioned problems, however, it is
unnecessary to automatically select the appropriate method. The AHOU method is the most modern, reliable method.
The static reduction will reduce the cost of solution and minimize the number of uninteresting high roots found at the
value of computed infinity due to the singularity of the mass matrix. Note that the static reduction used here does not
introduce any approximation, as the “omitted” that is w-set terms have null mass values. The introduction of unit diagonal
terms on null columns of  K ww  does not introduce approximations because the terms that multiply these diagonal terms
(the corresponding rows of  G ??  are null. The same results can be achieved at higher cost by first eliminating null rows
and columns of  K ww  by the single point constraint operations of Single Point Constraint Operations in SubDMAPs SEKR, SEMR2
and SEMRB, 368 followed by static reduction.

The solution for the system  K xx –   M xx    x = 0 is provided. Rigid body shapes D xr are provided when
SUPORTi entries exist, causing an r-set. The factor of K xx , named LLL, is used for computing approximate solutions for
buckling analysis, to better set the range of eigenvalues to be searched. For the tridiagonal methods, K and M have DOFs

Main Index
 412 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

with zero mass removed by an auto-omit method in a prior subDMAP. For other methods of analysis, K aa and M aa are
used directly for residual structure modes, and K vv and M vv for CMS (Component Mode Synthesis) modes.

The modes of the system are computed in the READ eigensolver module. If the solution appears to be of good quality the
subDMAP returns with no other operations performed. If the modes can not be found, or appear to be of low quality, the
module sets a flag (Neigv parameter). This flag causes the MMFIND subDMAP to be called. Given K and M, it finds a
high-ratio vector R and a matrix of constraint mode shapes MMECHA. These terms are defined in the MMFIND section
above.
There are nonzero terms in R at the end of massless mechanisms. These mechanisms can be removed with SPC-like
operations using R as a partitioning vector,

K xx = K22 K21 (7-216)

K12 K11

M xx is partitioned similarly, to obtain M22. Partitions involving the “1” subscript are discarded. The modified
eigenproblem solved is then  K22 –   M22    2 = 0 .  2 is merged to the size of the original K matrix by inserting
null rows for the “1-set” DOFs,

2
x = (7-217)
01
The auto-omit reduction is backed out by the operations
 w = G wx   x (7-218)

x
 = (7-219)
w

Real Eigenvalue Analysis in the READ Module

Real eigenvalues and their associated eigenvectors are determined in READ (Real Eigenvalue Analysis -- Displacement). In
the case of normal mode analysis, READ extracts real eigenvalues from the equation
 K – M   u  = 0 (7-220)
There are seven basic eigensolution methods. The Inverse (INV and SINV) Method obtains eigensolutions by iterations
based on the equation
 K – i M   i + 1  =  M   i  (7-221)

Main Index
 Chapter 7: Solution Sequences 413
Summary of Solution Sequence Operations

where  i is an estimate of the eigenvalue. This method is best suited to large problems with sparse matrices, where only a
few eigenvectors are desired.
The tridiagonal methods, Givens (GIV and MGIV) and Householder (HOU and MHOU), first mass-scale the equation
T
M = LL (7-222)

–1 –1 T
where  L  is a lower-triangular matrix. The  K  and  M  matrices are pre- and post-multiplied by  L  and  L  ,
respectively, to obtain the standard form

 J – I     = 0 (7-223)

–1 –1 T
where  J  =  L   K   L  and  I  is the identity matrix. Note that the mass matrix is in effect inverted, so that it
must be nonsingular. The  J  matrix is transformed to tridiagonal form for economical computation of eigenvalues and
eigenvectors. The eigenvectors in the tridiagonal basis are then back-transformed to the initial basis. The Modified Givens
(MGIV) and Modified Householder (MHOU) Methods use the shifted matrix
 K + s M  (7-224)

in place of  M  in the discussion above, thereby allowing the mass matrix to be singular if there are compensating terms
in the stiffness matrix. The GIV, MGIV, and MHOU methods are best suited to small problems, or for large problems
where many eigenvectors are required, after extensive static reduction or dynamic reduction is performed.
The Lanczos method uses a block shifted algorithm. Sets of vectors obtained by a recursive form are used to reduce the
problem to a reduced block tridiagonal form. The eigensolutions are computed in the reduced basis by a QL algorithm,
then back-transformed to the original basis. This is currently the most modern method, and should be considered for all
large-size problems. Since the method takes advantage of sparsity in the input matrices, it is most economical when used
without static or dynamic reduction. It is requested with either the EIGRL or EIGR Bulk Data entries.
The AHOU method combines the best features of all of the tridiagonal methods, and is the recommended method at
present. It first decomposes M. If the factors are stable, the HOU method is used. If the factors may be unstable, the
MHOU method is used instead.

Modal Solution Preparation in the MODERS subDMAP

MODERS is called when solving for the eigensolution of the residual structure, with a-set operations. It is used for both
structural mode and buckling analysis.
 In buckling, the r-set caused by SUPORTi entries are treated as equivalent SPCs. The rows and columns
associated with r-set DOFs are discarded before eigensolution.
 For structural analysis, the auto-omit operations take place for the tridiagonal methods. See the XREAD Section
for details.

Main Index
 414 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

There are also operations for ACMS and indexing of output tables present in MODERS. They can also be found in many
other SubDMAPs. Their description is beyond the scope of this manual, but briefly, ACMS (Automated Component Mode
Synthesis) is based on breaking a large model into superelements automatically, solving it in pieces, then re-assembling the
pieces so that the output appears to be from a conventional model. Some of the partitioning burden that was placed on the
user in manual superelement analysis proved difficult to automate, so more automatic alternate methods have been
developed for ACMS that require less user input. The automatic support system described in the RESVEC SubDMAP
section is a case in point. Most of the ACMS innovations are moved to conventional analysis, over time, after they have
been tested in field usage.
Table indexing is a method to allow direct access to data in large sequential files, to speed up interactive analysis. This work
is being done as part of the API project.

Calculation of Residual Vectors in MSC Nastran

Modal solutions are commonly used to reduce large-size models to relatively smaller representations that include the
important dynamics of the original model. In addition to reducing the model size, modal reduction has the added advantage
of producing diagonal mass and stiffness matrices for solution calculation. Although modal reduction may capture most of
the dynamic response of the structure, the static response may not be complete. This ‘error’ is due to modal truncation of
higher frequency modes that may contribute ‘statically’ to the total response. Residual vectors are used to improve the results
of modal solutions by attempting to account for the response of these higher frequency modes.
Residual vectors can be determined from any set of base vectors. Base vectors are the raw material used to generate the
residual vectors in MSC Nastran. Although any vector can be used as a residual vector, as long as it is partially independent
of the modal vectors, it may produce coupling with the modal vectors when added to the modes for matrix reduction. This
coupling destroys the diagonal properties of the reduced mass and stiffness matrices.
To ensure that the residual vector used in the reduction process result in diagonal matrices, and better results, the following
steps are performed.
1. Ensure that loads are linearly independent with the modal inertial forces.
2. Determine base vectors from static response due to loads.
3. Ensure that base vectors are linearly independent.
4. Orthogonalize the base vectors with respect to the modal vectors to produce residual vectors. These vectors will
result in diagonal mass and stiffness matrices.
The closer the base vectors approximate the actual deformation due to the load, the better the residual vectors and the better
the results. For this reason, base vectors in MSC Nastran are determined from the static response of the structure to the
following loads:
 Inertial forces due to rigid-body motion
 Applied loads
 Structural, viscous, and inertial forces due to enforced motion
 Forces at user specified discrete degrees of freedom (RVDOFi entries)
 Discrete damping forces due to viscous elements (CDAMPi and CVISC entries)

Main Index
 Chapter 7: Solution Sequences 415
Summary of Solution Sequence Operations

A new case control command, RESVEC, allows the user to specify the loads used for residual vector calculation. The format
for the RESVEC is as follows:

 SYSTEM/NOSYSTEM 
 
INRLOD APPLOD RVDOF DMPLOD  COMPONENT/NOCOMPONENT 
RESVEC (    ) = 
NOINRL NOAPPL NORVDO NODMP  BOTH or YES 
 NO 
 

Examples:

RESVEC=SYSTEM
RESVEC(NOINRL)=COMPONENT
RESVEC=NO

Describer Meaning
INRLOD/ Controls calculation of residual vectors based on inertia relief. (Default=INRLOD)
NOINRL
APPLOD/ Controls calculation of residual vectors based on applied loads.
NOAPPL (Default=APPLOD)
RVDOF/ Controls calculation of residual vectors based on RVDOFi entries
NORVDOF (Default=RVDOF)
DMPLOD/ Controls calculation of residual vectors based on viscous damping.
NODMP
SYSTEM/ Controls calculation of residual vectors for system (a-set) modes. For NOSYSTEM,
NOSYSTEM describers inside the parentheses are ignored. See Remark 2 for default.
COMPONENT/ Controls calculation of residual vectors for component (superelement or o-set)
NOCOMPONENT modes. For NOCOMPONENT, describers inside the parentheses are ignored. See
Remark 2 for default.
BOTH or YES Requests calculation of residual vectors for both system modes and component
modes. See Remark 2 for default.
NO Turns off calculation of residual vectors for both system and component modes and
describers inside the parentheses are ignored. See Remark 2 for default.

The inertial forces due to the normal mode shapes are removed from the loads used for residual vector calculation using the
following:
–1
P = P – MM P (7-225)
The modified loads are applied to the structure to determine the deformations that serve as the base vectors for residual
vector calculation.

Main Index
 416 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

 K   V load  =  P  (7-226)

Additionally, for transient response, initial conditions are also added to the set of base vectors. Before the initial condition
vectors are added, any mode shape content is removed.
–1
 V initial  =  V initial  –         M          M   V initial  (7-227)

The initial condition vectors are appended to the load vectors and the complete set of base vectors are made linearly
independent. This process removes vectors that are linear combinations of other vectors in the base vector set. The
remaining base vectors are then made orthogonal with respect to the mass and stiffness matrices. The resulting orthogonal
vectors are the residual vectors.
The residual vectors are appended to the modes and the reduced mass and stiffness matrices are modified to include the
residual vector additions.

   =   modes  residual  (7-228)

M = M (7-229)

K = K (7-230)

The augmented set of modes/residual vectors is used to reduce the damping, applied loads, and any external matrix input.
Residual vector calculation is available for all modal solutions in MSC Nastran. They can be calculated for the residual
structure and superelements to improve the analysis results.

New C-Set and R-Set Reduction with Residual Vectors

In previous version of MSC Nastran, the o-set (interior dofs) motion associated with the b-set, c-set, and r-set (boundary
dofs) was generated from the static reduction of the stiffness matrix.
–1
 G ot  =  G ob G ow  =  – K oo   K ob K ow  (7-231)

where the w-set = c-set + r-set.

The boundary motion determines the response of the o-set.

 ub 
 u 0  =  G ob G ow    (7-232)
 uw 
For dynamic analyses, the o-set motion can be augmented by the use of component modes to improve the response
estimate. The motion of the component modes is represented by the q-set.

Main Index
 Chapter 7: Solution Sequences 417
Summary of Solution Sequence Operations

 ub 
 
 u 0  =  G ob G ow  oq   u w  (7-233)
 
 q 

where the component modes,  oq , are determined from modes of the v-set (v-set = o-set + c-set + r-set).

The modes of the v-set are calculated in the standard fashion.

 –   M vv  +  K vv     v  =  0  (7-234)

The motion of the boundary grids are then removed from the calculated modes

  o  =   o  –  G ow    w  (7-235)

where

 o 
 v  =   (7-236)
 w 

  v  are modes of the v-set degrees of freedom.

  o  are the v-set modes with the w-set boundary motion removed.

The previous operation decouples the w-set boundary motion  G ow  from the v-set modes   v  . This decoupling allows
the boundary and component mode dofs to move independently and the total motion of the structure to be written in terms
of the a-set (b-set + w-set + q-set).

 ub  I 0 0  b
u
   
u
 w = 0 I 0  uw  (7-237)
   
 uo  G ob G ow  oq   q 

If residual vectors are requested, they are calculated using the o-set matrices and appended to the component modes.
The new w-set processing follows a slightly different path. The procedure is based on a method suggested by Dr. Arya Majed
and Ed Henkel. This method is the new default procedure for c-set reduction. The previous method for c-set reduction can
be selected by setting PARAM,MHRED to ‘NO’.
The new c-set procedure uses static reduction to determine the motion oft the v-set due to boundary b-set motion.
 u v  =  G vb   u b  (7-238)

Main Index
 418 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

where
–1
 G vb  =  – K vv   K vb  (7-239)

As previously was done, modes of the v-set are also calculated.

 –   M vv  +  K vv     v  =  0  (7-240)

If residual vectors are requested, they are calculated using the v-set size matrices and appended to the v-set modes.
At this point, the b-set boundary vectors and the v-set modes can describe the motion of the structure.

 ub  I 0  ub 
  =   (7-241)
 uv  G vb  vi   q 

or splitting the v-set into its o-set and w-set partitions,

 ub  I 0
   ub 
 u w  = G wb  wq   (7-242)
   q 
 uo  G ob  oq

In the above representation, the w-set dofs are not independent dofs, but are dependent on b-set and q-set motion. In order
to make the w-set independent, vectors that associate o-set motion with individual w-set dofs must be found. Once these
vectors are found, the components are removed from the b-set and q-set motion resulting in w-set vectors that are
independent.
The w-set vectors are determined using the flexibility of the v-set.

Y oo Y ow  P o   uo 
  =   (7-243)
Y wo Y ww  P w   uw 

If loads are applied only to the w-set, the displacements are dependent only on P w .

Y ow  uo 
 Pw  =   (7-244)
Y ww  uw 

–1
Vectors that associate o-set motion with individual w-set dofs can be found by setting P w = Y ww .

Main Index
 Chapter 7: Solution Sequences 419
Summary of Solution Sequence Operations

Y ow –1
–1 Y ow Y ww
 Y ww  = (7-245)
Y ww I ww

The above vectors are used to make the w-set an independent set.

 ub  I 0 0  ub 
   
 uw  = 0 I 0  uw  (7-246)
  –1 –1 –1  
 uo  G ob – Y ow Y ww G wb Y ow Y ww  oq – Y ow Y ww  wq  q 

The flexibility matrices Y are determined from displacements due to unit forces on the w-set.

–1 0
 K vv  =  Y vv  (7-247)
I ww

The modal flexibility is removed from the displacement response to determine the flexibility matrices used in (7-247)(7-246).

T –1 T
 Y vv  =  Y vv  –   vq     vq   M vv    vq     vq   M vv   Y vv  (7-248)

Y ow
 Y vv  = (7-249)
Y ww

Static Solution for Shape Functions

If the structure contains rigid body modes it is not stable for static analysis. This problem is solved by one of two methods:

Manual SUPORT (SubDMAP SEKRRS)

If the user supplies SUPORTi entries, the r-set DOFs it produces are used as “temporary” SPCs, applied to allow solving
for the static shapes, but removed before computing modes. Partitions involving “r” are discarded.

K ll K lr
K aa  (7-250)
K rl K rr

P l2
P a2  (7-251)
P r2

Main Index
 420 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

K ll  ul  P l2 (7-252)

u l2
u a2 = (7-253)
0 r2

Auto SUPORT (SubDMAPs MODEFSRS, SEMR3) ACMS Only

If no SUPORTi entries or an inadequate set are supplied for constraining all rigid body modes, and rigid body modes are
found when solving  K aa –   M aa    a = 0 , the rigid body modes  r are separated from the flexible modes based
on their natural frequency being smaller than a user parameter, PARAM, FZERO, (default = 1.E-4 Hz).  r is used to make
constraint equations MP ar ,

MP ar = M aa   r (7-254)

These equations are appended to K aa using Lagrange Multiplier Techniques, and solved directly,

K aa MP ar u a2 P a2
= (7-255)
T – q r2 O r2
MP ar O rr

The matrix on the left is named KLL in the subDMAP. It is similar to the KLL produced above by partitioning the larger
K aa matrix in that it is stable for static solution, but differs in that it is larger than K aa by the addition of the constraint
equations. These equations enforce the condition expressed by the bottom row,
T
MP ar  u a2 = 0 (7-256)

This is a statement that the c.g. of the structure is at rest on average, but the structure is free to deflect about it. This
approach will produce different u a2 vectors than those produced by the partitioned solution. In the partitioned solution
one set of points is grounded, and all loads are transmitted through the structure to them. A load on an explicit r-set point,
for example, will go directly to ground, and will cause zero deflection. This implies that some skill is required in selecting
DOFs for SUPORTi entries. With the LMT approach the constraints to ground are distributed, so that loads applied
anywhere will tend to load the entire structure.

Solution for Trial Vectors by Static Analysis

Inertia relief analysis is NOT performed, despite the presence of SUPORTi entries, usually used in the context of inertia
relief. In this context, SUPORTi data or its equivalent for the auto-support feature is used for SPC-type operations. In CMS

Main Index
 Chapter 7: Solution Sequences 421
Summary of Solution Sequence Operations

at present, K oo may not contain rigid body modes. There is no provision to constrain them with SUPORT-type inputs.
DOFs may be removed from the o-set to make it stable. For example, in superelement analysis, an interior point active in
the rigid body mode, which by default is in the o-set, can be made an exterior point. This changes it to the a-set, which is
held fixed when computing CMS modes. If the rigid body modes still exist downstream, they can be constrained in residual
structure calculations.
The u a2 trial vectors, or “raw static shapes”, as they are called in the SubDMAP comments, may not be linearly
independent of the mode shapes, and are certainly not orthogonal to the mode shapes or each other. An orthogonalization
process is now used to insure that the resulting shapes are orthogonal to the mode shapes and each other. Shapes that have
little information after this process are discarded.

Orthogonalization
The trial vectors u a2 may not be linearly independent of each other or the mode shapes. They are processed to find a set
of combined mode shapes and linear combinations of the trial shapes that are all orthogonal to each other with respect to
the mass matrix. This reduces the risk of ill-conditioning when using these matrices for solution later.
The trial vectors are appended to  a to form the combined set  ua1

 ua1 =  a u a1 (7-257)

An optional pre-sweeping of trial vectors is done when requested by PARAM, RESVSO, YES. The derivation of the
sweeping equation is similar to that for sweeping loads above, with the equation for x , the vector of linear combination
coefficients, being
x = M i – 1   a  M aa  u a2s (7-258)

The swept displacement shape vectors  ua1 are then determined by eliminating x ,

–1
 ua1 = u a2s –  a2  M i   a2  M aa  u a2s (7-259)

If the optional sweeping is not selected  a1 is merely equal to  a2 .

The u a1 vectors are mass-normalized into  ua such that the diagonal terms of the product on the left side are unity

 ua  M aa   ua = M zz (7-260)

diag  M zz  = I zz (7-261)

M zz is generally fully coupled at this point. It is decomposed with an LTL factorization,

Main Index
 422 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

M zz = L  D  L , where L is lower triangular and D is diagonal.

The ratio of D over the diagonal terms of M zz is calculated in the vector R ,

 diag  M zz  
R =  ------------------------  (7-262)
 diag  D  

Rows of R whose value exceeds PARAM, RESVRAT (default=1.e8) indicate columns of  ua that have little new content,
and are discarded to produce U a3 and M zz3 . M zz3 is diagonal in the region of the eigenvectors, but coupled in the
region of the auxiliary shapes. It is then re-named as M zz .

The generalized stiffness corresponding to M zz is calculated,

K zz = u a3  K aa  u a3 (7-263)

The eigensolution of the z-set is obtained,

 K zz –  z  M zz    z = 0 (7-264)

The eigenvectors  z of this reduced basis are used as a transformation to orthogonalize U a3 ,

 a =  ua   z (7-265)

The eigenvalue table of the combined system,  z , replace the prior eigenvalue table that contained modes only.  z includes
both structural modes and the effects of orthogonalized residual flexibility vectors.

Model Partitioning in the MODEFSRS subDMAP

MODEFSRS is used in the residual structure to separate the fluid and structural components of the a-set. The modes of
these components are computed independent of one another to find  sa and  fa , the structural and fluid modes,
respectively. Once computed, residual flexibility effects are added with calls to the RESVEC SubDMAP.
The eigenvectors with residual flexibility effects appended to them  a are then merged together to produce one eigenvector
matrix,

 sa 0
a = (7-266)
0  fa

Main Index
 Chapter 7: Solution Sequences 423
Summary of Solution Sequence Operations

For the cases where only fluid or only structural modes are requested, the identity matrix is substituted for the eigenvector
matrix that is not calculated.

Complex Eigenvalue Analysis in SubDMAP CEIGRS

Complex eigenvalues and the associated eigenvectors are calculated in CEAD (Complex Eigenvalue Analysis --
Displacement) using the inverse power method, the determinant method, the Hessenberg method, or the Lanczos method,
as requested by the user on the EIGC Bulk Data entry. For direct complex eigenvalue analysis and if PARAM,ASING = 0
d  ,  M d  , and
(default) then null rows and columns are discarded from  K dd  ,  M dd  , and  B dd  to form  K xx xx
d .
 B xx
In the case of a direct formulation CEAD extracts the eigenvalues from the following equation.
d p2 + B d p + K d  u d  = 0
 M xx (7-267)
xx xx x

 u xd  is then augmented with null rows to form  u d  .

In the case of a modal formulation the following equation is used:

 M hh p 2 + B hh p + K hh   u h  = 0 (7-268)

CEAD also normalizes the eigenvectors according to one of the following user requests:
 Unit magnitude of a selected coordinate (“POINT”).
 Unit magnitude of the largest component (“MAX”).
as specified on the EIGC Bulk Data entry.
The complex Lanczos method is recommended for large problems with sparse matrices. The HESS method is best suited
to small dense matrices, typical of small models or those typical in a modal formulation. The INV method is obsolete, and
is retained as an alternate method when all else fails.

Frequency Response and Random Analysis in SubDMAP FREQRS

The frequency-dependent loads are generated in FRLG (Frequency Response Load Generation) from the RLOAD1 and
RLOAD2 Bulk Data entries. Constraints are applied to reduce the loads to the d-set, and in the case of a modal formulation,
the loads are transformed to the h-set.
The solution for the response is performed in FRRD1 or FRRD2 (Frequency Response Solution). In the case of a direct
formulation the following equation is solved.

 – M dd  2 + iB dd  + K dd   u d  =  P d  (7-269)

and in the case of a modal formulation, the following equation is used in:

Main Index
 424 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

 – M hh  2 + iB hh  + K hh   u h  =  P h  (7-270)

The arithmetic used in the solution may be real or complex and the solution procedure may be symmetric or unsymmetric.
The choice of arithmetic and solution procedure is made by the program depending on the form of the dynamic matrices.
Power spectral density functions and autocorrelation functions are computed in RANDOM (Random Analysis) according
to the information on RANDPS and RANDT1 Bulk Data entries. Random analysis calculations are made for selected
displacements, loads, single point constraint forces, element stresses and element forces. The selection of items for random
analysis is made in the XY-output request part of the Case Control Section.

Transient Response in SubDMAPs SEDTRAN and SEMTRAN

The linear time-dependent loads are generated in TRLG (Transient Response Load Generation) from the TLOAD1 and
TLOAD2 Bulk Data entries. The nonlinear loads are generated from the NOLINi Bulk Data entries. Constraints are
applied to reduce the loads to the d-set, and in the case of a modal formulation, the loads are transformed to the h-set.
The solution for the response is performed in TRD1 (Transient Response Solution). In the case of a direct formulation the
following equation is integrated over the time periods specified via the TSTEP Bulk Data entry in subDMAP DTRANRS:

 M dd p 2 + B dd p + K dd
x   u  =  P  +  P nl 
d d d (7-271)

in order to determine the displacement velocity and acceleration response.

In the case of a modal formulation, the following equation is used in subDMAP MTRANRS:

 M hh p 2 + B hh p + K hh   u h  =  P h  +  P hnl  (7-272)

If there are no direct input matrices and no nonlinear loads, an analytical integration procedure is used rather than
numerical integration. The numerical integration procedure may use either symmetric or unsymmetric routines. The
choice is made by the program depending on the form of the dynamic matrices.

Buckling Analysis in SubDMAP MODERS

In linear buckling, a static load is applied to the structure and the following eigenvalue problem is solved subsequently,
d   = 0
 K aa +  i K aa (7-273)
i

d  is the differential stiffness,  are the eigenvalues and    are the

where  K aa  is the linear stiffness,  K aa i i
eigenvectors (buckling modes). The critical buckling loads are

 P icrit  =  i  P a  (7-274)

where  P a  is the vector of applied static loads including reaction forces.

Main Index
 Chapter 7: Solution Sequences 425
Summary of Solution Sequence Operations

d  is proportional to the applied load  P  .

The basic assumption in linear buckling is that the differential stiffness  K aa a
The assumption implies a linear material law and small deformations. For buckling in nonlinear problems see Additional
Topics, 511.

d  is derived from the

The linear stiffness  K aa  is derived from the variation of the stresses, the differential stiffness  K aa
variation of the strains keeping the stresses constant.
The Case Control Section must contain at least two subcases. A static loading condition is applied in the first subcase. The
load is defined with a LOAD, TEMP(LOAD), or DEFORM Case Control command. The load may also be defined with
enforced deformations on the SPC Bulk Data entries. In the second subcase, a METHOD command must appear to solve
the eigenvalue problem. The Lanczos method (EIGR Bulk Data entry) is the recommended eigenvalue solver for linear
buckling.
Linear buckling is available in the following solution sequences.

SOL 105 Linear buckling in statics with superelements, including original design sensitivity (DSA).

The linear buckling capability is available for the following elements: CONROD, CROD, CTUBE, CBAR, CBEAM,
CBEND, CQUAD4, CTRIA3, CTRIA6, CSHEAR, CHEXA, CPENTA, AND CTETRA.
The following recommendations apply to linear buckling analysis.
 Use at least five elements per half sine wave to get reasonable results in the buckling load.
 For buckling of 3-D shell structures, use PARAM,K6ROT,100. to assign a stiffness to the sixth degree of freedom.
The default is PARAM,K6ROT,0.
The following restrictions apply to linear buckling analysis.
 Offsets should not be used in beam, plate or shell elements. The buckling loads for structures with offsets are
incorrect.
 Follower force effects are not included in the differential stiffness. The calculated buckling loads are incorrect. The
Bulk Data entries FORCE1, FORCE2, MOMENT1, MOMENT2, PLOAD, PLOAD2, PLOAD4, RFORCE, TEMP,
TEMPD, TEMPP1, TEMPP3, TEMPRB describe follower forces.

Superelement Analysis
The theoretical aspects of superelement analysis are discussed in the MSC Nastran Superelements and Modules User’s Guide.
Superelements are mathematically equivalent to substructures. The major difference is in a more convenient interface for
the user and in the ability to process a much broader class of problems than is practical for most other substructuring
capabilities.
There are now two distinct paths through superelements in the solution sequences. There are also two major types of
superelements. When a BEGIN SUPER entry is present the program uses the more modern SEP1X module to make the
SEMAP table used to control partition of superelements. When there is a BEGIN BULK entry but no BEGIN SUPER
entries a parallel path using the older SEP1 module is used instead. While the newer module supports most features of both

Main Index
 426 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

part and grid list superelements, it does not yet have some of the advanced features supported in the older module. When
starting a new project you should decide at modeling initiation which path is most suitable because transitioning from one
to another can be painful, particularly when the model is in development. If one starts on the old path, for example, and
then adds a new feature that causes the program to automatically switch to the new path, some of your modeling input may
be invalidated. This may cause considerable, confusing, and unnecessary work to transition to the new path.
The short answer is that the new path should be the best path for most projects, based on the appeal of its advanced features
and the likelihood that future improvements in MSC Nastran may require use of the new path. The new path is forced by
replacing the BEGIN BULK entry with a BEGIN SUPER entry in your Main Bulk Data file. It should by the first entry
in your bulk data. It allows grid list superelements, but they need not be present.
The exception for the path chosen is when your project requires the use of features available in the old path but not yet
enabled in the new path. Some of these features may be used in the new paths by special avoidance techniques, but these
techniques can be cumbersome. It may be more convenient to stay with the older methods until the newer methods catch
up.

Terminology
 Newest Option: The partitioned bulk data file superelements (parts) are the form introduced in Version 69. They
are sometimes called substructures. A major feature of parts is ID independence. That is, element IDs and grid
points IDs may be duplicated across different parts, although they must be unique within a part.
 Intermediate: The Main Bulk Data Superelements (superelements) introduced in Version 47 are sometimes called
grid list superelements (list s.e.s). When used in the new path, they have more restrictive rules than in the old path.
 Oldest Path: List s.e.s only are used in the old path. There are no formal names at present to differentiate the
intermediate and old flavors. Some advanced features are available only in this path, at present.
A subtle difference between the two paths is that the new path will always use a single level configuration for list s.e.s if there
is no SETREE-type entry. It does this by defining new boundary points in the residual structure when some are needed but
not supplied by the user. The old path automatically converts to a multilevel configuration when required, regardless of the
presence of DTI, SETREE entries. It has no ability to generate new boundary points. This type of change when switching
from the old path to the new, on purpose or by accident can be confusing.

Criteria for Selecting Part Versus List Superelements

Each of the two major forms has overwhelming advantages in certain contexts. Some of these advantages are of a temporary
nature, as the rest of the technology in supporting FE tools catch up to the newest methods. However, other relative
advantages are due to the inherent difference in concept between the two major branches that will be relevant into the
foreseeable future. It is likely that both branches will be maintained and extended because of their unique capabilities. The
following criteria should be considered when making a choice:
 Converting an existing non-superelement model to a part configuration is a labor-intensive process at present. No
migration aides are provided. Converting it to a list superelement requires addition of only SESET entries, a trivial
task by comparison.

Main Index
 Chapter 7: Solution Sequences 427
Summary of Solution Sequence Operations

 When debugging a new model it is much easier to comprehend and more efficient when it is in a superelement
configuration. Errors in one area will not obscure errors in other areas, and the model solved is smaller, leading to
faster turnaround. It may be more economical to run it in a non-superelement configuration for production runs
after the model is stable and requires no more changes. If it is a list superelement this means merely removing the
SESET entries, after which it is automatically an exactly equivalent non-superelement model if advanced features
such as those provided by the RELEASE and CSUPEXT entries are avoided. If it is a part superelement there can
be a large labor cost for conversation, as all duplicated IDs will need to be changed, duplicate points at boundaries
will need to be resolved, etc. A list s.e. model, to make virtue of necessity, enforces rules such as unique IDs when
updating the model that make re-conversion to a non-s.e. configuration an almost automatic process.
 When starting a new modeling project where more than one person provides input, and most particularly when
more than one organization contributes to the model, the higher degree of data control for part superelements is
attractive. There is no need to legislate acceptable ID ranges for each part. It is possible to lock the description of a
part into its boundary matrices, in a data base marked with a time stamp, so that there is little questions about who
is responsible for the behavior of each part.
 Many GUIs can be used to modify list superelements because the rules for ID independence are the same for GUIs
as they are for list s.e.s Part superelements can be modeled in GUIs only one at a time. This makes boundary
matching, in particular, cumbersome.

Features Not Supported in the New Path at Present

 Fluid Structure Analysis: This capability is supported in the new path with the “old” path rules:
a. All fluid elements must be in a special list superelement, and/or in the residual structure
b. This special superelement may have fluid elements only; structural elements or grid points are not allowed
c. One or several fluid cavities may be in the fluid superelement
d. Fluid elements may be connected to structural elements in list superelements, or in the residual structure.
e. Elements in part superelements may not be connected to fluid elements.
 RELEASE Entries: Allowed only in the old path. There are semi-heroic modeling techniques that may be used to
simulate this feature in the new path, but they are not documented.
 SExSET Entries: Most set entries for non-superelement models such as ASET have synonyms for use in list s.e.s
such as SEASET. The synonym is not accepted in PARTs, but the simpler format of ASET, for example, is
accepted.

Features Not Supported in the Old Path

In general, features associated with automatic boundary control are not supported. The QRG is not explicit on this issue.
A list of entries not supported in the old path is:
SETREE, SENQSET, SEBNDRY, SEBULK, SECONCT, SEEXCLD, SELABLE, SELOC, SEMPLN
The DTI, SETREE entry is used in place of SETREE. The action of the SEELT entry is somewhat different. For the new
path, the SEMAP will be changed to account for the action of SEELT entries. For the old path, it is assumed that the
element being moved is one that can be moved without changing the SEMAP, such as an element attached only to boundary

Main Index
 428 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

points. If the SEELT requires change of the SEMAP the program processes conventionally until it encounters the s.e. with
the “illegal” SEELT entry, then gives a fatal error exit stating that the connectivity of the superelement is incorrect.

Check List for Transitioning an Old Model to the New Path

 Replace the BEGIN BULK entry with a BEGIN SUPER entry.
 Check the SEMAP table from an old run to determine if it is in a multilevel state. When it is, add SETREE entries
that define the tree in the old run, or convert it to a single level model if multilevel is not essential. If you have
many multilevel models to convert, there is a parameter that will request that SETREE entries be made from the
SEMAP, and placed in the.pch (punch) file.
 Discard all RELEASE and CSUPEXT entries. If you cannot find reasonable replacements for them you may be
forced to stay in the old path.
Decide if this is the right time to convert to a PART model, or whether the list s.e. capability in the new path is more
advantageous, based on the discussion above.
Solution sequences that support superelement analysis are listed in Output Description, 339.
The Case Control commands for superelement processing selection all begin with the letters SE (SEMG, SELG, SEKR,
SEMR, SELR, SEALL, and SEEX). The SEFINAL command is used to control the order of processing of the
superelements. A subcase structure is required that defines constraints, loads, and data recovery options for each
superelement. Each subcase includes the Case Control command SUPER in order to relate the subcase to a particular
superelement and load condition. Examples of Case Control (including plot requests) are given in the MSC Nastran
Superelements and Modules User’s Guide. A set of plotting commands is required for each superelement to be plotted, separated
by the plotter commands SEPLOT and SEUPPLOT.
Each superelement is given an identification number (SEID). The user assigns each grid point to a superelement on a GRID
or SESET Bulk Data entry. Identical image, mirror image, and external superelements are defined by the CSUPER and
optional SEQSEP Bulk Data entries. The SEELT Bulk Data entry may be used to change the superelement membership
of boundary elements. The RELEASE Bulk Data entry may be used to disconnect selected degrees of freedom from exterior
grid points of a superelement. Use of Parameters, 31 lists all PARAMeters used in superelement analysis.
The mesh generator option, called by MESH on the NASTRAN statement, supports the superelement capability fully, in
that all GRID entries generated for a field can be given the same SEID, and the SESET Bulk Data entry can change these
SEIDs at boundaries. The automatic grid point resequencing features are fully implemented for superelement models.
The superelement partitioning module passes on resequencing data from SEQGP Bulk Data entries only for grid points of
a superelement which are not exterior points for any superelement. This class of grid points can be determined from the
SEMAP table printed by the program, but may change if the order of processing is changed. The automatic resequencing
option processes each superelement separately.
Component mode synthesis uses SEBSETi, SECSETi, SESUP and SEQSETi entries to define the boundary condition
when computing component modes. Uncoupled solutions for statics and modes can be obtained by use of
PARAM,FIXEDB,-1.

Main Index
 Chapter 7: Solution Sequences 429
Summary of Solution Sequence Operations

Superelement Bulk Data entries may be specified in nonsuperelement solution sequences, but those entries pertaining to
the superelement capability (e.g., the CSUPER entry) are ignored, as well as some fields on other entries (e.g., Field 9 on
the GRID entry). However, better automatic resequencing may be possible if the SEID data is removed.

Superelement Processing in the Structured Solution Sequences

An overview of the superelement processing in SOLutions 101-200 is shown in Figure 7-3. Phase 0 performs the preparation
and bookkeeping function. The input data is stored on the database and scanned to prepare the SEMAP table. This table
lists the superelement to which each Bulk Data entry is to be applied. Phase 0 also partitions out the data that defines each
superelement and performs automatic restart processing. In Phase I this data is used to generate the boundary matrices to
be assembled when combining the superelements into the total structure. Phase II performs calculations done only on the
residual structure (the superelement processed last), such as transient response calculations or eigensolutions. Phase III
expands boundary solutions to the interior grid points of the superelement, performs the data recovery operations for
internal forces, stresses, and strain energy, and outputs these quantities. Phase IV combines plot vectors from upstream
superelements to provide deformed structural plots for all or part of the structure.

Main Index
 430 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

SOLs 101 through 200

Bulk Data Entries, Case Control Commands

Bookkeeping and Control (Phase 0)SubDMAP PHASE0

Generate IFP data blocks, resequence, and store.
Generate SEMAP SLIST (SEP1, SEP3)
Index to next superelement and repeat. (SEP2DR)
Partition out data for one superelement. (SEP2 and SEP2CT)
Undeformed plots.
Automatic restart processing.
Superelement model checkout.

Generation, Assembly, and Reduction (Phase I)SubDMAP PHASE1DR

Index to next superelement and repeat. (SEP2DR)
Generate system structural and load matrices in subDMAP PHASE1A.
Assemble upstream stiffness, loads (statics), mass and damping (dynamics) in SubDMAP PHASE1A.
Reduce to exterior points, dynamic reduction in subDMAPs SEKR, SEMRM, and SELR.
Eliminate ur set in subDMAPs SELRRS and SEMRRS. (residual structure only)

Residual Structure-Only Operations (Phase II)Various SubDMAPs

Formulate dynamic equations.
Eigensolutions, transient, frequency response calculations.

Data Recovery (Phase III)SubDMAP SUPER3

Generate DRLIST (SEP4)
Index to next superelement and repeat. (SEDRDR)
Partition out boundary solution. (SEDR)
Expand boundary displacements to interior points in subDMAP SEDISP.
Compute and add in fixed boundary displacements (statics only) in subDMAP SEDISP.
Compute and output stresses, strain energy, etc. in subDMAP SEDRCVR.

Upstream Plots (Phase IV)SubDMAP SUPER3

Index to next superelement and repeat. (PLVCDR)
Assemble upstream plot vectors (PLTVEC)
Plot deformed structure plots.

(Superelement module names are in parentheses)

Figure 7-3 Structured Solution Sequence Flow Chart

Main Index
 Chapter 7: Solution Sequences 431
Summary of Solution Sequence Operations

The input file processor operations are carried out in a conventional manner. The grid point resequencer module
resequences each superelement individually.
Some of the geometry processor operations of Geometry Processing in SubDMAP PHASE0, 362 are carried out to prepare tables of
all grid points in the entire model, useful for generation of the SEMAP table. After the SEMAP is available, the SLIST table
is assembled. This list controls the operations to be performed in that particular run for generating boundary matrices, as
selected with Case Control commands. The GP4 module generates constraint data in the form of the USET table, as
described in Geometry Processing in SubDMAP PHASE0, 362. A separate USET table is generated for each superelement. It is
similar to the USET table of the older rigid formats except that the exterior points are automatically placed in the a-set.
Assembly and Reduction (Phase I) in SubDMAP Phase1DR

The following operations may be carried out, as selected by the user with superelement-type Case Control commands. By
default, all superelements and the residual structure are given these operations serially. If requests for the SEMA (or SEMR)
operation are made on an upstream superelement and the residual structure, all superelements in the tree between the
selected superelement and the residual structure will also be given the SEMA operation.
Superelement Matrix Generation (SEMG) in SubDMAP PhaselA

The element structural matrices (stiffness, mass and damping) are generated and stored. The direct matrix input option
(DMIG Bulk Data entry) can be used to add stiffness, mass or damping terms to the system matrices in this operation.
These terms are selected by the Case Control commands K2GG, M2GG, and B2GG. The parameters CB1, CB2, CK1,
CK2, CM1 and CM2 can be used to scale the structural matrices.
The SEMA module adds in stiffness terms from any upstream superelements. The SEEX Case Control command may be
used to eliminate stiffness contributions from selected upstream superelements. The superelements that are upstream are
defined by the current SEMAP. The number of rows and columns of the upstream superelements must be consistent with
the number of exterior grid points, as defined by the SEMAP In order to complete execution of the SEMA module, the
boundary matrices of all upstream superelements must be present in the database, and consistent with the current SEMAP.
The SELA module adds loads from upstream superelements to the load vector for the current superelement. The SEEX
Case Control command may be used to eliminate load contributions from upstream superelements. The requirements for
conformity of load data in the database with the current state of the SEMAP are similar to those for stiffness, except that
only rows, not columns of the load vectors are mapped using data in the SEMAP table. The requirement on columns of all
load matrices is that they be equal to the number of columns (number of load condition) in the residual structure. If the
number of columns in all superelements and the residual structure is not the same, a user fatal error results. Static loads may
be applied to any grid point, whether exterior or interior to a superelement. Dynamic loads may be applied directly to
residual structure points. Dynamic load entries may also reference static load sets applied to the superelements via the
LOADSET Case Control command and LSEO Bulk Data entries.
Superelement Load Generation (SELG) in SubDMAP PHASE1A

The load generation module SSG1 is executed in the static and dynamic response solutions. This operation is not executed
in the normal modes and complex eigenvalue solution sequences. As loads may depend on structural matrices (i.e., thermal
or inertial loads), an SELG operation must be preceded by an SEMG operation. The resultants of these loads about the
grid point listed on the PARAM,GRDPNT Bulk Data entry are output.
Superelement Stiffness Reduction (SEKR) in SubDMAPs SEKR and SEKRRS

Main Index
 432 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

The user may change SPCs and/or MPCs between subcases (“boundary condition changes”). The GPSP module identities
grid point singularities, and at user's option provides SPCs for singular degrees of freedom. This is followed by the
constraint eliminations described in Multipoint Constraint Operations in SubDMAPs SEKR and SEMR2 and SEMRB, 366, Single Point
Constraint Operations in SubDMAPs SEKR, SEMR2 and SEMRB, 368, and Constraint and Mechanism Problem Identification in SubDMAP SEKR,
370. If the omitted degrees of freedom contain mechanism-type singularities, the grid and component number of all
downstream ends of the mechanism are output in the matrix named MECHS, and a user fatal error stops execution. The
approximate fixed-boundary natural frequencies of the superelement are printed in the FAPPROX matrix. The boundary
stiffness matrix  K aa  is stored in the database. Mass and damping matrices are not processed in this phase.

SUPORTi Bulk Data entries, which define the r-set, may reference only grid points in the residual structure. The rigid body
operations for the stiffness are performed as described in Static Solutions in SubDMAP SEKRRS, 376.
Superelement Load Reduction (SELR) in SubDMAP SELR

This operation processes loads in the static and dynamic response solution sequences. This operation is not executed in the
normal modes and complex eigenvalue solution sequences. The static load reduction of Static Condensation in SubDMAPs SEKR
and SEMR2, 373 is applied, stopping after generation of the  P a  matrix for superelements and  P l  for the residual
structure.
Superelement Mass Assembly and Reduction (SEMR) in SubDMAP SEMRM

In the dynamics solution sequences, the constraint operations of Multipoint Constraint Operations in SubDMAPs SEKR and SEMR2
and SEMRB, 366, Single Point Constraint Operations in SubDMAPs SEKR, SEMR2 and SEMRB, 368, and Constraint and Mechanism Problem
Identification in SubDMAP SEKR, 370 are applied to the mass and damping matrices. The rigid body operations of Static Solutions
in SubDMAP SEKRRS, 376 are applied to the mass of the residual structure only. Dynamic reduction and component mode
calculations are performed. This allows a mass-change-only type of restart by use of the SEMG and SEMR options and
avoids the costly reduction of the stiffness matrix in the SEKR operation if no stiffness changes are made.
The optional Bulk Data entry PARAM,ERROR,0 affects operations in Phase I. If it is not used, the default action is to
terminate the run after the first user fatal error, as in the older rigid formats. If the parameter is used, a user fatal error causes
a branch to the end of Phase I for the superelement, and continuation to the generation, assembly, and reduction operations
for any remaining superelements. The five SE-type commands may be executed on one run, or may be done separately. The
following sequence must be observed if these commands are executed separately:
SEKR after SEMG
SELG after SEMG
SELR after SEKR, SELG, SEMR
SEMR after SEKR
Residual Structure-Only (Phase II)

The operations on the superelements described above are also performed on the residual structure. This is followed by
operations unique to the residual structure. Eigensolutions and direct and modal forced response calculations take place at
this point. Direct input matrices, when selected by the Case Control commands M2PP, K2PP, and B2PP are constrained
and added as in the Formulation of Dynamic Equations in SubDMAP GMA, 387.

Main Index
 Chapter 7: Solution Sequences 433
Summary of Solution Sequence Operations

Bulk Data entries used for static loads (i.e., FORCE, GRAV, LOAD, etc.) may be used in direct and modal dynamic analysis.
They are interchangeable with DAREA Bulk Data entries and may be referenced by LSEQ Bulk Data entries.
Data Recovery Operations (Phase III) in SubDMAP SUPER3

The total solution vectors of the downstream superelement are partitioned to generate the boundary solution vectors for
the superelement being processed. The boundary solution is expanded to all interior points of the superelement. The fixed
boundary component is computed and added.
If SPC force output is requested, the resultants of the SPC forces about PARAM,GRDPNT or the basic coordinate system
origin are also output. The data recovery operations of Static Solutions in SubDMAP SEKRRS, 376 are carried out. (Note that
solution sequences involving complex eigenvalue or frequency response analysis may produce complex displacement
vectors. GPFDR module output (grid point force balance, element strain energy) may be requested for complex vectors,
but only the real component of the vector is processed.) Deformed structure plots of only elements in the superelement are
generated here.
Upstream Plots (Phase IV) in SubDMAP SUPER3

The displacement vectors of a superelement and all of its upstream members are combined and plotted at the user's request.

Superelement Processing
An overview of the superelement processing is shown in Figure 7-4. Phase 0 performs the preparation and bookkeeping
function. The input data is stored on the database and scanned to prepare the SEMAP table. This table lists the
superelement to which each Bulk Data entry is to be applied. Phase I partitions out the data that defines one superelement
and processes this data to generate the boundary matrices to be assembled when combining the superelements into the total
structure. Phase II performs calculations done only on the residual structure (the superelement processed last), such as
transient response calculations or eigensolutions. Phase III expands boundary solutions to the interior grid points of the
superelement, performs the data recovery operations for internal forces, stresses, and strain energy, and outputs these
quantities. Phase IV combines plot vectors from upstream superelements to provide deformed structural plots for all or part
of the structure.
The input file processor operations are carried out in a conventional manner. The grid point resequencer module
resequences each superelement individually.
Some of the geometry processor operations of Geometry Processing in SubDMAP PHASE0, 362 are carried out to prepare tables of
all grid points in the entire model, useful for generation of the SEMAP table. After the SEMAP is available, the SLIST table
is assembled.
This list controls the operations to be performed in that particular run for generating boundary matrices as selected with
Case Control options.
Assembly and Reduction (Phase I)

The following operations may be carried out as selected by the user on superelement-type Case Control commands. By
default, all superelements and the residual structure are given these operations serially. If requests for the SEMA (or SEMR)
operation are made on an upstream superelement and the residual structure, all superelements in the tree between the
selected superelement and the residual structure will also be given the SEMA operation.

Main Index
 434 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

Superelement Matrix Generation (SEMG)

The element structural matrices (stiffness, mass and damping) are generated and stored. The direct matrix input option
(DMIG Bulk Data entry) can be used to add stiffness, mass or damping terms to the system matrices in this operation.
These terms are selected by the Case Control commands K2GG, M2GG, and B2GG. The parameters CB1, CB2, CK1,
CK2, CM1, and CM2 can be used to scale the structural matrices.
Superelement Load Generation (SELG)

The load generation module SSGI is executed in the statics and dynamic response solutions. This operation is not executed
in the normal modes and complex eigenvalue solution sequences. As loads may depend on structural matrices (i.e., thermal
or inertial loads), an SELG operation must be preceded by an SEMG operation. The resultants of these loads about the
grid point listed on the PARAM,GRDPNT Bulk Data entry are output.

Main Index
 Chapter 7: Solution Sequences 435
Summary of Solution Sequence Operations

Bulk Data Entries, Case Control Commands

Bookkeeping and Control (Phase 0)

Generate IFP data blocks, resequence, and store.
Generate SEMAP, SLIST (SEP1, SEP3)

Generation, Assembly, and Reduction (Phase I)

Index to next superelement and repeat. (SEP2DR)
Partition out data for one superelement. (SEP2 and SEP2CT)
Undeformed plots.
Generate system structural and load matrices.
Assemble upstream stiffness, loads (statics), mass and damping (dynamics).
Reduce to exterior points, dynamic reduction.
Eliminate Uf set. (residual structure only)

Residual Structure-Only Operations (Phase II)

Forrnulate dynamic equations.
Eigensolutions, transient, frequency response calculations.

Data Recovery (Phase III)

Generate DRLIST (SEP4)
Index to next superelement and repeat. (SEDRDR)
Partition out boundary solution. (SEDR)
Expand boundary displacements to interior points.
Compute and add in fixed boundary displacements (statics only).
Compute and output stresses, strain energy, etc.

Upstream Plots (Phase IV)

Index to next superelement and repeat. (PLVCDR)
Assemble upstream plot vectors (PLTVEC)
Plot deformed structure plots.

(Superelement module names are in parentheses)

Figure 7-4 Solution Sequence Flow Chart

Superelement Stiffness Matrix Assembly and Reduction (SEKR)

Main Index
 436 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

The SEMA module adds in stiffness terms from any upstream superelements. The SEEX Case Control command may be
used to eliminate stiffness contributions from selected upstream superelements. The superelements that are upstream are
defined by the current SEMAP. The number of rows and columns of the upstream superelements must be consistent with
the number of exterior grid points, as defined by the SEMAP. In order to complete execution of the SEMA module, the
boundary matrices of all upstream superelements must be present in the database, and consistent with the current SEMAP.
The GP4 module generates constraint data in the form of the USET table, as described in Geometry Processing in SubDMAP
PHASE0, 362. A separate USET table is generated for each superelement. It is similar to the USET table of the older rigid
formats except that the exterior points are automatically placed in the a-set.
The user may change SPCs and/or MPCs between subcases (“boundary condition changes”) in SOLs 101 and 103 only.
The GPSP module identifies grid point singularities, and at user’s option provides SPCs for singular degrees of freedom.
This is followed by the constraint eliminations described in Multipoint Constraint Operations in SubDMAPs SEKR and SEMR2 and
SEMRB, 366, Single Point Constraint Operations in SubDMAPs SEKR, SEMR2 and SEMRB, 368, and Constraint and Mechanism Problem
Identification in SubDMAP SEKR, 370. If the omitted degrees of freedom contain mechanism-type singularities, the grid and
component number of all downstream ends of the mechanism are output in the matrix named MECHS, and a user fatal
error stops execution. The approximate fixed-boundary natural frequencies of the superelement are printed in the
FAPPROX matrix. The boundary stiffness matrix  KAA  is stored in the database. Mass and damping matrices are not
processed in this phase.
SUPORT Bulk Data entries, which define the r-set, may reference only grid points in the residual structure. The rigid body
operations for the stiffness are performed as described in Static Solutions in SubDMAP SEKRRS, 376.
Superelement Load Assembly and Reduction (SELR)

This operation processes loads in the static and dynamic response solution sequences. This operation is not executed in the
normal modes and complex eigenvalue solution sequences. The SELA module adds loads from upstream superelements to
the load vector for the current superelement. The SEEX Case Control command may be used to eliminate load
contributions from upstream superelements. The requirements for conformity of load data in the database with the current
state of the SEMAP are similar to those of the SEKR operation above, except that only rows, not columns of the load vectors
are mapped using data in the SEMAP table. The requirement on columns of all load matrices is that they be equal to the
number of columns (number of load condition) in the residual structure. If the number of columns in all superelements
and the residual structure is not the same, a user fatal error results. Static loads may be applied to any grid point, whether
exterior or interior to a superelement. Dynamic loads may be applied directly to residual structure points. Dynamic load
entries may also reference static load sets applied to the superelements via the LOADSET Case Control command and LSEQ
Bulk Data entries. The static load reduction of Static Condensation in SubDMAPs SEKR and SEMR2, 373 is applied, stopping after
generation of the  P a  matrix for superelements and  P l  for the residual structure.

Superelement Mass Assembly and Reduction (SEMR)

In the dynamics solution sequences, the constraint operations of Multipoint Constraint Operations in SubDMAPs SEKR and SEMR2
and SEMRB, 366, Single Point Constraint Operations in SubDMAPs SEKR, SEMR2 and SEMRB, 368 and Constraint and Mechanism Problem
Identification in SubDMAP SEKR, 370 are applied to the mass and damping matrices. The rigid body operations of Static Solutions
in SubDMAP SEKRRS, 376 are applied to the mass of the residual structure only. Dynamic reduction and component mode
calculations of Static and Dynamic Load Generation, 380 are performed. This allows a mass-change-only type of restart by use of

Main Index
 Chapter 7: Solution Sequences 437
Summary of Solution Sequence Operations

the SEMG and SEMR options and avoids the costly reduction of the stiffness matrix in the SEKR operation if no stiffness
changes are made.
The optional Bulk Data entry PARAM,ERROR,0 affects operations in Phase I. If it is not used, the default action is to
terminate the run after the first user fatal error, as in the older rigid formats. If the parameter is used, a user fatal error causes
a branch to the end of Phase I for the superelement, and continuation to the generation, assembly, and reduction operations
for any remaining superelements.
The five SE-type commands may be executed in one run, or may be done separately. The following sequence must be
observed if these commands are executed separately:
SEKR after SEMG
SELG after SEMG
SELR after SEKR, SELG, SEMR
SEMR after SFKR
Residual Structure-Only (Phase II)

The operations on the superelements described above are also performed on the residual structure. This is followed by
operations unique to the residual structure. Eigensolutions and direct and modal forced response calculations take place at
this point. Direct input matrices, when selected by the Case Control commands M2PP, K2PP, and B2PP are constrained
and added as in Formulation of Dynamic Equations in SubDMAP GMA, 387.
Bulk Data entries used for static loads (i.e., FORCE, GRAV, LOAD, etc.) may be used in direct and modal dynamic analysis.
They are interchangeable with DAREA Bulk Data entries and may be referenced by LSEQ Bulk Data entries.
Data Recovery Operations (Phase Ill)

The total solution vectors of the downstream superelement are partitioned to generate the boundary solution vectors for
the superelement being processed. The boundary solution is expanded to all interior points of the superelement. The fixed
boundary component is computed and added.
If SPC force output is requested, the resultants of the SPC forces about PARAM,GRDPNT or the basic coordinate system
origin are also output. The data recovery operations of Data Recovery Operations in SubDMAP SEDISP, 391 are carried out. (Note
that solution sequences involving complex eigenvalue or frequency response analysis may produce complex displacement
vectors. GPFDR module output (grid point force balance, element strain energy) may be requested for complex vectors,
but only the real component of the vector is processed.) Deformed structure plots of only elements in the superelement are
generated here.
Upstream Plots (Phase IV)

The displacement vectors of a superelement and all of its upstream members are combined and plotted at user's request.

Superelement Alternate Statics Solution

SOL 101 provides an alternate method of performing the superelement static condensation. PARAM,ALTRED,YES must
be entered in the Bulk Data Section. In some cases, the alternate solution can result in significant reduction in computer
cost for static analysis.

Main Index
 438 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

The stiffness and load reduction and displacement recovery are described in Static and Dynamic Load Generation, 380, Static
Condensation in SubDMAPs SEKR and SEMR2, 373 and Data Recovery Operations in SubDMAP SEDISP, 391.

Static Superelement Analysis With lnertia Relief

Inertia relief analysis is available in SOL 101. PARAM,INREL,-1 or PARAM,INREL, -2 must be present for either of
these options. The following method transforms static loads based on model geometry rather than the stiffness matrix as
described in Static Solutions in SubDMAP SEKRRS, 376.
The additional operations required to perform inertia relief are described in Static and Dynamic Load Generation, 380 and Static
Solutions in SubDMAP SEKRRS, 376.
Assumptions and Limitations

 The model must have six and only six rigid body modes for a continuous, three-dimensional structure.
 Masses on scalar points are ignored in calculating inertia loads.
 The structural model, when given enforced displacements at the reference point, is assumed to produce the same
motion as a rigid body model. This will be true for any structural or rigid element. It is possible, if unconventional,
to use CELASi or general elements, MPC equations, or DMIG coefficients that will not meet this assumption.
 Identical and mirror image superelements are not processed correctly.

Nonlinear Heat Transfer Analysis

Steady State Heat Transfer Analysis
The steady state heat balance equation is given by
4
 K   u  +     u + T abs  =  P  +  N  (7-275)

where:

K = a heat conduction matrix

 = a radiation exchange matrix
P = a vector of applied heat flows that are constant
N = a vector of nonlinear heat flows that depend on temperature
u = a vector of grid point temperature
T abs = the absolution temperature

The components of the applied heat flow vector,  P  , are associated either with surface heat transfer or with heat
generated inside the volume heat conduction elements. The vector of nonlinear heat flows,  N  , results from boundary
radiation, surface convection, and temperature dependent thermal loads.

Main Index
 Chapter 7: Solution Sequences 439
Summary of Solution Sequence Operations

The equilibrium equation is solved by a Newton iteration scheme, where the tangential stiffness matrix is approximated by
i
i i i  N 
3
 KT    K  + 4     ui + T abs  –  ------  (7-276)
 u 
and the residual vector is
i i i i i 4
 R  =  P  +  N  –  K   u  –     u i + T abs  (7-277)

Transient Heat Transfer Analysis

The general equation solved in transient analysis has the form
4
 B   u·  +  K   u  +     u + T abs  =  P  +  N  (7-278)

To take phase change into consideration, the heat diffusion equation is converted into

 H·  +  K   u  +     u + T abs  =  P  +  N 
4
(7-279)

where:

B = heat capacity matrix

K = heat conduction matrix
 = radiation matrix
P = vector of applied heat flows that are constant or functions of time
N = vector of nonlinear heat flows that depend on temperature
H = enthalpy vector
 H·  =  dH  dt 
u = grid point temperatures
 u·  =  du  dt 
T abs = absolute temperature scale factor

The equilibrium equation is solved by Newmark’s method with adaptive time stepping. Base on this one-step integration
scheme, the time derivative of the nodal temperatures at the  i + 1  th iteration of the time step  n + 1  is expressed as

i+1 1 1
 u· n + 1  = ---------  u ni ++ 11 – u n  +  1 – ---  u· n  (7-280)
t  
where

Main Index
 440 MSC Nastran Reference Guide
Summary of Solution Sequence Operations

i+1 i i
 un + 1  =  u n + 1  +  u n + 1  (7-281)

and

--1- = 2 – 2 (7-282)

The parameter  is specified on the PARAM,NDAMP Bulk Data entry. When  = 0   = 0.5  , no numerical
damping is requested. In this case, Newmark’s method is equivalent to the Crank-Nicolson method.

Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

8 Element Data Recovery

Resolved at Grid Points

 Introduction

Stress Recovery at Grid Points

Mesh Stress Discontinuities at Grid Points

Main Index
 442 MSC Nastran Reference Guide
Introduction

Introduction
Element stresses may be averaged at grid points and then used to calculate estimates of stress discontinuities in the model
mesh. Resolving element data at the grid points is supported only for selected linear elastic element types (see Table 8-1 for
supported element types). The calculation of average stresses at grid points and the calculation of estimates of stress
discontinuities use a common mesh definition method. The mesh is defined in the OUTPUT(POST) section of the Case
Control Section. There are two methods available for mesh definition: SURFACE input for 2-D shell element meshes and
VOLUME input for 3-D solid element meshes. The averaging of grid point stresses (and strains) and stress discontinuities
are calculated independently for meshes composed of only 2-D shell elements and those composed of only 3-D solid
elements. The element stress and strain data used in the calculation of grid point stresses and stress discontinuities must be
requested in the Case Control Section using the STRESS command.
The request for grid point stresses for reports is made by the GPSTRESS command in Case Control. The calculation of
strains may be substituted for stresses using the STRAIN command in place of the STRESS command in the Case Control
Section (see STRAIN (Case) in the MSC Nastran Quick Reference Guide command for element strain output requests).
The calculation of stress discontinuities is supported for both elements and grid points associated with an element mesh by
the ELSDCON and GPSDCON commands, respectively.
Stress Recovery at Grid Points and Mesh Stress Discontinuities at Grid Points explain in more detail the requesting and calculation of
grid point stresses and stress discontinuities, respectively.

Table 8-1 Elements Types Supported in Element Data Resolved at Grid Point Calculations
(Hyperelastic Elements Not Included)
Element 2-D Shell 3-D Solid Stress
Type Stress Stress Discontinuity
CQUAD4 x x
CQUADR x x
CQUAD8 x x
CTRIA3 x x
CTRIAR x x
CTRIA6 x x
CHEXA x x
CPENTA x x
CTETRA x x

Stress Recovery at Grid Points

Stresses may be computed at grid points on surfaces of two-dimensional plate elements, namely, the CQUAD4, CQUADR,
CQUAD4, CTRIA3, CTRIAR, and CTRIA6, and in volumes containing CHEXA, CPENTA, and CTETRA solid elements.
By default, stresses are output at the center of each of these elements and also at the vertices of CQUADR, CQUAD8,

Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points 443
Stress Recovery at Grid Points

CTRIAR, and CTRIA6 shell elements and the CHEXA, CPENTA, and CTETRA solid elements. The center output is not
universally satisfactory because in the case concerning the surfaces of CQUAD4 and/or CTRIA3 elements, the analyst often
has a requirement for realistic stress values at the grid points. Also, in the case of surfaces of CQUADR, CQUAD8, CTRIAR,
and CTRIA6 elements or of volumes containing CHEXA, CPENTA, and CTETRA elements, the vertex stresses output for
the elements that connect to a common grid point are often not identical. The option described in this section provides a
rational method that yields accurate and unique stresses at each vertex grid point in user-defined surfaces or volumes. Grid
point stress recovery is only available in linear static analysis, real eigenvalue analysis, and transient analysis. Grid point stress
recovery is not available for lamina stresses.
An option is also provided to estimate the probable error in these grid point stresses (see Mesh Stress Discontinuities at Grid
Points). Estimates are generated for each stress component. The root mean square error of these estimates is then computed
to provide a single measure of the error in the computation of the stresses at a grid point. This data provides a gross
indication of the adequacy of a finite element model to represent the physical phenomena under investigation. The error
estimator is available in SOLution Sequence 101.

Description of Method
A scheme to interpolate and/or extrapolate over a surface from a known set of stresses (the element stresses) requires that
these known stress components be transformed into a consistent coordinate system prior to interpolation and extrapolation.
This coordinate system is defined by the analyst and it should be generally compatible with the surface of interest in order
to obtain accurate grid point stresses. Generally, such a coordinate system will be the natural coordinate system that the
analyst used in the generation of the model of the surface or volume in question. Calculated grid point stress components
are output in this coordinate system, called the output coordinate system of the surface or volume.
Each stress component is treated independently of other stress components in the calculation of average grid point stress
components. The stress invariants at grid points (i.e., principal stresses, the vonMises stress, and the mean pressure) are
evaluated from the average values of the stress components at grid points. The grid point stress components are output in
the output coordinate system of the surface or volume and the stress invariants are oriented relative to this output coordinate
system.
Two methods are used to calculate grid point stress components for plate and shell elements: topological and geometric.
A single, simplified, topological method is used to calculate grid point stress components in volumes containing solid
elements. However, prior to a discussion of these methods, it is necessary to define the four categories into which grid points
are divided.
1. Interior Grid Points
Grid points that are connected only to interior line segments; i.e., line segments that are coincident with the edges
of two or more elements.
2. Corner Grid Points
Grid points that are only connected to a single element.
3. Edge Grid Points
Grid points that are connected to both interior and exterior line segments. An exterior line segment is coincident
with the edge of only one element.

Main Index
 444 MSC Nastran Reference Guide
Stress Recovery at Grid Points

4. Exception Grid Points

Grid points at which discontinuities in stress can occur. Exception points will be described more completely under
Treatment of Exception Points: BRANCHing.
Stresses at interior, corner, or edge grid points are considered to be continuous between directly connected elements.
Stresses at exception points may be discontinuous between connected elements and, as will be seen, a different grid point
stress may be output for each of the connected elements. Engineering judgment must be exercised in assessing the validity
of grid point stresses at exception points.

Topological Method: Plate Elements and Solid Elements

The topological method for the evaluation of grid point stresses only recognizes the existence of interior, edge, and corner
points; that is, exception points are not considered as a distinct category of grid point (exception points arise as a result of
geometrical rather than topological characteristics). The relations used to compute average grid point stresses in surfaces of
plate elements by the topological method are described in Table 8-2. The relations used to compute the average grid point
stresses in volumes containing solid elements are described in Table 8-3. Only the topological method is available for solid
elements.

Table 8-2 Evaluation of Average Grid Point Stresses for Plate and Shell Elements
by the Topological Method
Type of
Grid Point Average Grid Point Stress Component Remarks
Interior Ne If CQUAD4 and/or CTRIA3
elements connect to the same grid
1 point as CQUAD8 and/or CTRIA6
 I = ------  ei
Ne elements, then only the CQUAD8
i=1 and or CTRIA6 elements are used in
the calculation of average grid point
where: stresses.

 I = average stress at interior grid point

N e = number of directly connected elements

 element center stresses if only CQUAD4

 and/or CTRIA3 elements are connected

 ei = 
 element vertex stresses if CQUAD8 and/or
 CTRIA6 are connected


Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points 445
Stress Recovery at Grid Points

Table 8-2 Evaluation of Average Grid Point Stresses for Plate and Shell Elements
by the Topological Method (continued)
Type of
Grid Point Average Grid Point Stress Component Remarks
Edge Ne NI If an edge point is not connected to an
interior point by line segments, then
2 1 N I = 0 , the second term in the
 E = ------  ei – -----
N I  Ii

Ne relation  E is set to zero, and the
i=1 i=1 factor 2 in the first term is replaced by
1.
where:
Same as remark for interior grid
 E = average stress at edge grid point points.

N I = number of interior points connected by line

segments
 Ii = average stress at ineterior grid points
Corner Corner points connected to CQUAD4 element with the Same as remark for interior points
grid point diagonally opposite the corner point, an interior
point, or an edge point.
 c = 2 1 –  D

where:
 c = average stress at corner grid point
 i = CQUAD4 center stress

 D = average stress at grid point diagonally

opposite corner point
(If point diagonally opposite corner point is not an interior
or edge point, then  c =  1 .

Main Index
 446 MSC Nastran Reference Guide
Stress Recovery at Grid Points

Table 8-2 Evaluation of Average Grid Point Stresses for Plate and Shell Elements
by the Topological Method (continued)
Type of
Grid Point Average Grid Point Stress Component Remarks
Corner Corner points connected to CTRIA3 elements and the other two Same as remark for interior points
grid points, namely, E1 and E2, of the connected CTRIA3 are
edge points:

 c = 3 1 –   E1 +  E2 

where:

 c = average stress at corner grid point

 1 = CTRIA3 center stress
 E1 ,  E2 = stresses at edge points E1 and E2
(if either E1 or E2 is not an edge point, then  c =  1 )

Corner points connected to one CQUAD4 or one CTRIA6

element:

c = r
where:
 r = stress at element vertex connected to corner point
Corner points connected to two or more CTRIA6 elements:

Nc
1
 c = ------  r
Nc i
i=1
where:

N c = number of CTRIA6 elements connected to corner point.

Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points 447
Stress Recovery at Grid Points

Table 8-2 Evaluation of Average Grid Point Stresses for Plate and Shell Elements
by the Topological Method (continued)
Type of
Grid Point Average Grid Point Stress Component Remarks
Interior and Ne Mixtures of CHEXA, CPENTA, and
Edge 1 CTETRA elements may connect to
 I = ------  ei an interior or edge grid point.
Ne
i=1
where:

 I = average stress at interior or edge grid points

N e = number of directly connected elements
 ei = element vertex stresses
Corner c = r The same relation is used whether a
CHEXA, CPENTA, or CTETRA
element is connected to the grid
where: point.
 r = stress at element vertex connected to corner point

Table 8-3 Evaluation of Average Grid Point Stresses for Plate and Shell Elements by the Geometric Method
Type of
Grid Point Number of Elements Involved in Stress Calculation
Interior All directly connected elements
Edge Two directly connected elements:

Use the two known element stresses plus the stresses at all interior points connected by
line segments of the directly connected elements.

Three or more directly connected elements:

Use the known element stresses of the directly connected elements.

Corner Same as topological method (see Table 8-1).

Geometric Method: Surface Elements

In the geometric method, grid point stresses for points which connect to CQUAD8 or CTRIA6 elements use the average of
the element corner stresses. If any CQUAD4 or CTRIA3 elements connect to these points, their influence is not considered.

Main Index
 448 MSC Nastran Reference Guide
Stress Recovery at Grid Points

The geometric method for the determination of average grid point stresses which connect only to CQUAD4 or CTRIA3
elements utilizes the numerical technique of least squares. The “best fit” value of the stress  at a point ( x , y ) is
functionally denoted as:

  x,y  = A0 + A1 x + A2 y (8-1)

where the coefficients A 0 , A 1 x , and A 2 must be determined. The coordinates x and y are the locations of points at
which known stresses exist. In accordance with the least squares technique, it is required to minimize the function:
T
1
f = ---   –
2   
  –
 
  (8-2)

with respect to  A  . Here  represents the known element stresses.

To perform the minimization, equation (8-1) is written in matrix form as:




  =  T  A 
where:

T = 1 X1 X2
.. .. ..
. . .
i Xi Xi

and equation (8-2) is differentiated with respect to the coefficients A i .

f-
-------
T T
= T TA – T  = 0 (8-3)
A i

The coefficients  A  are then determined from equation (8-3) to be:

T –1 T
A = T T T  (8-4)
The average grid point stress at a given grid point can then be calculated by substituting equation (8-4) into the matrix
representation of equation (8-1) as follows:

 = 1 T –1 T
Xo Yo    T   T    T     =  Q     (8-5)

Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points 449
Stress Recovery at Grid Points

where:

Q = matrix of interpolation factors

X o Y o = locations of grid points at which stress is required

T –1
Note that if   T   T   is singular, the matrix of interpolation factors is taken to be 1  N I so that equation (8-5)
becomes:
NI

 1
= -----
NI  ei (8-6)
i=1

where N I = number of known stress points associated with the grid point. In most cases, equation (8-6) is equivalent to the
topological interpolation method.
The strategy used in the evaluation of equation (8-5) for interior, edge, and corner grid points is presented in Table 8-3.
As previously mentioned, the coordinates x and y of equation (8-1) are the locations of points at which known stresses
exist. These locations are actually taken as the projected coordinates of the known stress points on the average surface
evaluated at the grid point at which the stress is to be computed. This average surface is determined to be the surface whose
normal is the average of the unit normals to the elements connected to the grid point of interest. The local x-axis used in
the evaluation of the geometric coefficients of the matrix  T  is taken as the projection of the x-axis of the output
coordinate system on the average surface evaluated at the grid point of interest.

Treatment of Exception Points: BRANCHing

Exception points have been defined as grid points at which stress discontinuities may occur. Such discontinuities may occur
because a substantial difference in slope exists between adjacent elements directly connected to the same line segment or if
more than two elements are directly connected to the same line segment. Examples of these two cases are illustrated below.

Main Index
 450 MSC Nastran Reference Guide
Stress Recovery at Grid Points

Figure 8-1 Three or More Elements Connected to the Same Line Segment

Figure 8-2 Slope Difference between Adjacent Elements Connected to the Same Line

Note: That these two situations are geometric rather than topological in nature and, therefore, are only considered
under the geometric interpolation option.

Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points 451
Stress Recovery at Grid Points

Since exception points are defined as grid points at which stress discontinuities may exist, it is conceivable that a computed
average grid point stress at an exception point may be different for each of the directly connected elements. However, it is
more common to find that a subset of the elements that are directly connected to an exception point forms a subsurface on
which stresses may be considered to be continuous. A unique average grid point stress may then be computed for the
elements of the subsurface.
The calculation of average grid point stresses for a subsurface that contains two or more elements is performed with the
geometric interpolation method used for edge points. If a subsurface consists of a single element, the calculation of an
average grid point stress is performed in the same manner as utilized in the calculation of average grid point stresses at corner
points. It must be noted, however, that  D ,  E1 , and  E2 may be interior, exterior, or corner points (see Table 8-3).

User Information
The use of the grid point stress recovery option is demonstrated in this section through the consideration of several example
problems. These problems are discussed separately after the presentation of the requirements that must be met in the
Executive Control Section, the Case Control Section, and the Bulk Data Section.

Executive Control Section

A typical Executive Control Section would be of the following form:
SOL 101
CEND

Case Control Section

The option to obtain grid point stress output is exercised through commands in the Case Control Section. The general
request for the recovery of grid point stresses is accomplished using commands with the form:
GPSTRESS = ALL
or
SET j = k, l, m, . . .
GPSTRESS = j
The k, l, m, . . . reference the identification numbers of the surfaces or volumes of elements that are defined in the
OUTPUT(POST) section of the Case Control Section. Grid point stress data will be output only for those surfaces or
volumes referenced through the GPSTRESS Case Control command.

Note: If grid point stress data is desired, an ELSTRESS (or STRESS) request for all elements in the surfaces or volumes
of interest must be included in the Case Control Section. Such requests will cause the output of element stress
data as well as grid point stress data. If the output of element stress data is not desired, the requests
ELSTRESS(PLOT) = or STRESS(PLOT) = may be used.

Main Index
 452 MSC Nastran Reference Guide
Stress Recovery at Grid Points

In static analysis, one may place the GPSTRESS command above the subcase level or in individual subcases. If the
GPSTRESS command is above the subcase level, it will be utilized by all subcases and subcoms unless overridden by a
GPSTRESS command in a subcase or subcom.
The OUTPUT(POST) section of the Case Control Section contains specific requests for grid point stress data for the
surfaces or volumes of elements. Any number of element surfaces or volumes may be defined, but only those surfaces or
volumes that are referenced through the GPSTRESS command will have grid point stress data generated and output.
Element surfaces and volumes are defined through a SURFACE and VOLUME commands in the OUTPUT(POST)
section of the Case Control Section.

Grid Point Stress Output Description

Listing 8-1 through Listing 8-3 are examples of grid point stress output produced for two-dimensional plate elements
requested on the SURFACE Case Control command. Listing 8-1 is the most common form of output and is always the result
when the TOPOLOGICAL method is used, and when the GEOMETRIC method is used and no surface “BRANCHing”
is detected. Note that when the “ELEMENT ID” value is zero, it always indicates that all elements in the SURFACE
connected to the grid point contribute to the average grid point stress and lie in the “best” subsurface. Listing 8-2 and
Listing 8-3 show examples of the output when the GEOMETRIC method is used and when “BRANCHing” occurs. The
negative value for the “ELEMENT ID” indicates the “best” average grid point stress and the element identification number
of an element associated with “best” subsurface. Similarly a positive element identification numbers indicate another
average grid point stress for a “less best” grid point stress and the element identification number associated with it’s
subsurface. When TOLERANCE is set positive as in Listing 8-2 only the “best” grid point stress is calculated when
“BRANCHing” is detected and when TOLERANCE is set negative all subsurface average grid point stress are calculated.
See SURFACE (p. 743) in the .

Listing 8-1 Grid Point Stress Output – Typical

SQUARE PLATE WITH CIRCULAR HOLE JUNE 24, 1991 MSC/NASTRAN 6/14/91 PAGE 27
GRID POINT STRESS TEST CASE
UNIFORM LOAD ALONG X=5.0. SUBCASE 1
SUBCASE = 1
S T R E S S E S A T G R I D P O I N T S - - S U R F A C E 4
SURFACE X-AXIS X NORMAL(Z-AXIS) Z REFERENCE COORDINATE SYSTEM FOR SURFACE DEFINITION CID 1
GRID ELEMENT STRESSES IN SURFACE SYSTEM PRINCIPAL STRESSES MAX
ID ID FIBER NORMAL-X NORMAL-Y SHEAR-XY ANGLE MAJOR MINOR SHEAR VON MISES
10304 0 MID 4.965E+03 4.041E+03 -6.431E+03 -42.9450 1.095E+04 -1.944E+03 6.447E+03 1.204E+04
10305 0 MID 6.132E+03 3.839E+03 -5.518E+03 -39.1310 1.062E+04 -6.503E+02 5.636E+03 1.096E+04
10306 0 MID 7.033E+03 3.408E+03 -4.246E+03 -33.4429 9.837E+03 6.037E+02 4.617E+03 9.549E+03
10400 0 MID 6.101E+02 1.631E+04 -1.038E+03 -86.2353 1.638E+04 5.418E+02 7.919E+03 1.612E+04
10401 0 MID 1.463E+03 1.417E+04 -3.402E+03 -75.9185 1.503E+04 6.098E+02 7.209E+03 1.473E+04
10402 0 MID 2.228E+03 1.211E+04 -5.354E+03 -66.3558 1.446E+04 -1.156E+02 7.286E+03 1.452E+04
10403 0 MID 2.809E+03 1.019E+04 -6.381E+03 -60.0211 1.387E+04 -8.721E+02 7.371E+03 1.433E+04
10404 0 MID 3.159E+03 8.514E+03 -6.308E+03 -56.5007 1.269E+04 -1.016E+03 6.852E+03 1.323E+04
10405 0 MID 3.307E+03 7.318E+03 -5.482E+03 -55.0445 1.115E+04 -5.249E+02 5.837E+03 1.142E+04

Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points 453
Stress Recovery at Grid Points

Listing 8-2 Grid Point Stress Output with “BRANCHing” and TOLERANCE > 0
INTERSECTING PLATES JUNE 24, 1991 MSC/NASTRAN 6/14/91 PAGE 34
GRID POINT STRESS TEST CASE

SUBCASE = 1
S T R E S S E S A T G R I D P O I N T S - - S U R F A C E 4
SURFACE X-AXIS X NORMAL(Z-AXIS) Z REFERENCE COORDINATE SYSTEM FOR SURFACE DEFINITION CID 0
GRID ELEMENT STRESSES IN SURFACE SYSTEM PRINCIPAL STRESSES MAX
ID ID FIBER NORMAL-X NORMAL-Y SHEAR-XY ANGLE MAJOR MINOR SHEAR VON MISES
10304 0 Z1 9.107E+02 1.209E+03 2.799E+03 46.5224 3.863E+03 -1.743E+03 2.803E+03 4.969E+03
Z2 1.052E+03 1.571E+03 2.913E+03 47.5464 4.236E+03 -1.613E+03 2.925E+03 5.233E+03
MID 9.813E+02 1.390E+03 2.856E+03 47.0453 4.049E+03 -1.678E+03 2.863E+03 5.099E+03
10305 0 Z1 1.906E+03 5.699E+02 1.789E+03 34.7594 3.147E+03 -6.714E+02 1.909E+03 3.531E+03
Z2 2.032E+03 5.648E+02 1.934E+03 34.6114 3.367E+03 -7.698E+02 2.068E+03 3.811E+03
MID 1.969E+03 5.674E+02 1.861E+03 34.6824 3.257E+03 -7.206E+02 1.989E+03 3.671E+03
10306 0 Z1 2.771E+03 -1.983E+02 2.678E+02 5.1132 2.795E+03 -2.223E+02 1.508E+03 2.912E+03
Z2 3.024E+03 -1.118E+02 4.889E+02 8.6589 3.098E+03 -1.862E+02 1.642E+03 3.196E+03
MID 2.897E+03 -1.550E+02 3.783E+02 6.9615 2.944E+03 -2.012E+02 1.572E+03 3.049E+03
10400 -119 Z1 -5.775E+04 1.181E+03 -2.060E+04 -72.5215 7.667E+03 -6.423E+04 3.595E+04 6.839E+04
Z2 -5.821E+04 1.207E+04 -1.647E+04 -77.4454 1.573E+04 -6.188E+04 3.880E+04 7.106E+04
MID -5.798E+04 6.623E+03 -1.853E+04 -75.0777 1.156E+04 -6.292E+04 3.724E+04 6.942E+04
10401 -119 Z1 -1.856E+04 7.060E+03 -4.939E+03 -79.4558 7.979E+03 -1.948E+04 1.373E+04 2.446E+04
Z2 -1.714E+04 1.656E+04 -3.881E+03 -83.5131 1.700E+04 -1.758E+04 1.729E+04 2.995E+04
MID -1.785E+04 1.181E+04 -4.410E+03 -81.7177 1.245E+04 -1.849E+04 1.547E+04 2.696E+04

Listing 8-3 Grid Point Stress Output with “BRANCHing” and TOLERANCE < 0
INTERSECTING PLATES JUNE 24, 1991 MSC/NASTRAN 6/14/91 PAGE 40
GRID POINT STRESS TEST CASE

SUBCASE = 1
S T R E S S E S A T G R I D P O I N T S - - S U R F A C E 5
SURFACE X-AXIS X NORMAL(Z-AXIS) Z REFERENCE COORDINATE SYSTEM FOR SURFACE DEFINITION CID 0
GRID ELEMENT STRESSES IN SURFACE SYSTEM PRINCIPAL STRESSES MAX
ID ID FIBER NORMAL-X NORMAL-Y SHEAR-XY ANGLE MAJOR MINOR SHEAR VON MISES
10304 0 Z1 9.107E+02 1.209E+03 2.799E+03 46.5224 3.863E+03 -1.743E+03 2.803E+03 4.969E+03
Z2 1.052E+03 1.571E+03 2.913E+03 47.5464 4.236E+03 -1.613E+03 2.925E+03 5.233E+03
MID 9.813E+02 1.390E+03 2.856E+03 47.0453 4.049E+03 -1.678E+03 2.863E+03 5.099E+03
10305 0 Z1 1.906E+03 5.699E+02 1.789E+03 34.7594 3.147E+03 -6.714E+02 1.909E+03 3.531E+03
Z2 2.032E+03 5.648E+02 1.934E+03 34.6114 3.367E+03 -7.698E+02 2.068E+03 3.811E+03
MID 1.969E+03 5.674E+02 1.861E+03 34.6824 3.257E+03 -7.206E+02 1.989E+03 3.671E+03
10306 0 Z1 2.771E+03 -1.983E+02 2.678E+02 5.1132 2.795E+03 -2.223E+02 1.508E+03 2.912E+03
Z2 3.024E+03 -1.118E+02 4.889E+02 8.6589 3.098E+03 -1.862E+02 1.642E+03 3.196E+03
MID 2.897E+03 -1.550E+02 3.783E+02 6.9615 2.944E+03 -2.012E+02 1.572E+03 3.049E+03
10400 -119 Z1 -5.775E+04 1.181E+03 -2.060E+04 -72.5215 7.667E+03 -6.423E+04 3.595E+04 6.839E+04
Z2 -5.821E+04 1.207E+04 -1.647E+04 -77.4454 1.573E+04 -6.188E+04 3.880E+04 7.106E+04
MID -5.798E+04 6.623E+03 -1.853E+04 -75.0777 1.156E+04 -6.292E+04 3.724E+04 6.942E+04
219 Z1 -5.772E+04 1.152E+03 2.064E+04 72.4819 7.667E+03 -6.423E+04 3.595E+04 6.839E+04
Z2 -5.819E+04 1.204E+04 1.652E+04 77.4058 1.573E+04 -6.188E+04 3.880E+04 7.106E+04
MID -5.795E+04 6.597E+03 1.858E+04 75.0381 1.156E+04 -6.292E+04 3.724E+04 6.942E+04
10401 -119 Z1 -1.856E+04 7.060E+03 -4.939E+03 -79.4558 7.979E+03 -1.948E+04 1.373E+04 2.446E+04
Z2 -1.714E+04 1.656E+04 -3.881E+03 -83.5131 1.700E+04 -1.758E+04 1.729E+04 2.995E+04
MID -1.785E+04 1.181E+04 -4.410E+03 -81.7177 1.245E+04 -1.849E+04 1.547E+04 2.696E+04
120 Z1 -1.856E+04 7.060E+03 -4.939E+03 -79.4558 7.979E+03 -1.948E+04 1.373E+04 2.446E+04
Z2 -1.714E+04 1.656E+04 -3.881E+03 -83.5131 1.700E+04 -1.758E+04 1.729E+04 2.995E+04
MID -1.785E+04 1.181E+04 -4.410E+03 -81.7177 1.245E+04 -1.849E+04 1.547E+04 2.696E+04
219 Z1 -1.855E+04 7.053E+03 4.957E+03 79.4163 7.979E+03 -1.948E+04 1.373E+04 2.446E+04
Z2 -1.713E+04 1.655E+04 3.905E+03 83.4735 1.700E+04 -1.758E+04 1.729E+04 2.995E+04
MID -1.784E+04 1.180E+04 4.431E+03 81.6781 1.245E+04 -1.849E+04 1.547E+04 2.696E+04
220 Z1 -1.855E+04 7.053E+03 4.957E+03 79.4163 7.979E+03 -1.948E+04 1.373E+04 2.446E+04
Z2 -1.713E+04 1.655E+04 3.905E+03 83.4735 1.700E+04 -1.758E+04 1.729E+04 2.995E+04
MID -1.784E+04 1.180E+04 4.431E+03 81.6781 1.245E+04 -1.849E+04 1.547E+04 2.696E+04

General Remarks
Each element stress component (independently of all other stress components) is transformed into the output coordinate
system before interpolation. It is necessary that this output coordinate system vary slowly over the surface for the
interpolated stresses to be accurate. This approach is certainly valid for stress interpolation over elements that are in a
common plane and it should be adequate for engineering analysis when the difference in slope between adjacent elements

Main Index
 454 MSC Nastran Reference Guide
Mesh Stress Discontinuities at Grid Points

is small. However, when the elements do vary substantially from a smooth surface, the analyst must utilize engineering
judgment relative to the accuracy of the resultant stress data.
Average grid point stresses are only computed at element vertices. For elements with mid-edge grid points, one can estimate
the stress at these mid-edge grid points to be the average of the computed stresses at the grid point at the ends of the edge
in question.

Mesh Stress Discontinuities at Grid Points

Error Estimates for Grid Point Stress Data
Stress Recovery at Grid Points, 442 is devoted to a discussion of the averaging procedures utilized to provide meaningful stresses
at the grid points of finite element models in MSC Nastran. This section will focus on the description of error estimators
to assist the analyst in the identification of regions of his model that may require refinement.
The essence of the previous averaging procedures used to obtain grid point stress data is to:
1. Convert these local element stress components into a common user-defined coordinate system.
2. Average the several values of each stress component to obtain a unique value of the stress component that is to be
associated with the grid point in question.
3. Compute the stress invariants at the grid points from the stress components at the grid points.
In the general case, the stress components are  x ,  y ,  z ,  xy ,  xz , and  yz .

For discussion purposes, the averaging process used to compute the stress components at the grid points can be represented
in the form of equation (8-7).
Ne

g =   Wi ei  (8-7)
i=1
where:

g = weighted mean value of the stress component computed at the grid point.
 ei = value of the stress component in the ith element (i = 1,2,..., N e ) in the neighborhood of the grid
point.  gi is in the same coordinate system as  g .
Wi = weighting factor assigned to the ith element. The sum of the N e values of W i must equal 1. (This
requirement assures that all computed statistics will be unbiased.) The attribute of being unbiased
implies that the variance is equal to the mean square error. Equal weighting (i.e., W i = 1  N e ) is
assumed.

An estimate of the error in a particular component of stress at a grid point can then be computed by assuming that the
values of the corresponding stress components computed for the elements in the neighborhood of the grid point are data

Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points 455
Mesh Stress Discontinuities at Grid Points

points with uncorrelated random errors. It then follows that an estimate of the probable error in the stress component  g
at the grid point is:

Ne
2
g =  NWi ei 
e
i=1 2
  ei 
1 i=1
= ---------- ------------------------
-
Ne Ne
(8-8)

where  ei =  ei –  g . Thus, the probable error  g is the root mean square error in  ei divided by Ne .
It should be noted that the standard deviation or root mean square error is a reasonable measure of precision in many
practical cases, but it is easy to provide examples in which the standard deviation is a poor measure of the concentration of
the distribution about the mean. Equation (8-8) is assumed to provide an approximate error estimator for the grid point
stress data.

Element and Grid Point Stress Discontinuity Output Requests

It should be clear that element and grid point stress discontinuity output can only be obtained if the analyst has requested
grid point stress output via the GPSTRESS command in the Case Control Section of the MSC Nastran input file. The
GPSTRESS command, of course, requires that the analyst define all required SURFACEs and VOLUMEs in the
OUTPUT(POST) portion of the Case Control Section.
To output Grid Point Stress DisCONtinuities:
GPSDCON = ALL
or
SET j = k, l, m, ....
GPSDCON = j
To output ELement Stress DisCONtinuities:
ELSDCON = ALL
or
SET j = k, l, m, ....
ELSDCON = j
The following remarks should be noted relative to the use of the GPSDCON and/or ELSDCON commands.
1. The GPSDCON and ELSDCON commands are honored only in SOLution Sequences 101, 114, 144, and 200.

Main Index
 456 MSC Nastran Reference Guide
Mesh Stress Discontinuities at Grid Points

2. The GPSDCON and ELSDCON commands may be placed above the SUBCASE level or in individual
SUBCASEs and/or SUBCOMs. The use of these commands above the SUBCASE level causes stress discontinuity
data to be output for ALL SUBCASEs in the MSC Nastran input file. If the analyst wishes to restrict stress
discontinuity output to specific SUBCASEs, GPSDCON and/or ELSDCON commands should appear only under
these specific SUBCASEs.
3. The analyst is cautioned that these commands currently produce a substantial amount of output.
4. The GPSDCON and/or ELSDCON commands may be effectively used in restarts. For example, one could include
requests for GPSTRESS data during the initial run of a project and then obtain data on stress discontinuities on a
subsequent restart by including GPSDCON and/or ELSDCON commands in the Case Control Section.
5. Stress discontinuity output will only be provided when grid point stresses have been previously computed by the
methods described in this article. Thus, stress continuity will not be computed when, for example, both plate and
solid elements are connected to a grid point that is involved in stress discontinuity calculations.

Discussion of Error Measures

Large values of error estimators for CQUAD4 and CTRIA3 elements are not uncommon occurrences in statistical error
measures. For example, the mean and variance have direct analogs in engineering mechanics. The mean is analogous to
the centroid of a body and is thus a measure of where the mass is centered. The variance is the second moment about the
mean, and it tends to be small if the majority of the mass is concentrated about the centroid. As the mass is dispersed further
from the centroid, the moment of inertia tends to increase. If the mass is concentrated at the centroid (as it might be in
some idealized lumped mass models), the moment of inertia becomes zero.
Thus, the error estimates for CQUAD4 and CTRIA3 elements will tend toward larger values because the data used to
compute the estimators are dispersed relatively far from the mean. For those elements that provide stress data at vertices,
the estimators will tend toward smaller values because the stress data tend to be concentrated about the mean in well-
designed finite element models.
Generally, the goal is to design a finite element mesh for static analysis so that all important stress gradients will be
adequately represented. If the mesh is not sufficiently detailed, the stress data at element vertices or element centroids will
result in the relatively large values of the error estimators.
Very inaccurate values of these error estimators may be obtained at the edges of defined SURFACEs and on the faces of
defined VOLUMEs. Such edges and surfaces will often correspond with the boundaries of superelements.
In summary, error estimators can, in some cases, be highly inaccurate. Nevertheless, error estimate data is quite useful when
properly interpreted by the analyst.

Grid Point Stresses and Mesh Stress Discontinuities

Grid point stresses may be requested for grid points connected by CTRIA3, CTRIA6, CTRIAR, CQUAD4, CQUAD8,
CQUADR, CHEXA, CPENTA, and CTETRA elements (with linear material properties only). In linear static analysis only,
a mesh stress discontinuity error analysis may then be performed based on the computed element and grid point stresses.
Grid point stresses are not computed for laminar stresses.

Main Index
 Chapter 8: Element Data Recovery Resolved at Grid Points 457
Mesh Stress Discontinuities at Grid Points

The grid point stresses are calculated in the user-specified coordinate system. The following real membrane stresses are
output on request:
 Normal stresses in the x and y directions.
 Shear stresses on the x face in the y direction.
 Angle between the x axis and the major principal axis.
 Major and minor principal stresses.
 von Mises equivalent or maximum shear stress (see STRESS Case Control command).
The following real stresses are output on request for grid points connected to solid elements:
 The three normal stresses.
 The three shear stresses.
 Magnitude and direction of the three principal stresses.
 The mean pressure.
 The von Mises equivalent or octahedral stress (see STRESS Case Control command).
Only real stresses are available at the grid points. The user has the following options:
 The capability to vary output requests by SUBCASE.
 Choice of elements in stress field.
 Choice of topological or geometric interpolation (surface fields only).
 Choice of output coordinate system.
 Choice of output stress fiber locations (surface fields only).
The grid point stresses are requested in the Case Control Section for printing and postprocessing using the Case Control
commands GPSTRESS (Grid Point Stress) and STRFIELD (Stress Field). The user is required to request element stress
output (STRESS = N) for all elements referenced by selected SURFACE and VOLUME commands. The grid point stress
requests a set of surface or volume fields which are defined in the OUTPUT(POST) part of the Case Control Section on
the SURFACE or VOLUME commands, respectively. The postprocessing output requests are separated from case control
and plot requests by the OUTPUT(POST) command.
The general rules for Case Control syntax are the same as for plotting (see Plotting, 467). The grid point stress output is
defined on the SURFACE and VOLUME commands. The SURFACE command defines the following: output coordinate
system, list of elements in the stress surface field, output stress fiber location for stress evaluation and the choice of
topological or geometric interpolation method used to determine the grid point stresses from the elements. The VOLUME
command defines the following: output coordinate system, list of elements in the stress volume, and the choice of DIRECT
and/or PRINCIPAL stresses.
In the linear static structured solution sequences only (SOLutions 101, 114, 144, and 200) mesh stress discontinuities may
be requested to estimate errors due to the coarseness of the finite element mesh. The estimation of the error is computed
for the element and grid point stresses and requested by the ELSDCON and GPSDCON Case Control commands,
respectively. These commands select the desired surfaces and volumes which are defined in the OUTPUT(POST) Section
of Case Control. This output option also requires a corresponding GPSTRESS or STRFIELD request for the desired

Main Index
 458 MSC Nastran Reference Guide
Mesh Stress Discontinuities at Grid Points

surfaces and volumes and a STRESS Case Control command for the elements in those surfaces and volumes. If
PARAM,POST,0 is specified, then this output will also be written to the DBC database.
The following is computed and output for stress discontinuities on shell elements:
 Normal stress discontinuities in the x and y directions
 Shear stress discontinuities on the x-y direction
 Major and minor principal stress discontinuities
 Von Mises and maximum shear stress discontinuities
 Error estimate
The following is computed and output for stress discontinuities on solid elements:
 Normal stress discontinuities in the x, y, and z directions
 Shear stress discontinuities on the x-y, y-z, and x-z directions
 Major and minor principal stress discontinuities
 Von Mises stress and mean pressure discontinuities
 Error estimate

Main Index
 Chapter 9: DMAP
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

9 DMAP

 Introduction

Main Index
 460 MSC Nastran Reference Guide
Introduction

Introduction
MSC Nastran provides a wide variety of solution sequences. Each solution sequence consist of a series of DMAP (Direct
Matrix Abstraction Program) statements. MSC Nastran allows the user to modify the solution sequences or to write his or
her own solution sequences using DMAP. DMAP is a high level language with its own compiler and grammatical rules.
The MSC Nastran DMAP Programmer’s Guide provides detailed information on these modules.

Main Index
 Chapter 10: Database Concepts
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

10 Database Concepts

 Introduction

Delivery Databases

Dbset Deletion

Database Autoassignment

Database Project and Version

Database Archival, Compression, and Transfer

Database Compression

Database Transfer Across Different Computer Types

Main Index
 462 MSC Nastran Reference Guide
Introduction

Introduction
All MSC Nastran runs require a database. A database allows for the storage and retrieval of data: data blocks, parameters,
scratch files, and DMAP object and source files. During a run data may be written onto only one database. This is called
the primary database which is attached automatically or by the INIT MASTER statement. Additional databases may be
attached for “read-only” purposes. This includes a “delivery” database and one or more “located” databases. The delivery
database contains solution sequences and is automatically attached by the SOL Executive statement or the ACQUIRE FMS
statement. “Located” databases are attached by the DBLOCATE statement.
A database is divided into several DBsets which are initialized by the INIT statement. There are two types of DBsets:
permanent and scratch. Permanent DBsets may be saved at the end of the run and reused in a restart run. Scratch DBsets
are automatically deleted at the end of the run. A DBset may be composed of up to twenty concatenated physical files or
DBset members. Each member is assigned to a physical file by the ASSIGN statement.
By default, there are four DBsets - two permanent and two scratch - which are predefined and automatically assigned by
the program. The DBset names of the predefined permanent DBsets are: MASTER and DBALL. The DBset-names of the
predefined scratch DBsets are: SCRATCH and OBJSCR. The DBsets are briefly described below:
 The MASTER DBset is the directory of the database. It contains the names of all DBsets, DBset members and
their physical file names, and a directory of projects, versions, data blocks, parameters, DMAP source and object
files. It also contains the NDDL scheme used to describe the database.
 The DBALL DBset contains all the DMAP data blocks which may be saved permanently for reuse in a subsequent
run. In the solution sequences in which restarts are not possible, (for example, SOL 1) this DBset is empty.
 The SCRATCH DBset is a temporary DBset for all scratch data blocks and files. It has two partitions: one for
DMAP data blocks and one for DMAP module internal scratch files.
 The OBJSCR DBset is a temporary DBset for DMAP compilation.
 If the user desires to create his or her own DMAP source and object files, two other DBsets (USROBJ and
USRSOU) need to be allocated.
 The USROBJ and USRSOU DBsets contain DMAP source and object files that may be saved permanently for
execution in a subsequent run. See the COMPILE Executive Control statement.
If an INIT and ASSIGN statement are not specified for one or more of the predefined DBsets, then the corresponding
statement(s) are provided by default.
ASSIGN MASTER=’dbs-name.MASTER’
INIT MASTER(RAM) LOGICAL=(MASTER(5000))
ASSIGN DBALL=’dbs-name.DBALL’
INIT DBALL LOGICAL=(DBALL(25000))
ASSIGN USROBJ=’dbs-name.USROBJ’
INIT USROBJ LOGICAL=(USROBJ(5000))
ASSIGN USRSOU=’dbs-name.USRSOU’
INIT USRSOU LOGICAL=(USRSOU(5000))
ASSIGN SCRATCH=’temp-name.SCRATCH’
ASSIGN SCR300=’temp-name.SCR300’
INIT SCRATCH(MEM) LOGICAL=(SCRATCH(175000)),
SCR300=(SCR300(175000))

Main Index
 Chapter 10: Database Concepts 463
Introduction

ASSIGN OBJSCR=’temp-name.OBJSCR’
INIT OBJSCR LOGICAL=(OBJSCR(5000))
The numbers inside the parentheses are the maximum sizes of the DBsets in GINO blocks. By default, one member with
the same log-name as the DBset-name is assigned to each DBset. (For a description of a GINO block, DBset-name and
log-name, see the INIT statement). For example, the DBALL DBset has one member also called DBALL. The dbs-name
is determined from the dbs keyword on the nastran command (see Executing MSC Nastran (p. 1) in the MSC Nastran Quick
Reference Guide) and temp-name is generated by the command procedure (see the ). The scratch DBsets are assigned special
names on a scratch disk. (For a description of the dbs-name and temp-name, see the ). The filenames above are those that
would be generated on Linux-type computers. For example, if the name of the input file is called MYJOB.DAT and the
dbs keyword is not specified, then the default filenames for the permanent DBsets would be:
MYJOB.MASTER
MYJOB.DBALL
Also, they would have the same directory or path as the input file.
A physical filename is automatically created for any new DBsets specified by an INIT statement. For example, to create a
new DBset called DBUP for split database operations, specify:
INIT DBUP
The physical filename will then be: MYJOB.DBUP. For another example, to create two members for DBALL, specify
INIT DBALL LOGI=(DB1 ,DB2)
Their physical filenames will be MYJOB.DB1 and MYJOB.DB2.
The ASSIGN statement may be used to override these default names. For example, to rename MYJOB.DBUP to
FENDER.DBUP, specify
ASSIGN DBUP=’FENDER.DBUP’
INIT DBUP
To rename MYJOB.DB1 and MYJOB.DB2, specify
ASSIGN DB1 =’FENDER.DBALL1’
ASSIGN DB2=’FENDER.DBALL2’
INIT DBALL LOGI=(DB1,DB2)
The INIT statement may be used to change the maximum size of the predefined DBsets. (The size may be specified in
blocks, kilowords, megawords, kilobytes, or megabytes.) For example, to specify a size of 50,000 blocks for DBALL,
INIT DBALL LOGICAL=(DBALL(50000))
or 50,000 bytes,
INIT DBALL LOGICAL=(DBALL(50KB))
The INIT and ASSIGN statement may be used to specify more than one member for a DBset which may exist on different
physical devices.
ASSIGN DB1 =’physical file name of DB1 on disk 1’
ASSIGN DB2=’physical file name of DB2 on disk 2’
INIT DBALL LOGICAL=(DB1(5000),DB2(5000))

Main Index
 464 MSC Nastran Reference Guide
Delivery Databases

Delivery Databases
The Delivery Database contains the solution sequences. This database, in addition to the primary database, is automatically
assigned in order to execute an MSC Nastran solution sequence. For example, on Linux-type computers, the filenames of
the database are
SSS.MASTERA
SSS.MSCOBJ
SSS.MSCSOU
If the solution sequence is not an MSC Nastran solution sequence, as in the case of a user’s solution sequence; then the
ACQUIRE FMS statement may be used to select the appropriate database. ACQUIRE NDDL selects the SSS database.
The ACQUIRE statement is required when the DBLOAD FMS statement is used with the ENDJOB FMS statement.

Dbset Deletion
The disposition of permanent DBsets in the primary database at the end of the run is controlled in several ways.
 If scr=yes is specified on the nastran command (see Executing MSC Nastran (p. 2) in the MSC Nastran Quick Reference
Guide) or the INIT MASTER(S) statement appears in the FMS Section then all DBsets in the primary database
are deleted.
 An individual DBset member is automatically deleted at the end of the run by specifying the TEMP keyword on
the ASSIGN statement. For example, to delete the USROBJ and USRSOU DBsets at the end of the run, specify:
ASSIGN USROBJ=’A’ TEMP
ASSIGN USROBJ=’B’ TEMP
However, DBsets deleted in this manner or manually cannot be recreated in a future run.
 The DBSETDEL FMS statement provides a better method than the TEMP keyword on the ASSIGN statement
except that the DBsets are deleted at the beginning of the run. The deleted DBsets may be recreated in a future
run. For example:
DBSETDEL USROBJ,USRSOU

Database Autoassignment
When a database is created, the physical filenames of all DBset members are loaded into the MASTER DBset. Therefore,
when using DBLOCATE or RESTART, it is sufficient to assign the MASTER DBset member; and all other DBset
members previously initialized; that is, DBALL, DBUP, etc., are automatically assigned by the program. For example, on
restart, the following input will automatically assign the DBALL, USROBJ, and USRSOU DBsets:
ASSIGN MASTER=’filename of the MASTER DBset member’
RESTART
If DBLOCATE is used, then:
ASSIGN MSTR1 =’filename of the MASTER DBset member’
DBLOCATE LOGI=MSTR1

Main Index
 Chapter 10: Database Concepts 465
Database Project and Version

The autoassignment feature cannot assign the physical file of a DBset member that has been renamed. This is because the
MASTER DBset member contains the old physical filename. Therefore, an additional ASSIGN statement is required to
specify the new filename of the DBset member. If in the previous restart example DBALL is renamed, then the following
input is required:
ASSIGN MASTER=’filename of the MASTER DBset member’
ASSIGN DBALL=’new filename of the DBALL DBset member’
RESTART
In split database operations, it is often necessary to assign only some of the DBsets because the others are “offline.” The
NASTRAN AUTOASGN statement is used to specify the types of databases that will be autoassigned:
 AUTOASGN = 0: No databases will be autoassigned. This includes the primary database, Delivery Database, and
any located databases.
 AUTOASGN = 1: All databases will be autoassigned. This is the default.
 AUTOASGN = 2: Only the Delivery Database will be autoassigned.
 AUTOASGN = 4: Only located databases will be autoassigned.
If some combination of the options above is desired, then sum their values. For example, if the delivery and located
databases are to be autoassigned, then specify NASTRAN AUTOASGN = 6.

Database Project and Version

All data stored on the primary database is assigned to a particular project identification string (ID) and version identification
number (ID). The current project ID is specified on the PROJECT FMS statement. The current version ID is an integer
value assigned by the program each time the database is attached as the primary database. When a primary database is
initialized or a new project is specified, then all items stored on the database during the run are stored under a version ID
of integer 1. Then, in subsequent runs a new version is created with an ID equal to the last version ID incremented by one.
If the RESTART FMS statement is specified, then a new version of data is created which is equivalent to the last version.
In other words, all items in the old version are made available for reuse in the new version.

Database Archival, Compression, and Transfer

Database archival, compression and transfers are performed using the DBUNLOAD and DBLOAD FMS statements. The
user may select all items on the database or selected items according to the WHERE clause, see “WHERE and CONVERT
Clauses” in the File Management Statements (p. 43) in the MSC Nastran Quick Reference Guide.

Database Compression
When data is deleted from the database, its space is released for the storage of new data and this old space will be reused
before any new space is used. However, even if the space is not reused, the size of the database is not reduced. After several
restarts or the execution of the DBCLEAN FMS statement, the database may contain a significant amount of released or
“dead” space. There are two methods of removing dead space or “compressing” the database.

Main Index
 466 MSC Nastran Reference Guide
Database Transfer Across Different Computer Types

The first method involves the DBUNLOAD and DBLOAD FMS statements, and the examples shown above will also
perform a database compression. The second and simplest method involves the DBLOCATE statement specified with the
COPY keyword. For example, if the database was created with SOLs 100 through 200, then the input file is:
ACQUIRE NDDL
ASSIGN MYDB=’physical file name of the MASTER DBset’
DBLOCATE LOGI=MYDB COPY
ENDJOB
It is also possible to compress an individual DBset and change its maximum size. For example, the input file is:
ACQUIRE NDDL
ASSIGN MYDB=’physical file name of the MASTER DBset’
DBLOCATE LOGI=MYDB COPY WHERE(DBSET=’DBALL’)
INIT DBALL LOGICAL=(DBALL(10000))
ENDJOB

Database Transfer Across Different Computer Types

It is often desirable to copy a database from one computer to a dissimilar computer (e.g., copying a database created on a
Windows computer to a Linux workstation). A method is available using the DBUNLOAD and DBLOAD FMS
statements. The DBUNLOAD statement is first used to convert the database to a “neutral” file on the first computer. This
neutral file is now suitable for a simple copy to the second computer or may be accessed through a network if the two
computers are so connected. On the second computer, the DBLOAD statement is then used to convert the neutral file back
to a database. It is also necessary to execute the solution sequence DBTRANS (or 190) as described under Database
Migration, because some data blocks cannot be converted to a neutral format. (In SOL DBTRANS; data blocks EST,
EMAP, ETT, PTELEM, SLT, KDICT, KELM, MDICT, MELM, BDICT, BELM, and ACPT are regenerated.)
The ASSIGN statement is also required to assign the physical filename of the neutral file and specify that the file is
formatted since the default format (UNFORMATTED) is not transferable across dissimilar computers. The following
example unloads a database created by one of the unstructured solution sequences on the first computer to a neutral file:
ASSIGN DBUNLOAD=’physical filename of neutral file’ FORMATTED
DBUNLOAD FORMAT=NEUTRAL
ENDJOB
Then the neutral file is copied to, or accessed through a network by, the second computer:
ASSIGN DBLOAD=’physical filename of neutral file’ FORMATTED
DBLOAD FORMAT=NEUTRAL
SOL DBTRANU
CEND
With the DBTRANS solution sequence it is assumed that the neutral file contains all of the data from a completed and
successful analysis. If not, then the solution sequences may fail and the transfer may not be successful. In this case, it is
possible to alter DBTRANS with the ALTER statement to obtain the appropriate data.
If the database was created in a heat transfer analysis then NASTRAN HEAT = 1 must be specified.
If the database on the first computer was not created by a MSC Nastran solution sequence, then it may be necessary to
develop a new transfer solution sequence.

Main Index
 Chapter 11: Plotting
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

11 Plotting

 General Capability

Superelement Plotting

Postprocessor

Main Index
 468 MSC Nastran Reference Guide
General Capability

General Capability
MSC Nastran can generate the following kinds of plots:
 Undeformed geometric projections of the structural model.
 Static deformations of the structural model by either displaying the deformed shape (alone or superimposed on the
undeformed shape), or displaying the displacement vectors at the grid points (superimposed on either the
deformed or undeformed shape).
 Modal deformations resulting from real or complex eigenvalue analysis by the same options stated in 2 above.
Complex modes for flutter analysis may be plotted for any user-chosen phase lag.
 Deformations of the structural model for transient response or frequency response by displaying either vectors or
the deformed shape for specified times or frequencies.
 X-Y graphs of transient response or frequency response.
 V-F and V-G graphs for flutter analysis.
 Contour plots of displacements, temperature and stress on the structure.
Structural plots (items 1 through 4) are discussed in the Case Control Commands in the MSC Nastran Quick Reference Guide.
X-Y plots (item 5) are discussed in X-Y PLOT Commands. Requests for structure plots or X-Y plots are made in the Case
Control Section by submitting a structure plot request or an X-Y output request. The optional PLOTID command is
considered to be part of the plot request although it must precede any OUTPUT(PLOT), OUTPUT(XYOUT) or
OUTPUT(XYPLOT) commands. See PLOTID (Case) in the MSC Nastran Quick Reference Guide for a description of the
command.
Plot requests are separated from Case Control by the OUTPUT(PLOT), OUTPUT(XYPLOT) or OUTPUT(XYOUT)
commands. Data above this command (except PLOTID) will not be recognized by the plotter, even though it may have
the same name (for example, the SET command).

Superelement Plotting
Plotting can occur at four different places in the superelement solution sequences. There are two plot commands, SEPLOT
and SEUPPLOT, used with other Case Control and PARAM commands to control the type of plot to be prepared. A flow
chart of the solution process is given in Figure 11-1.
Undeformed structure plots are made early during Phase 0 in the structured solution sequences and during Phase l in the
unstructured solution sequences. They can be made for either one superelement only, or for a superelement and all of its
upstream members, as controlled by PARAM,PLOTSUP (see Figure 11-1). Plots are made for superelements selected by the
SEMG command, and listed in a plot request headed by SEPLOT SElD, where SEID is the superelement identification
number. In the structured solution sequences, the SEMG command is not required and all superelements which appear on
SEPLOT commands will be plotted (see Figure 11-1, block 0). Undeformed plots can be used to check geometry and
connectivity, and do not require the presence of property or material commands. In the unstructured solution sequences, a
branch to the end of the loop immediately after the plot module can be made by using the PARAM,PLOT,-1 command.
In dynamic analysis, solution set XY-plots are requested by use of SEPLOT 0. In the data recovery phase, XY-plots and
deformed structure plots for elements in one superelement only are requested by the SEPLOT command. Deformed
structure plots for a superelement and all its upstream plots are requested with the SEUPPLOT command.

Main Index
 Chapter 11: Plotting 469
Superelement Plotting

XY-plots and deformed structure plots are regarded as output requests and will result in automatic execution of the data
recovery loop as is required to produce the plots requested. For example, the command SEUPPLOT 0 will result in data
recovery being performed on the entire model, even in the absence of any other output requests. For SEUPPLOT requests,
if SUBCOM or SYMCOM subcases are used, each superelement must have identical SUBCOM and SYMCOM structure.

Main Index
 470 MSC Nastran Reference Guide
Superelement Plotting

START

0. Bookkeeping and Control Case

Control
Commands

PARAM,PLOTSUP,0 PARAM,PLOTSUP,-1

Prepare plotting data for Prepare plotting data for

SEPLOT
superelement and all upstream one superelement only
SEID
members (default)

1. Undeformed Plots

I. Generation, Assembly,
Reduction Loop

Generate, assemble, and

reduce one superelement

(Continued on following pages)

Figure 11-1 Phase 0 and I Superelement Plot Control In the Structured Solution Sequences

Main Index
 Chapter 11: Plotting 471
Superelement Plotting

I. Generation, Assembly, Case

Reduction Loop Control
Commands

PARAM,PLOTSUP,0 PARAM,PLOTSUP,-1

Prepare plotting data for Prepare plotting data for SET n=

superelement and all upstream one superelement only SEID(s)
members (default) SEMG=n
SEPLOT
SEID(s)

1. Undeformed Plots

PARAM,PLOT,-1
PARAM,PLOT,0
(Default)

Generate, assemble, and

reduce one superelement

(Continued on following page)

Figure 11-2 Phase l Superelement Plot Control In the Unstructured Solution Sequences (cont.)

Main Index
 472 MSC Nastran Reference Guide
Superelement Plotting

II. Residual Structure Only Phase

Eigensolutions, Forced
Response
SEPLOT 0

2. Solution Set XY Plots

III. Data Recovery Loop

Recover interior
displacements and
stresses
SEPLOT
SEID
2. XYPLOTS
Deformed structure
plots (local)

IV. Upstream Plot Loop

Assemble upstream plot

vectors SEUPPLOT
SEID

2. Deformed structure
plots (upstream)

END
Figure 11-3 Phase II, Ill, and IV Superelement Plot Control in All Solution Sequences (cont.)

Main Index
 Chapter 11: Plotting 473
Superelement Plotting

Structure Plotter Coordinate System and Orthographic Projection

In order to define the coordinates of an orthographic projection of the structural model, an R, S, T plotter coordinate
system is defined as shown in Figure 11-4.




Figure 11-4 Plotter Coordinate System

The S-T plane is taken as the plan of projection. The structural model is defined in the basic coordinate system, which is
denoted as the X, Y, Z coordinate system. The user specifies the position of the structural model with respect to the S-T
projection plane by the angles  ,  , and  . These angles position the X, Y, Z coordinate systems with respect to the R,
S, T coordinate system. The two coordinate systems are coincident for  =  =  = 0 . The sequence in which the
rotations are taken is crucial and has been arbitrarily chosen as  , the rotation about T-axis followed by  , the rotation
about the S-axis, followed by  , the rotation about the R-axis. Normally,  is not used since it does not affect the
appearance of the S-T projection, but only its orientation on the page.
The orthographic projection is obtained by computing the S and T coordinates of each point having coordinates X, Y, Z
from the transformation equation

 R   X 
   
 S  =  A   A   A   Y 
   
 T   Z 
where:

Main Index
 474 MSC Nastran Reference Guide
Superelement Plotting

cos  – sin  0
 A  = sin  cos  0 ,
0 0 1

cos  0 sin 
 A  = 0 1 0 ,
– sin  0 cos 
and

1 0 0
 A   = 0 cos  – sin 
0 sin  cos 
In order to illustrate clearly the orientation process, Figure 11-5 shows a rectangular parallelepiped as it is rotated through
the  ,  ,  sequence. The final S-T plane shown in Figure 11-5b contains the desired orthographic projection.

Main Index
 Chapter 11: Plotting 475
Superelement Plotting

S Y S
Y


X
R X
R

T Z Z T T Z Z T

Y X X Y X Y

S R S R

(a) Initial Position (b)  = 30

Main Index
 476 MSC Nastran Reference Guide
Superelement Plotting

Y S Y S

X X
Z Z

R
R

Z T
T Z T T



Y Z Y
Y
(c)  = 30  = 30  = 30
(c) S = 30R  = 30 S R X
X X
(Plotter will plot on the S-T Plane)

Figure 11-5 Plotter -- Model Orientation

Perspective Projection
In addition to the three angular relationships required for orthographic projection, the perspective projection requires
knowledge of the vantage point in the R-S-T system (that is, the three coordinates of the observer), and the location of the
projection plane (plotter surface). The vantage point is selected by the user (or automatically by the program), and lies in
the positive R half space as shown in Figure 11-6. The projection plane is chosen to lie between the observer and the S-T
plane.

Main Index
 Chapter 11: Plotting 477
Superelement Plotting

T S

Structural Model

T' Projection Plane S'

do

Vantage Point
 R o S o T o 

Figure 11-6 Perspective Projection Geometry

For each point, the coordinates S and T T’ on the projection plane (see Figure 11-6) are obtained from the orthographic
projection coordinates R, S, T by

   
 S'   S o  do  S – So 
 = ---------------- 
+ R 
 T'   T o  o – R  T – To 
   

where R o , S o , T o are the coordinates of the vantage point and d o is the separation distance between the vantage point
and the projection plane.

Main Index
 478 MSC Nastran Reference Guide
Superelement Plotting

Stereoscopic Projection
The stereoscopic effect is obtained through the differences in images received by the left and right eyes. Each is a perspective
image, but with a different vantage point. The two vantage points are separated horizontally by 70 mm (2.756 inches), the
nominal ocular separation standard used in commercially available stereoscopic cameras and viewers. To plots are produced
for viewing with a stereoscopic viewer.

Projection Plane to Plotter Transformations

Since the plotter surface is defined differently for each of the plotters used in MSC Nastran, and since it is desirable to
minimize the amount of special coding for each plotter, a common interface with the plotter routines is provided wherein
the plotted surface is assumed to have a lower left corner defined by x, y coordinates (0,0) and an upper right corner defined
by x, y coordinates (1,1). Plotter utility routines are utilized to convert from this system to each individual plotter.
Since the coordinates of points in any of the projection planes previously discussed may have arbitrary numerical values, a
linear mapping of the form

 
 x   S   xo 
  =  A ij   + 
 y   T   yo 
 

is used, where the transformation constants  A ij  and x o , y o are determined as shown below so as to fill that portion of
the x, y space required by the suer or automatically selected by the program. In the case of a perspective projection,  S 
and  T  are replaced by  S '  and  T'  . Since we desire no distortion of the plotted object and the orientation has
already been specified, we set

 A ij  = A 1 0
01
Let
1
M = ---   T max – T min  –  S max – S min  
2

Then, to fill the available x, y space we require, for M  0 ,

Main Index
 Chapter 11: Plotting 479
Postprocessors

0 = AT min + y o
1 = AT max + y o
(11-1)
m = AS min + x o
1 – m = AS max + x o

or, for M  0

m = AT min + y o
1 – m = AT max + y o
(11-2)
0 = AS min + x o
1 = AS max + x o

where m is the horizontal margin for M  0 and the vertical margin for M  0 .

From equations (11-1) and (11-2) the quantities m , A, x o , y o , and thus the plotter coordinates, are determined.

Deformation Scaling
In plotting deformed structures, the components of displacement are added to the coordinates of the undeformed grid
points to obtain the coordinates of the grid points of the deformed structure. Since the numerical magnitude of the physical
structural displacements is usually much smaller than the size of the structure itself, additional scaling must be performed
on the displacement vectors in order to obtain a viewable plot. In MSC Nastran this is done by the user who specified a
value for the magnitude of the maximum structural deflection in units of length of the undeformed structural model. Thus,
if max  T max – T min S max – S min  were, say, 1000 units, a specification of 50 would result in a plotted maximum
deformation equal to 5% of the maximum plot size. In addition, the scale of the deformed structure, described in Projection
Plane to Plotter Transformations, 478, is reduced by 5% to accommodate the deformation vectors.

Postprocessors
The plotps postprocessor reads plotting commands from a single MSC Nastran binary or neutral format plot file and
produces a file which can be printed or viewed on a PostScript device. Each image will be oriented to best correlate the plot’s
page size with the printable page size (assumed to be 7.5 inches wide and 10.0 inches tall). Command line arguments can
be specified in any order, command keywords can be abbreviated by truncation. These command keywords are discussed
in PLOTPS (p. 35) in the .

Main Index
 480 MSC Nastran Reference Guide
Postprocessors

Examples
 Translate a binary format plot file into PostScript:
mscplotps example
 Translate a neutral format plot file into PostScript:
mscplotps example

Main Index
Chapter 12: Interface With Other Programs
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

12 Interface With Other Programs

 Introduction 482

POST Case Control Command 482

PUNCH Output 488

Patran Support of Heat Transfer Analysis 519

The Nastran HDF5 Result Database (NH5RDB) 521

NH5RDB Data Access 532
 482 MSC Nastran Reference Guide
Introduction

Introduction
MSC Nastran provides for direct support of interface to MSC’s interactive graphics products, MSC Access, Patran, and
SDRC I-DEAS programs, and to other outside vendor programs.
For Patran, the support is provided via the MSC Nastran DBC module, which creates a “graphics” database. The interface
to the DBC module is described in the POST (p. 979) in the MSC Nastran Quick Reference Guide.
The graphics database created by the MSC Nastran DBC module, can also be read by the MSC Access library of object
routines. This object library can be linked with a user-created program that extracts data from the database. This is an
indirect method for outside vendors to obtain MSC Nastran model and results information. See the MSC Access Reference
Guide for more information.
Communication with Patran, SDRC I-DEAS, LMS International’s MSC_NF interface, and Dynamic Design Solutions’
SYSTUNE interface, is through the MSC Nastran OUTPUT2 module, which creates a FORTRAN readable file that is
converted by these outside programs to their own particular data formats. The user interface and the supported data for
these outside programs are described in POST (p. 979) in the MSC Nastran Quick Reference Guide (-1, -2, -4, and -5).
Sometimes just a direct punch output file is useful for creating input to external programs.

POST Case Control Command

Introduction
Finite element analysis of large structural components sometimes requires the consideration of many different loading
scenarios to complete the simulation. This can be especially true for static analysis, which may require several thousand
different load cases to be evaluated. During this process, recovery and postprocessing of large amounts of response data
usually occurs. Many analysis organizations use commercially available pre-/postprocessors, like Patran, to graphically
display results data. These programs typically read the results data from a fortran file generated by the OUTPUT2 module
in MSC Nastran. This file contains all of the output requested by the case control data recovery requests in each and every
subcase. Occasionally, output data is required for a subcase that does not need to be handled by the post-processing
program.

Theory
The amount of data recovered for each subcase is dependent upon the case control commands that are present. If a request
is placed above all subcases, that request applies to all of the subcases. Data is generated and produced one time for all
subcases by examining the case control requests in each subcase. Each separate type of response is stored in its own data
block. For example, displacement results are computed and stored separately from element stress results. These data blocks
will include results for any subcase that contains a data recovery request with a print, plot or punch destination. The data
blocks are then written to an external file by the OUTPUT2 module. There is no way to control the placement of the results
data on the output file except through the case control request for that data. For example, if there are three subcases and
displacement results for the first and third subcases are required for post-processing, but not the second subcase, one could
simply request displacement output for only the first and third subcases. However, if punched displacement results for the
second subcase were also required, displacements for all three subcases would be placed on the output file, even though

Main Index
 Chapter 12: Interface With Other Programs 483
POST Case Control Command

results for the second subcase are not needed nor wanted in the postprocessor. This applies to all output quantities requested
by case control commands. There is no way to handle this situation except to run the analysis twice and change the data
recovery requests. A simple method is needed to exercise more control over the contents of the output file.
The new POST Case Control command gives the user more control over the contents of the output file produced by the
MSC Nastran program when a PARAM,POST Bulk Data entry is present in the input. It does not affect the contents of
the results data blocks, only the amount of data that is transferred from those data blocks to the output file. This is
accomplished by giving the OUTPUT2 module access to the case control requests for each subcase. The POST command
is examined for the output requested in each subcase. Only that output requested by the POST command is placed on the
output file for the subcase. Note that the data must be made available by the presence of a data recovery command in the
subcase. For example, the POST command cannot output stress data if no stress data has been requested.
The operations associated with the POST Case Control command are closely related to the operations described under the
PARAM POST discussion in Parameters (p. 793) in the MSC Nastran Quick Reference Guide. All of the DMAP parameters
discussed in the QRG related to the POST parameter are honored and can be used to control the output produced. The
POST parameter value itself is supplied by the POST Case Control command through selection of one of the supported
post-processor programs shown in the following table.

ppname Product PARAM,POST,Value

PATRAN MSC/PATRAN V3 -1
SDRC SDRC IDEA-S -2
NF MSC/LMS NF -4
FEMTOOLS DDS/FemTools -5
UNIGRAHICS EDS/Unigraphics -6

Inputs
The POST Case Control command is designed to give the user a way to limit the output of data recovery results for a
particular subcase. The general format of the POST command is:
POST [ ppname ] [ TOFILE furn ] oplist
The POST command recognizes only a sub-set of all of the results data that can be produced by MSC Nastran. This sub-
set is summarized in the following table.

Output Item oplist Keyword Case Command

Displacements [NO]DISPLACE DISP
Forces of Single Point Constraint [NO]SPCFORCE SPCFORCE
Element Forces [NO]FORCES ELFO/FORCE
Element Stresses [NO]STRESS ELST/STRESS

Main Index
 484 MSC Nastran Reference Guide
POST Case Control Command

Output Item oplist Keyword Case Command

Element Strain Energy [NO]ESE ESE
Grid Point Force Balance [NO]GPFORCE GPFORCE
Stress at Grid Points [NO]GPSIGMA STRESS
Strain/Curvature at Grid Points [NO]GPEPSILON STRAIN
Composite Element Failure Indices [NO]PLYFAILURE STRESS
Element Kinetic Energy [NO]EKE EKE
Element Energy Loss [NO]EDE EDE
Multi-point Constraint Forces [NO]MPCFORCE MPCFORCE
Composite Lamina Stresses [NO]PLYSIGMA STRESS
Composite Lamina Strains [NO]PLYEPSILON STRAIN
Element Strains [NO]STRAIN STRAIN
Grid Point Stresses [NO]GPSTRESS GPSTRESS
Grid Point Strains [NO]GPSTRAIN GPSTRAIN
Applied Loads [NO]LOAD OLOAD
No items to be output NONE ----------------

The POST command permits the optional selection of the target postprocessor program as well as the target fortran file
unit reference number where the output will be stored.

Outputs
There are no special output considerations when using the POST Case Control command.

Guidelines and Limitations

The new POST Case Control command is intended primarily for static analysis. It is considered to be an extension of the
parameters associated with the POST DMAP parameter discussed in Parameters in the MSC Nastran Quick Reference Guide.
The POST command recognizes specification of any of the commercial post-processor products currently supported by the
MSC Nastran solution sequences. The following guidelines should be kept in mind when using the POST command.
 Use of the POST command is entirely optional. If it is not used and a PARAM,POST,x entry is present in the
input file, a postprocessor output file is generated just as in previous releases of the program.
 If any POST command operations are to be performed, a POST command must be specified above all subcases.
 When a POST command is encountered, a “PARAM,POST,x” entry is inserted in the Case Control at that point.
All other POST-related DMAP parameters are available to the user to control output of data as before.
 Once a particular postprocessor program is specified, it cannot be changed in any subsequent subcases.

Main Index
 Chapter 12: Interface With Other Programs 485
POST Case Control Command

 The POST command does not allow a change in specification of the output fortran unit reference number if it is
associated with a form=formatted OUTPUT2 file.
 POST command options are limited to SORT1 formatted output data at the present time.
 As is presently the case, only data that has been generated by the presence of a case control data recovery request
command will be available for postprocessing output.
 There is no control over any geometry data output or over any data blocks that are not generated from a case
control request. These data blocks will continue to be placed on the OUTPUT2 unit specified by the OUNIT2
parameter depending upon the value of the POST parameter (and the values of any related parameters).
 The placement of the POST command above the subcase level causes a cumulative effect on POST commands in
subsequent subcases. Any options specified above the subcase level propagate down into the POST command
within a subsequent subcase. Thus, if a POST command specifies NODISP (no displacement output wanted)
above the subcase level, then a POST command with the DISP option would be required within a subcase to
produce any output on the OUTPUT2 file for displacements. This also implies that changing the OUTPUT2 file
unit reference number with the TOFILE option in a subcase causes all output quantities currently scheduled for
output to be switched to the new unit number, not just those in the option list for the current POST command.

Demonstration Example
A simple model is presented to demonstrate the usage of the POST Case Control command to eliminate unwanted output
from being placed on the postprocessor data file generated by MSC Nastran. The model data itself is not important for this
example. It is the subcase structure that is the essence of the discussion. As an example of how this new feature might be
used, consider the case of a static simulation performed on a model requiring three load cases. The output required is
displacements at three grid points in the model. Furthermore, the output for the second load case must also include
punched displacement data for all grid points. Displacement results for subcases 100 and 300 are to be placed on the default
OUTPUT2 file for use by the Patran postprocessor. No output for subcase 200 is required for post-processing by Patran.
The subcase structure might look something like:
SET 1000 = 10, 11, 12
DISP = 1000
$
POST$ use all defaults: Patran v3.0, unit=12
$
SUBCASE 100
LOAD = 100
$
SUBCASE200
LOAD = 200
DISP(PUNCH) = ALL
POST NODISP $ stop any displacement output from going to POST file
$
SUBCASE300
LOAD = 300
In this example, the presence of the POST command above all of the subcases indicates that all output requested (DISP for
SET 1000 is the only output requested and generated) is to be stored on the default OUTPUT2 fortran unit for use by

Main Index
 486 MSC Nastran Reference Guide
POST Case Control Command

MSC.Patran V3.0. The presence of the POST command in SUBCASE 200 with the NODISP option prevents the large
output produced by the DISP(PUNCH)=ALL request in SUBCASE 200 from being placed on the OUTPUT2 file.

Model Description
The model used for this example is very simple as the intent is to focus on the subcase structure. It is a cantilever plate
structure consisting of only six CQUAD4 shell elements and fourteen grid points. There are three subcases used to apply
loads and recover displacements.

Input File
$
$***********************************************************************
$
$Minor Enhancement A01130/A01131 - new POST case control capability
$VERSION: 2004
$TEST DECK NAME: postrnex.dat
$
$PURPOSE:
$ Simple case control setup to demonstrate usage of the new POST Case
$ Control Command for the Release Notes
$
$DESCRIPTION:
$ An OP2 file is to be generated for the MSC.Patran V3.0 post-
$ processor. Three subcases are present in case control.
$ Displacement output is requested above the subcase level.
$ Displacements for subcases 100 and 300 are to be placed on the
$ default OP2 file. Displacements for subcase 200 are to be placed on
$ system punch file but not on the OP2 file.
$
$EXPECTED RESULTS:
$ The following UIM 4114 should be present in the .f06 output:
$
$*** USER INFORMATION MESSAGE 4114 (OUTPX2)
$ DATA BLOCK OUGV1 WRITTEN ON FORTRAN UNIT 12, TRL =
$ 101 0 160 0 0 0 0
$ (MAXIMUM POSSIBLE FORTRAN RECORD SIZE = xxxxx WORDS.)
$ (MAXIMUM SIZE OF FORTRAN RECORDS WRITTEN = 146 WORDS.)
$ (NUMBER OF FORTRAN RECORDS WRITTEN = 30 RECORDS.)
$ (TOTAL DATA WRITTEN FOR DATA BLOCK = 381 WORDS.)
$ (DATA FOR 2 CASE(S) WAS OUTPUT TO THIS UNIT. CASE ID LIST FOLLOWS.)
$ 100 300
$
$***********************************************************************
$
ID MSC,POSTRNEX
SOL 101
TIME 5
CEND
TITLE = POST CASE CONTROL COMMAND EXAMPLE FOR RELEASE NOTES
SUBTITLE = DEMONSTRATE POST COMMAND USAGE
$
SPC = 1
SEALL = ALL

Main Index
 Chapter 12: Interface With Other Programs 487
POST Case Control Command

SET 1000 = 10,11,12

DISP = 1000
$
$ place post command above subcase level. Nothing specified causes
$ all defaults to be taken:
$ POST PATRAN TOFILE 12 (all outputs requested via CC commands)
$ or, for the example here where only DISP is requested,
$ POST PATRAN TOFILE 12 DISP
$ Note that POST command options specified above the subcase level
$ flow down into all of the subcases.
$
POST patran tofile 12
$
$---------------------------------------------------------------
subcase 100
LOAD = 100
$---------------------------------------------------------------
subcase 200
LOAD = 200
DISP(PUNCH)=ALL $ request punched displacement output for all grids
POST NODISP $ no displacement output to default OP2 file.
$---------------------------------------------------------------
subcase 300
LOAD = 100
$---------------------------------------------------------------
BEGIN BULK
CQUAD4 1 1 1 4 5 2
CQUAD4 2 1 5 6 3 2
CQUAD4 3 1 4 7 8 5
CQUAD4 4 1 8 9 6 5
CQUAD4 5 1 7 10 11 8
CQUAD4 6 1 11 12 9 8
$
PSHELL 1 100 0.001 100
MAT1 100 10.7+6 0.33
$
FORCE 1 10 1. 10.
FORCE 1 11 1. 20.
FORCE 1 12 1. 10.
GRID 1 0. 1.75 0.
GRID 2 0. 2.275 0.
GRID 3 0. 2.8 0.
GRID 4 0.6 1.75 0.
GRID 5 0.6 2.275 0.
GRID 6 0.6 2.8 0.
GRID 7 1.2 1.75 0.
GRID 8 1.2 2.275 0.
GRID 9 1.2 2.8 0.
GRID 10 1.8 1.75 0.
GRID 11 1.8 2.275 0.
GRID 12 1.8 2.8 0.
GRID 13 0.0 1.75 1.
GRID 14 1.0 1.75 1.
LOAD 100 1. 1. 1
LOAD 200 1. 2. 1
LOAD 300 1. 3. 1
PARAM AUTOSPC YES

Main Index
 488 MSC Nastran Reference Guide
PUNCH Output

SPC 1 1 13456
SPC 1 2 123456
SPC 1 3 13456
ENDDATA

PUNCH Output
Description of the MSC Nastran System Cell
System cell 210 is used to control the punch formatting, OFP tables being processed, and determine if the line number is
placed in columns 73 through 80. It should be noted that this only effects the OFP punch capability and does not impact
other punch options such as ECHO = PUNCH, Composite Element, synthesized element, and material property values,
and other areas addressed outside of the OFP environment. The system cell values are:

Value Selection
0 “OLD” Punch, default in V2001 and earlier
1 “NEW” Punch, default in V2004 and uses NDDL
2 Same as 1 except the line number is eliminated

The term “OLD” punch refers to the algorithm used to process the entries of the output table. In MSC Nastran releases
prior to the current release, assumed was that the first item of an entry could be either integer or real depending upon the
table sort condition. The remaining entry items were assumed to be real. This led to incorrect translation of the binary when
the entry item was either integer or character in data value. This could cause “not a number” or numerical range value
exceptions in the translated punch file.
The “New” punch has been implemented in the time frame of the new executive system and uses the Nastran Data
Definition Language (NDDL) interface to correctly translate the entry data items. The implication is that the reading
program must better understand the true format of the entry within an output table class.

Review of OFP Table Codes and Punch Header Conventions

Table 12-1 Structural Punch Table Support and Descriptor Entry

Table Code Table Support Table Content and Descriptor
1 Old $DISPLACEMENTS
2 Old $OLOADS
3 Old $SPCF
4 Old $ELEMENT FORCES

Main Index
 Chapter 12: Interface With Other Programs 489
PUNCH Output

Table 12-1 Structural Punch Table Support and Descriptor Entry (continued)
Table Code Table Support Table Content and Descriptor
5 Old $ELEMENT STRESSES
6 None $EIGENVALUE SUMMARY
7 Old $EIGENVECTOR
8 None $GRID POINT SINGULARITY TABLE
9 None $EIGENVALUE ANALYSIS SUMMARY
10 Old $VELOCITY
11 Old $ACCELERATION
12 Old $NON-LINEAR-FORCES
13 None $GRID POINT WEIGHT OUTPUT
14 Old $EIGENVECTOR (SOLUTION SET)
15 Old $DISPLACEMENTS (SOLUTION SET)
16 Old $VELOCITY (SOLUTION SET)
17 Old $ACCELERATION (SOLUTION SET)
18 Old $ELEMENT STRAIN ENERGIES
19 Old $GRID POINT FORCE BALANCE
20 None $STRESS AT GRID POINTS
21 None $STRAIN/CURVATURE AT GRID POINTS
22 None $ELEMENT INTERNAL FORCES AND MOMENTS
23 None $ELEMENT ORIENTED FORCES
24 None $ELEMENT PRESSURES
25 None $COMPOSITE FAILURE INDICIES
26 New $GRID POINT STRESS/PLANE STRESS
27 New $GRID POINT STRESS VOLUME DIRECT
28 New $GRID POINT STRESS VOLUME PRINCIPAL
29 New $ELEMENT STRESS DISCONTINUITIES
30 New $ELEMENT STRESS DISCONTINUITIES DIRECT
31 New $ELEMENT STRESS DISCONTINUITIES PRINCIPAL
32 New $GRID POINT STRESS DISCONTINUITIES
33 New $GRID POINT SRESS DISCONTINUITIES DIRECT
34 New $GRID POINT STRESS DISCON PRINCIPAL

Main Index
 490 MSC Nastran Reference Guide
PUNCH Output

Table 12-1 Structural Punch Table Support and Descriptor Entry (continued)
Table Code Table Support Table Content and Descriptor
35 New $GRID POINT STRESS/PLAIN STRAIN
36 Old $ELEMENT KINETIC ENERGY
37 Old $ELEMENT ENERGY LOSS PER CYCLE
38 New $MAX/MIN SUMMARY INFORMATION
39 Old $MPCF
40 Old $MODAL GRID POINT KINETIC ENERGY

Note:  The “new” table format supports all of the “old” table formats for punch.
 The word, “stress” in referenced table codes can be replaced by “strain”.
 The “none” table format implies that there is no Case Control interface to select punch as an option,
only print is supported.
 When RANDOM is selected, then the Table Descriptor of Table codes 1 through 5 can be augmented
by the strings of - PSDF, - AUTO, - RMS, - NO, or - CRMS, to denote the entry formulation.

Other impacted aspects of RANDOM analysis are the replacement of the $SUBCASE ID entry with $RANDOM ID entry
and two additional Table Descriptions that have no base table augmentation. The table descriptors are:
$CROSS-PSDF
$CROSS-CORRELATION FUNCTION
Starting with the current version, when thermal analysis has been selected, the structural table descriptor strings have been
replaced by their correct thermal designation. Table 12-2 associates the table code to the thermal table descriptor string. In
addition, with the exception of temperature, which formats the punch output to bulk data conventions, the scalar class
entries have been shortened to a single data item.

Table 12-2 Alternate Punch Descriptor for Thermal Analysis

Table Code Thermal Descriptor Entry
1 $TEMPERATURE
2 $HEAT FLOW AT LOAD POINTS
3 $HEAT FLOW AT CONTSTRAINT POINTS
4 $ELEMENT GRADIENTS AND FLUXES
10 $ENTHALPY
11 $H DOT

Main Index
 Chapter 12: Interface With Other Programs 491
PUNCH Output

Description of an OFP Punch Header Block

The OFP Punch header is composed of six to nine entries, depending upon sort conditions, analysis approach, grid versus
element responses, and other conditions. In addition, column one has the following meaning. A “$” indicates the entry
belongs to the header section. A “-” indicates the continuation of item information for an entry. A blank in column one
implies the start of an entry. In general, the punch header entries are:
$TITLE = Case Control Title
$SUBTITLE = Case Control Subtitle
$LABEL = Case Control Lable
$Table Descriptor
$MAXMIN (Table code = 38) additional information
$Output Format
$SUBCASE ID or RANDOM ID, missing for SORT 2 STATICS
$ELEMENT TYPE = missing for grid table codes

Approach or Sort Specific Information

Note that entries one through four are always be present. The MAXMIN additional information entry is described in
“MAX/MIN/RMS Methodology in Data Recovery” in MSC.Nastran 2004 Release Guide.
The Table Descriptor entry has been augmented for the energy recovery tables 18, 36, and 37 to delineate the element type
for non frequency response analysis approaches, and to include the calculation option for the Frequency Response
formulation.

$ELEMENT STRAIN ENERGIES BAR OPT=AVERAGE

=AMPLITUDE
=PEAK

The Output Format entry has been augmented, when necessary, to include the superelement identification. The area
remains blank when a residual structure is detected.

$REAL OUTPUT SEID=10

The SUBCASE entry has been augmented to include the SURFACE or VOLUME designations when Grid Point Stress or
Discontinuity tables are being processed. The keyword FIBER or STRCUR will be present for strain table recovery.

$SUBCASE ID = 1 SURFACE= 1 FIBER

VOLUME=

The Element Type line has the most significant augmentation. The element name has now been appended to further
identify the element enumeration. Because certain elements may contain variable user selectable responses, additional
keywords are appended, when necessary, to reflect the entry content. The keywords VONM and SHEAR deal with the
invariant selected; STRCUR or FIBER for the strain recovery, CUBIC for the QUAD4 corner extrapolation method, and
MATERIAL for shell elements, when not in the element reference system.

Main Index
 492 MSC Nastran Reference Guide
PUNCH Output

$ELEMENT TYPE = 144 QUAD4C VONM SRCUR BILIN MATERIAL

SHEAR FIBER SGAGE
CUBIC

The last augmentation is for the sort 2 table condition and deals with the issues of approach and whether the entry
identification is integer or real.

$SPOINT ID = 1 IDENTIFIED BY SUBACASE

FREQUENCY
TIME
LOAD FACTOR

Example
The following are examples of header and entry changes between the MSC Nastran 2004 and later releases and previous
releases of the OFP punch formats.

Sort2 Statics, the SUBCASE line was deleted since the SUBCASE identifies the entry and SUBCASE is not
global for all entries.
$TITLE =CANTILEVER BEAM CONSTRUCTED OF PLATES, STRESS CONTOURS F10 1
$SUBTITLE= REF: SINGER,STRENGTH OF MATERIALS, ART 5-2,PG 133 2
$LABEL = 3
$DISPLACEMENTS 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$POINT ID = 1 6
1 G 0.000000E+00 0.000000E+00 0.000000E+00 7
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 8

$TITLE =CANTILEVER BEAM CONSTRUCTED OF PLATES, STRESS CONTOURS F10 1

$SUBTITLE= REF: SINGER,STRENGTH OF MATERIALS, ART 5-2,PG 133 2
$LABEL = 3
$DISPLACEMENTS 4
$REAL OUTPUT 5
$POINT ID = 1 IDENTIFIED BY SUBCASE 5
1 G 0.000000E+00 0.000000E+00 0.000000E+00 6
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 7

Main Index
 Chapter 12: Interface With Other Programs 493
PUNCH Output

In superelement processing the block content was not identified.

$TITLE =RESULT CASE NAME 1

$SUBTITLE=LOAD CASE NAME 1 2
$LABEL =RESTRAINT CASE NAME 1 3
$DISPLACEMENTS 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
101 G 0.000000E+00 9.994075E-04 0.000000E+00 7
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 8

$TITLE =RESULT CASE NAME 1

$SUBTITLE=LOAD CASE NAME 1 2
$LABEL =RESTRAINT CASE NAME 1 3
$DISPLACEMENTS 4
$REAL OUTPUT SEID=10 5
$SUBCASE ID = 1 6
101 G 0.000000E+00 9.994075E-04 0.000000E+00 7
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 8

When Multi-point Constraint Forces were recovered, they were incorrectly identified.
$SUBTITLE=LOAD CASE NAME 1 2
$LABEL =RESTRAINT CASE NAME 1 3
$SPCF 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
1 G 0.000000E+00 0.000000E+00 0.000000E+00 7
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 8

$TITLE =RESULT CASE NAME 1

$SUBTITLE=LOAD CASE NAME 1 2
$LABEL =RESTRAINT CASE NAME 1 3
$MPCF 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
1 G 0.000000E+00 0.000000E+00 0.000000E+00 7
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 8

Main Index
 494 MSC Nastran Reference Guide
PUNCH Output

To assist the user with the element type enumeration, the GPTABD name values were appended to the white
area of the ELEMENT TYPE entry.

$TITLE =RESULT CASE NAME 1

$SUBTITLE=LOAD CASE NAME 1 2
$LABEL =RESTRAINT CASE NAME 1 3
$ELEMENT FORCES 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$ELEMENT TYPE = 33 7
1 -3.906332E+01 -9.898468E+02 -3.342360E+01 8
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 9
-CONT- 0.000000E+00 0.000000E+00 10

$TITLE =RESULT CASE NAME 745

$SUBTITLE=LOAD CASE NAME 1 746
$LABEL =RESTRAINT CASE NAME 1 747
$ELEMENT FORCES 748
$REAL OUTPUT 749
$SUBCASE ID = 1 750
$ELEMENT TYPE = 33 QUAD4 751
1 -3.906332E+01 -9.898468E+02 -3.342360E+01 752
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 753
-CONT- 0.000000E+00 0.000000E+00 754

Main Index
 Chapter 12: Interface With Other Programs 495
PUNCH Output

Main Index
 496 MSC Nastran Reference Guide
PUNCH Output

When element classes had optional output selections the user could select from the
Case Control entry, these selections were not indicated in the punch heading
block.

$TITLE =RESULT CASE NAME 287

$SUBTITLE=LOAD CASE NAME 1 288
$LABEL =RESTRAINT CASE NAME 1 289
$ELEMENT STRAINS 290
$REAL OUTPUT 291
$SUBCASE ID = 1 292
$ELEMENT TYPE = 33 293
1 -5.000000E-02 8.596358E-05 -3.260426E-04 294
-CONT- -2.896712E-05 -2.010852E+00 8.647210E-05 295
-CONT- -3.265511E-04 2.515292E-04 5.000000E-02 296
-CONT- 8.596358E-05 -3.260426E-04 -2.896712E-05 297
-CONT- -2.010852E+00 8.647210E-05 -3.265511E-04 298
-CONT- 2.515292E-04 299

$TITLE =RESULT CASE NAME 287

$SUBTITLE=LOAD CASE NAME 1 288
$LABEL =RESTRAINT CASE NAME 1 289
$ELEMENT STRAINS 290
$REAL OUTPUT 291
$SUBCASE ID = 1 292
$ELEMENT TYPE = 33 QUAD4 VONM FIBER 293
1 -5.000000E-02 8.596358E-05 -3.260426E-04 294
-CONT- -2.896712E-05 -2.010852E+00 8.647210E-05 295
-CONT- -3.265511E-04 2.515292E-04 5.000000E-02 296
-CONT- 8.596358E-05 -3.260426E-04 -2.896712E-05 297
-CONT- -2.010852E+00 8.647210E-05 -3.265511E-04 298
-CONT- 2.515292E-04 299

$TITLE =RESULT CASE NAME 287

$SUBTITLE=LOAD CASE NAME 1 288
$LABEL =RESTRAINT CASE NAME 1 289
$ELEMENT STRAINS 290
$REAL OUTPUT 291
$SUBCASE ID = 1 292
$ELEMENT TYPE = 33 QUAD4 VONM STRCUR 293
1 0.000000E+00 8.596358E-05 -3.260426E-04 294
-CONT- -2.896712E-05 -2.010852E+00 8.647210E-05 295
-CONT- -3.265511E-04 2.515292E-04 -1.000000E+00 296
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 297
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 298
-CONT- 0.000000E+00 299

Main Index
 Chapter 12: Interface With Other Programs 497
PUNCH Output

Main Index
 498 MSC Nastran Reference Guide
PUNCH Output

When an element had other than real data items, the values were incorrectly formatted. In this example the integer are
outside the IEEE exponent range.

$TITLE =RESULT CASE NAME 1

$SUBTITLE=LOAD CASE NAME 1 2
$LABEL =RESTRAINT CASE NAME 1 3
$ELEMENT STRAINS 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$ELEMENT TYPE = 144 7
1 1.973054E+02 5.605194E-45 0.000000E+00 8
-CONT- 8.596358E-05 -3.260426E-04 -2.896712E-05 9
-CONT- -2.010852E+00 8.647210E-05 -3.265511E-04 10
-CONT- 2.515292E-04 -1.000000E+00 0.000000E+00 11
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 12
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 13
-CONT- 1.401298E-45 0.000000E+00 8.924844E-05 14
-CONT- -4.056804E-04 -1.781213E-04 -9.896718E+00 15
-CONT- 1.047868E-04 -4.212187E-04 3.214852E-04 16
-CONT- -1.000000E+00 0.000000E+00 0.000000E+00 17
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 18
-CONT- 0.000000E+00 0.000000E+00 2.802597E-45 19
-CONT- 0.000000E+00 8.894981E-05 -2.488935E-04 20
-CONT- -1.678175E-04 -1.320753E+01 1.086421E-04 21
-CONT- -2.685857E-04 2.242232E-04 -1.000000E+00 22
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 23
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 24
-CONT- 0.000000E+00 1.261169E-44 0.000000E+00 25
-CONT- 8.297735E-05 -2.488935E-04 1.065242E-04 26
-CONT- 8.897819E+00 9.131589E-05 -2.572320E-04 27
-CONT- 2.086957E-04 -1.000000E+00 0.000000E+00 28
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 29
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 30
-CONT- 1.121039E-44 0.000000E+00 8.267872E-05 31
-CONT- -4.056804E-04 1.236546E-04 7.104448E+00 32
-CONT- 9.038459E-05 -4.133863E-04 3.101407E-04 33
-CONT- -1.000000E+00 0.000000E+00 0.000000E+00 34
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 35
-CONT- 0.000000E+00 0.000000E+00 36

$TITLE =RESULT CASE NAME 1

$SUBTITLE=LOAD CASE NAME 1 2
$LABEL =RESTRAINT CASE NAME 1 3
$ELEMENT STRAINS 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$ELEMENT TYPE = 144 QUAD4C VONM STRCUR BILIN 7
1 CEN/ 4 0.000000E+00 8
-CONT- 8.596358E-05 -3.260426E-04 -2.896712E-05 9
-CONT- -2.010852E+00 8.647210E-05 -3.265511E-04 10
-CONT- 2.515292E-04 -1.000000E+00 0.000000E+00 11
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 12
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 13
-CONT- 1 0.000000E+00 1.131398E-04 14
-CONT- -4.066659E-04 -2.896712E-05 -1.594806E+00 15
-CONT- 1.135430E-04 -4.070691E-04 3.160993E-04 16
-CONT- -1.000000E+00 0.000000E+00 0.000000E+00 17
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 18
-CONT- 0.000000E+00 0.000000E+00 2 19
-CONT- 0.000000E+00 6.580506E-05 -2.497893E-04 20
-CONT- -2.896712E-05 -2.622126E+00 6.646837E-05 21
-CONT- -2.504526E-04 1.929787E-04 -1.000000E+00 22
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 23
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 24
-CONT- 0.000000E+00 9 0.000000E+00 25
-CONT- 5.983260E-05 -2.479976E-04 -2.896712E-05 26
-CONT- -2.687879E+00 6.051255E-05 -2.486775E-04 27
-CONT- 1.892093E-04 -1.000000E+00 0.000000E+00 28
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 29
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 30
-CONT- 8 0.000000E+00 1.065701E-04 31
-CONT- -4.046950E-04 -2.896712E-05 -1.621391E+00 32
-CONT- 1.069800E-04 -4.051049E-04 3.119066E-04 33
-CONT- -1.000000E+00 0.000000E+00 0.000000E+00 34
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 35
-CONT- 0.000000E+00 0.000000E+00 36

Main Index
 Chapter 12: Interface With Other Programs 499
PUNCH Output

Again the coordinate system was also not correctly indicated.

$TITLE = MSC/NASTRAN JOB CREATED ON 11-NOV-99 AT 14:47:27 24

$SUBTITLE=SUBCASE1 25
$LABEL = 26
$ELEMENT STRESSES 27
$REAL OUTPUT 28
$SUBCASE ID = 1 29
$ELEMENT TYPE = 33 30
1 -5.000000E-02 3.225000E+04 3.225000E+04 31
-CONT- -2.775000E+04 -4.500000E+01 6.000000E+04 32
-CONT- 4.500000E+03 5.788134E+04 5.000000E-02 33
-CONT- -3.225000E+04 -3.225000E+04 2.775000E+04 34
-CONT- 4.500000E+01 -4.500000E+03 -6.000000E+04 35
-CONT- 5.788134E+04 36

$TITLE = MSC/NASTRAN JOB CREATED ON 11-NOV-99 AT 14:47:27 24

$SUBTITLE=SUBCASE1 25
$LABEL = 26
$ELEMENT STRESSES 27
$REAL OUTPUT 28
$SUBCASE ID = 1 29
$ELEMENT TYPE = 33 QUAD4 VONM MATERIAL 30
1 -5.000000E-02 3.225000E+04 3.225000E+04 31
-CONT- -2.775000E+04 -4.500000E+01 6.000000E+04 32
-CONT- 4.500000E+03 5.788134E+04 5.000000E-02 33
-CONT- -3.225000E+04 -3.225000E+04 2.775000E+04 34
-CONT- 4.500000E+01 -4.500000E+03 -6.000000E+04 35
-CONT- 5.788134E+04 36

Main Index
 500 MSC Nastran Reference Guide
PUNCH Output

This checks that SHEAR versus VONM is correctly identified.

$TITLE =RESULT CASE NAME 7509

$SUBTITLE=LOAD CASE NAME 1 7510
$LABEL =RESTRAINT CASE NAME 1 7511
$ELEMENT STRESSES 7512
$REAL OUTPUT 7513
$SUBCASE ID = 1 7514
$ELEMENT TYPE = 33 7515
1 -5.000000E-02 -3.906332E+02 -9.898468E+03 7516
-CONT- -3.342360E+02 -2.010852E+00 -3.788981E+02 7517
-CONT- -9.910203E+03 4.765653E+03 5.000000E-02 7518
-CONT- -3.906332E+02 -9.898468E+03 -3.342360E+02 7519
-CONT- -2.010852E+00 -3.788981E+02 -9.910203E+03 7520
-CONT- 4.765653E+03 7521

$TITLE =RESULT CASE NAME 8433

$SUBTITLE=LOAD CASE NAME 1 8434
$LABEL =RESTRAINT CASE NAME 1 8435
$ELEMENT STRESSES 8436
$REAL OUTPUT 8437
$SUBCASE ID = 1 8438
$ELEMENT TYPE = 33 QUAD4 SHEAR 8439
1 -5.000000E-02 -3.906332E+02 -9.898468E+03 8440
-CONT- -3.342360E+02 -2.010852E+00 -3.788981E+02 8441
-CONT- -9.910203E+03 4.765653E+03 5.000000E-02 8442
-CONT- -3.906332E+02 -9.898468E+03 -3.342360E+02 8443
-CONT- -2.010852E+00 -3.788981E+02 -9.910203E+03 8444
-CONT- 4.765653E+03 8445

Main Index
 Chapter 12: Interface With Other Programs 501
PUNCH Output

Check if element class name exists for Laminated Composite and that the invariant
is eliminated since SHEAR context is only available.

$TITLE =SQUARE PLATE WITH CIRCULAR HOLE - COMPOSITES SOLUTION 1

$SUBTITLE=MEMBRANE PROPERTIES OF QUAD4 AND TRIA3 ELEMENTS 2
$LABEL =UNIFORM LOAD ALONG X=5.0. 3
$ELEMENT STRESSES 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$ELEMENT TYPE = 95 7
1 1 6.713920E+03 -2.537445E+03 8
-CONT- 1.848406E+02 0.000000E+00 0.000000E+00 9
-CONT- 1.144151E+00 6.717611E+03 -2.541137E+03 10
-CONT- 4.629374E+03 11

$TITLE =SQUARE PLATE WITH CIRCULAR HOLE - COMPOSITES SOLUTION 1

$SUBTITLE=MEMBRANE PROPERTIES OF QUAD4 AND TRIA3 ELEMENTS 2
$LABEL =UNIFORM LOAD ALONG X=5.0. 3
$ELEMENT STRESSES 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$ELEMENT TYPE = 95 QUAD4LC 7
1 1 6.713920E+03 -2.537445E+03 8
-CONT- 1.848406E+02 0.000000E+00 0.000000E+00 9
-CONT- 1.144151E+00 6.717611E+03 -2.541137E+03 10
-CONT- 4.629374E+03 11

Main Index
 502 MSC Nastran Reference Guide
PUNCH Output

The GPSDCON entry caused unidentified output to be created in the punch stream.

$TITLE =RESULT CASE NAME 831

$SUBTITLE=LOAD CASE NAME 1 832
$LABEL =RESTRAINT CASE NAME 1 833
$REAL OUTPUT 834
$SUBCASE ID = 1 835
1 1.358237E-19 3.892658E+02 2.825602E+03 836
-CONT- 5.154498E+02 3.298798E+02 2.884989E+03 837
-CONT- 1.277554E+03 2.729644E+03 1.673444E+03 838
1 1.358237E-19 3.892658E+02 2.825602E+03 839
-CONT- 5.154498E+02 3.298798E+02 2.884989E+03 840
-CONT- 1.277554E+03 2.729644E+03 1.673444E+03 841
1 1.738842E-19 3.892658E+02 2.825602E+03 842
-CONT- 5.154498E+02 3.298798E+02 2.884989E+03 843
-CONT- 1.277554E+03 2.729644E+03 1.673444E+03 844

$TITLE =RESULT CASE NAME 831

$SUBTITLE=LOAD CASE NAME 1 832
$LABEL =RESTRAINT CASE NAME 1 833
$GRID POINT STRESS DISCONTTINUITIES 834
$REAL OUTPUT 835
$SUBCASE ID = 1 SURFACE= 1 836
1 Z1 3.892658E+02 2.825602E+03 837
-CONT- 5.154498E+02 3.298798E+02 2.884989E+03 838
-CONT- 1.277554E+03 2.729644E+03 1.673444E+03 839
1 Z2 3.892658E+02 2.825602E+03 840
-CONT- 5.154498E+02 3.298798E+02 2.884989E+03 841
-CONT- 1.277554E+03 2.729644E+03 1.673444E+03 842
1 MID 3.892658E+02 2.825602E+03 843
-CONT- 5.154498E+02 3.298798E+02 2.884989E+03 844
-CONT- 1.277554E+03 2.729644E+03 1.673444E+03 845

Main Index
 Chapter 12: Interface With Other Programs 503
PUNCH Output

The ELSDCON entry caused unidentified output to be created in the punch stream.

$TITLE =RESULT CASE NAME 1867

$SUBTITLE=LOAD CASE NAME 1 1868
$LABEL =RESTRAINT CASE NAME 1 1869
$REAL OUTPUT 1870
$SUBCASE ID = 1 1871
1 5.724519E+10 1.491285E-07 1.358237E-19 1872
-CONT- 3.240547E+02 2.101422E+03 6.456028E+02 1873
-CONT- 2.751211E+02 2.147283E+03 9.770737E+02 1874
-CONT- 2.051576E+03 1.282938E+03 1875

$TITLE =RESULT CASE NAME 1869

$SUBTITLE=LOAD CASE NAME 1 1870
$LABEL =RESTRAINT CASE NAME 1 1871
$ELEMENT STRESS DISCONTINUITIES 1872
$REAL OUTPUT 1873
$SUBCASE ID = 1 SURFACE= 1 1874
1 QUAD4 Z1 1875
-CONT- 3.240547E+02 2.101422E+03 6.456028E+02 1876
-CONT- 2.751211E+02 2.147283E+03 9.770737E+02 1877
-CONT- 2.051576E+03 1.282938E+03 1878

Main Index
 504 MSC Nastran Reference Guide
PUNCH Output

The GPSTRESS entry caused unidentified output to be created in the punch stream.

$TITLE =RESULT CASE NAME 287

$SUBTITLE=LOAD CASE NAME 1 288
$LABEL =RESTRAINT CASE NAME 1 289
$REAL OUTPUT 290
$SUBCASE ID = 1 291
1 0.000000E+00 1.358237E-19 -7.940824E+02 292
-CONT- -1.270989E+04 1.011802E+03 4.819154E+00 293
-CONT- -7.087778E+02 -1.279519E+04 6.043207E+03 294
-CONT- 1.245593E+04 295

$TITLE =RESULT CASE NAME 2913

$SUBTITLE=LOAD CASE NAME 1 2914
$LABEL =RESTRAINT CASE NAME 1 2915
$GRID POINT STRESS/PLANE STRESS 2916
$REAL OUTPUT 2917
$SUBCASE ID = 1 SURFACE= 1 2918
1 0 Z1 -7.940824E+02 2919
-CONT- -1.270989E+04 1.011802E+03 4.819154E+00 2920
-CONT- -7.087778E+02 -1.279519E+04 6.043207E+03 2921
-CONT- 1.245593E+04 2922
1 0 Z2 -7.940824E+02 2923
-CONT- -1.270989E+04 1.011802E+03 4.819154E+00 2924
-CONT- -7.087778E+02 -1.279519E+04 6.043207E+03 2925
-CONT- 1.245593E+04 2926
1 0 MID -7.940824E+02 2927
-CONT- -1.270989E+04 1.011802E+03 4.819154E+00 2928
-CONT- -7.087778E+02 -1.279519E+04 6.043207E+03 2929
-CONT- 1.245593E+04 2930

Main Index
 Chapter 12: Interface With Other Programs 505
PUNCH Output

Main Index
 506 MSC Nastran Reference Guide
PUNCH Output

The GPSTRAIN entry caused unidentified output to be created in the punch stream. The
GPSTRESS phase II module also did not react to FIBER/STRCUR correctly.

$TITLE =RESULT CASE NAME 5665

$SUBTITLE=LOAD CASE NAME 1 5666
$LABEL =RESTRAINT CASE NAME 1 5667
$REAL OUTPUT 5668
$SUBCASE ID = 1 5669
1 0.000000E+00 1.358237E-19 1.006295E-04 5670
-CONT- -4.157221E-04 8.768948E-05 4.819154E+00 5671
-CONT- 1.043260E-04 -4.194186E-04 5.237446E-04 5672
-CONT- 4.801584E-04 5673
1 0.000000E+00 1.358237E-19 1.006295E-04 5674
-CONT- -4.157221E-04 8.768948E-05 4.819154E+00 5675
-CONT- 1.043260E-04 -4.194186E-04 5.237446E-04 5676
-CONT- 4.801584E-04 5677
1 0.000000E+00 1.738842E-19 1.299615E+04 5678
-CONT- -2.920491E+03 -1.808289E+03 -6.400696E+00 5679
-CONT- 1.319900E+04 -3.123345E+03 8.161175E+03 5680
-CONT- 1.500647E+04 5681

$TITLE =RESULT CASE NAME 5673

$SUBTITLE=LOAD CASE NAME 1 5674
$LABEL =RESTRAINT CASE NAME 1 5675
$GRID POINT STRAIN/PLANE STRESS 5676
$REAL OUTPUT 5677
$SUBCASE ID = 1 SURFACE= 1 FIBER 5678
1 0 Z1 1.006295E-04 5679
-CONT- -4.157221E-04 8.768948E-05 4.819154E+00 5680
-CONT- 1.043260E-04 -4.194186E-04 5.237446E-04 5681
-CONT- 4.801584E-04 5682
1 0 Z2 1.006295E-04 5683
-CONT- -4.157221E-04 8.768948E-05 4.819154E+00 5684
-CONT- 1.043260E-04 -4.194186E-04 5.237446E-04 5685
-CONT- 4.801584E-04 5686
1 0 MID 1.006295E-04 5687
-CONT- -4.157221E-04 8.768948E-05 4.819154E+00 5688
-CONT- 1.043260E-04 -4.194186E-04 5.237446E-04 5689
-CONT- 4.801584E-04 5690

$TITLE =RESULT CASE NAME 5673

$SUBTITLE=LOAD CASE NAME 1 5674
$LABEL =RESTRAINT CASE NAME 1 5675
$GRID POINT STRAIN/PLANE STRESS 5676
$REAL OUTPUT 5677
$SUBCASE ID = 1 SURFACE= 1 STRCUR 5678
1 0 MID 1.006295E-04 5679
-CONT- -4.157221E-04 8.768948E-05 4.819154E+00 5680
-CONT- 1.043260E-04 -4.194186E-04 5.237446E-04 5681
-CONT- 4.801584E-04 5682
1 0 CURV 0.000000E+00 5683
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 5684
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 5685
-CONT- 0.000000E+00 5686

Main Index
 Chapter 12: Interface With Other Programs 507
PUNCH Output

The GPSTRESS entry caused unidentified output to be created in the punch stream.

$TITLE = LINEAR ELEMENT CODE CHECK PROBLEM 198

$SUBTITLE= ELASTIC LOADING BY FORCE 199
$LABEL = CHECK GRID POINT STRESS OUTPUT FORMATS 200
$REAL OUTPUT 201
$SUBCASE ID = 2 202
3401 9.471862E+02 5.677020E+01 203
-CONT- 1.723746E+01 -NaNQ 204

$TITLE = LINEAR ELEMENT CODE CHECK PROBLEM 199

$SUBTITLE= ELASTIC LOADING BY FORCE 200
$LABEL = CHECK GRID POINT STRESS OUTPUT FORMATS 201
$GRID POINT STRESS/PLAIN STRAIN 202
$REAL OUTPUT 203
$SUBCASE ID = 2 SURFACE= 34 204
3401 9.471862E+02 5.677020E+01 205
-CONT- 1.723746E+01 206

The GPSTRESS entry caused unidentified output to be created in the punch stream.

$TITLE = LINEAR ELEMENT CODE CHECK PROBLEM 235

$SUBTITLE= ELASTIC LOADING BY FORCE 236
$LABEL = CHECK GRID POINT STRESS OUTPUT FORMATS 237
$REAL OUTPUT 238
$SUBCASE ID = 2 239
3901 1.979894E+03 6.972885E+01 -2.529273E-14 240
-CONT- 2.071350E-13 -3.198958E+00 2.251117E+01 241
-CONT- -6.832076E+02 1.946365E+03 242

$TITLE = LINEAR ELEMENT CODE CHECK PROBLEM 237

$SUBTITLE= ELASTIC LOADING BY FORCE 238
$LABEL = CHECK GRID POINT STRESS OUTPUT FORMATS 239
$GRID POINT STRESS VOLUME DIRECT 240
$REAL OUTPUT 241
$SUBCASE ID = 2 VOLUME= 39 242
3901 1.979894E+03 6.972885E+01 -2.529273E-14 243
-CONT- 2.071350E-13 -3.198958E+00 2.251117E+01 244
-CONT- -6.832076E+02 1.946365E+03 245

Main Index
 508 MSC Nastran Reference Guide
PUNCH Output

The GPSTRESS entry caused unidentified output to be created in the punch stream.

$TITLE = LINEAR ELEMENT CODE CHECK PROBLEM 336

$SUBTITLE= ELASTIC LOADING BY FORCE 337
$LABEL = CHECK GRID POINT STRESS OUTPUT FORMATS 338
$REAL OUTPUT 339
$SUBCASE ID = 2 340
3901 1.980150E+03 -4.018156E-01 6.987475E+01 341
-CONT- 9.999354E-01 -1.135504E-02 -5.369717E-04 342
-CONT- -1.903498E-05 4.556393E-02 -9.989614E-01 343
-CONT- 1.136771E-02 9.988969E-01 4.556077E-02 344
-CONT- -6.832076E+02 1.946365E+03 345

$TITLE = LINEAR ELEMENT CODE CHECK PROBLEM 339

$SUBTITLE= ELASTIC LOADING BY FORCE 340
$LABEL = CHECK GRID POINT STRESS OUTPUT FORMATS 341
$GRID POINT STRESS VOLUME PRINCIPAL 342
$REAL OUTPUT 343
$SUBCASE ID = 2 VOLUME= 39 344
3901 1.980150E+03 -4.018156E-01 6.987475E+01 345
-CONT- 9.999354E-01 -1.135504E-02 -5.369717E-04 346
-CONT- -1.903498E-05 4.556393E-02 -9.989614E-01 347
-CONT- 1.136771E-02 9.988969E-01 4.556077E-02 348
-CONT- -6.832076E+02 1.946365E+03 349

Main Index
 Chapter 12: Interface With Other Programs 509
PUNCH Output

The element class did not identify the Element Strain Energy output.

$TITLE =RESULT CASE NAME 3043

$SUBTITLE=LOAD CASE NAME 1 3044
$LABEL =RESTRAINT CASE NAME 1 3045
$ELEMENT STRAIN ENERGIES 3046
$REAL OUTPUT 3047
$SUBCASE ID = 1 3048
1 2.976484E-03 6.479465E-02 1.632466E+00 3049
2 1.780166E-03 3.875217E-02 5.816491E-01 3050

$TITLE =RESULT CASE NAME 8977

$SUBTITLE=LOAD CASE NAME 1 8978
$LABEL =RESTRAINT CASE NAME 1 8979
$ELEMENT STRAIN ENERGIES QUAD4 8980
$REAL OUTPUT 8981
$SUBCASE ID = 1 8982
1 2.976484E-03 6.479465E-02 1.632466E+00 8983
2 1.780166E-03 3.875217E-02 5.816491E-01 8984

In eigenvalue analysis the eigenvalue was wrong and element class identifier
missing.

$TITLE =RESULT_CASE_NAME 2871

$SUBTITLE=LOAD CASE NAME 2872
$LABEL =RESTRAINT CASE NAME 2873
$ELEMENT STRAIN ENERGIES 2874
$REAL OUTPUT 2875
$SUBCASE ID = 1 2876
$EIGENVALUE = .3390093E+13 MODE = 1 2877
1 5.414903E+02 6.852106E-02 4.061989E+02 2878
2 3.771790E+04 4.772884E+00 2.829408E+04 2879

$TITLE =RESULT_CASE_NAME 4101

$SUBTITLE=LOAD CASE NAME 4102
$LABEL =RESTRAINT CASE NAME 4103
$ELEMENT STRAIN ENERGIES TETRA 4103
$REAL OUTPUT 4104
$SUBCASE ID = 1 4105
$EIGENVALUE = 0.1580508E+07 MODE = 1 4106
1 5.414903E+02 6.852106E-02 4.061989E+02 4107
2 3.771790E+04 4.772884E+00 2.829408E+04 4108

Main Index
 510 MSC Nastran Reference Guide
PUNCH Output

The EKE entry did not identify the output, the eigenvalue was incorrect and the
element class type was missing.

$TITLE =RESULT_CASE_NAME 3389

$SUBTITLE=LOAD CASE NAME 3390
$LABEL =RESTRAINT CASE NAME 3391
$REAL OUTPUT 3392
$SUBCASE ID = 1 3393
$EIGENVALUE = .3390093E+13 MODE = 1 3394
1 6.581564E+04 8.328419E+00 4.937160E+04 3395
2 3.031419E+02 3.836007E-02 2.274019E+02 3396

$TITLE =RESULT_CASE_NAME 4821

$SUBTITLE=LOAD CASE NAME 4822
$LABEL =RESTRAINT CASE NAME 4823
$ELEMENT KINETIC ENERGY TETRA 4823
$REAL OUTPUT 4824
$SUBCASE ID = 1 4825
$EIGENVALUE = 0.1580508E+07 MODE = 1 4826
1 6.581565E+04 8.328419E+00 4.937161E+04 4827
2 3.031419E+02 3.836007E-02 2.274019E+02 4828

Main Index
 Chapter 12: Interface With Other Programs 511
PUNCH Output

ESE in Frequency Response did not identify the method used for calculating the
response quantity. The element class was also missing.

$TITLE = 1
$SUBTITLE= 2
$LABEL = 3
$ELEMENT STRAIN ENERGIES 4
$REAL OUTPUT 5
$SUBCASE ID = 2 6
$FREQUENCY = .1000000E+01 7
101 2.102882E-05 1.820742E-06 1.051441E-07 8
102 2.893353E-02 2.505157E-03 1.446677E-04 9
103 2.029759E-01 1.757429E-02 1.014879E-03 10

$TITLE = 1
$SUBTITLE= 2
$LABEL = 3
$ELEMENT STRAIN ENERGIES BAR OPT=AVERAGE 4
$REAL OUTPUT 5
$SUBCASE ID = 2 6
$FREQUENCY = 1.0000000E+00 7
101 3.086982E-05 2.673122E-06 1.543491E-07 8
102 2.927053E-02 2.534634E-03 1.463527E-04 9
103 2.036543E-01 1.763512E-02 1.018272E-03 10

Main Index
 512 MSC Nastran Reference Guide
PUNCH Output

In EKE output in Frequency Response the block was not identified, the method was
not indicated, and the element class was missing.

$TITLE = 5666
$SUBTITLE= 5667
$LABEL = 5668
$REAL OUTPUT 5669
$SUBCASE ID = 2 5670
$FREQUENCY = .1000000E+01 5671
101 4.627901E+01 5.389437E+00 2.313951E-01 5672
102 3.740206E+01 4.355669E+00 1.870103E-01 5673
103 2.966045E+01 3.454118E+00 1.483023E-01 5674

$TITLE = 5668
$SUBTITLE= 5669
$LABEL = 5670
$ELEMENT KINETIC ENERGY BAR OPT=AVERAGE 5671
$REAL OUTPUT 5672
$SUBCASE ID = 2 5673
$FREQUENCY = 1.0000000E+00 5674
101 4.627080E+01 5.390246E+00 2.313540E-01 5675
102 3.739541E+01 4.356322E+00 1.869771E-01 5676
103 2.965508E+01 3.454624E+00 1.482754E-01 5677

Main Index
 Chapter 12: Interface With Other Programs 513
PUNCH Output

In EDE output in Frequency Response the block was not identified, the method was
not indicated, and the element class was missing.

$TITLE = 9635
$SUBTITLE= 9636
$LABEL = 9637
$REAL OUTPUT 9638
$SUBCASE ID = 2 9639
$FREQUENCY = .1000000E+01 9640
3042 6.970902E+02 2.664647E+01 9641
5042 1.918979E+03 7.335353E+01 9642

$TITLE = 9799
$SUBTITLE= 9800
$LABEL = 9801
$ELEMENT ENERGY LOSS PER CYCLE DAMP2 OPT=AVERAGE 9802
$REAL OUTPUT 9803
$SUBCASE ID = 2 9804
$FREQUENCY = 1.0000000E+00 9805
3042 6.970200E+02 2.665085E+01 9806
5042 1.918356E+03 7.334915E+01 9807

Main Index
 514 MSC Nastran Reference Guide
PUNCH Output

Main Index
 Chapter 12: Interface With Other Programs 515
PUNCH Output

Nonlinear output blocks were not identified by the LOAD FACTOR.

$TITLE =RESULT_CASE_1 1
$SUBTITLE= 2
$LABEL = 3
$ELEMENT STRESSES 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$ELEMENT TYPE = 89 7
1 1.905082E+02 1.905082E+02 6.350274E-06 8
-CONT- 6.350274E-06 0.000000E+00 0.000000E+00 9

$TITLE =RESULT_CASE_1 1
$SUBTITLE= 2
$LABEL = 3
$ELEMENT STRESSES 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$ELEMENT TYPE = 89 RODNL 7
$LOAD FACTOR = 1.0000000E-01 7
1 1.905082E+02 1.905082E+02 6.350274E-06 8
-CONT- 6.350274E-06 0.000000E+00 0.000000E+00 9
___________________________________________________________________________Punch
format maintained, however block identifier changed to reflect content.

$TITLE = RADIATION BOUNDARY CONDITION 1

$SUBTITLE= 2
$LABEL = 3
$DISPLACEMENTS 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$POINT ID = 1 7
TEMP* 1 1 7.270000E+02 8
TEMP* 2 1 7.270001E+02 9
TEMP* 3 1 7.270000E+02 10

$TITLE = RADIATION BOUNDARY CONDITION 1

$SUBTITLE= 2
$LABEL = 3
$TEMPERATURE 4
$REAL OUTPUT 5
$SUBCASE ID = 1 6
$POINT ID = 1 IDENTIFIED BY TIME 7
TEMP* 1 1 7.270000E+02 8
TEMP* 2 1 7.270001E+02 9
TEMP* 3 1 7.270000E+02 10

Main Index
 516 MSC Nastran Reference Guide
PUNCH Output

When THERMAL analysis, the problem is a single degree of freedom system, therefore
when the true content is identified, the scale of the output can be reduced.

$TITLE = RADIATION BOUNDARY CONDITION 305

$SUBTITLE= 306
$LABEL = 307
$VELOCITY 308
$REAL OUTPUT 309
$SUBCASE ID = 1 310
$POINT ID = 1 311
0.000000E+00 S 1.817500E+00 0.000000E+00 0.000000E+00 312
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 313

$TITLE = RADIATION BOUNDARY CONDITION 1409

$SUBTITLE= 1410
$LABEL = 1411
$ENTHALPY 1412
$REAL OUTPUT 1413
$SUBCASE ID = 1 1414
$POINT ID = 1 IDENTIFIED BY TIME 1415
0.000000E+00 S 1.817500E+00 1416

Main Index
 Chapter 12: Interface With Other Programs 517
PUNCH Output

Same issues as previous.

$TITLE = RADIATION BOUNDARY CONDITION 801

$SUBTITLE= 802
$LABEL = 803
$ACCELERATION 804
$REAL OUTPUT 805
$SUBCASE ID = 1 806
$POINT ID = 1 807
0.000000E+00 S 1.066395E-06 0.000000E+00 0.000000E+00 808
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 809

$TITLE = RADIATION BOUNDARY CONDITION 2817

$SUBTITLE= 2818
$LABEL = 2819
$H DOT 2820
$REAL OUTPUT 2821
$SUBCASE ID = 1 2822
$POINT ID = 1 IDENTIFIED BY TIME 2823
0.000000E+00 S 2.730277E-04 2824

Same issues as ENTHLAPY.

$TITLE =LARGE PROBLEM 85

$SUBTITLE= 86
$LABEL = 87
$OLOADS 88
$REAL OUTPUT 89
$SUBCASE ID = 1 90
101 S 1.500000E+03 0.000000E+00 0.000000E+00 99
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 100

$TITLE =LARGE PROBLEM 85

$SUBTITLE= 86
$LABEL = 87
$HEAT FLOW AT LOAD POINTS 88
$REAL OUTPUT 89
$SUBCASE ID = 1 90
$LOAD FACTOR = 1.0000000E+00 90
101 S 1.500000E+03 95

Main Index
 518 MSC Nastran Reference Guide
PUNCH Output

Same issues as ENTHLAPY.

$TITLE =LARGE PROBLEM 247

$SUBTITLE= 248
$LABEL = 249
$SPCF 250
$REAL OUTPUT 251
$SUBCASE ID = 1 252
98 S -2.466861E+02 0.000000E+00 0.000000E+00 257
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 258
99 S -1.051691E+04 0.000000E+00 0.000000E+00 259
-CONT- 0.000000E+00 0.000000E+00 0.000000E+00 260

$TITLE =LARGE PROBLEM 169

$SUBTITLE= 170
$LABEL = 171
$HEAT FLOW AT CONTRAINT POINTS 172
$REAL OUTPUT 173
$SUBCASE ID = 1 174
$LOAD FACTOR = 1.0000000E+00 174
98 S -2.466861E+02 177
99 S -1.051691E+04 178

Main Index
 Chapter 12: Interface With Other Programs 519
Patran Support of Heat Transfer Analysis

Thermal element data recovery contains mis-labeled and possible numerical errors.

$TITLE = RADIATION BOUNDARY CONDITION 1649

$SUBTITLE= 1650
$LABEL = 1651
$ELEMENT FORCES 1652
$REAL OUTPUT 1653
$SUBCASE ID = 1 1654
$ELEMENT TYPE = 33 1655
$ELEMENT ID = 4 1656
0.000000E+00 5.724519E+10 1.491285E-07 0.000000E+00 1657
-CONT- 0.000000E+00 1.401298E-45 0.000000E+00 1658
-CONT- 0.000000E+00 1.401298E-45 1659

$TITLE = RADIATION BOUNDARY CONDITION 6233

$SUBTITLE= 6234
$LABEL = 6235
$ELEMENT GRADIENTS AND FLUXES 6236
$REAL OUTPUT 6237
$SUBCASE ID = 1 6238
$ELEMENT TYPE = 33 QUAD4 6239
$ELEMENT ID = 4 IDENTIFIED BY TIME 6240
0.000000E+00 QUAD4 0.000000E+00 6241
-CONT- 0.000000E+00 1 0.000000E+00 6242
-CONT- 0.000000E+00 1 6243

Patran Support of Heat Transfer Analysis

Patran offers a comprehensive and integrated graphics interface to the linear, nonlinear, steady-state, and transient thermal
analysis capabilities in MSC Nastran. From within the Patran environment, you can completely define your thermal
analysis model, submit it to MSC Nastran for analysis, import MSC Nastran thermal analysis results into the Patran
database, and visualize the results. Complete support is provided for analyses including conduction, convection, advection,
and radiation.
The interface is designed for both new and advanced users of MSC Nastran and Patran. The online help system provides
several guided example problems and describes all of the Patran menu interface to MSC Nastran’s thermal solver. Basic and
Advanced modes are included for most of the menu forms for thermal loads and boundary conditions. For example, the
Basic convection form only requires you to define the convection coefficient and ambient fluid temperature, while the
Advanced form includes options for a user-defined convection equation, time-dependent and temperature-dependent
coefficients, and variable ambient temperatures.
A brief overview of capabilities and limitations is given in the sections that follow.

Main Index
 520 MSC Nastran Reference Guide
Patran Support of Heat Transfer Analysis

Loads and Boundary Conditions

The following thermal loads and boundary conditions are supported, each of which can vary both spatially and temporally:
 Prescribed nodal temperatures
 Heat flux normal to a surface
 Vector heat flux from a distant radiant heat source
 Heat applied directly to a node (nodal source)
 Volumetric heat generation within conduction elements
 Basic convection
 Advection with forced convection boundaries
 Radiation exchange with space
 Radiation exchange within enclosures (cavity radiation)
 Multiple enclosures

Results Processing
 The following results quantities can be postprocessed within Patran:
 Nodal temperatures
 Applied heat loads
 Heats of constraint for temperature boundary conditions
 Heat fluxes within conduction elements
 Heat fluxes at boundaries due to applied heat loads, free/forced convection, advection, and radiation
 Temperature gradients
 Enthalpies
 Rate of change of enthalpies

Limitations
When using Patran to pre- and postprocess heat transfer analyses in MSC Nastran, keep in mind the following limitations:
 There is currently no support of MSC Nastran heat transfer Bulk Data entries through Patran’s NASTRAN Input
File Reader.
 Emissivity as a function of wavelength is currently not supported. The RADBND Bulk Data entry is used to
specify the wavelength break points for radiation problems.
 Automatic restart is not supported for the Patran interface to MSC Nastran heat transfer analysis; this includes:
• Restarting from a previously converged time step (SOL 159 to SOL 159)
• Using a steady-state analysis as a restart to a transient analysis (SOL 159 from SOL 153)

Main Index
 Chapter 12: Interface With Other Programs 521
The Nastran HDF5 Result Database (NH5RDB)

• Using a steady-state analysis as a restart to a thermal stress analysis (SOL 106 from SOL 153)
• Using various time steps from a transient thermal analysis as a restart to a quasi-static thermal stress analysis
(SOL 106 from SOL 159)
To perform many of the restarts listed above, you must modify the cold start run and manually insert the bulk data
entries for the restart.
 The global view factor is currently not supported. Specifically, the RADCAV Bulk Data entry, SETij, are not
supported. Basic view-factor calculations are supported.
 NOLIN Bulk Data entries are not supported. The NOLIN entries are used for various thermal-state control
functions (e.g., the nonlinear transient power function).
 The TLOAD2 Bulk Data entry is not supported. This entry enables you to specify various functions such as SIN,
COS, and EXP when applying a thermal load. One work-around is to use the Field capability to create these
functions. For example, you can create a PCL (PATRAN Command Language) function that can then be mapped
to a table using the Create/Non-Spatial/Tabular Input field capability.
 The ambient element (ELEAMB) field for the RADCAV Bulk Data entry is not supported. For radiation
enclosure problems, this means that for an incomplete cavity, all energy exchange carried on outside the enclosure
surface is lost to zero-degree space. Patran will not automatically generate an ambient element that could have a
nonzero temperature and a distinct area associated with it.

Documentation
For detailed documentation, refer to the Patran Interface to MSC Nastran Thermal,. This document is available online as
part of the help system for Patran.

The Nastran HDF5 Result Database (NH5RDB)

Introduction
The NH5RDB is a Nastran database file in Hierarchical Data Format (HDF5). The database organizes Nastran input and
output data in a hierarchical structure. Nastran data is stored as dataset in database and can be accessed through standard
HDF5 APIs or third party packages in multiple programming languages. The main features of NH5RDB include:
 A data type schema in XML defines database structure and datasets formats.
 All main Nastran input and output data are covered
 Supported by multiple Nastran solutions.
 Optional output by MDLPRM parameter and system cell.
 New DMAP modules to write Nastran data blocks in database.
 An open format and multiple programming language support for data access.

Main Index
 522 MSC Nastran Reference Guide
The Nastran HDF5 Result Database (NH5RDB)

The NH5RDB supports wide range of data types and is capable of defining complicated data structures. In the database,
the datasets are stored in a hierarchical structure, making it easy to add, remove or update datasets in applications. The
database supports high precision, compression and unlimited amount of data. Its open format and multiple programming
languages support make it ideal for FEA applications.
The NH5RDB has modal input and result data. It organizes data in groups and datasets. The input data has groups like
NODE, ELEMENT, CONSTRAINT, LOAD, etc. The result data includes ACOUSTIC, CONTACT, NODAL,
ELEMTENTAL and other groups. In each group, data are stored in table like datasets. A dataset in NH5RDB is defined
by a compound structure with member fields. For example, the GRID dataset in the input NODE group has fields for grid
number and coordinates, it will have all grid point definitions from the modal. The DISPLACEMENT dataset in RESULT
group has fields for displacement components and will contain all nodal displacement from all output steps. A dataset in
NH5RDB could be identified by its dataset name and group path defined in the schema.
There are some options available for NH5RDB output. The HDF5 parameter in MDLPRM is used to control NH5RDB
generation. By default, NH5RDB will not be generated. Data compression can be selected for NH5RDBB. Also the input
data can be selected not to be included in NH5RDB by HDF5 parameter options.
Since HDF5 is an open format architecture. The NH5RDB data can be accessed by standard HDF5 library APIs or third
party packages. Multiple programming languages, such as C/C++, FORTRAN, Python and Java, are supported. The
example section gives examples of NH5RDB data access in multiple languages.

Data Schema
The data schema is used to define database structure and its datasets formats. The database is in a tree structure with datasets
as its leaf nodes. The dataset path in the tree forms a unique identifier. The dataset data is in a table format with member
fields as columns. The schema is defined in XML with elements <typedef>, <group>, <dataset>, etc. The following schema
definitions show the top levels of the schema and the input GRID and stress output HEXA datasets.
<?xml version="1.0" encoding="utf-8"?>
<crdb schema="0">
<typedefs>...</typedefs>
<groups>
<group name="NASTRAN">
<group name="INPUT">
<group name="CONSTRAINT">...</group>
<group name="CONTACT">...</group>
<group name="COORDINATE_SYSTEM">...</group>
<group name="DESIGN">...</group>
<dataset name="DOMAINS">...</dataset>
<group name="DYNAMIC">...</group>
<group name="ELEMENT">...</group>
<group name="FATIGUE">...</group>
<group name="LOAD">...</group>
<group name="MATERIAL">...</group>
<group name="MATRIX">...</group>
<group name="NODE">...</group>
<group name="PARAMETER">...</group>
<group name="PARTITION">...</group>

Main Index
 Chapter 12: Interface With Other Programs 523
The Nastran HDF5 Result Database (NH5RDB)

<group name="PROPERTY">...</group>
<group name="TABLE">...</group>
<group name="UDS">...</group>
</group>
<group name="RESULT">
<group name="ACOUSTIC">...</group>
<group name="CONTACT">...</group>
<dataset name="DOMAINS">...</dataset>
<group name="ELEMENTAL">...</group>
<group name="FATIGUE">...</group>
<group name="MODAL">...</group>
<group name="NLOUT">...</group>
<group name="NODAL">...</group>
<group name="OPTIMIZATION">...</group>
<group name="SUMMARY">...</group>
<group name="MONITOR">...</group>
</group>
</group>
</groups>
</crdb>
Schema 1: The top level groups
<?xml version="1.0" encoding="utf-8"?>
<crdb schema="0">
<typedefs>
<typedef name="GRID_SS" description ="Grid strain and stress structure">
<field name="GRID" type="integer"/>
<field name="X" type="double"/>
<field name="Y" type="double"/>
<field name="Z" type="double"/>
<field name="TXY" type="double"/>
<field name="TYZ" type="double"/>
<field name="TZX" type="double"/>
</typedef>
</typedefs >
<groups>
<group name="NASTRAN">
<group name="INPUT">
<group name="NODE">
<dataset name="GRID" version="1">
<field name="ID" type="integer"/>
<field name="CP" type="integer"/>
<field name="X" type="double" size="3"/>
<field name="CD" type="integer"/>
<field name="PS" type="integer"/>
<field name="SEID" type="integer"/>
<field name="DOMAIN_ID" type="integer"/>
</dataset>
</group>
</group>

Main Index
 524 MSC Nastran Reference Guide
The Nastran HDF5 Result Database (NH5RDB)

<group name="RESULT">
<group name="ELEMENTAL">
<group name="STRESS">
<dataset name="HEXA">
<field name="EID" type="integer"/>
<field name="CID" type="integer"/>
<field name="CTYPE" type="character" size="4"/>
<field name="NODEF" type="integer"/>
<field name="SS" type="GRID_SS" size="9"/>
<field name="DOMAIN_ID" type="integer"/>
</dataset>
</group>
</group>
</group>
</group>
</groups>
</crdb>
Schema 2: The input GRID and output HEXA stress dataset definitions

Nastran Input and Output in NH5RDB

The NH5RDB includes Nastran input and output data. The supported data types are summarized in the following table.

Table 12-3 NH5RDB data categories

Category Description
Input:
Case control and parameters Case control option and parameter values
Constraint Constraint entries like SPC and MPC
Contact Contact entry like BSURF
Coordinate system Coordinate system entry like CORD1C
Design optimization Design optimization entry like BEADVAR
Dynamic Dynamic entries like EIGC
Element Element entries like CHEXA and CQUAD4
Fatigue Fatigue related entries like MATFTG
Load Load entries like FORCE and PLOAD4
Material Materials like MAT1, MAT2 and MAT8
Matrix Direct matrix input like CONM1
Node Node entries like GRID and SPOINT
Partition Define set or list like SET1

Main Index
 Chapter 12: Interface With Other Programs 525
The Nastran HDF5 Result Database (NH5RDB)

Category Description
Property Property like PSHELL and PSOLID
Table Table values in TABLEM1 and TABLEM2
Output:
Acoustic output Acoustic power, participation factor, etc.
Contact output Contact force and stress output
Elemental output Stress, strain, element force, etc.
Fatigue output Fatigue analysis output
Modal Modal response
Monitor Monitor output
Nodal output Displacement, velocity, acceleration, etc.
Optimization data Objective function, constraint, history values, etc.
Special nonlinear format data NLOUT format data

Nastran Solution Support

The NH5RDB is supported by multiple Nastran solutions. The table below shows all supported solution sequences.

Table 12-4 NH5RDB support solutions

Solution Number Description

101 Statics
103 Normal modes
105 Buckling
107, 110 Complex eigenvalues
108, 111 Frequency response
109, 112 Transient response
144, 145, 146 Aerodynamic analysis
200 Design Optimization
400 Nonlinear analysis

Main Index
 526 MSC Nastran Reference Guide
The Nastran HDF5 Result Database (NH5RDB)

NH5RDB Optional Output

The NH5RDB is an optional output. By default, the database will not be generated by Nastran. A parameter HDF5 is
added in MDLPRM for NH5RDB generation. In BULK data section, setting HDF5 parameter in MDLPRM will ask
Nastran to create the database.

Table 12-5 The HDF5 parameter in MDLPRM

Parameters Name Description, Type, and Default Values

HDF5 Control NH5RDB database creation.
-1 Do not create NH5RDB database (Default).
0 Create NH5RDB database without compression.
1 Create NH5RDB database with compression.
2 Create uncompressed NH5RDB database without input data.
3 Create compressed NH5RDB database without input data.

The usage of HDF5 parameter of MDLPRM in input file is shown as the following:

BEGIN BULK
MDLPRM,HDF5,1

In general, particularly for solid models, the user does not need rotational components in nodal vector type datasets as they
are not easy or useful to visualize in post processor, suppressing them can reduce result file size. An option is provided to
active this.
The option is an integer expressed by the MDLPRM keyword H5NORDOF or system cell 730.

Table 12-6 The H5NORDOF parameter in MDLPRM

Parameters Name Description

H5NORDOF Parameter to suppress output rotational components to NH5RDB database
-1 Do not create NH5RDB database (Default).
0 Create NH5RDB database without compression.

The option is only used when MDLPRM keyword HDF5 (or system cell 702) is 0 or 1, and applies to datasets for
displacement, velocity, acceleration and eigenvector.
To distinguish datasets with both translational and rotational components, the datasets without rotational have “_TRANS”
post prefix. For example, the corresponding dataset of DISPLACEMENT dataset is DISPLACEMENT_TRANS, and the
one for DISPLACEMENT_CPLX is DISPLACEMENT_TRANS_CPLX.

Main Index
 Chapter 12: Interface With Other Programs 527
The Nastran HDF5 Result Database (NH5RDB)

The NH5RDB is also optional for intermediate output through OUTCTRL keyword in NLOPRM. If this option is
selected, multiple NH5RDB files will be generated for output steps. Each NH5RDB file will contain data from its
corresponding output step. The data in these generated NH5RDB files can be accessed during Nastran run, allowing post
processing applications to access available data during a time consuming model run.

NH5RDB DMAP modules

There are several DMAP modules to write Nastran input and output data into database. The descriptions of these modules
are as the following:
CRDB_IN Write IFP data blocks into the NH5RDB database

Write IFP data blocks in an HDF5 format database. A maximum of five data blocks can be passed in one call.

Format:
CRDB_IN DB1,DB2,DB3,DB4,DB5//
NDDLNAM1/ DDLNAM2/NDDLNAM3/NDDLNAM4/NDDLNAM5/
SEID/AFPMID/TRIMID/ $

Input Data Blocks:

DBi Input data blocks, see remark 1 for supported data blocks.

Parameters:
NDDLNAMi Char8, default=blank NDDL data block name corresponding to DBi
SEID Integer, default=0 Super element id of the data block
AFPMID Integer, default=0 Acoustic field point mesh id
TRIMID Trim id, default=0 Trim component id

Remarks:
1. The supported input data blocks are GEOM1, GEOM2, GEOM3, GEOM4, EPT, EDT, MPT, DIT,
DYNAMIC, CONTACT, MATPOOL, EDOM, FATIGUE, PVT, CASECC and SPECSEL.
2. The SEID, AFPMID and TRIMID indicate which super element, acoustic field point mesh or trim component the
data blocks are from.
3. NDDLNAMi must be provided for all data blocks.

Main Index
 528 MSC Nastran Reference Guide
The Nastran HDF5 Result Database (NH5RDB)

CRDB_OUT Write OFP data blocks into NH5RDB database.

Write OFP data blocks in an HDF5 format database. A maximum of five data blocks can be passed in one call.

Format:
CRDB_OUT DB1,DB2,DB3,DB4,DB5//
NDDLNAM1/NDDLNAM2/NDDLNAM3/NDDLNAM4/NDDLNAM5/
SEID/AFPMID/TRIMID/DESCYCLE/$

Input Data Blocks:

DBi OFP data blocks, see remark 1 for supported data blocks.

Parameters:
NDDLNAMi Char8, default=blank NDDL data block name corresponding to DBi
SEID Integer, default=0 Super element id of the data block
AFPMID Integer, default=0 Acoustic field point mesh id
TRIMID Trim id, default=0 Trim component id
DESCYCLE Integer, default=0 Design cycle number in design optimization

Remarks:
1. The supported output data blocks are OUG. OVG, OAG, OES, OEF, OPG, OQG, OEFTG, OEFTGM,
OUGFP, OMPF, OERP, OARPWR, OVGFP, OAPWR2, OAIG, OCOMP, OCOMPM, OMCFRAC,
DBCOPT, NLOUT, OFVCCT and OFCON3D.
2. The SEID, AFPMID, TRIMID and DESCYCLE indicate which super element, acoustic field point mesh, trim
component and design cycle the data blocks are from.
3. NDDLNAMi must be provided for all data blocks.

Main Index
 Chapter 12: Interface With Other Programs 529
The Nastran HDF5 Result Database (NH5RDB)

CRDB_DBK Write special data blocks into the NH5RDB database

For some special datasets in NH5RDB, several Nastran data blocks and parameters are involved when writing them. The
CRDB_DBK is defined for these kinds of data output. The module accepts up to ten data blocks, five integer and five real
type parameters in its arguments. A string parameter is used in this module to indicate what data these input arguments are
for. The implementation of the module will check the string indicator and interpret the input arguments accordingly.
Currently only CLAMA and DSCM are supported in the string parameter, they are for complex eignvalue summary and
design sensitivity coefficients output.

Format:
CRDB_DBK DB1,DB2,DB3,DB4,DB5,DB6,DB7,DB8,DB9,DB10//
TYPE/PI1/PI2/PI3/PI4/PI5/
PR1/PR2/PR3/PR4/PR5/ $

Input Data Blocks:

DBi Data blocks to be processed.

Parameters:
TYPE Char8, default=blank Option indicating the to-do process action
PIi Integer, default=0 Integer parameters used with data blocks
PRi Real, default=0.0 Real parameters used with data blocks

Main Index
 530 MSC Nastran Reference Guide
The Nastran HDF5 Result Database (NH5RDB)

Remarks:
1. The module will interpret the TYPE parameter and process data blocks accordingly. Only the complex eigenvalue
CLAMA and design sensitivity coefficient DSCM are supported. The input arguments for the module are defined
as the following.

Type Data Block Integer Parameter Real Parameter Note

CLAMA DB1=CLAMA N/A PR1=Spin speed CLAMA=Complex eigenvalue
summary data block
DSCM DB1=DSCMCOL PI1=SEID N/A DSCMCOL=Design sensitivity
parameters data block
DB2=DSCM PI2=AFPMID
DSCM=Design sensitivity
DB3=DESTAB PI3=TRIMID coefficient matrix

DB4=R1VALRG PI4=DESCYCLE DESTAB=Design variables

attributes
DB5=R2VALRG
R1VALRG=Response type1 value
DB6=R3VALRG matrix

R2VALRG=Response type 2 value

matrix

R3VALRG=Response type3 value

matrix

SEID=Super element id

AFPMID=Acoustic field point id

TRIMID=Trim component id

DESCYCLE=Design cycle

Main Index
 Chapter 12: Interface With Other Programs 531
The Nastran HDF5 Result Database (NH5RDB)

CRDB_MCR Create intermediate HDF5 format database

Create intermediate HDF5 format database. The following calls of HDF5 output module, like CRDB_IN and
CRDB_OUT, will write data into the created intermediate database.

Format:
CRDB_MCR //NUM $

Parameters:
NUM Integer, default=0 Sequence number of the intermediate database

Remarks:
1. This module will create intermediate database with the job name and sequence number. All following HDF5 output
will be put into the intermediate database.
2. The intermediate database close module CRDB_MCL should be used with CRDB_MCR as pair.
CRDB_MCL Close intermediate HDF5 format database

Close intermediate HDF5 format database.

Format:
CRDB_MCL //

Remarks:
1. This module will close intermediate database. All following HDF5 output will be put into the primary database.
2. The intermediate database creation module CRDB_MCR should be used with CRDB_MCL as pair.

Test Case
The test model below shows the usage of HDF5 parameter and structure of NH5RDB database, the database is opened in
HDF5 browser HDFView.
Test model file: test.dat
SOL 101
TIME 5
CEND
DISP = ALL
STRESS = ALL
FORCE = ALL

Main Index
 532 MSC Nastran Reference Guide
NH5RDB Data Access

SPC = 1
SUBCASE 1
LOAD = 1
BEGIN BULK
MDLPRM,HDF5,1
GRID,1,,10.
GRID,2,,8.
GRID,3,,6.
GRID,4,,4.
GRID,5,,2.
GRID,6
CBAR,1,1,1,2,,1.
CBAR,2,1,2,3,,1.
CBAR,3,1,3,4,,1.
CBAR,4,1,4,5,,1.
CBAR,5,1,5,6,,1.
PBAR,1,1,4.,1.3333,1.3333,2.2496,,,PBAR1
+BAR1,1.,1.,1.,-1.,-1.,1.,-1.,-1.
MAT1,1,30.E06,,.3
SPC,1,6,123456,0.
FORCE,1,1,,4000.,,-1.
FORCE,1,1,,4000.,-1.
ENDDATA
Figure 12-1 NH5RDB in HDF View browser

NH5RDB Data Access

The NH5RDB is a Hierarchical Data Format (HDF5) database; it organizes and stores data in groups and datasets. For
Nastran data, the groups and datasets are defined in a schema file. The NH5RDB data structures will be created from these
definitions in the schema.
The HDF5 is an open format architecture; it provides multiple language programming interfaces for data operations. The
following shows how to access data in NH5RDB by examples in Python and C languages, a Java program using Java HDF
object package is also given.

Data schema
The data schema is an XML file. Its elements and attributes define file structure and data format in NH5RDB. A brief
summary of the schema elements and attributes is shown here.

Main Index
 Chapter 12: Interface With Other Programs 533
NH5RDB Data Access

Table 12-7 The schema XML elements and attributes

Element Description Element Attribute Attribute Value

<crdb> Root element schema Schema version number
<typedefs> Parent element of <typedef> description Description string
<typedef> Define structure type used in schema name Structure name
description Description string
<groups> Parent element of <group> description Description string
<group> Define a group name Group name
description Description string
<dataset> Define dataset format name Dataset name
description Description string
<field> Structure member field name Field name
type Field type
size Field size
description Description string

Like folders in a file system, the <group> element contains subgroups and datasets belong to it. The <dataset> element is
used to define storing data format, it has <field> element to define each field properties. The <typedef> defines data
structure element that will be used in the field type attribute in schema.
Below is a snippet of data schema that defines GRID input entry, nodal displacement and HEXA element stress output. A
special note is for DOMAINS dataset and DOMAIN_ID field in input and output data. NH5RDB will store same type
data in one dataset, for example, all displacement data will be put in the DISPLACEMENT dataset, no matter from which
sub case and time step the displacement is. To distinguish data from different cases, the domain concept is introduced. The
DOMAINS dataset consists of fields like ID, SUBCASE, STEP, etc. The ID field gives a unique number of the domain.
The combination of all other fields makes one domain entry, which indicates the source of data. Correspondingly, a
DOMAIN_ID field is attached to input and output data. This DOMAIN_ID is the ID number in DOMAINS dataset.
From the DOMAIN_ID, the data source information can be obtained. There is an example that shows how to relate data
with domains in example section.
<?xml version="1.0" encoding="utf-8"?>
<crdb schema="0">
<typedefs>
<typedef name="GRID_SS" description ="Grid strain and stress structure">
<field name="GRID" type="integer"/>
<field name="X" type="double"/>
<field name="Y" type="double"/>
<field name="Z" type="double"/>
<field name="TXY" type="double"/>

Main Index
 534 MSC Nastran Reference Guide
NH5RDB Data Access

<field name="TYZ" type="double"/>

<field name="TZX" type="double"/>
</typedef>
</typedefs >
<groups>
<group name="NASTRAN">
<group name="INPUT">
<group name="NODE">
<dataset name="GRID" version="1">
<field name="ID" type="integer"/>
<field name="CP" type="integer"/>
<field name="X" type="double" size="3"/>
<field name="CD" type="integer"/>
<field name="PS" type="integer"/>
<field name="SEID" type="integer"/>
<field name="DOMAIN_ID" type="integer"/>
</dataset>
</group>
</group>
<group name="RESULT">
<dataset name="DOMAINS">
<field name="ID" type="integer"/>
<field name="SUBCASE" type="integer"/>
<field name="STEP" type="integer"/>
<field name="ANALYSIS" type="integer"/>
<field name="TIME_FREQ_EIGR" type="double"/>
<field name="EIGI" type="double"/>
<field name="MODE" type="integer"/>
<field name="DESIGN_CYCLE" type="integer"/>
<field name="RANDOM" type="integer"/>
<field name="SE" type="integer"/>
<field name="AFPM" type="integer"/>
<field name="TRMC" type="integer"/>
<field name="INSTANCE" type="integer"/>
<field name="MODULE" type="integer"/>
</dataset>
<group name="NODAL">
<dataset name="DISPLACEMENT">
<field name="ID" type="integer"/>
<field name="X" type="double" description="X component"/>
<field name="Y" type="double" description="Y component"/>
<field name="Z" type="double" description="Z component"/>
<field name="RX" type="double" description="RX component"/>
<field name="RY" type="double" description="RY component"/>
<field name="RZ" type="double" description="RZ component"/>
<field name="DOMAIN_ID" type="integer"/>
</dataset>
</group>
<group name="ELEMENTAL">
<group name="STRESS">
<dataset name="HEXA">

Main Index
 Chapter 12: Interface With Other Programs 535
NH5RDB Data Access

<field name="EID" type="integer"/>

<field name="CID" type="integer"/>
<field name="CTYPE" type="character" size="4"/>
<field name="NODEF" type="integer"/>
<field name="SS" type="GRID_SS" size="9"/>
<field name="DOMAIN_ID" type="integer"/>
</dataset>
</group>
</group>
</group>
</group>
</groups>
</crdb>
Schema 3. Dataset and domains

NH5RDB
The NH5RDB database stores datasets as defined in the schema. A NH5RDB database snapshot is shown below. The tables
in the snapshot show the nodal displacement and HEXA element stress datasets.

Figure 12-2 NH5RDB snapshot

Main Index
 536 MSC Nastran Reference Guide
NH5RDB Data Access

Dataset domain index

As explained above about DOMAIN_ID, NH5RDB associates all data with domain id and put the same type of data in
one dataset. Therefore, a dataset may have data with multiple domain ids. For a large dataset, one common operation is to
access partial data from interested domain. To improve data retrieval efficiency, NH5RDB will generate domain id index
for corresponding output datasets. Under the root group, an INDEX group will be created for domain index datasets.
Corresponding to each output dataset, a domain index dataset with the same group and dataset name will be generated in
the INDEX group. For example, the nodal displacement dataset is located at
/NASTRAN/RESULT/NODAL/DISPLACEMENT in NH5RDB, its domain index dataset will be created as
/INDEX/NASTRAN/RESULT/NODAL/DISPLACEMENT. All the index datasets have the same format with three fields
DOMAIN_ID, POSITION and LENGTH. The DOMAIN_ID is the domain id number, the POSITION is the domain
id start position in corresponding output dataset, and the LENGTH is the number of rows for the domain id. From the
start position and number of rows, the data for the domain id can be located and accessed.
Records of a domain index dataset are sorted based on domain id, records with the same domain in a result dataset are
sorted, which are similar to MSC Nastran sort1 option. Client code can use this feature to implement binary search
algorithm for fast search. In most cases, the sort key is entity id; however there are exceptions, for example, result dataset
/NASRAN/RESULT/ELEMENTAL/VIBRATION_INTENSITY has two keys for sorting, the first key is element type
and the second is element id.
The figure below shows the displacement and its index datasets. From the index dataset, the displacement data for domain
id 5 can be found from position 76 and has 44 rows.

Main Index
 Chapter 12: Interface With Other Programs 537
NH5RDB Data Access

Figure 12-3 Domain index dataset

Database access examples

The HDF5 is platform independent and provides multiple programming language APIs for data operations. The examples
here are given in Python, C\C++ and Java to show dataset access in NH5RDB. In addition to source files, each example
gives build script and readme files for reference.
The Python examples uses PyTable package for HDF5 file. The PyTable and NumPy packages are required to be installed
with Python. The examples show how easy it is to access the datasets.
 Requirements of C and C++ examples
To build and run C and C++ examples, HDF5 library and a C++ compiler is required. All below examples are built
with Visual C++ 2013 on Windows and Intel C++ 2015 and HDF5 library 1.8.14 or 1.8.15-patch1, a later version
of the C++ compiler and HDF5 library should work.
 Example 1: Read nodal displacements
This example reads nodal displacement dataset and prints the grid id and displacements from the second row of the
table.

Main Index
 538 MSC Nastran Reference Guide
NH5RDB Data Access

import tables

# Open database file

file = tables.open_file("test.h5")

# Get displacement table

disp = file.root.NASTRAN.RESULT.NODAL.DISPLACEMENT

# Get row number 2 of the table

row = disp[2]

# Print grid id and its displacement

print row['ID'], row['X'], row['Y'], row['Z'], row['RX'], row['RY'],
row['RZ']

# Close file
file.close()
Example 1: Read nodal displacement in Python

 Example 2: Extract nodal displacements at time 3.0 and print in CSV format
This example has two steps, it first gets domain id belongs to time step 3.0, then extracts displacement for this time
step and writes to a CSV file. In the example, the where clause is used to extract data for specified criterion. The
time step and its corresponding displacements are shown in figure 4 and the converted CSV file is shown in figure 5.

import tables as pt

# -----------------------------------------------------------------------
--
# Extract nodal displacements at time = 3.000 and print as CSV
# -----------------------------------------------------------------------
--

# --- open h5 input file

fname = "ldr2s400utp02.h5"
h5 = pt.open_file(fname)

# --- step 1: get "domain id" for t = 3.000

# -- path to table with domain ids
tblDOMAINS = h5.root.NASTRAN.RESULT.DOMAINS

# -- where clause
t = 3.000
tol = 0.005
swhere = "(TIME_FREQ_EIGR > %6.4f)&(TIME_FREQ_EIGR < %6.4f)"%(t-tol,
t+tol)

# -- extract domain id
for row in tblDOMAINS.where(swhere) :
did = row["ID"]

Main Index
 Chapter 12: Interface With Other Programs 539
NH5RDB Data Access

break

# --- step 2: extract the nodal displacements

# -- path to table with nodal displacements
tblDISPS = h5.root.NASTRAN.RESULT.NODAL.DISPLACEMENT

# -- open CSV output file

fp = open("displacements.csv","w")

# -- where clause
swhere = "DOMAIN_ID == %d"%(did)

# -- loop over table

for row in tblDISPS.where(swhere):
gid = row["ID"]
x = row["X"]
y = row["Y"]
z = row["Z"]
rx = row["RX"]
ry = row["RY"]
rz = row["RZ"]
fp.write("%d,%14.5e,%14.5e,%14.5e,%14.5e,%14.5e,%14.5e\n"%\
(gid,x, y, z, rx, ry, rz) )

# -- close files
fp.close()

h5.close()
Example 2: Extract displacement and convert to CSV format

Main Index
 540 MSC Nastran Reference Guide
NH5RDB Data Access

Figure 12-4 Time step and displacement data

Main Index
 Chapter 12: Interface With Other Programs 541
NH5RDB Data Access

Figure 12-5 Converted CSV data

 Example 3: Extract time history of nodal translation displacements

This example extracts displacements of node number 63 for all time steps, the data is shown in figure 6.
import tables as pt
import numpy as np

# -----------------------------------------------------------------------
--
# Extract time history of translational displacements for node 63
# Modified to correct domain ids; IDs are read from DOMAIN table
# Correctly extracted values are: 5 to 86 for 2017, 2018
# Modified to accommodate 2016 or 2017++ DISPLACEMENT table format

# -----------------------------------------------------------------------
--

# --- open h5 input file

fname = "ldr2s400_20160.h5"
h5 = pt.open_file(fname)

# --- step 1: extract times from domains table

# -- path to table with domain ids
tblDOMAINS = h5.root.NASTRAN.RESULT.DOMAINS

Main Index
 542 MSC Nastran Reference Guide
NH5RDB Data Access

# -- create empty numpy arrays for domain ids and times

ids = np.array([], int)
t = np.array([], float)

# -- construct where clause

# (time steps have domain ids 5 to 86 as found in HDFVIEW)
# Modify for loop to use TABLE.where() to use ANALYSIS=6 (MTRAN steps),
# instead of hard coded Domain ids.
swhere = "(ID >= 5) & ( ID <= 86)"

# -- loop over domains table

for row in tblDOMAINS.where("ANALYSIS==6"):
t = np.append(t, row["TIME_FREQ_EIGR"])
ids = np.append(ids,row["ID"])

# -- need to get first/last domain ids from above for later use

min_mtran_id = ids[0]
max_mtran_id = ids[-1]

# -- create numpy arrays for results

nt = t.size
disp = np.zeros((nt,3), float)

# --- step 2: extract the nodal displacements

# -- path to table with nodal displacements
tblDISPS = h5.root.NASTRAN.RESULT.NODAL.DISPLACEMENT

# -- create where clause for node 63

node = 63
swhere = "(ID == %d) & ( DOMAIN_ID >= %d) & (DOMAIN_ID <= %d)"%( node,
min_mtran_id, max_mtran_id )

# Note that 2016 Displacement table has a VALUE column instead of X,Y,Z.
# Code modified to work for 2016 or 2017++ Displacement table
nast_ver = h5.root._v_attrs.SCHEMA
# -- loop over table
i = 0
for row in tblDISPS.where(swhere):
if (nast_ver>0 ) :
disp[i,0] = row["X"]
disp[i,1] = row["Y"]
disp[i,2] = row["Z"]
else :
disp[i,0] = row["VALUE"] [0]
disp[i,1] = row["VALUE"] [1]
disp[i,2] = row["VALUE"] [2]

i += 1

Main Index
 Chapter 12: Interface With Other Programs 543
NH5RDB Data Access

# -- close file
h5.close()

# --- write extracted time history data to output file

# -- open time history output file
fname = "disp_node_%d_mod3.csv"%(node)
fp = open(fname,"w")
fp.write(" TIME, XDISP, YDISP, ZDISP\n")
for i in np.arange(nt) :
fp.write("%6.3f,%14.6e,%14.6e,%14.6e\n"%\
(t[i], disp[i,0], disp[i,1], disp[i,2]))

# -- close output file

fp.close()
Example 3: Extract nodal displacement at all time steps

Figure 12-6 Nodal displacement at time steps

The following examples are programs in C language. The HDF5 table APIs are used in these examples. For
description of these APIs, please refer to HDF5 reference manual.
 Example 4: Read nodal displacements

Main Index
 544 MSC Nastran Reference Guide
NH5RDB Data Access

This example reads all data in the displacement dataset and prints the translation displacements. The dataset format
is as its schema definition; its C type structure definition is shown in the example and can be found in dataset header
file generated from schema.
#include "hdf5.h"
#include "hdf5_hl.h"
#include <cstdlib>

/*
* Read all nodal displacements
*/
void printDisplacement()
{
// displacement dataset name
const char* table = "/NASTRAN/RESULT/NODAL/DISPLACEMENT";

// displacement structure
typedef struct {
long long ID; // grid id
double X; // displacement
double Y;
double Z;
double RX;
double RY;
double RZ;
long long DOMAIN_ID;// domain id
} Type;

// displacement structure size

size_t ts = sizeof(Type);

// field offset
size_t offset[] = {
HOFFSET(Type, ID),
HOFFSET(Type, X),
HOFFSET(Type, Y),
HOFFSET(Type, Z),
HOFFSET(Type, RX),
HOFFSET(Type, RY),
HOFFSET(Type, RZ),
HOFFSET(Type, DOMAIN_ID)
};

Type type;
// field size
size_t size[] = {
sizeof(type.ID),
sizeof(type.X),
sizeof(type.Y),
sizeof(type.Z),
sizeof(type.RX),

Main Index
 Chapter 12: Interface With Other Programs 545
NH5RDB Data Access

sizeof(type.RY),
sizeof(type.RZ),
sizeof(type.DOMAIN_ID)
};

// open file
hid_t fid = H5Fopen("test.h5", H5F_ACC_RDONLY, H5P_DEFAULT);
if (fid >= 0) {
hsize_t nField, nRecord;
// get field and record number
if (H5TBget_table_info(fid, table, &nField, &nRecord) >= 0) {
if (nRecord > 0) {
// allocate buffer
Type* p = (Type*) malloc(ts * nRecord);
if (p) {
// read displacement dataset
if (H5TBread_table(fid, table, ts, offset, size, p) >= 0) {
// print translation displacement
for (size_t i = 0; i < nRecord; i++) {
printf("GRID %d : %e, %e, %e\n", p[i].ID, p[i].X,
p[i].Y, p[i].Z);
}
}
free(p);
}
}
}
// close file
H5Fclose(fid);
}
}

int main()
{
printDisplacement();
}
Example 4: Read and print nodal displacements

 Example 5: Read records (rows) of nodal displacements

The above example reads all data in nodal displacement dataset. This example shows reading only specified records
(rows) from a dataset. It will read the first 10 records of displacements.
/*
* Read the first 10 records of nodal displacements
*/
void printDisplacement2()
{
// displacement dataset name
char* table = "/NASTRAN/RESULT/NODAL/DISPLACEMENT";

Main Index
 546 MSC Nastran Reference Guide
NH5RDB Data Access

// displacement structure
typedef struct {
long long ID; // grid id
double X; // displacement
double Y;
double Z;
double RX;
double RY;
double RZ;
long long DOMAIN_ID;// domain id
} Type;

// displacement structure size

size_t ts = sizeof(Type);

// field offset
size_t offset[] = {
HOFFSET(Type, ID),
HOFFSET(Type, X),
HOFFSET(Type, Y),
HOFFSET(Type, Z),
HOFFSET(Type, RX),
HOFFSET(Type, RY),
HOFFSET(Type, RZ),
HOFFSET(Type, DOMAIN_ID)
};

Type type;
// field size
size_t size[] = {
sizeof(type.ID),
sizeof(type.X),
sizeof(type.Y),
sizeof(type.Z),
sizeof(type.RX),
sizeof(type.RY),
sizeof(type.RZ),
sizeof(type.DOMAIN_ID)
};

// open file
hid_t fid = H5Fopen("test.h5", H5F_ACC_RDONLY, H5P_DEFAULT);
if (fid >= 0) {
hsize_t nField, nRecord;
// get field and record number
if (H5TBget_table_info(fid, table, &nField, &nRecord) >= 0) {
if (nRecord > 0) {
// set record number to 10
if (nRecord > 10) {
nRecord = 10;
}

Main Index
 Chapter 12: Interface With Other Programs 547
NH5RDB Data Access

// allocate buffer
Type* p = (Type*) malloc(ts * nRecord);
if (p) {
// read the first 10 records
if (H5TBread_records(fid, table, 0, nRecord, ts, offset, size, p)
>= 0) {
// print translation displacement
for (size_t i = 0; i < nRecord; i++) {
printf("GRID %d : %e, %e, %e\n", p[i].ID, p[i].X,
p[i].Y, p[i].Z);
}
}
free(p);
}
}
}
// close file
H5Fclose(fid);
}
}
Example 5: Read nodal displacement records

 Example 6: Read dataset fields (columns) of HEXA element stress by name

The above example shows reading records (rows) of dataset. This example will show how to read some fields
(columns) in dataset. It will read data fields from HEXA element stress output. The specified fields are given by
their field names, which are defined in schema.
/*
* Read HEXA element stress data by field name
*/
void printHexaStress()
{
/*
The HEXA stress dataset structure in database

typedef struct {
long long EID; // Element identification number
long long CID; // Stress Coordinate System
char CTYPE[4]; // Coordinate System Type (BCD)
long long NODEF; // Number of Active Points
long long GRID[9]; // Number of active grids or corner grid ID
double X[9]; // Normal X
double Y[9]; // Normal Y
double Z[9]; // Normal Z
double TXY[9]; // Shear xy
double TYZ[9]; // Shear yz
double TZX[9]; // Shear zx
long long DOMAIN_ID; // Domain identifier
} Type;
*/

Main Index
 548 MSC Nastran Reference Guide
NH5RDB Data Access

// HEXA stress dataset name

char* table = "/NASTRAN/RESULT/ELEMENTAL/STRESS/HEXA";

// fields to read from HEXA stress

typedef struct {
long long EID;
long long X[9];
long long Y[9];
long long Z[9];
long long TXY[9];
long long TYZ[9];
long long TZX[9];
long long DOMAIN_ID;
} Type;

size_t ts = sizeof(Type);
size_t offset[] = {
HOFFSET(Type, EID),
HOFFSET(Type, X),
HOFFSET(Type, Y),
HOFFSET(Type, Z),
HOFFSET(Type, TXY),
HOFFSET(Type, TYZ),
HOFFSET(Type, TZX),
HOFFSET(Type, DOMAIN_ID)
};

Type type;
size_t size[] = {
sizeof(type.EID),
sizeof(type.X),
sizeof(type.Y),
sizeof(type.Z),
sizeof(type.TXY),
sizeof(type.TYZ),
sizeof(type.TZX),
sizeof(type.DOMAIN_ID)
};

// field names of HEXA stress to read

char* name = "EID,X,Y,Z,TXY,TYZ,TZX,DOMAIN_ID";

hid_t fid = H5Fopen("test.h5", H5F_ACC_RDONLY, H5P_DEFAULT);

if (fid >= 0) {
hsize_t nField, nRecord;
if (H5TBget_table_info(fid, table, &nField, &nRecord) >= 0) {
if (nRecord > 0) {
Type* p = (Type*) malloc(ts * nRecord);
if (p) {
// read fields by name

Main Index
 Chapter 12: Interface With Other Programs 549
NH5RDB Data Access

if (H5TBread_fields_name(fid, table, name, 0, nRecord, ts,

offset, size, p) >= 0){
for (size_t i = 0; i < nRecord; i++) {
printf("EID %d : DOMAIN %d\n", p[i].EID, p[i].DOMAIN_ID);
for (int j = 0; j < 9; j++) {
printf("\t%e, %e, %e\n", p[i].X[j], p[i].Y[j], p[i].Z[j]);
}
}
}
free(p);
}
}
}
H5Fclose(fid);
}
}

Example 6: Read HEXA element stress data by field name

 Example 7: Read data fields (columns) of HEXA element stress by index

In addition to names, field index can be used to specify the fields to read. This example uses an index array to specify
the fields.
/*
* Read HEXA element stress data by field index
*/
void printHexaStress2()
{
/*
The HEXA stress dataset structure in database

typedef struct {
long long EID; // Element identification number
long long CID; // Stress Coordinate System
char CTYPE[4]; // Coordinate System Type (BCD)
long long NODEF; // Number of Active Points
long long GRID[9]; // Number of active grids or corner grid ID
double X[9]; // Normal X
double Y[9]; // Normal Y
double Z[9]; // Normal Z
double TXY[9]; // Shear xy
double TYZ[9]; // Shear yz
double TZX[9]; // Shear zx
long long DOMAIN_ID; // Domain identifier
} Type;
*/

// HEXA stress dataset name

char* table = "/NASTRAN/RESULT/ELEMENTAL/STRESS/HEXA";

Main Index
 550 MSC Nastran Reference Guide
NH5RDB Data Access

// fields to read from HEXA stress

typedef struct {
long long EID;
long long X[9];
long long Y[9];
long long Z[9];
long long TXY[9];
long long TYZ[9];
long long TZX[9];
long long DOMAIN_ID;
} Type;

size_t ts = sizeof(Type);
size_t offset[] = {
HOFFSET(Type, EID),
HOFFSET(Type, X),
HOFFSET(Type, Y),
HOFFSET(Type, Z),
HOFFSET(Type, TXY),
HOFFSET(Type, TYZ),
HOFFSET(Type, TZX),
HOFFSET(Type, DOMAIN_ID)
};

Type type;
size_t size[] = {
sizeof(type.EID),
sizeof(type.X),
sizeof(type.Y),
sizeof(type.Z),
sizeof(type.TXY),
sizeof(type.TYZ),
sizeof(type.TZX),
sizeof(type.DOMAIN_ID)
};

// field index of HEXA stress to read

int index[] ={0, 5, 6, 7, 8, 9, 10, 11};
int num = sizeof(index) / sizeof(int);

hid_t fid = H5Fopen("test.h5", H5F_ACC_RDONLY, H5P_DEFAULT);

if (fid >= 0) {
hsize_t nField, nRecord;
if (H5TBget_table_info(fid, table, &nField, &nRecord) >= 0) {
if (nRecord > 0) {
Type* p = (Type*) malloc(ts * nRecord);
if (p) {
// read fields by index
if (H5TBread_fields_index(fid, table, num, index, 0, nRecord, ts,
offset, size, p) >= 0){
for (size_t i = 0; i < nRecord; i++) {

Main Index
 Chapter 12: Interface With Other Programs 551
NH5RDB Data Access

printf("EID %d : DOMAIN %d\n", p[i].EID, p[i].DOMAIN_ID);

for (int j = 0; j < 9; j++) {
printf("\t%e, %e, %e\n", p[i].X[j], p[i].Y[j], p[i].Z[j]);
}
}
}
free(p);
}
}
}
H5Fclose(fid);
}
}
Example 7: Read HEXA element stress data by field index

Main Index
 552 MSC Nastran Reference Guide
NH5RDB Data Access

 Example 8: Write displacement, HEXA element stress and domain datasets

The example writes the nodal displacement, hexahedral element stress and domain datasets. These datasets use the
same formats as defined in NH5RDB schema and table data is appended row by row. Both Python and C programs
are given below.
# Write displacement, hexa stress and domain datasets in Python
from tables import *
from numpy import *

# Domain structure
class DOMAINS(IsDescription):
ID = Int64Col(pos = 0)
SUBCASE = Int64Col(pos = 1)
STEP = Int64Col(pos = 2)
ANALYSIS = Int64Col(pos = 3)
TIME_FREQ_EIGR = Float64Col(pos = 4)
EIGI = Float64Col(pos = 5)
MODE = Int64Col(pos = 6)
DESIGN_CYCLE = Int64Col(pos = 7)
RANDOM = Int64Col(pos = 8)
SE = Int64Col(pos = 9)
AFPM = Int64Col(pos = 10)
TRMC = Int64Col(pos = 11)
INSTANCE = Int64Col(pos = 12)
MODULE = Int64Col(pos = 13)

# Displacement structure
class DISPLACEMENT(IsDescription):
ID = Int64Col(pos = 0)
X = Float64Col(pos = 1)
Y = Float64Col(pos = 2)
Z = Float64Col(pos = 3)
RX = Float64Col(pos = 4)
RY = Float64Col(pos = 5)
RZ = Float64Col(pos = 6)
DOMAIN_ID = Int64Col(pos = 7)

# Hexa element stress structure

class HEXA(IsDescription):
EID = Int64Col(pos = 0)
CID = Int64Col(pos = 1)
CTYPE = StringCol(4, pos = 2)
NODEF = Int64Col(pos = 3)
GRID = Int64Col(shape = 9, pos = 4)
X = Float64Col(shape = 9, pos = 5)
Y = Float64Col(shape = 9, pos = 6)
Z = Float64Col(shape = 9, pos = 7)
TXY = Float64Col(shape = 9, pos = 8)
TYZ = Float64Col(shape = 9, pos = 9)
TZX = Float64Col(shape = 9, pos = 10)

Main Index
 Chapter 12: Interface With Other Programs 553
NH5RDB Data Access

DOMAIN_ID = Int64Col(pos = 11)

# Open file
h5 = open_file('test.h5', mode = 'w')
# Create domain table
domain = h5.create_table('/NASTRAN/RESULT', 'DOMAINS',
DOMAINS, createparents = True)
# Set domain values
row = domain.row
for i in xrange(1, 2):
row['ID'] = i
row['SUBCASE'] = 1
row['STEP'] = 0
row['ANALYSIS'] = 1
row['TIME_FREQ_EIGR'] = 0
row['EIGI'] = 0
row['MODE'] = 0
row['RANDOM'] = 0
row['DESIGN_CYCLE'] = 0
row['SE'] = 0
row['AFPM'] = 0
row['TRMC'] = 0
row['INSTANCE'] = 0
row['MODULE'] = 0
row.append()
domain.flush() # Flush table to file

# Create displacement table

displacement = h5.create_table('/NASTRAN/RESULT/NODAL',
'DISPLACEMENT', DISPLACEMENT, createparents = True)
# Populate displacement, the values are for example only.
row = displacement.row
for i in xrange(1, 10):
row['ID'] = i # grid id
row['X'] = i * 1.0
row['Y'] = i * 2.0
row['Z'] = i * 3.0
row['RX'] = 0.0
row['RY'] = 0.0
row['RZ'] = 0.0
row['DOMAIN_ID'] = 1 # domain id
row.append()
displacement.flush()

# Create hexa element stress table

hexa = h5.create_table('/NASTRAN/RESULT/ELEMENTAL/STRESS',
'HEXA', HEXA, createparents = True)
# Populate hexa stress, the values are for example only.
row = hexa.row
for i in xrange(1, 6):
row['EID'] = i

Main Index
 554 MSC Nastran Reference Guide
NH5RDB Data Access

row['CID'] = 0
row['CTYPE'] = 'GAUS' # Gauss point
row['NODEF'] = 8
row['GRID'] = array(i * arange(9)) # grid ids including center
row['X'] = array(i * arange(9)) # x normal stress of grids
row['Y'] = array(i * arange(9))
row['Z'] = array(i * arange(9))
row['TXY'] = array(i * arange(9)) # xy shear stress of grids
row['TYZ'] = array(i * arange(9))
row['TZX'] = array(i * arange(9))
row['DOMAIN_ID'] = 1 # domain id
row.append()
hexa.flush()

# Close file
h5.close()
Example 8.a: Write nodal displacement, hexahedral element stress and domain datasets in Python

// Write displacement, hexa stress and domain datasets in C

void writeStressHexa()
{
const char* result = "/NASTRAN/RESULT";
const char* nodal = "/NASTRAN/RESULT/NODAL";
const char* stress = "/NASTRAN/RESULT/ELEMENTAL/STRESS";
const char* dptable = "DISPLACEMENT";
const char* hxtable = "HEXA";
const char* dmtable = "DOMAINS";
hsize_t chunk = 50;

// displacement structure
typedef struct {
long long ID; // grid id
double X; // displacement
double Y;
double Z;
double RX;
double RY;
double RZ;
long long DOMAIN_ID;// domain id
} Displacement;

// displacement structure size

size_t dp = sizeof(Displacement);

// displacement field offset

size_t dpo[] = {
HOFFSET(Displacement, ID),
HOFFSET(Displacement, X),
HOFFSET(Displacement, Y),

Main Index
 Chapter 12: Interface With Other Programs 555
NH5RDB Data Access

HOFFSET(Displacement, Z),
HOFFSET(Displacement, RX),
HOFFSET(Displacement, RY),
HOFFSET(Displacement, RZ),
HOFFSET(Displacement, DOMAIN_ID)
};

// displacement field size

Displacement disp;
size_t dps[] = {
sizeof(disp.ID),
sizeof(disp.X),
sizeof(disp.Y),
sizeof(disp.Z),
sizeof(disp.RX),
sizeof(disp.RY),
sizeof(disp.RZ),
sizeof(disp.DOMAIN_ID)
};

// displacement field name

const char* dpn[] = {
"ID",
"X",
"Y",
"Z",
"RX",
"RY",
"RZ",
"DOMAIN_ID"
};

// displacemen field type

hid_t dpt[] = {
H5T_NATIVE_LLONG,
H5T_NATIVE_DOUBLE,
H5T_NATIVE_DOUBLE,
H5T_NATIVE_DOUBLE,
H5T_NATIVE_DOUBLE,
H5T_NATIVE_DOUBLE,
H5T_NATIVE_DOUBLE,
H5T_NATIVE_LLONG
};

// hexa stress structure

typedef struct {
long long EID; // Element identification number
long long CID; // Stress Coordinate System
char CTYPE[4]; // Coordinate System Type (BCD)
long long NODEF; // Number of Active Points
long long GRID[9]; // Number of active grids or corner grid ID

Main Index
 556 MSC Nastran Reference Guide
NH5RDB Data Access

double X[9]; // Normal X

double Y[9]; // Normal Y
double Z[9]; // Normal Z
double TXY[9]; // Shear xy
double TYZ[9]; // Shear yz
double TZX[9]; // Shear zx
long long DOMAIN_ID; // Domain identifier
} Hexa;

// hexa stress structure size

size_t hx = sizeof(Hexa);

// hexa field offset

size_t hxo[] = {
HOFFSET(Hexa, EID),
HOFFSET(Hexa, CID),
HOFFSET(Hexa, CTYPE),
HOFFSET(Hexa, NODEF),
HOFFSET(Hexa, GRID),
HOFFSET(Hexa, X),
HOFFSET(Hexa, Y),
HOFFSET(Hexa, Z),
HOFFSET(Hexa, TXY),
HOFFSET(Hexa, TYZ),
HOFFSET(Hexa, TZX),
HOFFSET(Hexa, DOMAIN_ID)
};

// hexa stress field size

Hexa hexa;
size_t hxs[] = {
sizeof(hexa.EID),
sizeof(hexa.CID),
sizeof(hexa.CTYPE),
sizeof(hexa.NODEF),
sizeof(hexa.GRID),
sizeof(hexa.X),
sizeof(hexa.Y),
sizeof(hexa.Z),
sizeof(hexa.TXY),
sizeof(hexa.TYZ),
sizeof(hexa.TZX),
sizeof(hexa.DOMAIN_ID)
};

// hexa stress field name

const char* hxn[] = {
"EID",
"CID",
"CTYPE",
"NODEF",

Main Index
 Chapter 12: Interface With Other Programs 557
NH5RDB Data Access

"GRID",
"X",
"Y",
"Z",
"TXY",
"TYZ",
"TZX",
"DOMAIN_ID"
};

// domain structure
typedef struct {
long long ID; // Domain identifier
long long SUBCASE; // Subcase number
long long STEP; // Step number
long long ANALYSIS; // Analysis type
double TIME_FREQ_EIGR; // Time, frequency or real part of eigen
value
double EIGI; // Imaginary part if eigen value (if applicable)
long long MODE; // Mode number
long long DESIGN_CYCLE; // Design cycle
long long RANDOM; // Random code
long long SE; // Superelement number
long long AFPM; // acounstic field point mesh id
long long TRMC; // trim component id
long long INSTANCE; // Instance
long long MODULE; // Module
} Domain;

// domain structure size

size_t dm = sizeof(Domain);

// domain field offset

size_t dmo[] = {
HOFFSET(Domain, ID),
HOFFSET(Domain, SUBCASE),
HOFFSET(Domain, STEP),
HOFFSET(Domain, ANALYSIS),
HOFFSET(Domain, TIME_FREQ_EIGR),
HOFFSET(Domain, EIGI),
HOFFSET(Domain, MODE),
HOFFSET(Domain, DESIGN_CYCLE),
HOFFSET(Domain, RANDOM),
HOFFSET(Domain, SE),
HOFFSET(Domain, AFPM),
HOFFSET(Domain, TRMC),
HOFFSET(Domain, INSTANCE),
HOFFSET(Domain, MODULE)
};

// domain field size

Main Index
 558 MSC Nastran Reference Guide
NH5RDB Data Access

Domain domain;
size_t dms[] = {
sizeof(domain.ID),
sizeof(domain.SUBCASE),
sizeof(domain.STEP),
sizeof(domain.ANALYSIS),
sizeof(domain.TIME_FREQ_EIGR),
sizeof(domain.EIGI),
sizeof(domain.MODE),
sizeof(domain.DESIGN_CYCLE),
sizeof(domain.RANDOM),
sizeof(domain.SE),
sizeof(domain.AFPM),
sizeof(domain.TRMC),
sizeof(domain.INSTANCE),
sizeof(domain.MODULE)
};

// domain field name

const char* dmn[] = {
"ID",
"SUBCASE",
"STEP",
"ANALYSIS",
"TIME_FREQ_EIGR",
"EIGI",
"MODE",
"DESIGN_CYCLE",
"RANDOM",
"SE",
"AFPM",
"TRMC",
"INSTANCE",
"MODULE"
};

// domain field type

hid_t dmt[] = {
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,
H5T_NATIVE_DOUBLE,
H5T_NATIVE_DOUBLE,
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,
H5T_NATIVE_LLONG,

Main Index
 Chapter 12: Interface With Other Programs 559
NH5RDB Data Access

H5T_NATIVE_LLONG
};

// create file
hid_t fid = H5Fcreate("test.h5", H5F_ACC_TRUNC, H5P_DEFAULT,
H5P_DEFAULT);
if (fid >= 0) {
hid_t lpid = H5Pcreate(H5P_LINK_CREATE);
if (lpid >= 0) {
if (H5Pset_create_intermediate_group(lpid, 1) >= 0) {
// create domain dataset
memset(&domain, 0, dm);
hid_t gid = H5Gcreate2(fid, result, lpid, H5P_DEFAULT,
H5P_DEFAULT);
if (gid >= 0) {
if (H5TBmake_table("", gid, dmtable, 14, 0, dm, dmn, dmo, dmt,
chunk, NULL, 1, NULL) >= 0) {
domain.ID = 1;
domain.SUBCASE = 1;
domain.ANALYSIS = 1;
H5TBappend_records(gid, dmtable, 1, dm, dmo, dms, &domain);
}
H5Gclose(gid);
}

// create displacement dataset

gid = H5Gcreate2(fid, nodal, lpid, H5P_DEFAULT, H5P_DEFAULT);
if (gid >= 0) {
if (H5TBmake_table("", gid, dptable, 8, 0, dp, dpn, dpo, dpt,
chunk, NULL, 1, NULL) >= 0) {
// populate displacement
for (int i = 0; i < 10; i++) {
disp.ID = i + 1;
disp.DOMAIN_ID = domain.ID;
disp.X = i * 1.0;
disp.Y = i * 2.0;
disp.Z = i * 3.0;
disp.RX = 0.0;
disp.RY = 0.0;
disp.RZ = 0.0;
H5TBappend_records(gid, dptable, 1, dp, dpo, dps, &disp);
}
}
H5Gclose(gid);
}

// create hexa dataset

gid = H5Gcreate2(fid, stress, lpid, H5P_DEFAULT, H5P_DEFAULT);
if (gid >= 0) {
hsize_t dim[1] = {9};
hid_t ai = H5Tarray_create2(H5T_NATIVE_LLONG, 1, dim);

Main Index
 560 MSC Nastran Reference Guide
NH5RDB Data Access

hid_t ad = H5Tarray_create2(H5T_NATIVE_DOUBLE, 1, dim);

hid_t ct = H5Tcopy(H5T_C_S1);
if (ai >= 0 && ad >= 0 && ct >= 0 && H5Tset_size(ct, 4) >= 0) {
// data types
hid_t hxt[12] = {H5T_NATIVE_LLONG, H5T_NATIVE_LLONG, ct,
H5T_NATIVE_LLONG, ai, ad, ad, ad, ad, ad, ad,
H5T_NATIVE_LLONG};
if (H5TBmake_table("", gid, hxtable, 12, 0, hx, hxn, hxo, hxt,
chunk, NULL, 1, NULL) >= 0) {
for (int i = 0; i < 10; i++) {
memset(&hexa, 0, hx);
hexa.EID = i + 1;
hexa.NODEF = 8;
strncpy(hexa.CTYPE, "GAUS", 4);
hexa.DOMAIN_ID = domain.ID;
for (int j = 1; j < 9; j++) {
hexa.GRID[j] = i * 10 + j;
hexa.X[j] = i * 10 + j;
hexa.Y[j] = i * 20 + j;
hexa.Z[j] = i * 30 + j;
hexa.TXY[j] = i * 10 + j;
hexa.TYZ[j] = i * 20 + j;
hexa.TZX[j] = i * 30 + j;
}
H5TBappend_records(gid, hxtable, 1, hx, hxo, hxs, &hexa);
}
}
}
if (ai >= 0) H5Tclose(ai);
if (ad >= 0) H5Tclose(ad);
if (ct >= 0) H5Tclose(ct);
H5Gclose(gid);
}
}
H5Pclose(lpid);
}
H5Fclose(fid);
}
}
Example8. Write nodal displacement, hexahedral element stress and domain datasets in C

 Example 9: Query and print result value by user specified domain and entities
This C++ example shows how to search records by a specified domain id and entity ids from a result dataset in the
database. This search procedure can be implemented by reading records one by one and check if ids are agree,
however it is very slow when the dataset becomes huge. To improve performance, the program can first search for
where the domain is located by using domain index dataset, which is prepared when creating NH5RDB database,
then read all the records with the specified domain into memory and find the records with the specified entity ids.
Both search procedures should use binary search because

Main Index
 Chapter 12: Interface With Other Programs 561
NH5RDB Data Access

1. Domain ids in index dataset are sorted.

2. Within the same domain id, the entity ids are also sorted.
Here is an NH5RDB example mceig.h5 generated by a sol 400 perturbation analysis job (mceig.dat). It has one
subcase with a nonlinear static step (analysis=nlstat) followed by a complex eigenvalue analysis step
(analysis=mceig). This model includes 44 grids and 12 elements.
The contents of the NH5RDB are shown in Figure 12-7, the first 11 result domains are from the nonlinear static
step. Suppose we want to query displacement values of grid 25, 31 to 38 in the 4th increment of this step, which
has domain id 4, the procedure is shown in Figure 12-7.

Figure 12-7 Process flow of the query

For details to build and run the program, read README.txt in example 9.

Main Index
 562 MSC Nastran Reference Guide
NH5RDB Data Access

Here shows source code of below two functions, check nh5qry.cpp for other details in example 9. Below is the
function to search domain:
//
// Purpose: search the start position and length (number of rows) of a
domain
// in a dataset
// Input:
// hid_t file: hdf5 file id
// const std::string& restype0: result type
// long long domain_id: domain id to be searched
// output:
// long long& pos: the start row position of the domain
// long long& length: number of rows of the domain
// return:
// 0: successfull
// !0: fail
//
int search_domain(hid_t file, const std::string& restype0, long long
domain_id,
// output
long long &pos, long long& length )
{
hid_t index_dset = 0;
std::string fullname;
int err = get_dset(file, restype0, true, index_dset, fullname);
if (err != 0)
return err;

//
// read domain index dataset into memory
//
// get number of rows of domain index dataset
hid_t space_id = H5Dget_space(index_dset);
hsize_t dims_out[4];
int status_n = H5Sget_simple_extent_dims(space_id, dims_out, NULL);
hsize_t total_row = dims_out[0];

// read all rows in one call

std::vector<DomainIndex_T> indics;
err = read_h5(index_dset, 0, total_row, indics);
if (err != 0)
return err;

// binary search the domain, domain dataset is sorted based on

DOMAIN_ID
DomainIndex_T index = { domain_id, 0, 0 };
auto bound = std::lower_bound(indics.begin(), indics.end(), index,
DomainIndex_Cmp());
if (bound != indics.end() && bound->DOMAIN_ID == domain_id)
{

Main Index
 Chapter 12: Interface With Other Programs 563
NH5RDB Data Access

pos = bound->POSITION;
length = bound->LENGTH;

return 0;
}
else
{
fprintf(stderr, "No domain %d found in the damain index dataset
%s\n",
domain_id, fullname.data());
return -1;
}
}

Below is the function to search records of specified ids in a domain:

//
// Purpose: search and print out records of specified ids in a domain
// Input:
// hid_t file: hdf5 file id
// long long domain_id: domain id to be searched
// const std::string& restype0: result type
// DISPLACEMENT, DISPLACEMENT_CPLX, QUAD4, QUAD_CN are valid words
// const std::string& id_str: a srting to specify ids
// "1:6 8 10" means 1 through 6, 8 and 10.
// return:
// 0: successful
// Non 0: fail
//
template<typename T>
int do_search(hid_t file, long long domain_id, const std::string&
restype0,
const std::string& id_str)
{
long long pos, length;
int err = search_domain(file, restype0, domain_id, pos, length);

if(err == 0 && length > 0)

{
hid_t dset = 0;
std::string fullname;
int err = get_dset(file, restype0, false, dset, fullname);
if (err != 0)
return err;

printf("Row range of domain %ld, %ld %ld\n", domain_id,

pos, pos + length);
std::vector<T> displacements;
print_head(displacements);

//
// read all rows into memory and call binary search algorithm.

Main Index
 564 MSC Nastran Reference Guide
NH5RDB Data Access

// if the number of rows is huge, it is better to implement a

binary
// search algorithm to the hdf5 file directory for memory
efficiency
//
// std::vector<typename T> displacements;
err = read_h5(dset, pos, length, displacements);
if (err != 0)
{
H5Fclose(file);
return err;
}

std::vector<long long> ids = parse_id_string(id_str);

for(auto id_it = ids.begin(); id_it != ids.end(); ++id_it)
{
long long id = *id_it;
//records in a dataset is sorted based on Nastran sort1,binary
search
// the interesting ID
T disp0 = { id };
auto bound_rec = std::lower_bound(displacements.begin(),
displacements.end(), disp0, CompareID<T>());
if (bound_rec != displacements.end() && match_id(*bound_rec,
id))
{
auto const & disp = *bound_rec;
size_t dist = std::distance(displacements.begin(),
bound_rec);
print_row(disp);
}
else
fprintf(stderr, "Cannot find record of id %ld from dataset
\n", id);
}
}

return 0;
}
Example 9. Query and print result value by user specified domain and entities

 Example 10: Read nodal displacement dataset in Java

This example demonstrates how to access dataset in NH5RDB in Java. The Java HDF Object Package, which wraps
HDF Java interfaces in an object model, is used in the program. The package consists of several Java jar files and
shared libraries. It can be obtained from HDF group website.
In the example, the nodal displacement dataset is read in first. Then each field data, including ID, displacement
component and DOMAIN_ID, is obtained and print out to the system output.
/*

Main Index
 Chapter 12: Interface With Other Programs 565
NH5RDB Data Access

* This example uses Java HDF Object Package to access NH5RDB.

*
*/
import hdf.object.*;
import java.util.*;

/**
* The NH5RDB class defines methods to access NH5RDB.
*
*/
public class NH5RDB {
/**
* Read nodal displacement dataset and print to system output
*
* @param file the NH5RDB file name
* @throws Exception
*/
public void printDisplacement(String file) throws Exception {
FileFormat ff = null;
try {
ff =
FileFormat.getFileFormat(FileFormat.FILE_TYPE_HDF5).createInstance(
file, FileFormat.READ);
ff.open();
CompoundDS ds = (CompoundDS)
ff.get("NASTRAN/RESULT/NODAL/DISPLACEMENT");
List data = (List) ds.getData();
long[] ID = (long[]) data.get(0);
double[] X = (double[]) data.get(1);
double[] Y = (double[]) data.get(2);
double[] Z = (double[]) data.get(3);
double[] RX = (double[]) data.get(4);
double[] RY = (double[]) data.get(5);
double[] RZ = (double[]) data.get(6);
long[] DOMAIN_ID = (long[]) data.get(7);
for (int i = 0; i < ID.length; i++) {
System.out.println("ID: " + ID[i] + ", Domain: " + DOMAIN_ID[i]);
System.out.println("\tX = " + X[i]);
System.out.println("\tY = " + Y[i]);
System.out.println("\tZ = " + Z[i]);
System.out.println("\tRX = " + RX[i]);
System.out.println("\tRY = " + RY[i]);
System.out.println("\tRZ = " + RZ[i]);
}
} finally {
if (ff != null) {
ff.close();
}
}
}

Main Index
 566 MSC Nastran Reference Guide
NH5RDB Data Access

/**
* @param args the command line arguments
*/
public static void main(String[] args) {
try {
NH5RDB h5 = new NH5RDB();
h5.printDisplacement("test.h5");
} catch (Exception e) {
e.printStackTrace();
}
}
}

Example 10: Read nodal displacement in Java

 Python utility function

For post processing demonstration, python functions to calculate principle and Von Mises stress are given here. Also
a program to find the maximum Von Mises stress in frequency response analysis is shown below. These python
functions can be used as utility functions for user in post processing applications.
• Calculate principal and Von Mises stress from stress dataset
import sys
import os
import math
import numpy

class ZeroStressException(Exception) :
# exception needed to jump out when all stress/strain components
# are zero.
def __str__(self):
return repr(self.value)

def _cal_prstrs(stress) :
# principalStress(stress, pstr, dirc, idg33, strn, sn, so, iloc) :
# --------------------------------------------------------------
#

#***********************************************************************
#
# Description: Principal and Invarient Values in 3D
#
# Purpose: Given normal and shear stresses in STR this routine
# attempts to find principal stresses PSTR and their direction
# cosines DCOS. Invarient for Mean Pressure and VonMises will also
# be returned.
#
# Arguments:
# STRESS[6] double,input Normal and Shear Stresses
# PSTR[3] double,output Principal Stresses

Main Index
 Chapter 12: Interface With Other Programs 567
NH5RDB Data Access

# DIRC[3][3] double,outputDirection Cosines

# idg33 integer,inputIndicator for Invariant Type
# 0->Oct 1->VonMises
# strn logical,input Indicator for strain versus stress
# false->stress true->strain
# sn double,output Mean Pressure
# so double,output Oct Shear or VonMises
# iloc integer,input Location in entry for MAXMIN processing
#
# Method: Cubic Equation for Principal Stresses in 3 dimensions is:
#
# 3 2 2 2 2
# S -(S +S +S )S + (S S + S S + S S - T - T - T )S
# X Y Z X Y Y Z Z X XY YX ZX
#
# 2 2 2
# - (S S S + 2 T T T - S T - S T - S T ) = 0
# X Y Z XY YZ ZX X YZ Y ZX Z XY
#
# The three roots S , S , and S are the Principal Stresses.
# 1 2 3
#

#***********************************************************************
#
# --- constants
power = 1.0 / 3.0
degrad = math.pi / 180.

# --- local arrays

prin = [0.0, 0.0, 0.0 ]
s = [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
temp = [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0]
# --- removed parameters, now used as local variables
dirc = [ [0.0, 0.0, 0.0 ],[0.0, 0.0, 0.0 ],[0.0, 0.0, 0.0 ]]
pstr = [0.0, 0.0, 0.0 ]
idg33 = 0
strn = False

# --- scaling stresses to be in range -1.0 to +1.0

equal = False
rm = 0.0
fl = ( stress[0] + stress[1] + stress[2] ) / 3.0
for i in (0, 1, 2) :
if ( abs(stress[i] - fl) > rm ) : rm = abs(stress[i] - fl)
if ( abs(stress[i+3]) > rm ) : rm = abs(stress[i+3])
#
# --- create exception for zero stress/strain values
try :
if rm <= 0.0 : # former goto 180
raise ZeroStressException

Main Index
 568 MSC Nastran Reference Guide
NH5RDB Data Access

# --- special cases

# --- normalize stresses
# --- former label 180
sx = (stress[0] -fl )/rm
sy = (stress[1] -fl )/rm
sz = (stress[2] -fl )/rm
txy = stress[3] / rm
tyz = stress[4] / rm
tzx = stress[5] / rm
#
# The above cubic is of the form,
#
# 3 2
# Y + PY + QY + R = 0
#
# and substituting for Y the value X - P/3 gives,
#
# 3
# X + AX + B = 0
#
# Now finding the three roots (Values of X)
#
txysq = txy**2
tyzsq = tyz**2
tzxsq = tzx**2
sxsy = sx * sy
p = -sx -sy -sz
q = sxsy + sy*sz + sz*sx - txysq - tyzsq - tzxsq
r = -sxsy*sz -2.0 *txy*tyz*tzx + sx*tyzsq + sy*tzxsq +
sz*txysq
smalla = q - p**2/3.0
smallb = (2.0*p**3 - 9.0*p*q)/27.0 + r
#
# --- for principal stresses we assume no imaginary roots
pover3 = p / 3.0
quan = smallb**2/4.0 + smalla**3/27.0
if (quan < -0.0000001 ) : # --- former goto 400
# --- former label 400
# here three real unequal roots (trig solution)
quan = 2.0 * math.sqrt(abs(smalla**3/27.0))
if (quan <= 0.0 ) :
# -- former goto 150
raise ZeroStressException
#
phiov3 = math.acos(-smallb/quan)/3.0
quan = 2.0 * math.sqrt(abs(smalla/3.0))
prin[0] = quan * math.cos(phiov3) - pover3
prin[1] = quan * math.cos(phiov3 + 120.0 * degrad) - pover3
prin[2] = quan * math.cos(phiov3 + 240.0 * degrad) - pover3
else :
# here at least two roots are equal

Main Index
 Chapter 12: Interface With Other Programs 569
NH5RDB Data Access

# kpr will point to the unequal root after "label 1000" sort
kpr = 0
bover2 = -smallb/2.0
aplusb = 2.0 * abs(bover2)**power
if bover2 < 0.0 :
aplusb = -aplusb
kpr = 1
prin[0] = aplusb - pover3
prin[1] = -aplusb/2.0 - pover3
prin[2] = prin[1]
equal = True
# --- former goto 1000
# --- here principal stresses are known
# --- former label 1000
pstr[0] = prin[0]
pstr[1] = prin[1]
pstr[2] = prin[2]
prin[0] = max( [ pstr[0], pstr[1], pstr[2] ] )
prin[1] = min( [ pstr[0], pstr[1], pstr[2] ] )
prin[2] = pstr[0] + pstr[1] + pstr[2] - prin[0] - prin[1]
pstr[0] = rm * prin[0] + fl
pstr[1] = rm * prin[1] + fl
pstr[2] = rm * prin[2] + fl
#
# ---- Solving for Direction COSINES. --------------------------
------
#
# Where S is a Principal Stress and K,L,M are the Direction
Cosines
#
# A solution of the Direction Cosines can be found by solving any
# two of the following three equations with the fourth.
#
#
# (S -S)K + T L + T M = 0 (1)
# X XY ZX
#
#
# T K + (S -S)L + T M = 0 (2)
# XY Y YZ
#
#
# T K + T L + (S -S)M = 0 (3)
# ZX YZ Z
#
#
# 2 2 2
# K + L + M = 1
#
# --------------------------------------------------------------
------

Main Index
 570 MSC Nastran Reference Guide
NH5RDB Data Access

#
for k in range(3) : # former do 2000 k=1,3
if ( equal and ( k != kpr ) ) : # former goto 1400
for j in ( 0, 1, 2) :
dirc[j][k] = 0.0
else :
s[0] = sx - prin[k]
s[1] = txy
s[2] = tzx
s[3] = txy
s[4] = sy - prin[k]
s[5] = tyz
s[6] = tzx
s[7] = tyz
s[8] = sz - prin[k]
#
# Select 2 vectors whose cross products magnitude is non-
zero.
# itk: calling a daxb equivalent will not gonna work,
because
# it depends on fortran call by reference convention. we
will
# have to write out what's done inside daxb.
# ( may be array striping could also have been used )
# D(1) = A(2)*B(3) - A(3)*B(2)
# D(2) = A(3)*B(1) - A(1)*B(3)
# D(3) = A(1)*B(2) - A(2)*B(1)
# CALL DAXB( S(1), S(4), TEMP(1) )
temp[0] = s[1] * s[5] - s[2] * s[4]
temp[1] = s[2] * s[3] - s[0] * s[5]
temp[2] = s[0] * s[4] - s[1] * s[3]
# CALL DAXB( S(4), S(7), TEMP(4) )
temp[3] = s[4] * s[8] - s[5] * s[7]
temp[4] = s[5] * s[6] - s[3] * s[8]
temp[5] = s[3] * s[7] - s[4] * s[6]
# CALL DAXB( S(7), S(1), TEMP(7) )
temp[6] = s[7] * s[2] - s[8] * s[1]
temp[7] = s[8] * s[0] - s[6] * s[2]
temp[8] = s[6] * s[1] - s[7] * s[0]
#

jj = 0
fmax = 0.0
# -- start former 1380 loop
# for j in range(0,8,3) :
for j in (0,3,6) :
fl = math.sqrt(temp[j]**2 + temp[j+1]**2 +
temp[j+2]**2)
if fl > fmax :
jj = j
fmax = fl

Main Index
 Chapter 12: Interface With Other Programs 571
NH5RDB Data Access

# -- end of former 1380 loop

# IF( FMAX .LE. 0.0000001D0 ) GO TO 1400

if not (fmax <= 0.0000001 ) :
for j in (0,1,2) :
dirc[j][k] = temp[jj] / fmax
jj += 1
# former goto 2000
else :
# Ill condition (can not determine unique direction
of this
# Principal Stress.)
# former 1400 label
for j in ( 0, 1, 2) :
dirc[j][k] = 0.0
# former 2000 label

except ZeroStressException :
# --- special case when all six components are zero
for i in range(3) : # former 150 do-loop
pstr[i] = fl
dirc[0][i] = 0.0
dirc[1][i] = 0.0
dirc[2][i] = 0.0

# former 2500 label

sn = -(pstr[0] + pstr[1] + pstr[2])/3.0
voct =
math.sqrt(((pstr[0]+sn)**2+(pstr[1]+sn)**2+(pstr[2]+sn)**2)/3.0)
vonm = voct*3.0/math.sqrt(2.0)
# if ( strn ):
# vonm = 2.0*vonm/3.0
#
# if idg33 == 0 :
# so = voct
# else :
# so = vonm
return pstr

def _cal_vonmises_stress(str) :
""" Calculates Von Mises stress from six stress components.
Arguments:
str tuple containing the six stress components
Returns:
vonmises stress
"""
aa = (str[0] - str[1])**2 + (str[1] - str[2])**2 + (str[2] -
str[0])**2 +\
6*( str[3]**2 + str[4]**2 + str[5]**2)
if ( aa > 0.0 ) :

Main Index
 572 MSC Nastran Reference Guide
NH5RDB Data Access

vonm = math.sqrt(aa / 2.0)

else :
# -- in theory we should never get here
vonm = 0.0
return vonm

def _cal_vonmises_prinstress(pstr):
""" Calculates Von Mises stress from principal stresses.
Arguments:
pstrs array containing the three principal stresses
Returns:
vonmises stress
"""
sn = -(pstr[0] + pstr[1] + pstr[2])/3.0
voct =
math.sqrt(((pstr[0]+sn)**2+(pstr[1]+sn)**2+(pstr[2]+sn)**2)/3.0)
vonm = voct*3.0/math.sqrt(2.0)
return vonm

if __name__ == "__main__" :

def _check_stress( stresses ) :

""" calculates and prints principal stresses and vonmises stress
from six stress components passed as tuple in stresses
"""
pstr = _cal_prstrs( stresses)
vonm = _cal_vonmises_prinstress(pstr)
vonm2 = _cal_vonmises_stress(stresses)
print("%15.6e %15.6e %15.6e --> %15.6e %15.6e"% \
(pstr[0],pstr[1],pstr[2], vonm,vonm2))

_check_stress((2.0, 0.0, 0.0, 0.0, 0.0, 0.0))

# pstr = _cal_prstrs((-38971184.0, -450032.6875, 3.2782554626464844e-
07, \
# -1299021.125, -17.771186828613281, -200944.828125))
_check_stress((2.0, 0.0, 0.0, 0.0, 0.0, 0.0))
_check_stress((0.0, 2.0, 0.0, 0.0, 0.0, 0.0))
_check_stress((0.0, 0.0, 2.0, 0.0, 0.0, 0.0))
print(" --- ")
_check_stress((1.0, 1.0, 0.0, 0.0, 0.0, 0.0))
_check_stress((0.0, 1.0, 1.0, 0.0, 0.0, 0.0))
_check_stress((1.0, 0.0, 1.0, 0.0, 0.0, 0.0))
print(" --- ")
_check_stress((0.0, 0.0, 0.0, 1.0, 0.0, 0.0))
_check_stress((0.0, 0.0, 0.0, 0.0, 1.0, 0.0))
_check_stress((0.0, 0.0, 0.0, 0.0, 0.0, 1.0))
print(" --- ")
_check_stress((0.0, 0.0, 0.0, 2.0, 0.0, 0.0))
_check_stress((0.0, 0.0, 0.0, 0.0, 2.0, 0.0))
_check_stress((0.0, 0.0, 0.0, 0.0, 0.0, 2.0))
print(" --- ")

Main Index
 Chapter 12: Interface With Other Programs 573
NH5RDB Data Access

_check_stress((1.0, -1.0, 0.0, 0.0, 0.0, 0.0))

• Calculate maximum Von Mises stress in frequency response analysis

The python program calculates maximum Von Mises stress from complex stress tensor in frequency analysis. Both
2D and 3D element calculation methods are given.
from math import *
import tables
# import numpy
#
# script to calculate true von Mises stress from complex stress tensor
# in frequency response analysis.
# CTRIA3 (top&bottom), CQADU4 (top&bottom), CTETRA (centroid),
# CPENTA(centroid) and CHEXA(centroid) are supported.
# Results are written to .csv file.
#
# Based on paper from MSC World Conference in 1993:
# "EXACT CALCULATION OF MINIMUM MARGIN OF SAFETY FOR FREQUENCY RESPONSE
# ANALYSIS STRESS RESULTS USING YIELDING OR FAILURE THEORIES"
#
# Usage :
# python calc_vm_from_h5_2d3d.py <jobname.h5> <vonmisesstress.csv>
# with
# jobname.h5 the Hdf5 file resulting from the Nastran job
# vonmises.csv the csv output file with the Von Mises stresses
# default is the name of the Hdf5 file with the
# extension .h5 replaced by .csv

#
# 2D element complex von mises function
#
def vM2D(strD) :
# bottom
s1=sqrt(strD[0]**2+strD[1]**2)
a1=atan(strD[1]/strD[0])
if strD[0] < 0.0 :
a1=a1+pi
s2=sqrt(strD[2]**2+strD[3]**2)
a2=atan(strD[3]/strD[2])
if strD[2] < 0.0 :
a2=a2+pi
s12=sqrt(strD[4]**2+strD[5]**2)
a12=atan(strD[5]/strD[4])
if strD[4] < 0.0 :
a12=a12+pi
rhs=(1.0*( (s1*s1*sin(2.0*a1)) + (s2*s2*sin(2.0*a2)) +
3.0*(s12*s12*sin(2.0*a12)) - (abs(s1)*abs(s2)*sin(a1+a2)) )
/
( (s1*s1*cos(2.0*a1)) + (s2*s2*cos(2.0*a2)) +

Main Index
 574 MSC Nastran Reference Guide
NH5RDB Data Access

3.0*(s12*s12*cos(2.0*a12)) - (abs(s1)*abs(s2)*cos(a1+a2)) )
)
o1=-0.5*atan(rhs)
svm1=sqrt( (s1*s1*cos(a1+o1)*cos(a1+o1)) +
(s2*s2*cos(a2+o1)*cos(a2+o1)) +
3.0*(s12*s12*cos(a12+o1)*cos(a12+o1)) -
(abs(s1)*abs(s2)*cos(a1+o1)*cos(a2+o1)) )
o1=o1+pi/2.0
svm2=sqrt( (s1*s1*cos(a1+o1)*cos(a1+o1)) +
(s2*s2*cos(a2+o1)*cos(a2+o1)) +
3.0*(s12*s12*cos(a12+o1)*cos(a12+o1)) -
(abs(s1)*abs(s2)*cos(a1+o1)*cos(a2+o1)) )
cvm1=max(svm1,svm2)
# top
s1=sqrt(strD[6]**2+strD[7]**2)
a1=atan(strD[7]/strD[6])
if strD[6] < 0.0 :
a1=a1+pi
s2=sqrt(strD[8]**2+strD[9]**2)
a2=atan(strD[9]/strD[8])
if strD[8] < 0.0 :
a2=a2+pi
s12=sqrt(strD[10]**2+strD[11]**2)
a12=atan(strD[11]/strD[10])
if strD[10] < 0.0 :
a12=a12+pi
rhs=( 1.0 * ( (s1*s1*sin(2.0*a1)) + (s2*s2*sin(2.0*a2)) +
3.0*(s12*s12*sin(2.0*a12)) - (abs(s1)*abs(s2)*sin(a1+a2))
) /
( (s1*s1*cos(2.0*a1)) + (s2*s2*cos(2.0*a2)) +
3.0*(s12*s12*cos(2.0*a12)) - (abs(s1)*abs(s2)*cos(a1+a2))
) )
o1=-0.5*atan(rhs)
svm1=sqrt( (s1*s1*cos(a1+o1)*cos(a1+o1)) +
(s2*s2*cos(a2+o1)*cos(a2+o1)) +
3.0*(s12*s12*cos(a12+o1)*cos(a12+o1)) -
(abs(s1)*abs(s2)*cos(a1+o1)*cos(a2+o1)) )
o1=o1+pi/2.0
svm2=sqrt( (s1*s1*cos(a1+o1)*cos(a1+o1)) +
(s2*s2*cos(a2+o1)*cos(a2+o1)) +
3.0*(s12*s12*cos(a12+o1)*cos(a12+o1)) -
(abs(s1)*abs(s2)*cos(a1+o1)*cos(a2+o1)) )
cvm2=max(svm1,svm2)

return [cvm1,cvm2]
#
# 3D elements complex von mises function
#
def vM3D(strD) :
s1=sqrt(strD[0]**2 + strD[1]**2)
a1=atan(strD[1]/strD[0])

Main Index
 Chapter 12: Interface With Other Programs 575
NH5RDB Data Access

if strD[0] < 0.0 :

a1=a1+pi
s2=sqrt(strD[2]**2 + strD[3]**2)
a2=atan(strD[3]/strD[2])
if strD[2] < 0.0 :
a2=a2+pi
s3=sqrt(strD[4]**2 + strD[5]**2)
a3=atan(strD[5]/strD[4])
if strD[4] < 0.0 :
a3=a3+pi
s12=sqrt(strD[6]**2 + strD[7]**2)
a12=atan(strD[7]/strD[6])
if strD[6] < 0.0 :
a12=a12+pi
s23=sqrt(strD[8]**2 + strD[9]**2)
a23=atan(strD[9]/strD[8])
if strD[8] < 0.0 :
a23=a23+pi
s31=sqrt(strD[10]**2 + strD[11]**2)
a31=atan(strD[11]/strD[10])
if strD[10] < 0.0 :
a31=a31+pi
#
rhs=(1.0*( (s1*s1*sin(2.0*a1)) + (s2*s2*sin(2.0*a2)) +
(s3*s3*sin(2.0*a3)) +
3.0*((s12*s12*sin(2.0*a12)) + (s23*s23*sin(2.0*a23)) +
(s31*s31*sin(2.0*a31))) - (abs(s1)*abs(s2)*sin(a1+a2)) -
(abs(s1)*abs(s3)*sin(a1+a3)) - (abs(s2)*abs(s3)*sin(a2+a3))
) /
( (s1*s1*cos(2.0*a1)) + (s2*s2*cos(2.0*a2)) +
(s3*s3*cos(2.0*a3)) +
3.0*((s12*s12*cos(2.0*a12)) + (s23*s23*cos(2.0*a23)) +
(s31*s31*cos(2.0*a31))) - (abs(s1)*abs(s2)*cos(a1+a2)) -
(abs(s1)*abs(s3)*cos(a1+a3)) - (abs(s2)*abs(s3)*cos(a2+a3))
) )
o1=-0.5*atan(rhs)
svm1=sqrt( (s1*s1*cos(a1+o1)*cos(a1+o1)) +
(s2*s2*cos(a2+o1)*cos(a2+o1)) +
(s3*s3*cos(a3+o1)*cos(a3+o1)) +
3.0*((s12*s12*cos(a12+o1)*cos(a12+o1)) +
(s23*s23*cos(a23+o1)*cos(a23+o1)) +
(s31*s31*cos(a31+o1)*cos(a31+o1))) -
(abs(s1)*abs(s2)*cos(a1+o1)*cos(a2+o1)) -
(abs(s1)*abs(s3)*cos(a1+o1)*cos(a3+o1)) -
(abs(s2)*abs(s3)*cos(a2+o1)*cos(a3+o1) ) )
o1=o1+pi/2.0
svm2=sqrt( (s1*s1*cos(a1+o1)*cos(a1+o1)) +
(s2*s2*cos(a2+o1)*cos(a2+o1)) +
(s3*s3*cos(a3+o1)*cos(a3+o1)) +
3.0*((s12*s12*cos(a12+o1)*cos(a12+o1)) +
(s23*s23*cos(a23+o1)*cos(a23+o1)) +

Main Index
 576 MSC Nastran Reference Guide
NH5RDB Data Access

(s31*s31*cos(a31+o1)*cos(a31+o1))) -
(abs(s1)*abs(s2)*cos(a1+o1)*cos(a2+o1)) -
(abs(s1)*abs(s3)*cos(a1+o1)*cos(a3+o1)) -
(abs(s2)*abs(s3)*cos(a2+o1)*cos(a3+o1) ) )
cvm1=max(svm1,svm2)

return cvm1
# main function
def main(infile,outfile) :
#
# Open database file
#
fh5 = tables.open_file(infile)
outf=open(outfile,'w')
# Get element stress tables
try :
stressQ4 = fh5.root.NASTRAN.RESULT.ELEMENTAL.STRESS.QUAD4_CPLX
hasQ4 = True
except :
hasQ4 = False
try :
stressT3 = fh5.root.NASTRAN.RESULT.ELEMENTAL.STRESS.TRIA3_CPLX
hasT3 = True
except :
hasT3 = False
try :
stressTet = fh5.root.NASTRAN.RESULT.ELEMENTAL.STRESS.TETRA_CPLX
hasTet = True
except :
hasTet = False
try :
stressPen = fh5.root.NASTRAN.RESULT.ELEMENTAL.STRESS.PENTA_CPLX
hasPen = True
except :
hasPen = False
try :
stressHex = fh5.root.NASTRAN.RESULT.ELEMENTAL.STRESS.HEXA_CPLX
hasHex = True
except :
hasHex = False
domain1 = fh5.root.NASTRAN.RESULT.DOMAINS
freq=[]
doma=[]
subc=[]
outf.write('FREQ,SUBCASE ID')
for row in domain1 :
freq.append(row["TIME_FREQ_EIGR"])
doma.append(int(row["ID"]))
subc.append(int(row["SUBCASE"]))
#
# QUAD4

Main Index
 Chapter 12: Interface With Other Programs 577
NH5RDB Data Access

#
if hasQ4 == True :
strDQ4={}
elemQ4=[]
for row in stressQ4 :
strDQ4[int(row["DOMAIN_ID"]), int(row["EID"])]=[
float(row["X1R"]), float(row["X1I"]),
float(row["Y1R"]), float(row["Y1I"]),
float(row["XY1R"]), float(row["XY1I"]),
float(row["X2R"]), float(row["X2I"]),
float(row["Y2R"]), float(row["Y2I"]),
float(row["XY2R"]), float(row["XY2I"]) ]
if int(row["EID"]) not in elemQ4:
elemQ4.append(int(row["EID"]))
for el1 in elemQ4 :
tmp=',%i BOTTOM,%i TOP' %(el1,el1)
outf.write(tmp)
#
# TRIA3
#
if hasT3 == True :
strDT3={}
elemT3=[]
for row in stressT3 :
strDT3[int(row["DOMAIN_ID"]), int(row["EID"])]=[
float(row["X1R"]), float(row["X1I"]),
float(row["Y1R"]), float(row["Y1I"]),
float(row["TXY1R"]),float(row["TXY1I"]),
float(row["X2R"]), float(row["X2I"]),
float(row["Y2R"]), float(row["Y2I"]),
float(row["TXY2R"]),float(row["TXY2I"]) ]
if int(row["EID"]) not in elemT3:
elemT3.append(int(row["EID"]))
for el1 in elemT3 :
tmp=',%i BOTTOM,%i TOP' %(el1,el1)
outf.write(tmp)
#
# TETRA
#
if hasTet == True :
elemTet=[]
strDTet={}
for row in stressTet :
strDTet[int(row["DOMAIN_ID"]), int(row["EID"])]=[
float(row["XR"][0]), float(row["XI"][0]),
float(row["YR"][0]), float(row["YI"][0]),
float(row["ZR"][0]), float(row["ZI"][0]),
float(row["TXYR"][0]), float(row["TXYI"][0]),
float(row["TYZR"][0]), float(row["TYZI"][0]),
float(row["TZXR"][0]), float(row["TZXI"][0]) ]
if int(row["EID"]) not in elemTet:

Main Index
 578 MSC Nastran Reference Guide
NH5RDB Data Access

elemTet.append(int(row["EID"]))
for el1 in elemTet :
tmp=',%i CENTROID' %(el1)
outf.write(tmp)
#
# PENTA
#
if hasPen == True :
elemPen=[]
strDPen={}
for row in stressPen :
strDPen[int(row["DOMAIN_ID"]), int(row["EID"])]=[
float(row["XR"][0]), float(row["XI"][0]),
float(row["YR"][0]), float(row["YI"][0]),
float(row["ZR"][0]), float(row["ZI"][0]),
float(row["TXYR"][0]), float(row["TXYI"][0]),
float(row["TYZR"][0]), float(row["TYZI"][0]),
float(row["TZXR"][0]), float(row["TZXI"][0]) ]
if int(row["EID"]) not in elemPen:
elemPen.append(int(row["EID"]))
for el1 in elemPen :
tmp=',%i CENTROID' %(el1)
outf.write(tmp)
#
# HEXA
#
if hasHex == True :
elemHex=[]
strDHex={}
for row in stressHex :
strDHex[int(row["DOMAIN_ID"]),int(row["EID"])]=[
float(row["XR"][0]), float(row["XI"][0]),
float(row["YR"][0]), float(row["YI"][0]),
float(row["ZR"][0]), float(row["ZI"][0]),
float(row["TXYR"][0]), float(row["TXYI"][0]),
float(row["TYZR"][0]), float(row["TYZI"][0]),
float(row["TZXR"][0]), float(row["TZXI"][0]) ]
if int(row["EID"]) not in elemHex:
elemHex.append(int(row["EID"]))
for el1 in elemHex :
tmp=',%i CENTROID' %(el1)
outf.write(tmp)
cnt=0
outf.write('\n')
#
while cnt < len(freq) :
fr1=freq[cnt]
dom1=doma[cnt]
subc1=subc[cnt]
cnt=cnt+1
outf.write('%12.6f,%i,' %(fr1,subc1))

Main Index
 Chapter 12: Interface With Other Programs 579
NH5RDB Data Access

stressVM=[]
#
# QUAD4
#
if hasQ4 == True :
for el1 in elemQ4 :
cvm=vM2D(strDQ4[dom1,el1])
stressVM.append(cvm[0])
stressVM.append(cvm[1])
#
# TRIA3
#
if hasT3 == True :
for el1 in elemT3 :
cvm=vM2D(strDT3[dom1,el1])
stressVM.append(cvm[0])
stressVM.append(cvm[1])
#
# TETRA
#
if hasTet == True :
for el1 in elemTet :
cvm=vM3D(strDTet[dom1,el1])
stressVM.append(cvm)
#
# PENTA
#
if hasPen == True :
for el1 in elemPen :
cvm=vM3D(strDPen[dom1,el1])
stressVM.append(cvm)
#
# HEXA
#
if hasHex == True :
for el1 in elemHex :
cvm=vM3D(strDHex[dom1,el1])
stressVM.append(cvm)
for st in stressVM :
outf.write('%12.6f,' %(st))
outf.write('\n')

if __name__ == '__main__' :
import sys
infile=sys.argv[1]
if len(sys.argv) > 2 :
outfile=sys.argv[2]
else :
outfile=infile[:-3] + '_stress.csv'
main(infile,outfile)

Main Index
 580 MSC Nastran Reference Guide
NH5RDB Data Access

Example 11: Print nodal displacement by module

This example extracts nodal displacement in a multiple module model and prints out the displacement by each module.
import os
import sys
import tables
# Usage: python print_module_disp.py test.h5

if __name__ == '__main__' :
if len(sys.argv) < 2:
print('Please give the file name')
sys.exit(1)

try:
# Open h5 file
file = tables.open_file(sys.argv[1])

# Get domain ID list for module number

domains = file.root.NASTRAN.RESULT.DOMAINS
d = {}
for domain in domains:
m = domain['MODULE']
id = domain['ID']
if m in d:
d[m].append(id)
else:
d[m] = [id]

# Print nodal displacement for each module

disp = file.root.NASTRAN.RESULT.NODAL.DISPLACEMENT
for m in sorted(d.keys()):
l = d[m]
# Print module number
print('Module = {0}:'.format(str(m)))
for id in l:
s = "DOMAIN_ID == {0}".format(str(id))
# Find nodal displacement of the module and print
for row in disp.where(s):

Main Index
 Chapter 12: Interface With Other Programs 581
NH5RDB Data Access

gid = row['ID']
x = row['X']
y = row['Y']
z = row['Z']
# Print node id, X, Y, Z and domain
print('ID = {0}, {1}, {2}, {3}, {4}'.format(gid, x, y, z,
id))

# Close file
file.close()
except Exception as e:
print('Exception: {0}'.format(str(e)))

Schema document
The XML schema is distributed with Nastran. With XML schema, an HTML document is provided for dataset
descriptions. See reference section for these documents.

Figure 12-8 NH5RDB schema HTML document

Main Index
 582 MSC Nastran Reference Guide
NH5RDB Data Access

Reference
MSC Nastran Quick Reference Guide
MSC Nastran DMAP Programmer’s Guide
HDF5 Software Documentation

Examples Files
 DataType_v20214.html
 DataType_v20214.xml
 Example 1
 Example 2
 Example 3
 Example 4
 Example 5
 Example 6
 Example 7
 Example 8a
 Example 8b
 Example 9
 Example 10
 Utility Function

Main Index
 Appendix A: Numerical Accuracy Considerations
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

A Numerical Accuracy
Considerations
 Introduction 584
 Linear Equation Solution 584

Eigenvalue Analysis 584

Matrix Conditioning 585

Numerical Accuracy Issues 586
 Sources of Mechanisms 587

Sources of Nonpositive Definite Matrices 588

Detection and Avoidance of Numerical Problems 589

Main Index
 584 MSC Nastran Reference Guide
Introduction

Introduction
MSC Nastran is an advanced finite element analysis (FEA) program that can solve models in excess of 50 million degrees
of freedom. The numerical analysis capabilities of MSC Nastran are continually enhanced to provide the highest level of
speed and accuracy.
This section provides a brief overview for detecting and avoiding numerical ill-conditioning problems, especially as they
relate to dynamic analysis. For more information regarding the numerical analysis algorithms of MSC Nastran, see the MSC
Nastran Numerical Methods User’s Guide.

Linear Equation Solution

The basic statement of the linear equation solution is:
Ax = b
where  A  is a square matrix of known coefficients (and usually symmetric for structural models),  b  is a known vector,
and  x  is the unknown vector to be determined.
MSC Nastran has two methods to solve the set of linear equations: the direct and the iterative method. For the remainder
of this section, only the direct method will be considered.
The methods used for solution in MSC Nastran are based on a decomposition of  A  to triangular matrices and are
followed by forward-backward substitution to get  x  . The equations for this solution technique are:

A = LU
where  L  is a lower-triangular matrix and  U  an upper-triangular matrix, and

Ly = b (1-1)

where  y  is an intermediate vector. Equation (A-1) is called the forward pass because the solution starts with the first row
where there is only one unknown due to the triangular form of  L  . The backward pass starts with the last row and provides
the solution
Ux = y

Eigenvalue Analysis
The general eigensolution equation is
2
 K + pB + p M   u  = 0 (1-2)

Main Index
 Appendix A: Numerical Accuracy Considerations 585
Matrix Conditioning

where p is the complex eigenvalue. This equation can always be transformed to a special eigenvalue problem for a matrix
A
A –  I = 0 (1-3)

where  I  is the identity matrix. Equation (A-3) is the basis of all the transformation methods of MSC Nastran (HOU, GIV,
etc.). The iterative methods (INV, SINV) work directly from equation (A-2). The Lanczos method uses both. If  A  is a
symmetric matrix, the eigenvectors are orthogonal, and they can be normalized such that
T
  = I

where    is a square matrix whose columns contain the eigenvectors   i  . With this normalization convention, then

T
 A = 
and

T
A =  

where    is the eigenvalue diagonal matrix.

Matrix Conditioning
Reordering the previous equations, any matrix  A  can be expressed as a sum of its eigenvalues multiplied by dyadic
eigenvector products

T T  T
 A  = 1  1   1  + 2  2   2  +  + n  n   n 
 

 T
Defining  B i  =   i    i  , which is a full rank 1 matrix, then
 
n
A =  k  Bk 
k=1

where n is the dimension of  A  . On the average, an element of  B k  has the same magnitude as the corresponding
element of  B k + 1  . Let B max be the magnitude of the largest coefficient of all  B k  matrices. Then

Main Index
 586 MSC Nastran Reference Guide
Numerical Accuracy Issues

a ij  B max   1 +  2 +  +  n 

This equation shows that the terms of  A  are dominated by the largest eigenvalues. Unfortunately, the smallest
eigenvalues are those of greatest interest for structural models. These small eigenvalues must be calculated by taking the
differences of coefficients that are dominated by the largest eigenvalue. For this reason, the ratio  n   1 is called a
numerical conditioning number. If this number is too large, numerical truncation causes a loss of accuracy when computing
the lowest eigenvalues of a system.
The assumptions that allow this simple analysis are often pessimistic in practice, that is, the bounds predicted by the error
analysis are conservative. However, the effects it predicts do occur eventually so that models that produce acceptable results
for one mesh size may produce unacceptable results with a finer mesh size due to the higher eigenvalues included in the
larger-sized matrices occurring from the finer mesh.

Definiteness of Matrices
A matrix whose eigenvalues are all greater than zero is said to be positive definite. If some eigenvalues are zero but none are
less than zero, the matrix is positive semi-definite. An elastic stiffness matrix assembled from elements is at least positive
semi-definite. If all of the structure’s rigid-body modes are constrained, the stiffness matrix is positive definite. Another
category is the indefinite matrices category. These matrices have zeroes or blocks of zeroes on the diagonal.
Although definiteness is most concisely defined in terms of eigenvalues, it is not a practical test for large matrices because
of the computational cost of extracting all of the eigenvalues. However, other operations, particularly linear equation
solution and dynamic reduction, may detect nonpositive definite matrices and provide diagnostics using these terms as
described later in this appendix.

Numerical Accuracy Issues

The numerical operations of MSC Nastran are executed in a finite 64-bit floating point arithmetic. Depending on the
specific computer’s word structure (number of bits for mantissa versus exponent), different roundoff errors may occur. To
attain the most numerical accuracy possible, the following strategies are used in MSC Nastran.
In the decomposition of positive definite matrices, the Gaussian elimination process (which is the basis of all decomposition
algorithms in MSC Nastran) does not require numerical pivots. Since some of the matrices are not positive definite, sparse
decomposition (both symmetric and unsymmetric) employs a numerical pivoting (row/column interchange) algorithm.
These methods consider a pivot term suitable if

THRESH
a kk  10  MAX a ki
i
which means that a diagonal term is accepted as a pivot if it is greater than the maximum term in that row multiplied by
THRESH
10 (the default for THRESH is –6).
To ensure numerical accuracy in eigenvalue calculations, most MSC Nastran methods use a spectral transformation of

Main Index
 Appendix A: Numerical Accuracy Considerations 587
Sources of Mechanisms

1
 = --------------
 – s

where  s is an eigenvalue shift.

This transformation is shown in Figure A-1.



3

2
1

S 1 2 3

Figure 1-1 Spectral Transformation

The spectral transformation ensures uniform accuracy throughout the frequency by shifting to the area of interest. Another
effect of this transformation is the welcomed  space separation of closely-spaced eigenvalues. When  s is close to an
eigenvalue  i , the decomposition of the shifted matrix

T
 K – s M  =  L   D   L 

may produce high MAXRATIO messages. Automatic logic to perturb the  s value in this case is implemented in MSC
Nastran.

Sources of Mechanisms
In all of the decomposition methods, a null row or column in  A  causes a fatal error message. All other causes of
singularity are not distinguishable from near-singularity because of the effects of numerical truncation. Only warning
messages are usually provided for these cases.

Main Index
 588 MSC Nastran Reference Guide
Sources of Nonpositive Definite Matrices

In standard decomposition,
T
A = LDL 

the process starts to compute the first term of  D  with the first internal degree of freedom and then processes each
additional degree of freedom and its associated terms. It can be shown that when processing the k-th row, the k-th row and
all rows above it are in effect free, and all rows below it are constrained to ground.
The term of  D  at the k-th row is proportional to the amount of coupling between that degree of freedom and the degree
of freedom with a higher value of k. If the terms of the k-th row and above are not connected to the remaining rows, the
k-th term of  D  goes to zero. Because of numerical truncation, the term may be a small positive or negative number
instead. If a term of  D  is calculated to be identically zero, it is reset to a small number because of the indeterminacy of
its calculation.
The existence of such a small term defines a mechanism in static analysis in MSC Nastran. A mechanism is a group of
degrees-of-freedom that may move independently from the rest of the structure as a rigid body without causing internal
loads in the structure. A hinged door, for example, is a mechanism with one rigid-body freedom. If the hinges are
disconnected, the door mechanism has six rigid-body freedoms.
Mechanisms are characterized by non-dimensional numbers derived by dividing the terms of  D  into the corresponding
diagonal term of  A  . If these matrix diagonal to factor diagonal ratios are large numbers, various warning and fatal messages
are produced, depending on the context.

Sources of Nonpositive Definite Matrices

A negative semi-definite element stiffness matrix, whose eigenvalues are all negative or zero, implies that the element has an
energy source within it. MSC Nastran discourages using such elements by giving fatal messages when negative element
thicknesses or section properties are input. However, there are applications where such elements are useful, such as when
using an element with negative stiffness in parallel with a passive element to model a damaged or thinned-down element.
For this reason, negative material property coefficients and negative stiffnesses for scalar elements are allowed. Also, some
incorrect modeling techniques, such as the misuse of the membrane-bending coupling term on PSHELL entries (MID4),
can lead to negative eigenvalues of the element stiffness matrix.
Stiffness matrices with negative eigenvalues cause negative terms in  D  . The number of such terms is automatically output
by the standard symmetric decomposition subroutine. Their existence again causes various warning and fatal messages,
depending on the context. The most common cause of negative terms in  D  is true mechanisms, whose terms are small
negative numbers and are actually computational zeroes.
Indefinite matrices occur when using the Lagrange multiplier method for constraint processing. In this method, the
constraints are not eliminated from but are concatenated to the g-size system of equations. The constraint equations have
zeroes on the diagonal, resulting in indefinite matrices. These systems can be solved using the block pivoting scheme of the
sparse decomposition.

Main Index
 Appendix A: Numerical Accuracy Considerations 589
Detection and Avoidance of Numerical Problems

Detection and Avoidance of Numerical Problems

Static Analysis
Models used in static analysis must be constrained to ground in at least a statically determinate manner even for unloaded
directions. For example, a model intended for only gravity loading must be constrained in horizontal directions as well as
vertical directions. The evidence of unconstrained directions is that the entire model is a mechanism, that is, the large ratio
occurs in the following scenarios:
 At the last grid point in the internal sequence.
 Connecting soft elements to stiff elements. Local stiffness is a function of element thickness (moment per unit
rotation) through element thickness cubed (force per unit deflection) and is inversely proportional to mesh
spacing, again in linear through cubic ratios. Some relief is possible by sequencing the soft degrees-of-freedom first
in the internal sequence, although this is difficult to control in the presence of automatic re-sequencing. Reliable
corrections are:
• Replace the very stiff elements with rigid elements.
• Place the soft and stiff elements in different superelements.
 The elements omitted through oversight. The corrective action here is to start with the grid points listed in the
diagnostics and track back through the elements connected to them through the upstream grid points. The
missing elements may be anywhere upstream. PARAM,GPECT,1 output and undeformed structure plots all
provide useful data for detecting missing elements.
There are two major methods of identifying large ratios and non-positive definite matrices. In some solutions, the largest
matrix diagonal to factor diagonal ratio greater than 105 (MAXRATIO default) is identified by its internal sequence number,
and the number of negative factor diagonal terms is output. The best method to identify mechanisms here is to apply
checkout loads that cause internal loads in all of the elements. Then inspect the displacement output for groups of grid
points that move together with implausibly large displacements and common values of grid point rotation. The only
condition that causes a fatal error is a true null column, and NASTRAN SYSTEM(69)=16 avoids this fatal error by placing
a unit spring coefficient on the degrees of freedom with null columns. This option is recommended only for diagnostic runs
because it may mask modeling errors.
5
In other solution sequences, all ratios greater than 10 are printed in a matrix format named the MECH table. The
external sequence number of each large ratio is also printed, which is the grid point and degree of freedom number. If any
such ratios exist, the action taken depends on the value of PARAM,BAILOUT. In the conventional solution sequences, its
default value causes the program to continue after printing the MECH matrix. In the superelement solution sequences, a
different default causes a fatal error exit after printing the MECH matrix. For both types of solution sequences, the opposite
action may be requested by setting the value of PARAM,BAILOUT explicitly. Also, the criterion used for identifying large
ratios may be changed using PARAM,MAXRATIO.
3 6 7 8
For static analysis, values between 10 and 10 are almost always acceptable. Values between 10 and 10 are
questionable. When investigating structures after finding these values, some types of structures may be found to be properly
modeled. It is still worthwhile to investigate the structures with questionable values.

Main Index
 590 MSC Nastran Reference Guide
Detection and Avoidance of Numerical Problems

The solutions with differential stiffness effects offer another method to obtain nonpositive definite stiffness matrices. For
example, a column undergoing compressive gravity loading has a potential energy source in the gravity load. A lateral load
that is stabilizing in the absence of gravity (that is, a decreased load causes a decreased deflection) is destabilizing when
applied in a postbuckled state.

Eigensolutions Using the Inverse Iteration and Lanczos Methods

The matrix  K –  s M  is decomposed where  s is an eigenvalue shift and  K  and  M  are the stiffness and mass
matrices, respectively. This condition allows the solution of models with rigid-body modes since the singularities in  K 
are suppressed if there are compensating terms in  M  . The only conditions that should cause fatal messages due to
singularity are
 The same null column in both  K  and  M  . These columns and rows are given an uncoupled unit stiffness by
the auto-omit operation if the default value (0) for PARAM,ASING is used. If this value is set to -1, a null column
in both matrices is regarded as an undefined degree of freedom and causes a fatal error exit.
 A massless mechanism. One commonly encountered example is a straight-line, pinned rod structure made from
CBAR elements with no torsional stiffness (J) defined. The structural mass generated for CBAR elements does not
include inertia for the torsion degrees-of-freedom. The natural frequency for this torsion mode approaches the
limit of zero torsional stiffness divided by zero mass, which is an undefined quantity. If the elements lie along a
global coordinate axis, the mass term is identically zero, which leads to very large negative or positive eigenvalues
and is usually beyond any reasonable search region. If the elements are skewed from the global axes, the eigenvalues
may be computed at any value (including negative) because of the indeterminacy caused by numerical truncation.
The negative terms on factor diagonal message generally occurs for every decomposition performed in the
iteration. It can be shown from Sturm sequence theory that the number of negative terms is exactly equal to the number of
eigenvalues below  s . This condition is a means of determining if all roots in the range have truly been found.

Eigensolutions Using Transformation Methods

If the GIV or HOU methods (and their variations) are selected, one of the first operations performed is a Choleski
decomposition. It is used to reduce the problem to a special eigenvalue problem. The mass matrix is given this operation
for the straight GIV or HOU method. Identically null columns are eliminated by the auto-omit operation. Poor
conditioning can result from several sources. An example is a point mass input on a CONM2 entry that uses offsets in three
directions. The grid point to which it is attached has nonzero mass coefficients for all six degrees-of-freedom. However, only
three independent mass degrees-of-freedom exist, not six. Another example results in a superelement analysis when most of
the elements in a superelement do not have mass and all interior masses are restricted to only a few structural or mass
elements. The boundary matrix produced for the superelement is generally full, no matter what its rank; that is, regardless
of the number of independent mass degrees of freedom that it contains.
The presence of a rank-deficient mass matrix when using the GIV or HOU method produces fatal messages due to the
singularity of the mass matrix or produces solutions with poor numerical stability. Poor stability is most commonly detected

Main Index
 Appendix A: Numerical Accuracy Considerations 591
Detection and Avoidance of Numerical Problems

when making small changes to a model and then observing large changes in the solution. Either the MGIV and MHOU
methods or the AGIV and AHOU methods are the preferred methods to use when nearly singular mass matrices are expected
because these methods decompose the matrix  K –  s M  instead of the mass matrix.

The shift parameter  s is automatically set to be large enough to control rigid-body modes. Better modeling practices also
reduce the costs and increase the reliability for the two examples cited above. If an offset point mass is significant in any
mode, it is better to attach it to an extra grid point at its center of gravity and model the offset with a rigid element. The
auto-omit feature then eliminates the rotational degrees of freedom. Similarly, if only a few interior points of a superelement
have mass, it may be more economical to convert them to exterior points, which also eliminates the singular boundary mass
matrix.
It is possible to input negative terms into a mass matrix directly with DMIG terms or scalar mass elements. This class of
problem causes fatal errors due to non-positive definite mass matrices for the transformation methods, fatal errors with the
Lanczos method, and wrong answers with the INV method. The complex eigenvalue methods should be used for this type
of problem, which infers that modal or dynamic reduction methods may not be used.
A similar but different problem arises because Cholesky decomposition is used on the generalized mass matrix (named MI
in the diagnostics) when orthogonalizing the eigenvectors with respect to the mass matrix. The existence of negative terms
here indicates that poor eigenvectors were computed. The row number where the negative term occurs is printed in the
diagnostic. This row number does not refer to a physical degree of freedom but refers instead to an eigenvector number.
The usual cause is computing eigenvectors for computational infinite roots in the modified transformation methods. This
problem can be avoided by setting ND on the EIGR entry to a value less than the row number that appears in the diagnostics
or before the fact by setting F2 on the EIGR entry to a realistic value instead.

Frequency Response Analysis

Negative terms on the factor and high factor to matrix diagonal ratios can be expected when using coupled methods and
may often be safely ignored. These messages are merely an indication that the excitation frequency is above one or some of
the natural frequencies of the system (negative terms), or is near a natural frequency (high ratio).

Transient Response Analysis

A well-conditioned model should have neither negative terms nor high ratios on its factor terms. The causes of such
messages include all of the effects described above. Negative mass terms can be detected by rapidly diverging oscillations in
the transient solution.

The Large Mass Method

The large mass method is one of the methods for enforcing motion. It can be used with direct methods and with all the
reduction methods. The basis of the method is to attach artificial masses to the structure at the degrees-of-freedom where
the motion is to be enforced. These large masses should be orders of magnitude larger than the total mass or the moment

Main Index
 592 MSC Nastran Reference Guide
Detection and Avoidance of Numerical Problems

3 8
of inertia of the structure. A survey of the literature shows recommendations for mass ratios ranging from 10 to 10
6
with a value of 10 as the most common recommendation.
The mass ratio affects both the accuracy and numerical conditioning, and must be adjusted in a compromise that meets
both criteria. With regard to load accuracy, the error in the approximation is inversely proportional to the ratio. A ratio of
3
10 causes an error of ten percent at resonance for a mode with one-half percent of damping—it represents an extreme
case. Off-resonance excitation or higher damping ratios result in lower errors. Numerical conditioning problems are much
more difficult to predict.

The Lagrange Multiplier Method

Another method to enforce motion is the Lagrange multiplier method. In this method, the input motion function is
described by a constraint equation. This method provides better accuracy than the large mass method where the numerical
error introduced is proportional to the large mass.
The formulation introduces an indefinite system of linear equations where some numerical problems may arise from the
fact that the system matrix contains terms that have dimensions of stiffness as well as nondimensional (constraint) terms.
The decomposition of this matrix with the sparse decomposition methods pivoting strategy is stable.
You may control the pivoting strategy with the THRESH DMAP parameter. The default value (–6) is adequate for most
Lagrange multiplier solutions. In fact, higher (ranging up to –2) values provide better accuracy, while increasing the number
of pivots may result in performance degradation.

Main Index
 Appendix B: Theory of Eigenvalue Extraction
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

B Theory of Eigenvalue
Extraction
 Sturm Sequence Theory 594
 Eigensolution Methods 594

Main Index
 594 MSC Nastran Reference Guide
Sturm Sequence Theory

Sturm Sequence Theory

Methods of Computation
A numerical technique which is very helpful in a modal extraction is Sturm Sequence. In this approach, a frequency is
selected and computations are performed to determine the number of roots which exist below that frequency. In some cases
this frequency is chosen manually by user input, other times it is selected automatically by MSC Nastran. This choice
depends upon which modal extraction technique is being used.
Starting with the basic characteristic equation
 K + s M  = 0 (2-1)

a frequency,  2 =  s is substituted. This frequency is called the shift point. The determinant K –  s M is evaluated for
T
the frequency  s . The resulting matrix is factored into it  L   D   L  components. The diagonal matrix,  D  , is known
as the Factor Diagonal Matrix. The number of negative terms in the Factor Diagonal Matrix, known as the Sturm Number,
are the number of frequencies that exist below the shift point frequency,  s .

Sturm Sequence logic is a very important part of the many of the modal extraction and dynamic reduction techniques.
When it is utilized, the Sturm Sequence logic issues a number of messages which indicate the various shifts within the
frequency range of interest. These messages help the various techniques to determine how many roots are in specific ranges
and also if all modes have been found in the range.

Eigensolution Methods
Methods of Eigensolution
Seven methods of real eigenvalue extraction are provided in MSC Nastran. These methods are numerical approaches to
solving the equation of motion for its natural frequencies and modes shapes. The reason for seven different numerical
techniques is related to the fact that no one method is the best for all problems. While most of the methods can be applied
to all problems, the difference between the methods many times reduces to the efficiency of the solution process.
Most methods of algebraic eigenvalue extraction belong to one of two groups:
 Transformation methods
 Tracking methods.
In a transformation method, the eigenvalue equation is first transformed into a special form (such as tridiagonal) from
which eigenvalues may easily be extracted. In a “tracking” method, the roots are extracted from the original dynamic matrix
one at a time using an iterative procedure.
Four of the real eigenvalue extraction methods available in MSC Nastran are transformation methods. These are:
 Givens Methods

Main Index
 Appendix B: Theory of Eigenvalue Extraction 595
Eigensolution Methods

 Householder Method
 Modified Givens Method
 Modified Householder Method.
These transformation methods use a “tridiagonal” solution method.
Two of the real eigenvalue extraction methods available in MSC Nastran are classified as tracking methods. These are:
 Inverse Power Method
 Sturm Modified Inverse Power Method.
The final real eigenvalue extraction method available in MSC Nastran is the Lanczos method. The Lanczos method
combines the best characteristics of the tracking and transformation methods.

The Standard Tridiagonal Methods (GIV and HOU)

The Givens(GIV) and Householder (HOU) modal extraction methods require a positive definite mass matrix. There are
no restrictions on the stiffness matrix other than that it must be symmetric. These matrices will always result in real
(positive) roots. The tridiagonal methods are the most efficient methods for small problems and problems with dense
matrices when a large portion of the eigenvectors are needed. These methods find all of the eigenvalues in the structure as
a result of the transformation process. The user must make a specific request for the required eigenvectors (mode shapes).
The tridiagonal methods do not take advantage of the sparse matrices, but they are efficient with the dense matrices
sometimes created using dynamic reduction. The steps used in the tridiagonal methods are as follows:
1. Perform a Cholesky decomposition of the mass matrix:
T
M = LL (B-2)
where  L  is a lower triangular matrix.
2. The symmetric eigenvalue equation can be written as:
 J – I   w  = 0 (B-3)
–1
where  J  is the symmetric matrix and  I  is the identity matrix. To do this, premultiply equation (2-1) by  L 
and substitute for  M  from equation (B-2) to set the following:
–1 –1 T
L Ku – L LL u = 0 (B-4)
Then make the transformation:
w = Lu (B-5)
which reduces equation (B-4) to equation (B-3) using the following:
–1 – 1,T
J = L KL (B-6)

Main Index
 596 MSC Nastran Reference Guide
Eigensolution Methods

3. The  J  matrix is converted to a tridiagonal matrix by a transformation method according to the Givens or
Householder method. (A tridiagonal matrix is a matrix where the only nonzero terms in the i-th column are the i –
1, i, i + 1 row.)
4. All the eigenvalues of the tridiagonal matrix are extracted using a modified QR algorithm.
5. The eigenvectors are computed over a given frequency range or, at user option, for a given number of eigenvalues
using inverse iteration. Since the roots are known to high accuracy at this point, the vectors converge rapidly.
6. Physical eigenvectors are recovered by performing the inverse transformation to that made in Step 3 and then
solving equation (B-5) as follows:
T,–1
u = L w
7. If a SUPORT entry is used, the first N r eigenvectors computed are discarded. The rigid body modes from the
SUPORT computation are put in their place, and the corresponding eigenvalues are set to zero. The SUPORT
operation does cause some redundant operations to be performed as the rigid body modes are computed by two
methods and then one set is discarded. This is generally a minor issue.
8. The back-transformed vectors  PHI  are given a final stage of orthogonalization with respect to the mass matrix.
The modal mass matrix is calculated as follows:
T
 M modal  =     M    
then decomposed,
T
 M modal  =  L   L 
T
The refined eigenvectors,    are found by a forward pass on the equation:
T T
L = 
Note that the tridiagonal methods fail in Step 1 if the mass matrix is not positive definite. In order to minimize this
eventuality, degrees of freedom with null columns are removed from ( u ) by the automatic application of static
condensation. This is called AUTO-OMIT. (This procedure can be bypassed by using PARAM,ASING). The application
of the AUTO-OMIT process is a precaution and may not remove all possible causes of mass matrix singularity such as, a
point mass offset from a grid point, but it greatly improves the reliability and convenience of the standard tridiagonal
methods.

The Modified Tridiagonal Methods (MGIV, MHOU, AGIV, AHOU)

The modified tridiagonal methods, Modified Givens (MGIV) and Modified Householder (MHOU), are similar to the
standard tridiagonal methods (GIV, HOU) with the exception that the mass matrix can be singular.
The steps used in the modified methods are as follows:
1. Perform a Cholesky decomposition of the positive definite matrix

Main Index
 Appendix B: Theory of Eigenvalue Extraction 597
Eigensolution Methods

T
 K + s M  =  L   L  (B-7)

where  s is a positive number automatically selected by the program to optimize the reliability and accuracy of
eigenvalue extraction. It is large enough to stabilize all rigid body modes but also be as small as possible to reduce
the numerical contamination. The goal of the algorithm is to find a value near the first flexible frequency of the
system. It uses the diagonal terms of the mass and stiffness matrices as follows:
1
 s = -------------------------------
N
1  2
M ii
N  -------K ii
-
i=1
where N is the dimension of the matrices. If M ii or X ii = 0 (an under defined degree-of-freedom) or
M ii  X ii  10**8 (a likely “very large” mass term used to approximate a rigid body mode), the corresponding
term is omitted from the summation. If the value calculated by this approximate method does not result in a stable
Cholesky decomposition,  s is increased by adding the arbitrary amount of 0.01. If this value also results in failure,
the original value is increased by 0.02. If this value also fails, a fatal error is issued.
2. Write the eigen-equation standard form as:

 J – I   w  = 0 (B-8)
where:
1
 = --------------- (B-9)
 + s
and:
–1 – 1,T
J = L ML (B-10)
by rearranging the terms to read as follows:
 K +  s M –   +  s M   u  = 0 (B-11)

Then premultiply equation (B-11) by – 1    +  s  and substitute for  K +  s M  from equation (B-7) to obtain:

T
LL
M – ---------------  u  = 0
 + s

–1
Premultiply by  L  and then substitute:
T
w = L u

Main Index
 598 MSC Nastran Reference Guide
Eigensolution Methods

resulting in equations (B-8), (B-9), and (B-10).

3. The remaining steps are identical to those for the unmodified methods where:

 J  =  J  and  w  =  w 
Although the mass matrix is not required to be nonsingular in the modified methods, a singular mass matrix can
produce one or more zero eigenvalues,  s , which are equivalent to infinite eigenvalues,  i . Due to roundoff error,
these will appear in the output as very large positive or negative eigenvalues. To reduce the incidence of such
meaningless results, degrees of freedom with null masses are eliminated by the AUTO-OMIT static condensation
as in the case of the unmodified methods. This step can be bypassed by using PARAM,ASING.
Many times the user is not conversant in the status of the mass matrix prior to the actual matrix assembly. To assist the user
in choosing the appropriate method, two options, Automatic Givens (AGIV) and Automatic Householder (AHOU) are
available. The automatic methods, AHOU and AGIV, initially use the standard methods. In the first step of the method,
if the mass matrix is not well-conditioned for decomposition, the method shifts to the corresponding modified method.
The modified methods are more expensive and introduce numerical “noise” due to the shift, but they resolve most of the
numerical problems of the ill-conditioned mass matrix.
Operations involved in the computation of the natural frequencies and mode shapes is proportional to the number of
degrees of freedom ( N ). The time-consuming operations for the tridiagonal methods are mass matrix decomposition and
reduction to standard form ( 5  2 N 3 for a full mass matrix, negligible for a diagonal matrix) tridiagonalization ( 2  3 N 3
for HOU, 4  3 N 3 for GIV), QR root iteration ( 9 N 2 ), inverse iteration for vectors ( N 2 per vector), and inverse
transformation of vectors ( N 2 per vector for HOU, 2 N 2 for GIV). As N becomes large (greater than 100, perhaps),
terms in N 3 dominate.

The Inverse Power Methods (INV, SINV)

The inverse power method with shifts is particularly effective when the mass and stiffness matrices are sparse and when only
a few of the eigenvalues are required. In MSC Nastran, the method is used as a stand-alone method to find all of the
eigenvalues within a user-specified domain.
If we define the eigenvalue problem as:
 K – M   u  = 0 (2-12)
Let:
 = o + 

where  o is called the shift point.

The iteration algorithm is:

 K – o M   wn  =  M   un – 1  (2-13)

Main Index
 Appendix B: Theory of Eigenvalue Extraction 599
Eigensolution Methods

1
 u n  = ------  w n  (2-14)
Cn

where C n is equal to the value of the element of  w n  with the largest absolute value.  u n  is then placed on the right
side of equation (2-13) and iteration continues. It is easy to prove that 1  C n converges to  1 , the shifted eigenvalue nearest
to the shift point, and that  u n  converges to the corresponding eigenvector of equation (2-12). The following observations
are pertinent.
 A triangular decomposition of the matrix  K –  o M  is required in order to evaluate  w n  from (2-13). The
effort required to perform the triangular decomposition is greatly reduced if  K –  o M  is a narrow band matrix.
Structural analysis is characterized by narrowly banded stiffness and mass matrices. From Sturm’s theorem, the
number of negative terms on the factor diagonal is exactly equal to the number of roots below  o . This number is
printed out at every shift and is identified as the Sturm sequence number.
 It is unnecessary for  K  to be nonsingular, so that rigid body modes cause no special difficulty. Similarly, null
rows in  M  cause no problems. If they are removed by static condensation, they will increase the cost of the
solution unnecessarily. This occurs primarily due to the increase in bandwidth inherent in static condensation
 The shift point  o may be changed (at the cost of an additional triangular decomposition) in order to improve the
rate of convergence toward a particular eigenvalue or to improve the accuracy and convergence rate after several
roots have been extracted from a given shift point. The next shift point is computed from the residual terms
remaining after computing the present root by a root estimation algorithm. See the MSC Nastran Numerical Methods
User’s Guide for details.
  o can be placed to obtain the eigenvalues within a desired frequency band and not just those that have the
smallest absolute value.
 The Sturm sequence data can be used to ensure that all eigenvalues within the requested frequency band have been
found. This can be done manually with DIAG 16 for the INV method or automatically with the SINV method
(see the MSC Nastran Numerical Methods User’s Guide for details). In the SINV method, the requested band is broken
into subregions, and the Sturm number is determined at each boundary of the subregion. If the root estimation
algorithm predicts that the next root is out of the subregion but the Sturm sequence data states that more roots
remain in the subregion, a bisection method is used to iterate to the missing roots. All roots for a subregion must
be found before the method is allowed to enter the next subregion. The INV method tends to compute shifts that
wander more about the requested frequency band and computes more roots outside the band than the SINV
method does.
In the practical application of the tracking methods, most of the numerical efficiency issues (trial eigenvectors, convergence
criteria, etc.) are automatically handled by MSC Nastran. The shift points must be selected so that the numerical algorithms
can operate in an optimum manner. Generally, this is handled by MSC Nastran, but the user can assist MSC Nastran
through a careful choice of frequency range.

Main Index
 600 MSC Nastran Reference Guide
Eigensolution Methods

If the SUPORT entry is used, the rigid body modes calculated are used directly as the first set of eigenvectors with their
eigenvalues set to zero. Iteration then proceeds for the nonzero roots. Use of the SUPORT feature reduces costs of the INV
and SINV methods considerably besides giving the aesthetic benefits of exact zero-frequency rigid body modes. The
SUPORT entry is recommended for use with these methods.

The Shifted Block Lanczos Method

The Lanczos method (LAN) available in MSC Nastran is an extremely reliable and accurate method for solving the large
sparse eigenvalue problems typical of many structures.
A shifted block Lanczos algorithm is implemented in MSC Nastran. A summary of that method is given here with special
emphasis on the effect of user inputs.
For the standard eigenvalue problem written in the form:
Ax = x
the Lanczos method computes a sequence of vectors q called Lanczos vectors and scalars Q j , A . The basic recurrence
relation for the j-th step of the Lanczos algorithm can be written as follows:
 rj + 1  =  A   qj  – j  qj  – j  qj – 1 

j =  qi  T  A   qj 

j + 1 = rj + 1

 rj + 1 
 q j + 1  = ------------------
Bj + 1

The first j-steps of the method can be combined and written in matrix form:

 A   Qj  –  Q   Tj  = j + 1  qj + 1   ej  T

where  e j  = unit vector with the j-th component equal to one and all other components equal to zero

 Q j  =   q 1 ,  q 2 , ...,  q j   with  Q jT   Q j  =  I 

Main Index
 Appendix B: Theory of Eigenvalue Extraction 601
Eigensolution Methods

1 2
2 2 3
3 3 4

 Tj  =

j

j j

The eigenvalues of the tridiagonal matrix  T j  are called Ritz values. Some of these Ritz values are good approximations
to eigenvalues of  A  . The corresponding Ritz vectors  x i  are computed from the eigenvectors of  T j  and the Lanczos
vectors. An eigenpair is defined as an eigenvalue  s i  and its corresponding eigenvector   ik  . Let    ik , s i  be an
eigenpair of the tridiagonal matrix  T j  . Then an approximate eigenpair for the original problem is   i ,  x i   where:

i = i

 xi  =  Qj   si 

As implied by the formulation of the algorithm, the method computes a tridiagonal matrix and a set of Lanczos vectors that
become larger with each iteration. The algorithm in exact arithmetic can be terminated when B j + 1 = 0 or when
sufficiently many eigenvalues of  A  have been computed. Lanczos also handles the effects of roundoff.
The algorithm that has been incorporated in MSC Nastran is a shifted block version of the basic Lanczos recurrence
relation. The blocking approach to the Lanczos algorithm (working with blocks of p Lanczos vectors simultaneously)
enables the method to handle eigenvalues of high multiplicity including shifting in the algorithm which helps to speed the
convergence of the method.
In this method, the A matrix is replaced by the sum of the stiffness matrix plus shifted mass matrix, similar to the approach
in the modified tridiagonal methods. This allows solution of problems with rigid body modes and generally faster
convergence. The shifts (trial roots) are influenced by user input. The user is required to input either a frequency band of
interest, a desired number of roots, or a combination of this data.
As an example, suppose that both the lower bound (V1) and upper bound (V2) of the frequency band are defined by the
user as well as the number of roots desired (ND). The main objective for the shift strategy is to encompass the region
containing the eigenvalues of interest in which all of the eigenvalues have been computed and the number of eigenvalues
has been verified with a Sturm sequence check.

Main Index
 602 MSC Nastran Reference Guide
Eigensolution Methods

The first shift is at V1 to determine the Sturm sequence number. The factor matrix is stored for possible later use. If V1 is
not defined by the user, the first shift is made at  s . The second shift point is chosen at V2.

After a Lanczos iteration based on the selected shift and the computation of the corresponding Ritz values, some values will
be found to be acceptable eigenvalues while other eigenpairs will fail convergence tests. A decision is made as to whether it
is more efficient to re-shift or continue with further iteration. The re-shift point is chosen from the existing Ritz values
(smaller than V2) or from the stored factor from the shift at V2. At this point, the number of accepted eigenvalues and the
Sturm sequence numbers at each end of the region are known.
If iteration with the factor from V2 does not produce all of the roots desired, two more types of shift will be used. If it
appears that the missing roots are near the lowest shift, an approximate first mode eigenvalue is used. This is either a nonzero
SHFSCL field from the EIGRL entry or the a similar shift term used in the modified tridiagonal methods. If the missing
roots are elsewhere, bisection between prior shifts is used. If all requested eigenpairs still cannot be found, a user information
message is printed, and all accepted eigenpairs are output. If some roots have been found, the module exits normally. A fatal
error is issued if no roots have been found. If accepted Ritz values are found outside the requested region, they are discarded.
The strategy for selecting block size is to first find the minimum of N/2, 2E, p and MAXSET. N is the problem size, E the
number of roots desired, p the number of Lanczos vectors that will fit into memory, and MAXSET the user selection for
block size (which has a default of 7).
Another efficiency feature unique to the Lanczos method is selective orthogonalization of the trial eigenvectors. All of those
contained in the same block are made orthogonal to one another. Tests are made, at several steps, to determine if further
orthogonalization between vectors in different blocks is required. This avoids the cost term cubic on E (the number of
eigenvectors calculated) which becomes the dominant term for GDR and the SINV methods as E becomes large.
If the SUPORT entry is used, the  D ir  matrix is merged with an identity matrix  I rr  to form raw rigid body vectors
 D ar  . These raw vectors are used as initial vectors. If the r-set has been well-chosen, the rigid body modes will converge
rapidly. If they have been chosen poorly, the Lanczos iterations will continue until a good set of eigenvectors have been
found. The corresponding roots are set to the aesthetically pleasing value of zero rather than to the small numerical zeros
likely to be produced by the actual computations.
The Lanczos method is more forgiving than the other methods since it repairs the effects of a poor r-set selection. However,
the repeatability of rigid body modes computed by Lanczos with the SUPORT entry is not the same as with other extraction
methods. Rigid body modes computed using the SUPORT entry and a non-Lanczos method will produce identical modes,
irrespective of the method. If the rigid body modes produced by the Lanczos method with SUPORT entries are compared
with those produced by any of the other methods, the Lanczos modes will be a proper linear combination of the modes
from the other methods and may be markedly different in appearance. Furthermore, if small changes are made to the
stiffness matrix, the Lanczos modes again may be markedly different, although they are again a proper linear combination
of the set produced by other methods. This lack of repeatability makes the Lanczos method unusable with the scaled
response spectra capability of MSC Nastran.
The cost for the LAN method is dominated by the decomposition, inverse iteration, and orthogonalization operations at
the size of the input matrices N. The number of roots per decomposition varies by problem and is difficult to determine a
priori.

Main Index
 Appendix B: Theory of Eigenvalue Extraction 603
Eigensolution Methods

User Interface for Real Eigenvalue Analysis

The EIGR and EIGRL Bulk Data entries are the heart of the user interface because they define a method and select
parameters that control the eigenvalue extraction procedure. The EIGRL entry is used for the LANCZOS method, and the
EIGR entries are used for all of the other methods.
The data used for the EIGR entry depend on the method selected. The basic data required is related to the eigenvalue
extraction method chosen and to the frequency range or number of required roots. The basic format of the Bulk Data entry
is as follows:

1 2 3 4 5 6 7 8 9 10
EIGR SID METHOD F1 F2 NE ND +BC
+BC NORM GID CID

The SID (Field 2) is the Set ID associated with this EIGR entry. A particular EIGR entry is activated in an analysis using
the METHOD = SID entry in the Case Control Section. If the Set ID on the METHOD and EIGR match, the eigenvalues
will be extracted using the method referenced on that entry.
The METHOD Entry LANzos (Field 3) selects the eigenvalue method from the following list:

1. INV Inverse Power

2. AHOU Automatic Householder
3. SINV Sturm Modified Inverse Power
4. GIV Givens Method
5. HOU Householder Method
6. MGIV Modified Givens Method
7. MHOU Modified Householder Method
8. AGIV Automatic selection of Givens or Modified Givens

The F1 Entry (Field 4) is used to specify the lowest frequency of interest in the eigenvalue extraction.
The F2 Entry (Field 5) is used to specify the highest frequency of interest in the eigenvalue extraction.
The NE entry (Field 6) is used by the INV method only. It defines the estimated number of roots in the range. A good
estimate will result in a more efficient solution.
The ND entry (Field 7) is used to specify the desired number of roots in the range of interest.
The NORM (Field 2 – continuation) entry on the continuation card is used to specify the method of eigenvector
normalization. The choices are:
1. MASSMass normalization (Default – if used continuation entry not required)
2. MAXNormalization to maximum A-set component
3. POINTNormalization to user-defined degrees of freedom

Main Index
 604 MSC Nastran Reference Guide
Eigensolution Methods

The POINT entry (Field 3 – continuation) is used to specify the point of interest from the analysis set for POINT
normalization only.
The CID entry (Field 4 – continuation) is used to specify the component of displacement for POINT normalization only.
There is an interrelationship between the F1, F2, and ND field on the EIGR entry.

Table 2-1 EIGR Input Parameters

When METHOD = INV When METHOD = GIV, MGIV,
or SINV HOU, or MHOU
F1,F2 Frequency range of interest ( Real  0.0 ). F1 Frequency range of interest ( Real  0.0 ; F1 < F2. If ND is
must be input. If METHOD = SINV and ND blank, F1 and F2 are ignored. If ND is blank, eigenvectors are
is blank, then F2 must be input. found whose natural frequencies lie in the range between
F1and F2.
NE Estimate of number of roots in range (Required Not used.
for METHOD = “INV” (Integer > 0). Not
used by SINV method.
ND Desired number of roots (Default is 3*NE, Desired number of eigenvectors ( Integer  0 ). If ND is zero,
INV only) (Integer > 0). If this field is blank the number of eigenvectors is determined from F1 and F2
and METHOD = SINV, then all roots (Default = 0). If all three are blank, then ND is automatically
between F1 and F2 are searched and the limit is set to one more than the number of degrees of freedom listed
600 roots. on SUPORT entries.

The rules for METHOD = GIV, HOU, MGIV, MHOU, AGIV and AHOU are identical. If any of these methods are
selected, MSC Nastran will find all of the eigenvalues but will compute only the eigenvectors specified by F1 and F2 or
those specified by ND (the “desired” number). F1 and F2 specify the lower and upper bounds of the frequency range in
which eigenvectors are computed, and ND specifies the number of eigenvectors, starting with the lowest (or the first rigid
body mode, if present). If F1, F2, and ND entries are present, ND takes precedence. Note that the default for F1 is 0.0.
The following examples demonstrate the use of the EIGR data entry.

1 2 3 4 5 6 7 8 9 10
EIGR 1 AHOU 10

In the first example (SID = 1), the automatic Householder method is selected with 10 roots requested. Since the default
“MASS” eigenvector normalization is requested, no continuation entry is needed.

EIGR 2 AHOU 100.0 +2

+2 MAX

For the second example (SID = 2), the same method is requested, but all the roots below 100 cycles per unit time are
requested with “MAX” vector normalization.

Main Index
 Appendix B: Theory of Eigenvalue Extraction 605
Eigensolution Methods

1 2 3 4 5 6 7 8 9 10
EIGR 3 SINV 0.0 100.0 6 +3
+3 POINT 32 4

For the third example (SID = 3), the inverse iteration method with Sturm sequence checks is requested for the first six roots
in the range specified. The POINT normalization is requested for grid point 32 in the R1 direction.
If METHOD = SINV, the values of F1, F2, and ND determine both the number of eigenvalues and the eigenvectors that
will be computed. These entries also provide hints to help MSC Nastran find the eigenvalues. F1 and F2 specify the
frequency range of interest within which MSC Nastran will search for modes. MSC Nastran will attempt to find all of the
modes in the range F1, F2 or the number specified by ND, whichever is less. If searching stops because ND modes are
found, there is no guarantee that they will be the lowest eigenvalues. Iterative methods only guarantee the solution for
modes nearest the shift point. If ND modes are not found in the range of interest, SINV will usually find one mode outside
the range F1, F2 (or possibly more) before stopping the search.
The inverse power method is particularly efficient when only a small number of eigenvalues and eigenvectors are wanted.
Very often only the lowest mode is of interest. The following example illustrates an EIGR entry which will extract only the
lowest nonzero eigenvalue.

EIGR SID METHOD F1 F2 NE ND

EIGR 13 INV 0.0 .011 1

It is assumed in the example that the frequency of the lowest mode is greater than 0.01 cycles per unit time. The program
will find one eigenvalue outside the range F1, F2 and stop the search. The eigenvalue found is guaranteed to be the lowest
nonzero eigenvalue (or a member of the lowest closely spaced cluster of eigenvalues in cases with pathologically close roots)
provided that there are no negative eigenvalues and that the SUPORT entry has been used to specify the correct number
of zero eigenvalues. The NE entry (field 6) is used only by the INV method.
The fields of the EIGRL entry are designed to select and set parameters for the Lanczos method as well as request
diagnostics.

EIGRL SID V1 V2 ND MSGL VL MAXSET SHFSCL

EIGRL 1 0.1 3.2 10

The V1 field defines the lower frequency bound, the V2 field defines the upper frequency bound, and the ND field defines
the number of eigenvalues and eigenvectors desired in the region. The V1 and V2 are expressed in units of cycles per unit
time. Examples of the results of using explicit or default values for the V1, V2, and ND fields are shown in Table 2-2. The
defaults on the EIGRL entry are designed to provide the minimum number of roots in cases where the input is ambiguous.

Main Index
 606 MSC Nastran Reference Guide
Eigensolution Methods

Table 2-2 Model Outputs from EIGRL Input Options

Case V1 V2 ND Number and Type of Roots Found
1 V1 V2 ND Lowest ND or all in range; whichever is smaller
2 V1 V2 All in range
3 V1 ND Lowest ND in range [V1, +00]
4 V1 Lowest root in range [V1, +00]
5 ND Lowest ND roots in [-00, +00]
6 Lowest root
7 V2 ND Lowest ND roots below V2
8 V2 All below V2

The MSGLVL field of the EIGRL entry is used to control the amount of diagnostic output provided by Lanczos. The
default value of 0 produces no diagnostic output. The values 1, 2, or 3 provide more output with the higher values providing
increasingly more output. Typically, the default value is sufficient. In some cases, higher diagnostics levels may be in order
to help resolve difficulties with special modeling problems.
The MAXSET field is used to control the block size, the number of appropriate eigenvectors computed in the outermost
iteration loop. The default value of 7 is recommended for most applications. There may be special cases where a larger value
may result in quicker convergence of highly multiple roots or a lower value may result in more efficiency when the structure
is lightly coupled. However, the default value has been chosen after reviewing results from a wide range of problem types
on several different computer types with the goal of minimizing the sum of CPU and I/O cost.
One common occurrence is for the block size to be reset by MSC Nastran during the run because there is insufficient
memory for a block size of 7. Computational efficiency tends to degrade as the block size decreases. It is, therefore,
important to examine the EIGENVALUE ANALYSIS SUMMARY output to determine that the program has sufficient
memory to use an efficient block size. A smaller block size may be more efficient when only a few roots are requested. The
minimum recommended block size is 2.
The SHFSCL field allows a user-designated shift to be used when all other shifting strategies fail. This entry is only used
under special circumstances. However, when used, its value should be set to the expected first natural frequency.
The Lanczos method normalizes the computed eigenvectors using the MASS method. The other normalization methods
are not available.

Solution Control For Normal Modes

When used as an independent solution, normal modes analysis is available in SOL 103. The Executive Control Section can
also contain diagnostic DIAG 16 which will print the iteration information used in the INV or SINV methods.
One of the most important entries in the Case Control Section is the METHOD entry. This entry is required. The Set ID
referenced by the METHOD entry refers to the Set ID of an EIGR entry or an EIGRL entry in the Bulk Data.

Main Index
 Appendix B: Theory of Eigenvalue Extraction 607
Eigensolution Methods

When a modal extraction is performed, the MSC Nastran output file contains various diagnostic messages and an
Eigenvalue Table. Optional grid and element output are available using standard Case Control output requests. These
requests are summarized in Table 2-3.

Table 2-3 Eigenvalue Extraction Output Requests

Grid Output
DISPLACEMENT Requests the eigenvector (mode shape) for a set of grid points
(or VECTOR)
GPFORCE Requests the Grid Point Force Balance Table to be computed for each mode
GPSTRESS Requests grid point stresses to be computed for a set of grid points. This request must be
accompanied by the ELSTRESS Case Control request and the definition of Stress Surfaces
and/or Stress Volumes in the OUTPUT(POST) section of the Case Control. This request
also requires the use of Rigid Format Alter RF3D81 when used in SOL 3.
SPCFORCE Requests Forces of Single Point Constraint to be computed for a set of grid points for each
mode.
Grid Output
Element Output
ELSTRESS Requests the computation of modal stresses for a set of elements for each mode
(or STRESS)
ESE Requests the computation of modal element strain energies for a set of elements for each
mode
ELFORCE Requests the computation of modal element forces for a set of elements for each mode
(or FORCE)
STRAIN Requests the computation of the modal element strains for a set of elements
MODES A special Case Control request which permits selective output requests to be processed on
selective modes
Note: For large problems, if GPSTRESS is requested for all eigenvalues, it is likely that a large amount of time will be
spent on stress recovery.

In addition to Bulk Data entries required to define the structural model, the only required Bulk Data entry is the eigenvalue
selection entry, EIGR or EIGRL. The EIGR entry is used to select the modal extraction parameters for the Givens,
Householder, Modified Givens, Modified Householder, Automatic Givens, and Automatic Householder methods. The
EIGRL entry is used to select the modal extraction parameters for the Lanczos method.

Main Index
 Appendix C: Nonlinear Analysis
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

C Nonlinear Analysis


General Description of Nonlinear Capabilities 609
 Nonlinear Modal Analysis 698

Main Index
 Appendix C: Nonlinear Analysis 609
General Description of Nonlinear Capabilities

General Description of Nonlinear Capabilities

This section discusses the procedures in SOL 106, SOL 129, SOL 153, and SOL 159. Nonlinear effects in structures occur
mainly due to nonlinear materials and large deformations (rotations and/or strains). Contact problems exhibit nonlinear
effects due to changes in boundary conditions. All of these attributes may be represented by MSC Nastran nonlinear
elements.
The nonlinear element library consists of:
 CBUSH and CBUSH1D for generalized nonlinear springs and dampers.
 CROD, CONROD, and CTUBE for unidirectional truss members.
 CBEAM for axially and laterally deforming line members.
 CQUAD4 and CTRIA3 for membrane, plate and shell modeling and hyperelastic plane strain.
 CQUAD8, CQUAD, and CTRIA6 for hyperelastic plane strain.
 CQUADX and CTRIAX for hyperelastic axisymmetric modeling.
 CHEXA, CPENTA, and CTETRA for solid modeling.
 CGAP and slideline contact (BCONP, BLSEG, BFRIC, BWIDTH) for contact and friction modeling.
For advanced nonlinear analysis, we recommend you use SOL 400 (see MSC Nastran Nonlinear User’s Guide.)
Nonlinear elements may be combined with linear elements for computational efficiency if the nonlinear effects can be
localized. Primary operations for nonlinear elements are updating element coordinates and applying loads for large
displacements, and the internal relaxation iteration for material nonlinearity. Refer to Table 3-1 for a summary of the
nonlinear element properties.
The geometric nonlinearity becomes discernible when the structure is subjected to large displacement and rotation.
Geometric nonlinear effects are prominent in two different aspects: geometric stiffening due to initial displacements and
stresses, and follower forces due to a change in loads as a function of displacements. These effects are included, but the large
deformation effect resulting in large strains is implemented only for hyperelastic materials. Large strain effects are not
available for elastic-plastic, creep, or any other material types.
Material nonlinearity is an inherent property of any engineering material. Material nonlinear effects may be classified into
many categories and those included in MSC Nastran are plasticity, nonlinear elasticity, hyperelasticity, creep, and
viscoelasticity. Creep and viscoelasticity, implemented as a generalized nonlinear viscoelastic capability, may be coupled with
plasticity. Many options are available for yield criteria and hardening behavior in plasticity.
The primary solution operations are gradual load or time increments, iterations with convergence tests for acceptable
equilibrium error, and stiffness matrix updates. The iterative process is based on the modified-Newton's method combined
with optional expeditious methods such as the quasi-Newton (BFGS) update and the line search. The stiffness matrix
updates are performed occasionally to improve the computational efficiency, and may be overridden at the user's discretion.
A number of options for arc-length methods are also available for snap-through or post-buckling analysis of static problems.
For transient response analysis, a number of options are available for implicit direct time integration, combined with
adaptive and expedient iteration strategies similar to those implemented for static analysis.
Solution Sequences 106, 129, 153, and 159 consolidate all the nonlinear features described above. SOLs 106 and 153 are
applicable to static, quasi-static, and nonlinear buckling analyses. SOLs 129 and 159 are primarily applicable to dynamic

Main Index
 610 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

transient response analysis with some limited static analysis capability. SOLs 153 and 159 may also be used to perform a
nonlinear heat transfer analysis. See the MSC Nastran Thermal Analysis User’s Guide. These solution sequences can
accommodate superelements and provide easy restarts from the database.

Table 3-1 Summary of Properties of the Nonlinear Elements

Material Property Static Loads DATA Recovery

Geometric Nonlinearity
No. of Gauss Points
No. of GRID Points

Grid Point Stress

Connectivity

Properties

Structure Plot

Contour Plot
Anisotropic

Orthotropic
Nonlinear

Pressure
Isotropic

Thermal

Gravity

Stress

Force
BCONP BWIDTH variable X X X
BLSEG BFRIC
CBEAM PBEAM 3 16* X 1† X X X X X
PBCOMP
CBUSH PBUSH 1 or 2 X X X X X X
PBUSHT
CBUSH1D PBUSH1D 1 or 2 X X X X X X
CGAP PGAP 3 0 X X
CHEXA PSOLID 8 8 X 1 9 X X X X X X
PLSOLID 8 8 X MATHP X X X X X
20‡ 27
CONROD --- 2 1 X 1 X X X X X X
CPENTA PSOLID 6 6 X 1 9 X X X X X X
PLSOLID 6 6 X MATHP X X X X X

15‡ 21
CQUAD4 PSHELL 4 5* X** 1 2 8 X X X X X X X X
PCOMP
PLPLANE 4 4 X MATHP X X X X
CQUAD8 PLPLANE 8 ‡ 9 X MATHP X X X X
CQUAD PLPLANE 9‡ 9 X MATHP X X X X

Main Index
 Appendix C: Nonlinear Analysis 611
General Description of Nonlinear Capabilities

Table 3-1 Summary of Properties of the Nonlinear Elements (continued)

Material Property Static Loads DATA Recovery

Geometric Nonlinearity
No. of Gauss Points
No. of GRID Points

Grid Point Stress

Connectivity

Properties

Structure Plot

Contour Plot
Anisotropic

Orthotropic
Nonlinear

Pressure
Isotropic

Thermal

Gravity

Stress

Force
CROD PROD 2 1 X 1 X X X X X
CTETRA PSOLID 4,10 1 X 1 9 X X X X X X
PLSOLID 4 1 X MATHP X X X X X

10‡ 5
CTRIA3 PSHELL 3 5* X** 1 2 8 X X X X X X X X
PCOMP
PLPLANE 3 1 X MATHP X X X X
CTRIA6 PLPLANE 6‡ 3 X MATHP X X X X
CTUBE PTUBE 2 1 X 1 X X X X X
CQUADX PLPLANE 4 4 X MATHP X X X X X

9‡ 9
CTRIAX PLPLANE 3 1 X MATHP X X X X X

6‡ 3

*May be changed by the user.

†For the beam element, only elastic-perfectly plastic material is available.
‡Center grid, if applicable, and midside grids may be omitted. For hyperelastic elements it is recommended that either
all or no midside grids be omitted.
**Always in equilibrium in updated system.

Note: 1. Integers listed under Material properties identify (i) on MATi Bulk Data entries.
2. MATS1 and CREEP may be attached to MAT1, MAT2, and MAT9.
3. Edge nodes are not applicable to nonlinear elements, except hyperelastic elements and the TET10 element.

Main Index
 612 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Rudiments of User Interface

The input file consists of an optional NASTRAN statement, Executive Control Section, Case Control Section, and Bulk
Data Section. All the features and principles for the user interface are common in MSC Nastran; thus, all the features of
the user interface for nonlinear analysis are compatible with those for the linear analysis. Any exceptions for nonlinear
analysis will be explained in the relevant sections that follow.
Mechanical design is dictated by the strength, dynamic, and stability characteristics of the structure. MSC Nastran provides
the analysis capabilities of these characteristics with solution sequences, each of which is designed for specific applications.
The type of desired analysis is specified in the Executive Control Section by using a solution sequence identification. SOLs
106 and 153 are designed for static, quasi-static, and buckling analyses. SOLs 129 and 159 provide nonlinear transient
response analysis.
The basic input data required for a finite element analysis may be classified as follows:
 Geometric data
 Element data
 Material data
 Boundary conditions and constraints
 Loads and enforced motions
 Solution methods
The first three classes of data may not be changed during the course of an analysis whereas the last three classes of data may
be changed during the analysis as defined in the Case Control Section. Examples of input data are given in Listing 3-1 and
Listing 3-2 for nonlinear static and dynamic analysis, respectively.

Main Index
 Appendix C: Nonlinear Analysis 613
General Description of Nonlinear Capabilities

Listing 3-1 Example Input Data for Nonlinear Static Analysis

SOL 106 $ NONLINEAR STATIC ANALYSIS
DIAG 8,50 $ DIAGNOSTIC PRINTOUT
CEND $ END OF EXECUTIVE CONTROL DATA
TITLE = TEST OF CTETRA ELEMENT (CUBE SUBJECT TO UNIAXIAL LOADING)
DISP = ALL
STRESS = ALL
SPC = 100
SUBCASE 1
SUBTITLE = ELASTIC -- LOAD TO 850. PSI
LABEL = LOAD TO YIELD
LOAD = 50
NLPARM = 50
SUBCASE 2
SUBTITLE = PLASTIC -- LOAD TO 1000. PSI
LABEL = LOAD BEYOND YIELD
LOAD = 100
NLPARM = 100
SUBCASE 4
SUBTITLE = ELASTIC -- UNLOAD COMPLETELY TO 0. PSI
LABEL = FULL UNLOAD
LOAD = 200
NLPARM = 200
$OUTPUT(PLOT) $ PLOT DATA ADDED FOLLOWING THIS LINE IF NEEDED
$ END OF CASE CONTROL DATA
BEGIN BULK
$ PARAMETERS FOR NONLINEAR ITERATION
NLPARM 50 1 AUTO UPW NO
NLPARM 100 8 SEMI UPW NO
NLPARM 200 2 AUTO UPW NO
.
.
.

ENDDATA

Listing 3-2 Example Input Data for Nonlinear Dynamic Analysis

SOL 129 $ NONLINEAR TRANSIENT ANALYSIS
DIAG 8,50 $ DIAGNOSTIC PRINTOUT
CEND $ END OF EXECUTIVE CONTROL DATA
TITLE = GAP VERIFICATION BY NONLINEAR TRANSIENT ANALYSIS SEALL = ALL $ FOR ALL MATRICIES
SET 1 = 201 $ ,202,200
SET 3 = 22 $ ,23
SET 4 = 20 $ CONROD
DISP = 1
STRESS = 3 FORCE
$ FOR GAP S
FORCE
EL = 4
LOADSET = 20
`SPC = 100
SUBCASE 1 $ UP TO 0.5 SECONDS
DLOAD = 210
TSTEPNL = 22
SUBCASE 2 $ UP TO 1 SECOND
DLOAD.= 210
TSTEPNL = 21
OUTPUT(XYPLOT)
CSCALE = 1.3
XTITLE = TIME IN SECONDS
YTITLE = DISPLACEMENT IN INCHES
YMIN = 0.
YMAX = 15.
XYPLOT DISP RESP/201(Tl),202(Tl),200(Tl)
$ END OF CASE CONTROL DATA
BEGIN BULK
TSTEPNL 21 200 .0025 2 AUTO 10
.
.
.

ENDDATA

Main Index
 614 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Case Control Section

The Case Control Section is used to define input data sets, output features, and subcases. The subcase structure in MSC
Nastran provides a unique means of changing loads, boundary conditions, and solution methods by making selections from
the Bulk Data. Confining the discussion to SOL 106 and SOL 129, loads and solution methods may change from subcase
to subcase on an incremental basis. However, constraints can be changed from subcase to subcase only in the static solution
sequence. As a result, the subcase structure determines a sequence of loading and constraint paths in a nonlinear analysis.
The subcase structure also allows the user to select and change output requests for printout, plot, etc., by specifying set
numbers with keywords. Any selections made above the subcase specifications are applicable to all the subcases. Selections
made in an individual subcase supersede the selections made above the subcases. Table 3-2 summarizes the Case Control
commands for nonlinear analysis.

Table 3-2 Summary of Nonlinear Case Control Commands

Load Selection
LOAD Selects static loading condition.
CLOAD Selects static load combination for superelements.
DLOAD Selects dynamic loading conditions.
LOADSET Selects static load sets defined on the Bulk data LSEQ.
NONLINEAR Selects nonlinear loading (NOLINi) for transient response.
IC Selects initial conditions for transient response.
Solution Method Selection
METHOD Selects methods for eigenvalue analysis.
NLPARM Selects iteration methods for nonlinear static analysis.
TSTEPNL Selects iteration methods for nonlinear transient analysis.
Output Requests
DISPLACEMENT Requests output for displacements of physical points.
VELOCITY Requests output for velocities of physical points.
ACCELERATION Requests output for acceleration of physical points.
ELFORCE Requests output for element forces.
STRESS Requests output for element stresses.
BOUTPUT Requests output for slideline contact.
SPCFORCES Requests output for constraint forces of SPC points.
NLLOAD Requests output for NOLINi in transient response.
OUTPUT(PLOT) Requests the beginning of the plotter output.
Superelement Control

Main Index
 Appendix C: Nonlinear Analysis 615
General Description of Nonlinear Capabilities

Table 3-2 Summary of Nonlinear Case Control Commands (continued)

SUPER Specifies the superelement identification number and the load sequence
number.
SEALL Combines the functions of SEMG, SELG, SEKR, SEMR, and SELR.
SEKR Specifies the superelement identification numbers for which stiffness matrices
will be assembled and reduced.
SELG Specifies the superelement identification numbers for which load vectors will
be generated.
SELR Specifies the superelement identification numbers for which the static load
matrices will be assembled and reduced.
SEMG Specifies the superelement identification numbers for which stiffness, mass,
and damping matrices will be generated.
SEMR Specifies the superelement identification numbers for which the mass and
damping matrices will be assembled and reduced.

Bulk Data Section

All details of the input data are specified in the Bulk Data Section. Constraint, load, and solution method Bulk Data entries
are not used unless they are selected in the Case Control Section. The nonlinear Bulk Data entries are summarized in
Table 3-3. All the input data designed specially for nonlinear analysis will be described in the following sections.

Table 3-3 Summary of Nonlinear Bulk Data Entries

Element Connectivity
CBEAM Defines connection for beam element.
CBUSH Defines connections for generalized spring and damper.
CBUSH1D Defines connections for rod type spring and damper.
CGAP Defines connection for gap or frictional element.
CHEXA Defines connection for six-sided solid element.
CONROD Defines connection and properties for rod.
CPENTA Defines connection for five-sided solid element.
CQUAD4 Defines connection for quadrilateral element with plane strain or bending and membrane
stiffness.
CQUAD8 and Defines connection for plane strain hyperelastic quadrilateral element.
CQUAD
CROD Defines connection for rod with axial and torsional stiffness.
CTETRA Defines connection for four-sided solid element.

Main Index
 616 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-3 Summary of Nonlinear Bulk Data Entries (continued)

CTRIA3 Defines connection for triangular element with plane strain or bending and membrane
stiffness.
CTRIA6 Defines connection for hyperelastic plane strain triangular element.
CQUADX Defines connection for axisymmetric hyperelastic quadrilateral element.
CTRIAX Defines connection for axisymmetric hyperelastic triangular element.
CTUBE Defines connection for a tube.
Element Properties
PBCOMP Defines properties for composite CBEAM.
PBEAM Defines properties for CBEAM.
PBUSH and Defines properties for CBUSH.
PBUSHT
PBUSH1D Defines properties for CBUSH1D.
PCOMP Defines properties for composite material laminate.
PGAP Defines properties for CGAP.
PROD Defines properties for CROD.
PSHELL Defines properties for CTRIA3 and CQUAD4.
PSOLID Defines properties for CHEXA, CPENTA and CTETRA.
PTUBE Defines properties for CTUBE.
PLPLANE Defines properties for hyperelastic plane elements CTRIA3, CTRIA6, CQUAD4,
CQUAD8, CQUAD, CQUADX and CTRIAX.
PLSOLID Defines properties for hyperelastic solid elements CHEXA, CPENTA and CTETRA.
Material Properties
CREEP Defines creep material properties.
MAT2 Defines anisotropic material properties for shell elements.
MAT8 Defines orthotropic material properties for shell elements.
MAT9 Defines anisotropic material properties for solid elements.
MATS1 Defines properties for plastic and nonlinear elastic material.
MATHP Defines properties for hyperelastic material.
TABLEDi Defines a function for internal load versus disp or velocity for bush element.
TABLES1 Defines a function for stress-dependent material properties or experimental data for
hyperelastic material.
TABLEST Combines many TABLES1 entries for temperature dependent material properties.

Main Index
 Appendix C: Nonlinear Analysis 617
General Description of Nonlinear Capabilities

Table 3-3 Summary of Nonlinear Bulk Data Entries (continued)

Slideline Contact
BCONP Defines contact parameters.
BLSEG Defines boundary line segments.
BFRIC Defines frictional properties for slideline contact.
BWIDTH Defines boundary line segments width/thickness.
BOUTPUT Defines grid points where contact output is requested.
Constraints
SPC Defines single-point constraints and enforced displacements.
SPC1 Defines single-point constraints.
MPC Defines a linear relationship for two or more degrees of freedom.
Loads
CLOAD Defines a static load combination for superelement loads.
FORCEi Defines concentrated load at grid point.
LSEQ Defines static load sets for dynamic analysis.
MOMENTi Defines moment at a grid point.
NOLINi Defines nonlinear transient load.
PLOAD Defines pressure loads on CQUAD4, CTRIA3, CHEXA, CPENTA, and CTETRA. Should
not be used for hyperelastic plane elements CQUAD4, CQUAD8, CQUAD, CTRIA3, and
CTRIA6 or for hyperelastic CHEXA, CPENTA, CTETRA with midside nodes.
PLOAD2 Defines pressure loads on shell elements, CQUAD4 and CTRIA3. Not available for
hyperelastic elements.
PLOAD4 Defines pressure loads on surfaces of CHEXA, CPENTA, CTETRA, CTRIA3 and
CQUAD4 elements. Not available for hyperelastic plane elements CQUAD4, CQUAD8,
CQUAD, CTRIA3, and CTRIA6.
PLOADX1 Defines pressure loads on axisymmetric elements CQUADX and CTRIAX.
RFORCE Defines load due to centrifugal force field.
TEMP Defines temperature at grid points.
TEMPPi Defines temperature field for surface elements.
TEMPRB Defines temperature field for line elements.
TIC Specifies initial values for displacement and velocity.
TLOADi Defines loads as a function of time.
Solution Methods

Main Index
 618 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-3 Summary of Nonlinear Bulk Data Entries (continued)

NLPARM Defines iteration methods for nonlinear static analysis.
NLPCI Defines arc-length methods for nonlinear static analysis.
TSTEPNL Specifies integration and iteration methods for nonlinear transient analysis.
EIGB or EIGRL Defines eigenvalue extraction method for buckling analysis.

Parameters
Parameters constitute an important part of the input data because they maintain communications among the user, DMAP,
and modules. Parameters are single-valued input data used for requesting special features or specifying miscellaneous data.
Users can change the initial values of parameters by specifying the PARAM entry in the Bulk Data Section or in the Case
Control Section. Many options available in MSC Nastran are controlled by user-specified parameters. Table 3-4 lists the
parameters used in nonlinear Solution Sequences 106 and 129.

Table 3-4 PARAMeters Used in Nonlinear Solution Sequences

B = Must be specified in the Bulk Data Section only.
E = May be specified in either the Bulk Data or Case Control Section.
Solution
Sequence
Parameter
Name 106 129 Default Description
ADPCON E E 1.0 Adjusts automatic penalty values for slideline contact regions.
AUTOSPC E E NO Specifies automatic single-point-constraints.
AUTOSPCR E B NO Specifies automatic SPC's for residual structure.
BAILOUT E E 0 Terminates superelement program with singularities.
BETA B 0.33333 Specifies control factor for Newmark-Beta method.
BUCKLE -1 Selects nonlinear buckling analysis for restarts.
CB1,2 E 1.0 Selects superelement damping matrix coefficients.
CK1,2,3 E 1.0 Selects superelement stiffness matrix coefficients.
CM1,2 E 1.0 Selects superelement mass matrix coefficients.
CNSTRT 1 Controls restarts for constraint change.
COUPMASS E E -1 Generates coupled rather than lumped mass matrices.
CURV -1 Computes stress data in a material coordinate system.
CURVPLOT -1 Requests x-y (curve) plots.
DBDICT B B -1 Prints database dictionary at the end and beginning of the run.

Main Index
 Appendix C: Nonlinear Analysis 619
General Description of Nonlinear Capabilities

Table 3-4 PARAMeters Used in Nonlinear Solution Sequences (continued)

B = Must be specified in the Bulk Data Section only.
E = May be specified in either the Bulk Data or Case Control Section.
Solution
Sequence
Parameter
Name 106 129 Default Description
DBDRNL E E 0 Prints database dictionary in the nonlinear loop (SOLs 106 and 129 only).
DLOAD B B 0 Indicates no other changes except for loads (SOLs 106 and 129 only).
DOPT 0 Controls x spacing of curves for CURVPLOT.
ERROR E E -1 Terminates run for superelement errors.
EST E E 2 Requests element measure and volume computation.
FACTOR B B 10,000 Generates SEQID on SEQGP entry.
FKSYMFAC B 1.0 Controls symmetrization of stiffness.
FOLLOWK E E YES Include follower force stiffness.
GPECT E E -1 Prints all elements connected to each grid point.
GRDPNT E E -1 Executes Grid Point Weight Generator.
G E 0.0 Specifies uniform structural damping coefficient.
INRLM E 0 Requests appending of inertia relief modes.
KDIAG 1.0 Applies value of a spring constant to nonlinear stiffness matrix diagonals.
K6ROT E E 100.0 Assigns stiffness to normal rotation of CQUAD4 and CTRIA3 elements.
LANGLE E E 1 Selects gimbal angle approach for large rotations if 1 and rotation vector
approach if 2.
LGDISP E E -1 Selects large displacement effects.
LOOPID E E 0 Specifies loop ID in the database for restarts.
MAXLP E 5 Maximum number of iterations for internal loop.
MAXRATIO E E 1 .0E + 7 Determines the singularity of the stiffness matrix.
MODEL B B 0 Stores several models in the database.
MPCX B B 0 Controls MPC and rigid element processing on restarts.
NDAMP E 0.025 Specifies numerical damping in ADAPT method.
NINTPTS 10 Requests CURV to interpolate over all elements.
NLAYERS B B 5 Specifies number of integration points through thickness for CQUAD4 and
CTRIA3.
NLTOL B 2 Sets accuracy tolerances

Main Index
 620 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-4 PARAMeters Used in Nonlinear Solution Sequences (continued)

B = Must be specified in the Bulk Data Section only.
E = May be specified in either the Bulk Data or Case Control Section.
Solution
Sequence
Parameter
Name 106 129 Default Description
NMLOOP E 0 Saves tangent stiffness in SOL 106 for normal modes and other linear solution
sequences.
NOCOMPS E 1 Suppresses composite stress output.
NODATA B B 0 Indicates no changes in Bulk Data (SOLs 106 and 129 only).
NOTRED E 0 Controls t-set reduction (SOL 129 only).
OG 0 Calculates stress/strain data at grid points for CURV.
OLDELM 1 Computes element forces using old method.
OUTOPT 0 Sets format for output quantities computed by CURV.
OUTPUT 0 Jumps to data recovery.
PDRMSG E 1 Suppresses data recovery messages.
PLOT E E 1 Jumps to undeformed plot.
PLOTSUP E E 0 Makes undeformed plots for superelements.
PLTMSG E E 1 Suppresses undeformed plot messages.
POST E E 1 Stores post-processing data blocks in database.
PRGPST E E YES Suppresses singularity printout.
PROUT E E -1 Suppresses execution and printout from ELTPRT.
PRPHIVZ E 1.0E + 37 Prints eigenvector matrix for general dynamic reduction.
RESDUAL E E 1 Skips to nonlinear operations (Phase 2) (SOLs 106 and 9129 only).
SDATA E 1 Obtains additional solution set output (SOL 129 only).
SEFINAL B B -1 Indicates no data changes if change SEFINAL.
SEMAP B B SEMAP Determines partitioning of superelements.
SEMAPPRT B B 3 Prints SEMAP table.
SENAME E E SENAME Identifies superelement by name.
SEQOUT B B 0 Controls output options for OLDSEQ.
SKPLOAD 1 Skips load vector calculation.
SKPMTRX 1 Skips matrix reduction and decomposition.
SLOOPID B 0 Specifies loop ID from SOL129 database for restart.

Main Index
 Appendix C: Nonlinear Analysis 621
General Description of Nonlinear Capabilities

Table 3-4 PARAMeters Used in Nonlinear Solution Sequences (continued)

B = Must be specified in the Bulk Data Section only.
E = May be specified in either the Bulk Data or Case Control Section.
Solution
Sequence
Parameter
Name 106 129 Default Description
SOLID B B 0 Stores several solutions in same database.
SPCGEN E E 0 Places automatic SPC's in SPC1 format.
SSG3 0 Skips operations prior to FBS.
START B B 0 Defines number of grid points for OLDSEQ.
STIME E 0.0 Specifies time step for restarts.
SUBID E 1 Specifies subcase ID for restarts.
SUBSKP 0 Skips a number of subcases for restarts.
SUPER B B 0 Deletes ungrouped grid points.
TABID E 2 Controls punch for response spectra.
TABS E 0.0 Converts units of temperature input to absolute.
TESTNEG E -1/-2/1 Tests for negative terms on factor diagonal of matrix.
TESTSE 1.0E + 36 Tests for strain energy for iteration control.
TSTATIC E -1 Selects static solution with TSTATIC=1 using ADAPT method in TSTEPNL.
USETPRT E E -1 Selects output type for degrees of freedom.
USETSEL E E 0 Controls sets for row list output option.
WTMASS E E 1.0 Multiplies terms of structural mass matrix by value.
W3,W4 E 0.0 Selects frequency for conversion of damping.

Nonlinear Element Output Codes for Plotting

Nonlinear elements (which reference nonlinear geometry and/or material) have different plot codes than linear elements.
The plot codes be found in Item Codes (p. 1051) in the .

Nonlinear Characteristics and General Recommendations

Modeling for nonlinear analysis should follow the guidelines for good modeling practice pertaining to linear analysis, which
are summarized as follows:
 The analyst should have some insight into the behavior of the structure to be modeled; otherwise, a simple model
should be the starting point.

Main Index
 622 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

 Substructuring should be considered for the modularity of the model and/or synergism between projects and
agencies involved.
 The size of the model should be determined based on the purpose of the analysis, the trade-offs between accuracy
and efficiency, and the scheduled deadline.
 Prior contemplation of the geometric modeling will increase efficiency in the long run. Factors to be considered
include selection of coordinate systems, symmetric considerations for simplification, and systematic numbering of
nodal points and elements for easy classification of locality.
 Discretization should be based on the anticipated stress gradient; that is, a finer mesh in the area of stress
concentrations.
 Element types and the mesh size should be carefully chosen. For example, avoid highly distorted and/or stretched
elements (with high aspect ratio); use CTRIA3 and CTETRA only for geometric or topological reasons.
 The model should be verified prior to the analysis by some visual means, such as plots and graphic displays.
Nonlinear analysis requires better insight into structural behavior. First of all, the type of nonlinearities involved must be
determined. If there is a change in constraints due to contact during loading, the problem may be classified as a boundary
nonlinear problem and would require CGAP elements or slide line contact algorithm. The material nonlinearity is
characterized by material properties. However, the material nonlinear effects may or may not be significant depending on
the magnitude and duration of the loading, and occasionally on environmental conditions. The anticipated stress level
would be a key to this issue. The geometric nonlinearity is characterized by large rotations which usually cause large
displacements. Intuitively, geometric nonlinear effects should be significant if the deformed shape of the structure appears
distinctive from the original geometry without amplifying the displacements. There is no distinct limit for large
displacements because geometric nonlinear effects are related to the dimensions of the structure and the boundary
conditions. The key to this issue is to know where the loading point is in the load-deflection curve of the critical area.
SOL 106 or SOL 129 must be used if the model contains any of the following:
 A parameter LGDISP (for geometric nonlinearity).
 Any hyperelastic element (PLPLANE or PLSOLID property entry, pointing to a MATHP material).
 Any CGAP element (CGAP and PGAP).
 Slideline contact (BCONP, BLSEG, and/or BFRIC).
 Any active nonlinear material data (specified on MATS1 and/or CREEP).
 Any combination of the above.
The model may consist of superelements, but only the residual structure (superelement 0) may consist of nonlinear
elements mixed with any type of linear elements. All the gaps, slideline contact regions, and hyperelastic elements are always
actively nonlinear if included in the residual structure. However, other potentially nonlinear elements in the residual
structure become actively nonlinear only if a parameter, LGDISP, is used and/or if they use the nonlinear material data
specified on the MATS1 and/or CREEP data entries.
The subcase structure of MSC Nastran may be used in SOLs 106 and 129 to characterize a nonproportional loading path.
Cyclic loading is one example. Some nonlinear problems, such as plastic deformation, creep deformation, and contact
problems with frictional forces, are path-dependent; others, such as hyperelastic or purely geometric nonlinear, are path-

Main Index
 Appendix C: Nonlinear Analysis 623
General Description of Nonlinear Capabilities

independent. An incremental process may not be required for a purely geometric nonlinear problem. For computational
efficiency, however, the analysis of such a problem does frequently need incremental solutions.
With these points in mind, additional recommendations are imperative for nonlinear analysis:
Identify the type of nonlinearity and localize the nonlinear region for computational efficiency. If unsure, perform
a linear analysis by SOL 101 or SOL 109 prior to the nonlinear analysis by SOL 106 or SOL 129, respectively.
Segregate the linear region by using superelements and/or linear elements if possible. Notice that the potentially
nonlinear elements can be used as linear elements.
The nonlinear region usually requires a finer mesh. Use a finer mesh if severe element distortions or stress
concentrations are anticipated.
Be prepared for restarts with the database properly stored in cases of divergence and changing constraints or
loading paths via subcases.
The subcase structure should be utilized properly to divide the load or time history for convenience in restarts, data
recovery, and database storage control, not to mention changing constraints and loading paths.
 The load or time for the subcase should then be further divided into increments, not to exceed 20 load steps (for
SOL 106) or 200 time steps (for SOL 129) in each subcase.
 Many options are available in solution methods to be specified on NLPARM (for SOL 106) or TSTEPNL (for
SOL 129) data entries. The defaults should be used on all options before gaining experience.
 Caution should be exercised in specifying CGAP element properties. In particular, the closed gap stiffness should
not exceed the stiffness of the adjacent degree of freedom by 1000 times.
 Normal rotation for CQUAD4 and CTRIA3 elements should be restrained by a parameter K6ROT when the
geometric nonlinearity is involved. Default value for parameter K6ROT is 100.
 Understand the basic theory of plasticity or creep before using these capabilities.
 Caution should be exercised in preparing input data for creep, because it is unit dependent.
 For the transient analysis by SOL 129, some damping is desirable and the massless degree of freedom should be
avoided.
 The time step size for a transient response analysis should be carefully determined based on the highest natural
frequency of interest because it has significant effects on the efficiency as well as the accuracy.
For any anomalies, refer to the MSC Nastran error list and general limitations list.

Material Nonlinearity
Material nonlinearity becomes an issue in the analysis if the stress-strain relationship is nonlinear. Constitutive equations
characterize the macroscopic behavior of the material, which stems from its microscopic constitution. All engineering
materials are inherently nonlinear. It is, however, not feasible to characterize a nonlinear material by a single set of equations
for the entire range of environmental conditions, such as loading, temperature, and rate of deformation. Depending on the
application or the situation, the material behavior may be idealized or simplified to account for certain effects, which are
important in the analysis.

Main Index
 624 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

The linear elastic (Hookean) material is the simplest case of all. As the strain exceeds the infinitesimal range, the material
exhibits nonlinear behavior: nonlinear elastic if it is recoverable and plastic if it is irrecoverable. If the temperature effects
on the material properties become an important issue, coupling between thermal and mechanical behavior should be
properly taken into consideration, which is the realm of thermo-elasticity and thermo-plasticity. If the strain-rate has
significant effects on the material response, we have to resort to the theories of visco-elasticity and visco-plasticity.
Anisotropic effects add complexity to the material model.
Development of material constitutive models has been prolific in recent years. Detailed discussion in this section will be
confined to the capabilities provided in MSC Nastran. The following nonlinear materials are available:
 Nonlinear elastic
 Hyperelastic
 Elastic-plastic
 Visco-elastic
 Temperature dependent

Nonlinear Elastic Material

Typical stress-strain relationship for a nonlinear elastic material is shown in Figure 3-1. Notice that both the loading and
unloading occur along the same path.

loading unloading

Total Strain


Figure 3-1 Nonlinear Elastic Material

The nonlinear elastic theory implemented is useful to predict multiaxial stress state for materials for which only the tension
and compression stress-strain data is available. The stress-strain data in tension may be different from that in compression.
The theory implemented in MSC Nastran is adequate only for small strains (less than 20%) in multiaxial stress state as it

Main Index
 Appendix C: Nonlinear Analysis 625
General Description of Nonlinear Capabilities

is not based on the classical theory of finite elasticity. For large strains in multiaxial state, use of hyperelastic material is
strongly recommended.

Hyperelastic Material
The hyperelastic model is used to analyze elastomers up to a large strain. Elastomers are natural or synthetic rubberlike
materials that have the ability to undergo deformation under the influence of a force and regain their original shape once
the force is removed (unless damage occurs). They are used extensively in many industries, because of their wide availability
and low cost. Some examples of “real world” analyses in the automotive industry include o-rings, bushings, gaskets, seals,
rubber boots, and tires.
The hyperelastic constitutive model available in MSC Nastran is based on a generalized Rivlin strain energy function
expressed as follows:
Na Nd
j 2i
U =  A ij  I 1 – 3   I 2 – 3  +  Di  J – 1 – v  T – T0  
i+j = 1 i=1

where A ij are the material constants related to shear deformation (distortion) and D i are the material constants related to
volumetric deformation. The strain energy function is a potential for the stresses. For example, the Cauchy stresses; i.e.,
components of the stress vector t = dP  dA where dA is an infinitesimal area in the deformed configuration and P is
the force, are obtained as follows:
2 U
 = --- F ------- F T
J C

where F = dx  dX is the deformation gradient; x and X are the positions of a material point in the deformed and the
undeformed configurations, respectively; C = F T F is the right Cauchy-Green deformation tensor; J = det F is the
volume ratio dV  dV 0 ; and I 1 and I 2 are the first and second distortional strain invariants,

I 1 =  12 +  22 +  32 I 2 =  1– 2 +  2– 2 +  3– 2

where  I are principal stretches l  l 0 of the distortional deformation; l is the current and l 0 is the original length along

a principal direction;  12 are principal values of C = J – 2 / 3 C ;  v is the volumetric coefficient of thermal expansion;
 v = 3 , where  is the coefficient of linear thermal expansion of an isotropic material; T is the current and T 0 is the
initial temperature.
The use of a strain energy function implies isotropy in the undeformed configuration. The model can have up to 25 material
constants.
Special cases of the material law are:

Main Index
 626 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Mooney-Rivlin Material, N a = 1

Neo-Hookean Material, N a = 1 , A 01 = 0.0

A curve-fitting algorithm based on least squares fitting is available to determine the material constants A ij and D i based
on experimental data up to the fifth-order polynomial form ( N a and/or N d up to 5) for the following deformations of the
material:
 Simple tension/compression
The simple tension/compression test is characterized by a prescribed axial force or displacement. X values in the
TABLES1 entry which contains experimental data must be stretch ratios l  l 0 . Y values must be values of the
engineering stress F  A 0 .

 Equibiaxial tension
The equibiaxial tension test is characterized by prescribed equal tensile stresses or elongations in two directions and
is typically produced by the pressure of a spherical membrane. X values in the TABLES1 entry which contains
experimental data must be stretch ratios l  l 0 . Y values must be values of the engineering stress F  A 0 , l = current
length, F = current value of the force, l 0 and A 0 = initial length and cross-sectional area, respectively. In the case

of pressure of a spherical membrane, the engineering stress is given by Pr 0  2  2t 0 where P = current value of
the pressure and r 0 , t 0 = initial radius and thickness.

 Simple shear
The simple shear test is obtained by a prescribed shearing force or displacement of a cube. This is not a
homogeneous deformation. It is approximated by one that is homogeneous isochoric as follows:
x 1 = X 1 + X 2 x2 = X2 x3 = X3

Main Index
 Appendix C: Nonlinear Analysis 627
General Description of Nonlinear Capabilities

x2 , X2

tan – 1 

x1 , X1

where x 1 , x 2 , x 3 are the coordinates of the deformed body and X 1 , X 2 , X 3 are the coordinates of the body in
its undeformed configuration. The user should be aware of this approximation. X values in the TABLES1 entry
which contains experimental data must be values of the shear tangent  . Y values must be values of the engineering
shear stress F  A 0 . It is recommended that simple shear be used along with another test case; otherwise, the
coefficients cannot be determined uniquely.
 Pure shear
The pure shear test is a more direct way of imposing a state of shear deformation and is obtained by a prescribed
uniform stretch  1 of a specimen, which is long in the 2-direction leaving the 3-direction stress free, or by holding
the 2-direction fixed while applying strains in the 1-direction of a thin rectangular sheet, as in the classical
experiments of Rivlin and Saunders. If the material is incompressible, this results in  3 = 1   1 , which is
characteristic of a state of shear deformation. It is recommended that pure shear be used along with another test
case; otherwise, the coefficients cannot be determined uniquely. X and Y values, respectively, in the TABLES1 entry
which contains experimental data must be stretch ratios  1 = l  l 0 and values of the nominal stress F  A 0 l =
current length, F = current value of the force, l 0 and A 0 = initial length and cross-sectional area, respectively in
the 1-direction.
 Volumetric compression
Pure volumetric compression data may be used to obtain fitted values of the material constants related to volumetric
deformation. X values in the TABLES1 entry which contains experimental data must be values of the volume ratio
J =  3 , where  = l  l 0 is the stretch ratio in all three directions; Y values must be values of the pressure,
assumed positive in compression.

Main Index
 628 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

The first four test cases (experiments) are used to obtain the material constants A ij . The last test case (volumetric
compression) is used to obtain the material constants D i . In obtaining the material constants, compressibility is taken into
account, in the case of simple tension/compression and equibiaxial tension, when the volumetric part of the strain energy
function is of the order of one.

User Interface
The MATHP material entry defines the hyperelastic material. The MATHP entry consists of five continuation entries that
define the A ij and D i material constants of the strain energy function. The continuation entries are not required if a Neo-
Hookean or a Mooney-Rivlin material is modeled.
If the material constants are not known, the curve-fitting algorithm is invoked if the sixth continuation entry of the
MATHP entry contains nonblank fields, TAB1, TAB2, TAB3, TAB4, and TABD. The TABi fields reference experimental
data that is contained in TABLES1 entries. The order of the strain energy polynomial function is specified using the NA
and ND fields of the MATHP entry for the distortional and volumetric parts, respectively. The curve-fitting algorithm
computes the material constants from the supplied data, and the analysis continues using the produced constants.

Output Description
The curve-fitting algorithm computes the material constants and prints the results in an updated MATHP entry image in
the output file. The image is printed using a 16-character field and therefore is convenient to directly input the produced
MATHP entry into a data file in large-field format. An example of the produced output is shown in Listing 3-3.

Guidelines
If the strains are small, the results of the hyperelastic elements should compare with those of the linear elastic elements where
the shear modulus and the bulk modulus G  2  A 10 + A 01  and the bulk modulus K  2D 1 .

Conventional Mooney-Rivlin and Neo-Hookean materials are fully incompressible. Full incompressibility is not presently
available but may be simulated with a large enough value of D 1 . However, a value of D 1 higher than 10 3   A 10 + A 01 
is not recommended.
When determining the material constants from experimental data, it is best to use data from more than one test involving
different kinds of deformation. A rank deficiency warning may occur in the least squares fitting algorithm if there is
insufficient experimental data. It is best to try to avoid rank deficiency by providing more experimental data or by lowering
the order of the strain energy polynomial.

Limitations
It is recommended that Poisson's ratio for the hyperelastic material be no greater than .4995, or, equivalently, that
D 1   A 10 + A 01  does not exceed 1000. As incompressibility increases, convergence difficulties and/or ill conditioning
of the stiffness matrix occurs.

Main Index
 Appendix C: Nonlinear Analysis 629
General Description of Nonlinear Capabilities

The hyperelastic material cannot be defined for a small strain element; that is, it cannot be referenced by a PSHELL or
PSOLID Bulk Data entry.
Temperature dependence of the hyperelastic material constants is not supported.

Example: Curve-Fitting Algorithm

An example created to solve a fully nonlinear (finite deformation) CQUAD4 element under pure shear with material
constants obtained from the simultaneous fitting of experimental data in simple tension, equibiaxial tension, and pure
shear. Annotated Bulk Data input is shown in the following listing that illustrates the MATHP and TABLES1 entries,
which contain the experimental data for curve fitting.

Main Index
 630 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Listing 3-3 Example Using Material Curve Fitting

BEGIN BULK
$ PARAMETERS
PARAMDBDROPT0
PARAMLGDISP1
$
NLPARM1040AUTO1
$
$ DEFINE GEOMETRY
GRID 10.0.0.123456
GRID 21.0.0.23456
GRID 31.0.1.2456
GRID 40.0.1.12456
$ DEFINE ELEMENTS
CQUAD4 111234
$ CONSTRAINTS AND LOADING
MPC10211.31-1.
MPC10431.33-1.
SPC1031
SPCD 100315.
$ DEFINE PROPERTIES
$ LARGE STRAIN QUAD4 ELEMENT
PLPLANE11ISOP2
$ HYPERELASTIC MATERIAL
$ D1 SPECIFIED
$ AIJ MATERIAL CONSTANTS TO BE DETERMINED FROM EXPERIMENTAL DATA
MATHP11500.+MA101
+MA10131+MA102
+MA102+MA103
+MA103+MA104
+MA104+MA105
+MA105+MA106
$TAB1TAB2TAB3TAB4TABD
+MA106100200400
$
$ TAB1 DATA
$

$ Treloar [1944] data in simple tension. Nominal stresses in kgcm-2

$
TABLES1 100 +TA101
+TA101 1. 0. 1.125 1. 1.25 2. 1.5 3. +TA102
+TA102 1.525 4. 1.875 5. 2. 6. 2.25 7. +TA103
+TA103 2.5 8. 3. 9. 3.625 10. 4. 12.5 +TA104
+TA104 4.75 16. 5.25 20. 5.75 23. 6. 27. +TA105
+TA105 6.25 31. 6.5 34. 6.75 38.75 7. 42.5 +TA106
+TA106 ENDT
$
$ TAB2 DATA
$
$ Treloar [1944] data in equibiaxial tension. Nominal stresses in kg-cm-2
$
TABLES1 200 +TA201
+TA201 1.016 0.83 1.07 1.56 1.15 2.55 1.21 3.25 +TA202
+TA202 1.335 4.31 1.44 5.28 1.66 6.65 1.93 7.88 +TA203
+TA203 2.46 9.74 3. 12.64 3.4 14.61 3.77 17.33 +TA204
+TA204 4.1 20.11 4.32 22.40 4.54 24.41 ENDT
$
$ TAB4 DATA
$
$ Treloar* [1944] data in pure shear. Units of the nominal stress are kgcm-2

$
TABLES1 400 +TA401
+TA401 1.01 0.24 1.1 1.12 1.2 1.92 1.31 2.87 +TA402
+TA402 1.49 3.62 1.86 5.61 2.36 7.2 2.97 9.045 +TA403
+TA403 3.45 10.8 3.93 12.45 4.4 14.23 4.714 15.826 +TA404
+TA404 4.96 17.46 ENDT
ENDDATA

*See Experimental Data Fitting in the MSC Nastran Nonlinear User’s Guide.

The following listing illustrates the produced MATHP entry image that is printed in the output just after the ENDDATA
entry image.

Main Index
 Appendix C: Nonlinear Analysis 631
General Description of Nonlinear Capabilities

Listing 3-4 Effective Stress-Strain Law

*** SYSTEM INFORMATION MESSAGE 6410 (IFP8). BEGIN PROCESSING OF MATHP ENTRY ID = 1
DISTORTIONAL PARAMETER FITTING.
SQUARE ROOT OF SUM OF THE ERRORS SQUARED = 0.25629E+01
UPDATED MATHP ENTRY
1 1.54082 0.24046 1500.00000 0.00000 0.00000 0.00000 0.00000
3 1
-0.00375 -0.01743 0.00104 0.00000
0.00029 0.00033 -0.00004 0.00000 0.00000
0.00000 0.00000 0.00000 0.00000 0.00000 0.00000
0.00000 0.00000 0.00000 0.00000 0.00000 0.00000 0.00000

Figure 3-2(a) through Figure 3-2(c) illustrate the curves obtained from this and similar files containing a large strain CHEXA
in simple and equibiaxial tension with the material constants generated for the first-, second-, and third-order strain energy
polynomial functions from simultaneous fitting of experimental data in simple tension, equibiaxial tension, and pure shear.

Figure 3-2 MSC Nastran Curve-Fitting Algorithm for Hyperelastic Materials (a)

Main Index
 632 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Figure 3-3 MSC Nastran Curve-Fitting Algorithm for Hyperelastic Materials (b)

Figure 3-4 MSC Nastran Curve-Fitting Algorithm for Hyperelastic Materials (c)

Main Index
 Appendix C: Nonlinear Analysis 633
General Description of Nonlinear Capabilities

Elastic-Plastic Material
Before discussing the computational aspects associated with the various models, the definitions of some of the commonly
used terms in plasticity are given below.

Yield Stress Yield Stress is usually measured as the value of stress which produces the smallest measurable
permanent strain.
Bauschinger Effect When a plastically deformed specimen is unloaded, residual stresses on a microscopic scale
remain and influence the plastic yielding for the different loading. If the previous strain was a
uniform extension and the specimen is then reloaded in compression in the opposite direction,
it is observed that yielding occurs at a much reduced stress. This is known as the Bauschinger
Effect.
The phenomenon of Bauschinger effect can best be described with reference to Figure 3-5 which
shows an idealized stress-strain curve of a metal first deformed by uniform tension, the load
removed and then reloaded in compression. According to one model (which is one extreme
viewpoint), it is assumed that the elastic unloading range will be double the initial yield stress.
If the initial yield stress in tension is  y and the specimen is loaded up to a stress  1 and
unloaded, the plastic yielding will begin in compression at a stress  2 given by

 2 =  1 – 2 y

loading

unloading

Total Strain

Plastic
Figure 3-5 Material Nonlinear Models

Main Index
 634 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Yield Function
Associated with the yield criteria, there exist yield functions in the form of

F  , p  = f    – Y   p  = 0
where:

f    = an effective stress, a function of the stress state

Y   p  = a tensile yield stress, a function of the stress hardening.

Geometric representation of the yield function is a surface in the stress space (  1 –  2 –  3 coordinates) which is called
a yield surface. The von Mises yield surface is a circular cylinder and Tresca's is a hexagonal cylinder. The yield surface is
reduced to the yield locus in the case of plane stress, as shown in Figure 3-6 and Figure 3-7 for the von Mises and Tresca
criteria, respectively. The Mohr-Coulomb and Drucker-Prager yield functions represent conical surfaces as shown in
Figure 3-8 and Figure 3-9.

Figure 3-6 von Mises Yield Locus for Plane Stress

2

2 = Y
Y

2 – 1 = Y

1 = Y

1
–Y Y

–1 = Y 1 – 2 = Y

–2 = Y
–Y
Figure 3-7 Tresca's Yield Locus for Plane Stress

Main Index
 Appendix C: Nonlinear Analysis 635
General Description of Nonlinear Capabilities

2

Y

1
–Y Y

–Y

Main Index
 636 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Drucker Prager   0
1952 1 = 2 = 3

–3

von Mises
 = 0
3 c cot

–2
c cot
Tension 
Failure

62 sin 
tan  = --------------------------------
3  3 – sin  

–1 Drucker-Prager and von Mises

Figure 3-8 Some Isotropic Yield Surfaces in Principal Stress Space

Figure 3-9 Some Isotropic Yield Surfaces in Principal Stress Space

Viscoelastic Material in Nonlinear Analysis

The quasi-static behavior of the viscoelastic material may be analyzed using the creep capability of MSC Nastran.
Viscoelastic properties (linear or nonlinear) can be defined in terms of rheological model parameters ( K p , C p , and C s )
of the Maxwell-Kelvin model in the Bulk Data entries CREEP and TABLES1. Then the quasi-static analysis of the
viscoelastic materials can be performed by specifying time increment in the Bulk Data entry NLPARM (selected by the
subcase) in SOL 106.

Main Index
 Appendix C: Nonlinear Analysis 637
General Description of Nonlinear Capabilities

Mohr-Coulomb   0
(1882) (1733)

1 = 2 = 3

–3

Tresca (1864)
 = 0
3 c cot 

–2

Pressure-Dependent
–1

Mohr-Coulomb and Tresca

Temperature-Dependent Material
Currently only linear and nonlinear elastic materials can be temperature dependent. Temperature-dependent properties for
elastic and nonlinear elastic materials are specified by using the MATTi, TABLEST, and TABLES1 Bulk Data entries.

Main Index
 638 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Elastic-plastic and hyperelastic materials cannot be temperature dependent. A limited temperature dependence for creep
materials is available based on the Arrhenius equation.

Geometric Nonlinearity
Geometric nonlinearities are manifested in problems involving large rotations and large deformation. The characteristics
are follower forces due to large rotations, geometric stiffening due to initial stress effect (as a result of large rotations), and
large strains due to large deformation.
Geometric nonlinear effects should be significant if the deformed shape of the structure appears distinctive from the original
geometry by a visual inspection. A more rigorous and quantitative definition for the large displacements can be derived
from the plate theory of Kirchhoff and Love: the small deflection theory is valid for a maximum deflection of less than 20%
of the plate thickness or 2% of the small span length. However, this definition seems to be a little conservative for numerical
analysis, and there is no distinct limit for large displacements because geometric nonlinear effects are related to the boundary
conditions as well as the dimensions of the structure. If the load-deflection curve of the critical point can be estimated, the
loading point should be in the nonlinear portion of the curve.
Geometric nonlinear effects in the structure involving large rotations, whether rigid body rotations or deformation induced
rotations, are self-evident. Stiffening of a membrane, stiffness in a pendulum or snap-through of an arch belong to this
category. The motion of a pendulum under gravity is caused by geometric (differential) stiffness. Follower forces are
manifested when the applied loads are displacement dependent, such as pressure load and thermal load applied on the
surface that rotates. Centrifugal force is another example of follower forces. Large strain effects are pronounced in metal
forming, rubber and elastomer applications. In such applications the strains exceed 100%. Finite strain formulation is
required to treat the problems in this category. MSC Nastran currently supports the large strain capability for rubbers and
elastomers. This formulation contains the effect of large strain as well as geometric nonlinearity. However, it does not
support large strain formulation for metal forming applications. In most structural applications, however, moderately large
strains (20 to 30%) appear in local areas if there is any large deformation
Other geometric nonlinear effects are treated by updated element coordinates, gimbal angles (or rotation vector), and the
differential stiffness  K d  .

Nonlinear Elements
The nonlinear elements can be divided into two categories: physical and contact elements. Physical elements have material
properties associated with them. Contact elements do not have any material properties associated with them.

Physical Elements
The nonlinear physical elements are listed in Table 3-5 and Table 3-6.
The CBUSH element may have nonlinear force-deflection properties defined on the PBUSH and PBUSHT property
entries. The CBUSH element is geometric linear; the orientation is not updated in case of large deformation. The
CBUSH1D element is both geometric and material nonlinear.

Main Index
 Appendix C: Nonlinear Analysis 639
General Description of Nonlinear Capabilities

The elements CROD, CONROD, and CTUBE may have material nonlinear extensional properties, with linear torsion.
The user may supply plastic or nonlinear elastic material properties. Since the stress-strain curve for compression need not
be the same as for tension, this element can, for example, be used to model cables which cannot carry tension.
The CBEAM element has been modified to provide plastic hinges at the ends of an otherwise elastic element.This element
is intended for collapse analysis in frameworks with loads at the joints, and materials with small work hardening. The user
need not specify the cross-section axis about which the yielding occurs, since the implementation allows for combinations
of bending moments in two directions plus an axial load. The flexibility of the plastic hinge is based upon eight idealized
rods at each end, chosen to match the total area, center of gravity and moments of inertia of the cross section. As is shown
in Table 3-7, the calculated ultimate moment agrees reasonably well with the theoretical value for many cross-sections. The
material specified on a MATS1 Bulk Data entry should be elastic-perfectly plastic. Use of any nonlinear material other than
elastic-perfectly plastic will give inaccurate results.
The CQUAD4 and CTRIA3 plate elements are available for combined bending and membrane strains found in shell
structures as well as for plane strain applications. For shell type problems the state of stress is assumed to be constant in each
of a variable number of equal layers. One layer should be used for plane strain. The user input is the one-dimensional stress-
strain curve for tension. Higher order CQUAD8, CQUAD, and CTRIA6 elements exist for fully nonlinear (i.e., large strain
and large rotation) hyperelastic plane strain analysis.
The CHEXA, CPENTA, and CTETRA elements are used to model three dimensional solid structures. No midside nodes are
allowed for nonlinear analysis, except for the hyperelastic 20-noded CHEXA, 15-noded CPENTA, and 10-noded CTETRA
elements.
Axisymmetric elements CQUADX and CTRIAX are available for fully nonlinear analysis with the hyperelastic material.

Table 3-5 Physical Elements for Nonlinear Analysis -- Small Strain Elements
Geometric
Nonlinear Elasto- Only
Name Characteristics Elastic Plastic Nonlinear
CBUSH Generalized 3D spring Yes No No
damper with 6 stiffnesses
for 3 translations and 3
rotations.

CBUSH1D Rod-like spring and Yes No Yes

damper with generalized
force function f  u,u· 
for axial displacement u
and velocity u· .

Main Index
 640 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-5 Physical Elements for Nonlinear Analysis -- Small Strain Elements (continued)
Geometric
Nonlinear Elasto- Only
Name Characteristics Elastic Plastic Nonlinear
CROD, CONROD, CTUBE Extension may be plastic Yes Yes Yes
or nonlinear elastic.
Torsion is elastic (linear).

CBEAM Plastic hinge at each end, No Yes (only Yes

which couples axial elastic-
motion and rotations. perfectly
1/8 1/8 plastic)
3/4
Linear material for center
section, traverse shear,
and torsion.

Plastic
Hinges
CQUAD4 Shell and Plate

Two-dimensional Yes Yes Yes

plasticity with FIVE
layers for membrane and
bending. 12 layers
maximum; 1 layer
minimum.

Transverse shear is linear.

Plane Strain

PSHELL with MID2=-1 Yes Yes Yes

and PARAM,NLAYER=1
Plane Stress

MID2=0 Yes Yes Yes

NLAYER=1

Main Index
 Appendix C: Nonlinear Analysis 641
General Description of Nonlinear Capabilities

Table 3-5 Physical Elements for Nonlinear Analysis -- Small Strain Elements (continued)
Geometric
Nonlinear Elasto- Only
Name Characteristics Elastic Plastic Nonlinear
CTRIA3 Shell and Plate

Two-dimensional Yes Yes Yes

plasticity with FIVE
layers for membrane and
bending. 12 layers
maximum; 1 layer
minimum.

Transverse shear is linear.

Plane Strain

PSHELL with MID2=-1 Yes Yes Yes

and PARAM,NLAYER=1
Plane Stress

PSHELL with MID2=0 Yes Yes Yes

NLAYER=1
CHEXA Eight Gauss points for Yes Yes Yes
nonlinear analysis.

Reduced shear
integration.

Strain (bubble) function.

No mid side nodes

allowed.

Main Index
 642 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-5 Physical Elements for Nonlinear Analysis -- Small Strain Elements (continued)
Geometric
Nonlinear Elasto- Only
Name Characteristics Elastic Plastic Nonlinear
CPENTA Six Gauss points for Yes Yes Yes
nonlinear analysis.

Reduced shear
integration.

Strain (bubble) function. No No No

CTETRA For four-noded Yes Yes Yes

tetrahedron a single
Gauss point for nonlinear
analysis.

***
For 10-noded
tetrahedron four Gauss
points for nonlinear

***
analysis.

 One-dimensional stress-strain curves use the MAT1 entry.

 All other elements may be used for nonlinear analysis as long as they remain linear.

Main Index
 Appendix C: Nonlinear Analysis 643
General Description of Nonlinear Capabilities

Table 3-6 Physical Elements for Nonlinear Analysis -- Fully Nonlinear

(Finite Deformation) Hyperelastic Elements
Name Characteristics Gauss Point   
CQUAD, CQUAD4, CQUAD8, Plane strain or axisymmetric 1 – – --
CQUADX analysis.
 –
Linear isoparametric 2
 quadrilateral with constant – 
4 3 pressure. Compressible and 3
nearly incompressible  
applications. 4

 Four Gauss points.  = 1 3

1 2
CQUAD, CQAUD8, CQUADX Plane strain or axisymmetric 1 – – --
analysis.
2 0 –
 Quadratic isoparametric
4 7 3 quadrilateral with linear 3  –
pressure. Compressible and
nearly incompressible 4 – 0
8 9 6 applications. Center node not
 5 0 0
available on the CQUAD8
connectivity. 6  0
1 5 2 Nine Gauss points. 7 – 0

8 0 
9  

 = 15  5

Main Index
 644 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-6 Physical Elements for Nonlinear Analysis -- Fully Nonlinear

(Finite Deformation) Hyperelastic Elements (continued)
Name Characteristics Gauss Point   
CTRIA3, CTRIA6, CTRIAX Plane strain or axisymmetric 1 1/3 1/3 --
analysis.

1 Linear isoparametric triangle

with constant pressure.
Compressible and nearly
incompressible applications.

2 3 One Gauss point.

 
CTRIA6, CTRIAX Plane strain or axisymmetric 1 1/6 1/6 --
analysis.
2 2/3 1/6
1 Quadratic isoparametric triangle
with constant pressure. 3 1/6 2/3
Compressible and nearly
4 6 incompressible applications.

2 3 Three Gauss points.

5
 
CHEXA Linear isoparametric 1   
hexahedron with constant
pressure. Compressible and 2 –  
 nearly incompressible
8 7 applications. 3  – 
 4 – – 
5 6 Eight Gauss points.
 5   –
4
3 6 –  –

7 + – –
1 2
8 – – –
 = 1 3

Main Index
 Appendix C: Nonlinear Analysis 645
General Description of Nonlinear Capabilities

Table 3-6 Physical Elements for Nonlinear Analysis -- Fully Nonlinear

(Finite Deformation) Hyperelastic Elements (continued)
Name Characteristics Gauss Point   
CHEXA Quadratic ispoarametric 1   
hexahedron with linear pressure. 2 0
Compressible and nearly 3
 
incompressible applications. 4
–  
 5  0 
8 7
Twenty-seven Gauss points. 6 0 0
19
7 0

20 18 –
 8 – 
5 17 6
15 9

 0 – 
16 10
11 – – 
4 14 11
13 3
12 12   
10 13 0 0
14
 0
1 9 2 15
–  0
16  0 0
17 0 0 0
18 – 0 0
19 – 0
20
 0
21 0 – 0
22 – – –
23  
24
–
0 
25 –
26
–  –
27  0
0 0
–
 =  15   5 –
– 0
 – –
0 – –
– – –

Main Index
 646 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-6 Physical Elements for Nonlinear Analysis -- Fully Nonlinear

(Finite Deformation) Hyperelastic Elements (continued)
Name Characteristics Gauss Point   
CPENTA Linear isoparametric 1 1/2 0 –1  3
pentahedron with constant
pressure. Compressible and 2 1/2 1/2 –1  3
6 5 nearly incompressible
 applications. 3 0 1/2 –1  3
 
Six Gauss points. 4 1/2 0 1 3
4
5 1/2 1/2 1 3
6 0 1/2 1 3
3 2

Main Index
 Appendix C: Nonlinear Analysis 647
General Description of Nonlinear Capabilities

Table 3-6 Physical Elements for Nonlinear Analysis -- Fully Nonlinear

(Finite Deformation) Hyperelastic Elements (continued)
Name Characteristics Gauss Point   
CPENTA Quadratic isoparametric 1 1/3 1/3 – 15  5
pentahedron with linear
pressure. Compressible and 2 1 1 – 15  5
nearly incompressible
applications. 3 1 1 – 15  5
6 14 5
 Twenty-one Gauss points. 4 1 1 – 15  5
 15 13
 5 2 2 – 15  5
4
12 11 6 2 2 – 15  5

 1 = 0.0597158717 7 2 2 – 15  5
3 10 2
8  1 = 0.4701420641 8 1/3 1/3 0
9 7
 2 = 0.7974269853 9 1 1 0
1
 2 = 0.1012865073 10 1 1 0

11 1 1 0

12 2 2 0

13 2 2 0

14 2 2 0

15 1/3 1/3 + 15  5
16 1 1 + 15  5
17 1 1 + 15  5
18 1 1 + 15  5
19 2 2 + 15  5
20 2 2 + 15  5
21 2 2 + 15  5

Main Index
 648 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-6 Physical Elements for Nonlinear Analysis -- Fully Nonlinear

(Finite Deformation) Hyperelastic Elements (continued)
Name Characteristics Gauss Point   
CTETRA Linear isoparametric 1 1/4 1/4 1/4
tetrahedron with constant
pressure. Compressible and
 nearly incompressible
applications.
4
One Gauss point.

1
2 3

 
CTETRA Quadratic isoparametric 1 1/6 1/6 1/6
tetrahedron with linear pressure.
Compressible and nearly 2 1/2 1/6 1/6
 incompressible applications.
3 1/6 1/2 1/6
Five Gauss points.
4 4 1/6 1/6 1/2

5 1/4 1/4 1/4

9 10
8

1
5 7
2 3

 6 

Main Index
 Appendix C: Nonlinear Analysis 649
General Description of Nonlinear Capabilities

Table 3-7 Accuracy in Calculation of Ultimate Moment in Yielded State

Section

z
t
Moment Axis 
Any multiple of 45° .9481
y from y-axis

Moment Axis 
y or z .9856
y

z
h/w  for y-axis  for z-axis
0 .8536 .9856
t
.5 .9031 .9543
h y
1.0 .9295 .9295

2.0 .9543 .9031

w
 .9856 .8536
t « h, w

Main Index
 650 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-7 Accuracy in Calculation of Ultimate Moment in Yielded State (continued)

Section
ht w / wt f  for y-axis  for z-axis
z tf 0 .8536 .9856

.5 .8826 1.102

1.0 .9031 1.207

h y
2.0 .9295 1.394
tw
 .9856
w
t f , t w « h, w

Contact Elements
The contact (interface) elements are listed in Table 3-8.
The CGAP element is intended to model surfaces which may come into contact. When positive pressure exists, the gap can
carry any transverse shear load which is less than the coefficient of friction times the normal load. The CGAP element
connects two grid points which may be initially coincident. There is no geometric nonlinear behavior, which implies that
the orientation of the contact plane does not change during deflection. (The physical shape of the two contact surface would
have to be specified and it would require the solution of a difficult analytic problem to determine the location of the actual
contact point or points.)
Due to the requirements of the solution algorithm, there must be a finite compression stiffness for the CGAP element. The
value must be carefully chosen, since a very large value may lead to numerical problems. A finite extensional stiffness may
be used to prevent the drift of structures that are supported only by CGAP elements. The user may also supply an initial
opening of the gap. An orientation vector is required (see CGAP entry for details) to define the transverse axes, even if no
shear is to be carried.
The slideline contact is capable of modeling nonlinear contact geometry and inelastic material behavior including large
deformation. Slideline contact is useful for two-dimensional geometries, for example plane stress, plane strain and
axisymmetric. They can also be used for three-dimensional geometries provided the contact between the two bodies can be
defined in terms of parallel planes called the slideline planes. The sliding and separation of the two bodies is restricted to
the slideline planes. The bodies can have large relative motions within the slideline plane. However, relative motions outside
the slideline planes are ignored; therefore, they must be small compared to a typical slideline element.

Main Index
 Appendix C: Nonlinear Analysis 651
General Description of Nonlinear Capabilities

Table 3-8 Contact Elements for Nonlinear Analysis

Geometric
Name Characteristics Friction Law Nonlinearity
CGAP Contact between two points. Coulomb's Law with: No
kinetic friction 0 static friction
Extension has different springs for
compression and tension.

Shear force is less than the friction

coefficient times the compression.
Transverse finite compression stiffness is
required.
Slideline Contact between a point and a line. Coulomb's Law with: Yes
kinetic friction = static friction

Small Strain Elements

A subset of MSC Nastran's elements for linear analysis is available for nonlinear structural analysis (see Table 3-9). A more
general overview for all types of elements is presented in Structural Elements, 40. The nonlinear elements for small strain
analysis can handle material or geometric nonlinearities or both. In geometric nonlinear analysis, the elements may undergo
large total displacements and rotations but the net deformation of each element has to remain small, therefore these
elements are called “small strain” elements. For linear material and infinitesimal deformation, a nonlinear analysis with small
strain elements gives results which are identical to a linear analysis. The small strain elements may be combined with any
other type of elements, e.g., hyperelastic elements and/or linear elements.

Table 3-9 Small Strain Elements in Nonlinear Structural Analysis

Element Property Material4
Line Elements
CROD PROD MAT1 (MATT1), MATS1
CONROD -- MAT1 (MATT1)
CTUBE PTUBE MAT1 (MATT1)
CBEAM PBEAM MAT1 (MATT1), MATS1

PBCOMP1 (Composite Beam) (only elastic-perfectly plastic)

2-D Elements -- Plane Strain, Plane Stress, Shells
CQUAD4 PSHELL MAT1 (MATT1), MAT2 (MATT2),
MAT83, MATS1, CREEP
CTRIA3 PCOMP (Composite Shell)2

Main Index
 652 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Table 3-9 Small Strain Elements in Nonlinear Structural Analysis (continued)

Element Property Material4
Solid Elements
CHEXA 8 Node only PSOLID MAT1 (MATT1), MAT9 (MATT9),
MATS1, CREEP
CPENTA 6 Node only
CTETRA 4 & 10 Node

For all elements, geometric nonlinearity is turned on with PARAM, LGDISP,1.

 PBCOMP may refer to MAT1 only (linear material).
 PCOMP must refer to MAT1, MAT2, MAT8 (only linear material).
 MAT8 must not be referred by a MATS1 entry (only linear orthotropic material).
 MAT1linear elastic isotropic material.
MAT2linear elastic anisotropic material for plates.
MAT8linear elastic orthotropic material for plates.
MAT9linear elastic anisotropic material for solids.
MATTitemperature dependency for MAT1, MAT2, MAT9.
MATS1nonlinear material (piecewise linear, elasto-plastic) referred by MAT1, MAT2, MAT9.
CREEPcreep characteristics, referred by MAT1, MAT2, MAT9.

User Interface
Element-related user input is summarized in Table 3-9. The nonlinear element calculations are activated if a nonlinear
material exists (MATS1 or CREEP Bulk Data entries) or if a geometric nonlinear analysis of the whole model is requested
with PARAM,LGDISP,1. The default is geometric linear analysis, PARAM,LGDISP,-1.
For elements with rotational degrees of freedom (CBEAM, CTRIA3, CQUAD4), two different definitions for large rotations
are available, gimbal angles and rotation vector. The definitions are chosen with PARAM,LANGLE,i. The default is the
gimbal angle definition, i = 1. The three gimbal angle components are the three subsequent rotations around the global
axes. The second rotation component refers to the second global axis which has been rotated once by the first rotation. The
third rotation refers to the third global axis which has been rotated twice by the first and second rotation. The rotation
vector definition is chosen with i = 2. The components of the rotation vector refer to the fixed global axes. The direction
of the rotation vector is the principal direction of the 3-D rotation. The magnitude of the rotation vector is the principal
value of the 3-D rotation. The difference of the two angle definitions in the output of the elements appears only for large
3-D rotations.

Output
Element stress and strain output is requested with the STRESS Case Control command. The stress and strain output for
all nonlinear elements is divided into a nonlinear and a linear format. The nonlinear format outputs stresses together with
strains. Compared to the linear format, the nonlinear format provides more information with regard to nonlinear material

Main Index
 Appendix C: Nonlinear Analysis 653
General Description of Nonlinear Capabilities

laws (effective strain, equivalent stress etc.). In the linear format, strain output is not available. The stresses and strains of
the small strain elements refer to the undeformed area and length, respectively. The components of the stresses and strains
are in the deformed element coordinate system which correlates with the elements' rigid body deformation.
Averaged grid point stresses are requested with the GPSTRESS Case Control command. Averaged grid point stresses are
available in nonlinear static analysis but not in nonlinear transient analysis. Element forces are requested with the FORCE
or ELFORCE Case Control command. Element forces per unit length refer to the undeformed length. The components of
the element forces are in the deformed element system. Element forces are available in nonlinear static analysis but not in
nonlinear transient analysis.
Grid point forces are requested with the GPFORCE Case Control command. Grid point forces are not available in
nonlinear structural analysis.
Elforces are not available for solid elements.
Deformations are requested with the DISP Case Control command. The output of displacements and rotations is available
for all types of nonlinear structural analysis.

Corotational Formulation in Geometric Nonlinear Analysis

For small strain elements, a large deformation is split into element rigid body deformation and element net distortion. The
rigid body rotations of the points in the element are approximated by the rotation of the local element triad which follows
the element deformation, as shown in Figure 3-10. For the remaining element net distortions, a linear strain measure is used.
The method is called corotational formulation. The formulation is capable to simulate large total deformations if a fine
mesh is used and the stretches remain small.
In geometric nonlinear analysis with MSC Nastran, the tangent stiffness matrix is calculated in two steps. First, the linear
stiffness or nonlinear material stiffness is calculated. The linear stiffness or nonlinear material stiffness is constructed from
the variation of the stresses in the virtual work. Second, the differential stiffness is calculated. The differential stiffness is
constructed from the variation of the strains in the virtual work. The differential stiffness is also called geometric stiffness.

Main Index
 654 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Undeformed Element

e2

I Deformed Element
0 e1

d2
X0 XI
d
0 d1
b2 x0
uI

I
b1 xI

Figure 3-10 Corotational Concept for Small Strain Elements in Geometric Nonlinear Analysis

Guidelines
Large element net deformations should be avoided. In areas of the structure where large total deformations are expected,
the mesh must be fine enough to keep the element net deformations small. The element net rotation should not exceed 20
degrees and the element should not be stretched by more than 10%. If stretches exceed 20%, it is recommended to use
hyperelastic elements if applicable.

Limitations
 The element net distortions have to remain small.
 Rigid body elements (RBEi, RBAR, RROD entries, etc.) do not rotate in geometric nonlinear analysis.
 Multipoint constraints (MPCs) remain linear, the user defined constraint equations do not change automatically
in geometric nonlinear analysis. In nonlinear static analysis, the user may change the MPCs from subcase to
subcase. Then, the changes in the MPCs are accounted for incrementally.
 Offsets in the CBEAM, CTRIA3 and CQUAD4 elements are not allowed in combination with nonlinear material.
 The results in linear and nonlinear buckling with offsets may be incorrect.

Main Index
 Appendix C: Nonlinear Analysis 655
General Description of Nonlinear Capabilities

 A temperature variation of the nonlinear material constants in plasticity is not available.

 For the CBEAM, the only nonlinear material available is elastic-perfectly plastic.

Fully Nonlinear Hyperelastic Elements

The hyperelastic elements are fully nonlinear finite deformation elements, including large strain and large rotation. The
following elements are available: plane strain (CQUAD4, CQUAD8, CQUAD, CTRIA3, and CTRIA6), solid (CHEXA,
CPENTA, and CTETRA) and axisymmetric elements (CQUADX and CTRIAX). See Nonlinear Elements, 638, Table 3-6 for
more details. The element formulation is total Lagrangean in updated coordinates. See Figure 3-11 for an illustration of the
total Lagrangean concept. A nonlinear strain measure and point wise rotation computed at each Gauss point are employed.
Equilibrium is satisfied in the deformed configuration.

Main Index
 656 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

New Estimate
Last Estimate  ui + 1 
Previous  ui 
Converged Solution

Vi + 1
Vi

u i
u0
ui

V0 u i + 1 = u i + u i

Initial State
xi + 1 = X + ui + 1
y
X

x
Basic Coordinate System

Figure 3-11 Total Lagrangian Concept: Deformations are Measured from an Initial Undeformed State

The elements are especially designed to address volumetric locking and other related problems such as ill-conditioning of
the stiffness matrix, incorrect stresses etc., due to near incompressibility. To this end, a mixed formulation based on a three
field variational principle in which the pressure and the volume ratio are interpolated independently from the displacements
is employed. At the incompressible limit the bulk modulus becomes infinite and the volumetric strain J – 1 becomes zero
so that the pressure may no longer be obtained from the displacements through the constitutive equation. The pressure and
volume ratio interpolations are discontinuous between the elements.

Main Index
 Appendix C: Nonlinear Analysis 657
General Description of Nonlinear Capabilities

The hyperelastic elements are supported in the nonlinear solutions, SOLs 106 and 129. They can be run through SOLs
101 and 105 for model checkout in linear analysis, the results, however, will generally not be correct. For buckling analysis,
PARAM,BUCKLE in SOL 106 should be used instead.

User Interface
Because of no difference in geometry, the hyperelastic elements are defined using the current CQUAD4, CQUAD8,
CTRIA3, CTRIA6 CHEXA, CPENTA, and CTETRA connectivity entries. Additional CQUAD, CQUADX, and CTRIAX
connectivity entries are available for the hyperelastic elements. The PID field of the connectivity entries for the plane or
axisymmetric and for the solid hyperelastic elements references the PLPLANE and PLSOLID property entries, respectively.
The MID field references a MATHP material entry that defines the hyperelastic material. This is the only material model
available for fully nonlinear, finite deformation analysis.

Output Description
The nonlinear stress output for the hyperelastic elements differs from other material nonlinear stress output. The
hyperelastic stress output contains the following quantities, at each Gauss point, in the basic coordinate system for solid
and axisymmetric elements and in the CID coordinate system, defined on a PLPLANE entry, for plane strain elements:
Cauchy stresses, i.e., components of the stress vector t = dP  dA where dA is an infinitesimal area in the deformed
configuration, and P is the force. See Hyperelastic Material, 625 for a description of the constitutive relation.
 Pressure
1
p = ---   x +  y +  z 
3
 Logarithmic strains, i.e., the components of
3

 ln  l N l N lT
l=1

where  l are principal stretches l  l 0 in the principal directions N l . See Figure 3-12. For compatibility with other
MSC Nastran elements, twice the shear components of the logarithmic strain tensor are printed in the output.
N l = principal directions in undeformed configuration; n l = RN l = principal directions in deformed configuration,
obtained from N l by a rigid body rotation. The deformation along N l is a pure stretch, after the rigid body motion has
been factored out.

Main Index
 658 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

n1

N2 N1
n2 dx = ln 1
dX = l 0 N 1
u Deformed
Configuration
N3 n3
Undeformed
Configuration
x
z X

basic
x
Figure 3-12 Deformation Configurations

 Volumetric strain
dV – dV 0
J – 1 = ----------------------
dV 0

where J is an approximation to the determinant of the deformation gradient dx  dX .

This output is obtained with the STRESS (or ELSTRESS) Case Control command. Examples of the new output for the
plane, axisymmetric and solid elements are shown in Listing 3-5.
The linear output contains principal Cauchy stresses and directions with respect to the basic (or CID) coordinate system.
See Listing 3-6.

Main Index
 Appendix C: Nonlinear Analysis 659
General Description of Nonlinear Capabilities

Listing 3-5 Output for Continuum Elements

N O N L I N E A R S T R E S S E S I N H Y P E R E L A S T I C Q U A D R I L A T E R A L E L E M E N T S ( QUADFD )

ELEMENT GRID/ POINT CAUCHY STRESSES/ LOG STRAINS PRESSURE VOL. STRAIN
ID GAUSS ID X Y Z XY
1 GAUS 1 1.678579E+02 -1.188351E-03 5.456201E+00 1.560323E-16 5.777097E+01 1.925699E-02
1.791759E+00 -1.772686E+00 .0 .0
2 1.678579E+02 -1.188351E-03 5.456201E+00 -1.191492E-15 5.777097E+01 1.925699E-02
1.791759E+00 -1.772686E+00 .0 .0
3 1.678579E+02 -1.188351E-03 5.456201E+00 -3.368891E-16 5.777097E+01 1.925699E-02
1.791759E+00 -1.772686E+00 .0 .0
4 1.678579E+02 -1.188351E-03 5.456201E+00 -7.083336E-16 5.777097E+01 1.925699E-02
1.791759E+00 -1.772686E+00 .0 .0
N O N L I N E A R S T R E S S E S IN H Y P E R E L A S T I C A X I S Y M M. Q U A D R I L A T E R A L ELEMENTS (QUADXFD)

ELEMENT GRID/ POINT CAUCHY STRESSES/ LOG STRAINS PRESSURE VOL. STRAIN
ID GAUSS ID RADIAL(X) AXIAL(Y) THETA(Z) XY
1 GAUS 1 1.808360E-09 2.917973E+02 1.808360E-09 4.298727E-15 9.726578E+01 3.242192E-02
-9.420009E-01 1.930910E+00 -9.570014E-01 1.188717E-01
2 1.808132E-09 2.917973E+02 1.808132E-09 -2.820565E-17 9.726578E+01 3.242192E-02
-9.420009E-01 1.930910E+00 -9.570014E-01 1.188717E-01
3 1.808019E-09 2.917973E+02 1.808019E-09 -2.820565E-17 9.726578E+01 3.242192E-02
-9.420009E-01 1.930910E+00 -9.570014E-01 1.188717E-01
4 1.808218E-09 2.917973E+02 1.808218E-09 1.004509E-16 9.726578E+01 3.242192E-02
-9.420009E-01 1.930910E+00 -9.570014E-01 1.188717E-01
N O N L I N E A R S T R E S S E S I N H Y P E R E L A S T I C H E X A H E D R O N E L E M E N T S ( H E X A F D )

ELEMENT GRID / POINT CAUCHY STRESSES/ LOG STRAINS PRESSURE VOL. STRAIN
ID GAUSS ID X Y Z XY YZ ZX
1 GAUS 1 2.9180E+02 1.8047E-09 1.8047E-09 -7.8413E-15 -7.7424E-17 2.3330E-15 9.7266E+01 3.2422E-02
1.9459E+00 -9.5700E-01 -9.5700E-01 .0 .0 .0
2 2.9180E+02 1.8047E-09 1.8047E-09 -3.1028E-16 -1.4850E-16 9.6370E-16 9.7266E+01 3.2422E-02
1.9459E+00 -9.5700E-01 -9.5700E-01 .0 .0 .0
3 2.9180E+02 1.8047E-09 1.8047E-09 2.8235E-15 6.2553E-17 -6.4668E-15 9.7266E+01 3.2422E-02
1.9459E+00 -9.5700E-01 -9.5700E-01 .0 .0 .0
4 2.9180E+02 1.8047E-09 1.8047E-09 -5.9973E-16 9.1812E-17 -2.1878E-15 9.7266E+01 3.2422E-02
1.9459E+00 -9.5700E-01 -9.5700E-01 .0 .0 .0
5 2.9180E+02 1.8047E-09 1.8047E-09 2.3338E-15 -1.2646E-16 2.8235E-15 9.7266E+01 3.2422E-02
1.9459E+00 -9.5700E-01 -9.5700E-01 .0 .0 .0
6 2.9180E+02 1.8047E-09 1.8047E-09 -2.3359E-16 -1.1296E-16 2.1388E-15 9.7266E+01 3.2422E-02
1.9459E+00 -9.5700E-01 -9.5700E-01 .0 .0 .0
7 2.9180E+02 1.8047E-09 1.8047E-09 2.1626E-15 2.1405E-17 -6.2480E-15 9.7266E+01 3.2422E-02
1.9459E+00 -9.5700E-01 -9.5700E-01 .0 .0 .0
8 2.9180E+02 1.8047E-09 1.8047E-09 -4.0477E-16 2.4964E-17 -1.9690E-15 9.7266E+01 3.2422E-02
1.9459E+00 -9.5700E-01 -9.5700E-01 .0 .0 .0
SIMPLE TENSION DECEMBER 1, 1993 MSC/NASTRAN 11/30/93 PAGE 108

Main Index
 660 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Listing 3-6 Output for Continuum Elements in Basic Coordinate System

NONLINEAR
LOAD STEP = 1.00000E+00
S T R E S S E S I N H Y P E R E L A S T I C Q U A D R I L A T E R A L E L E M E N T S ( QUADFD )
ELEMENT GRID / POINT ---------CAUCHY STRESSES-------- PRINCIPAL STRESSES (ZERO SHEAR)
ID GAUSS ID NORMAL-X NORMAL-Y SHEAR-XY ANGLE MAJOR MINOR
1 GAUS 1 1.678579E+02 -1.188351E-03 1.560323E-16 .0000 1.678579E+02 -1.188351E-03
2 1.678579E+02 -1.188351E-03 -1.191492E-15 .0000 1.678579E+02 -1.188351E-03
3 1.678579E+02 -1.188351E-03 -3.368891E-16 .0000 1.678579E+02 -1.188351E-03
4 1.678579E+02 -1.188351E-03 -7.083336E-16 .0000 1.678579E+02 -1.188351E-03
S T R E S S E S I N H Y P E R E L A S T I C A X I S Y M M. Q U A D R I L A T E R A L E L E M E N T S (QUADXFD)
ELEMENT GRID / POINT ---------CAUCHY STRESSES-------- PRINCIPAL STRESSES (ZERO SHEAR)
ID GAUSS ID RADIAL(X) AXIAL(Y) SHEAR-XY ANGLE MAJOR MINOR
1 GAUS 1 1.808360E-09 2.917973E+02 4.298727E-15 90.0000 2.917973E+02 1.808360E-09
2 1.808132E-09 2.917973E+02 -2.820565E-17 -90.0000 2.917973E+02 1.808132E-09
3 1.808019E-09 2.917973E+02 -2.820565E-17 -90.0000 2.917973E+02 1.808019E-09
4 1.808218E-09 2.917973E+02 1.004509E-16 90.0000 2.917973E+02 1.808189E-09
NONLINEAR SUBCASE 100 $ UNIAXI
LOAD STEP = 1.00000E+00
S T R E S S E S I N H Y P E R E L A S T I C H E X A H E D R O N E L E M E N T S ( H E X A F D )
GRID / POINT --------CAUCHY STRESSES--------- DIR. COSINES MEAN
ELEMENT-ID GAUSS ID NORMAL SHEAR PRINCIPAL -A- -B- -C- PRESSURE
1 GAUS
1 X 2.917973E+02 XY -7.841313E-15 A 2.917973E+02 LX 1.00 .0 .0 9.726578E+01
Y 1.804707E-09 YZ -7.742389E-17 B 1.804693E-09 LY .00 .0 .0
Z 1.804707E-09 ZX 2.332978E-15 C 1.804693E-09 LZ .00 .0 .0
2 X 2.917973E+02 XY -3.102808E-16 A 2.917973E+02 LX 1.00 .0 .0 9.726578E+01
Y 1.804707E-09 YZ -1.485038E-16 B 1.804693E-09 LY .00 .0 .0
Z 1.804707E-09 ZX 9.636988E-16 C 1.804693E-09 LZ .00 .0 .0
3 X 2.917973E+02 XY 2.823466E-15 A 2.917973E+02 LX 1.00 .0 .0 9.726578E+01
Y 1.804707E-09 YZ 6.255328E-17 B 1.804707E-09 LY .00 .0 .0
Z 1.804707E-09 ZX -6.466815E-15 C 1.804707E-09 LZ .00 .0 .0
4 X 2.917973E+02 XY -5.997311E-16 A 2.917973E+02 LX 1.00 .0 .0 9.726578E+01
Y 1.804707E-09 YZ 9.181191E-17 B 1.804707E-09 LY .00 .0 .0
Z 1.804707E-09 ZX -2.187819E-15 C 1.804707E-09 LZ .00 .0 .0
5 X 2.917973E+02 XY 2.333811E-15 A 2.917973E+02 LX 1.00 .0 .0 9.726578E+01
Y 1.804707E-09 YZ -1.264558E-16 B 1.804707E-09 LY .00 .0 .0
Z 1.804707E-09 ZX 2.823466E-15 C 1.804707E-09 LZ .00 .0 .0
6 X 2.917973E+02 XY -2.335867E-16 A 2.917973E+02 LX 1.00 .0 .0 9.726578E+01
Y 1.804693E-09 YZ -1.129638E-16 B 1.804694E-09 LY .00 .0 .0
Z 1.804693E-09 ZX 2.138826E-15 C 1.804694E-09 LZ .00 .0 .0
7 X 2.917973E+02 XY 2.162630E-15 A 2.917973E+02 LX 1.00 .0 .0 9.726578E+01
Y 1.804665E-09 YZ 2.140466E-17 B 1.804636E-09 LY .00 .0 .0
Z 1.804665E-09 ZX -6.248027E-15 C 1.804636E-09 LZ .00 .0 .0
8 X 2.917973E+02 XY -4.047673E-16 A 2.917973E+02 LX 1.00 .0 .0 9.726578E+01
Y 1.804651E-09 YZ 2.496398E-17 B 1.804622E-09 LY .00 .0 .0
Z 1.804651E-09 ZX -1.969031E-15 C 1.804622E-09 LZ .00 .0 .0
LOAD STEP = 1.00000E+00

Guidelines
MSC Nastran performs a fully nonlinear analysis, including the effect of large strain and large rotation, if a PLPLANE
and/or PLSOLID Bulk Data entry are encountered. The hyperelastic material is defined by using the MATHP Bulk Data
entry. A nonlinear analysis is always performed except in linear solutions when a hyperelastic material is defined.
It is best that the differential stiffness PARAM,LGDISP,1 be used in all hyperelastic runs, but it is not the default. If the
strains and the rotations are small, then the hyperelastic analysis should compare to the linear elastic analysis. However, for
hyperelastic analysis, it is always best to use PARAM,LGDISP,1.
Due to the independent pressure interpolations, particularly at very high compressive values of the pressure (of the order of
the bulk modulus), indefiniteness of the stiffness matrix may occur, caused by an apparent material instability. By default
in such instances MSC Nastran will ignore the so-called differential stiffness. Although indefiniteness of the stiffness matrix
is probably not associated with a geometric instability, this may indeed help, as many of the terms, which heavily depend
on the pressure, are part of the differential stiffness.

Main Index
 Appendix C: Nonlinear Analysis 661
General Description of Nonlinear Capabilities

Limitations
The fully nonlinear hyperelastic CQUAD, CQUAD4, CQUAD8, CTRIA3, and CTRIA6 are plane strain elements only; they
do not contain bending properties, nor can they be used in plane stress analysis. Therefore, they cannot be subject to out-
of-plane loading conditions and do not sustain a thickness. The elements assume a thickness of 1.0 for mass calculation in
dynamic analyses.
The lower-order fully nonlinear hyperelastic elements CQUAD4, CTRIA3, 8-noded CHEXA, 6-noded CPENTA, and 4-
noded CTETRA are not immune to shear locking and therefore are not suitable for bending type loading. The higher order
elements CQUAD8, 9-noded CQUAD, CTRIA6, 20-noded CHEXA, 15-noded CPENTA, and 10-noded CTETRA should
be used in such situations. See Example 1: Rubber Bushing Problem, 662.
Even though the fully nonlinear elements are defined using the current CQUAD4, CQUAD8, CTRIA3, CTRIA6, CHEXA,
CPENTA, and CTETRA Bulk Data entries, internally, MSC Nastran designates them as the separate elements,
CQUAD4FD, CHEXAFD (for Finite Deformation) etc., because of the different element formulations. This designation is
printed in all the output in which the elements are referenced by name, such as PARAM,EST, PARAM,GPECT, the
sequence processor, and the element stress output.
GPSTRESS and FORCE output is not available for the fully nonlinear hyperelastic elements.
The element types for the fully nonlinear hyperelastic elements are as follows:

Element Element Number in

Element Number OESNL Data Block
CQUAD4 139 201
CQUAD8 (Reverts to CQUAD4 when all midside grids are omitted) 164 208
CQUAD (Reverts to CQUAD4 when center and all midside grids are omitted) 164 208
CTRIA3 162 206
CTRIA6 (Reverts to CTRIA3 when all midside grids are omitted) 167 211
CHEXA with 8 nodes 140 202
CHEXA (Reverts to 8-noded CHEXA when all midside grids are omitted) 163 207
CPENTA with 6 grids 160 204
CPENTA (Reverts to 6-noded CPENTA when all midside grids are omitted) 165 209
CTETRA 161 205
CTETRA (Reverts to 4-noded CTETRA when all midside grids are omitted) 166 210
CQUADX 170 214
CQUADX (Reverts to 4-noded CQUADX when all midside grids are omitted) 171 215
CTRIAX 168 212
CTRIAX (Reverts to 3-noded CTRIAX when all midside grids are omitted) 169 213

Main Index
 662 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Example 1: Rubber Bushing Problem

The cross section of a rubber bushing is shown in Figure 3-13. It is assumed that the frame and internal shaft are rigid and
the rubber bushing is perfectly bonded to these components. The rubber material is assumed to be of the Mooney-Rivlin
type with A 10 = 0.177 N  mm 2 , A 01 = 0.045 N  mm 2 , and D 1 = 333 N  mm 2 . Considering symmetric
conditions, only one-half of the rubber bushing is modeled with 72 finite deformation 4-noded CQUAD elements. The same
problem is also modeled with 72 hyperelastic 9-noded CQUAD elements. The grid points on the outer boundary are fully
constrained to simulate the rubber/frame interface. For the grid points on the inner boundary, only the horizontal degrees
of freedom were constrained, while the vertical degrees of freedom were tied together with MPCs. Force is applied to the
top grid point on the inner wall in the vertical direction.

Frame

Rubber

Shaft


15 mm

30 mm
Figure 3-13 Rubber Bushing

The goal of the analysis is to determine the force-displacement curve of the unit and compare it with the solution in the
reference. (Reference: Sussman, T., and Bathe, K. J., “A Finite Element Formulation for Nonlinear Incompressible Elastic
and Inelastic Analysis,” Computers and Structures (26), pages 357-409, 1987.)
Annotated Bulk Data input of interest is shown in Listing 3-7.

Main Index
 Appendix C: Nonlinear Analysis 663
General Description of Nonlinear Capabilities

Listing 3-7 Input for Rubber Bushing Example

ID MSC,BUSHING $
SOL 106
TIME 180
CEND
TITLE = RUBBER BUSHING BUSHING
SUBTITLE = HYPERELASTIC MODEL TEST
ECHO=UNSORT
DISP(PLOT)=ALL
STRESS(PLOT)=ALL
SUBCASE 100
MPC = 100
NLPARM = 100
LOAD = 100 $
.
.
.
SUBCASE 800
MPC = 100
NLPARM = 800
LOAD = 800 $
BEGIN BULK
$ PARAMETERS
PARAM,DBDROPT,0
PARAM,POST,0
PARAM,LGDISP,1
$
NLPARM1005ITER1NO+NL101
+NL101-3+NL102
.
.
.
NLPARM8005ITER1NO+NL801
+NL801-3+NL802
$ DEFINE GEOMETRY
CORD2C10.0.0.0.0.1.+COR1
+CORD1.0.0.
GRDSET3456
GRID 101 115.0-90.0.13456
GRID 102 117.5-90.0.13456
GRID 103 120.0 -90.0.13456
.
.
.
$ DEFINE ELEMENTS
CQUAD4 1011101102202201
CQUAD4 1021102103203202
CQUAD4 1031103104204203
.
.
.
$ PROPERTY ENTRY FOR 4-NODE PLANE STRAIN ELEMENT WITH MIXED ELEMENT FORMULATION
PLPLANE 1 1
$ HYPERELASTIC MATERIAL - MOONEY RIVLIN TYPE
MATHP 1 0.177 0.045 333.3
$ DEFINE LOADING CONDITIONS
FORCE 1001301-50.0.1.0.
.
.
.
FORCE 8001301-400.0.1.0.
$ DEFINE CONSTRAINTS
MPC10010121.13012-1.
.
.
.
ENDDATA

The resulting force-displacement curve at grid point 1301 is shown in Figure 3-14 and the original and deformed models at
a force of 800N are shown in Figure 3-15. The solution obtained with 4-noded CQUAD elements is overly stiff for a force
greater than 200N (shear locking). The shear locking problem is addressed with the higher order 9-noded CQUAD elements.

Main Index
 664 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Figure 3-14 Force-Displacement Curve of a Rubber Bushing

Figure 3-15 Initial and Deformed Shapes of a Rubber Bushing

The model file for this example is bushing.dat, which can be found in the tpl directory.

Main Index
 Appendix C: Nonlinear Analysis 665
General Description of Nonlinear Capabilities

Example 2: Pressure of a Circular Plate

A circular plate is 15 in. in diameter and 0.5 in. thick. It is simply supported along the edge and is subjected to a uniform
follower pressure. The plate is modeled with solid (CHEXA and CPENTA) and with axisymmetric elements. The solid
model represents a 10° segment. Material constants are A 10 = 80 psi , A 01 = 20 psi , D 1 = 50,000 psi .

Annotated Bulk Data input of interest for the wedge model is shown in Listing 3-8 and for the axisymmetric model is shown
in Listing 3-9.

Listing 3-8 Input of Circular Plate Wedge Model

BEGIN BULK
$ DEFINE PARAMETERS
PARAM,LGDISP,1
PARAM,DBDRNL,-1
PARAM,POST,0
NLPARM 10 10 ITER 1 100 +NL10
+NL10 -3
.
.
.
$ DEFINE GEOMETRY
CORD2C 1 0. 0. 0. 0. 0. 1. +COR1
+COR1 1. 0. 0.
GRDSET 2456
GRID 101 1 0. -5. -0.25 1 12456
GRID 102 1 0.75 -5. -0.25 1
GRID 103 1 1.50 -5. -0.25 1
.
.
.
$ DEFINE ELEMENTS
CHEXA 102 1 102 103 1103 1102 202 203 +HX102
+HX102 1203 1202
CHEXA 103 1 103 104 1104 1103 203 204 +HX103
+HX103 1204 1203
CHEXA 104 1 104 105 1105 1104 204 205 +HX104
+HX104 1205 1204
.
.
.
CPENTA 101 1 101 102 1102 201 202 1202
CPENTA 201 1 201 202 1202 301 302 1302
CPENTA 301 1 301 302 1302 401 402 1402
CPENTA 401 1 401 402 1402 501 502 1502
.
.
.
$ DEFINE ELEMENT AND MATERIAL PROPERTIES
PLSOLID 1 1
MATHP 1 80. 20. 5.+4
$ DEFINE LOADING CONDITIONS
LOAD 3 1. 3. 1
LOAD 15 1. 15. 1
.
.
.
LOAD 45 1. 45. 1
PLOAD4 1 101 1. 101
PLOAD4 1 102 1. 102 1103
.
.
.
$ DEFINE CONSTRAINTS
MPC 10 1102 1 1. 102 1 -1.
MPC 10 1102 3 1. 102 3 -1.
.
.
.
ENDDATA

Main Index
 666 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Listing 3-9 Input for Axisymmetric Model

$ INFLATION OF A CIRCULAR PLATE
$ AXISYMMETRIC QUAD4 ELEMENTS
BEGIN BULK
$ DEFINE PARAMETERS
PARAM,LGDISP,1
PARAM,POST,0
NLPARM 10 10 AUTO1 +NL10
+NL10 -3
.
.
.
$ DEFINE GEOMETRY
GRDSET 3456
GRID 101 0. -0.25 0. 13456
GRID 102 0.75 -0.25 0.
GRID 103 1.50 -0.25 0.
.
.
.
$ DEFINE ELEMENTS
CQUADX 101 1 101 102 202 201
CQUADX 102 1 102 103 203 202
CQUADX 103 1 103 104 204 203
.
.
.
$ DEFINE ELEMENT AND MATERIAL PROPERTIES
PLPLANE1 1
MATHP 1 80. 20. 5.+4
$ DEFINE LOADING CONDITIONS
PLOADX1 3 101 3. 101 102
PLOADX1 3 102 3. 102 103
.
.
.
PLOADX1 45 110 45. 110 111
ENDDATA

Original and deformed models are shown in Figure 3-16 and Figure 3-17.

Main Index
 Appendix C: Nonlinear Analysis 667
General Description of Nonlinear Capabilities

Figure 3-16 Deformed Shape of Circular Plate Under Pressure — CHEXA and CPENTA Hyperelastic Elements

Figure 3-17 Deformed Shape of Circular Plate Under Pressure — Axisymmetric Hyperelastic Elements

The model files for this example are prplth.dat and prpltg.dat, which can be found in the examples directory.

Bushing Elements
The CBUSH element is a generalized spring and damper. Stiffness and damping can be specified for six degrees of freedom,
three translations and three rotations. In SOL 106 and 129, the user can specify six nonlinear force-deflection curves instead
of combined stiffness values. The properties of the CBUSH element are defined on the PBUSH and PBUSHi entries. The
PBUSHT entry refers to TABLEDi entries for force-deflection curves. The CBUSH element is geometric linear only. The
circulation is not updated.
The CBUSH1D element is a rod-type spring and damper, it is a 1D version of the CBUSH element. The CBUSH1D
element can be geometric and material nonlinear. General nonlinear functions of the axial forces versus axial displacement
and for velocity are defined with the PBUSH1D entry. The CBUSH1D element is the only element which can model
rotation damping. The effect is demonstrated in the following example.
The CBUSH and CBUSH1D elements are recommended over the CELASi, CVISC, and CDAMP elements. The bushing
elements always insure rigid body invariance and they offer nonlinear capabilities.

Swinging Pendulum
The solution to a swinging pendulum is calculated using SOL 129. The pendulum has a length of 1.0 [m]. and has a
concentrated mass of M = 1,000.0 [kg] at the free end, see Figure 3-18. We start the analysis with the pendulum at rest in
horizontal position. The free end is loaded with a gravity induced load of G = 10,000.0 [N]. The leg of the pendulum is

Main Index
 668 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

very stiff, K = 1.e+7 [N/m], so that the relative axial deformation is small compared to the overall motion of the pendulum.
Large deformation effects are turned on with PARAM,LGDISP,1.

1.0 m

M = 1,000 kg
x

z
G = 10,000 N
G

G
Figure 3-18 Swinging Pendulum

Main Index
 Appendix C: Nonlinear Analysis 669
General Description of Nonlinear Capabilities

Listing 3-10 MSC Nastran Input File for the Swinging Pendulum
$
$ Name: ar29src1
$
$ A swinging pendulum is modeled. The pendulum is loaded with
$ gravity load. The analysis goes through one and a half cycles,
$ starting in horizontal position. Large deformation is turned on.
$
$ The pendulum is modeled with two separate systems.
$ 1. CROD and CVISC,CROD rotates, CVISC stays in horizontal
$ orientation.
$ 2. CBUSH1D with SPRING and DAMPER, spring and damper rotate.
$
TIME 10 $
SOL 129 $
CEND
$
TITLE= ar29src1 swinging elastic pendulum
SUBTITLE= demonstrate rotating damping
.
.
.
$
BEGIN BULK
$
param, lgdisp, 1
$
$ GEOMETRY
$
.
.
.
$ CBUSH1D spring and rotating viscous damper
$
$CBUSH1D,EID,GA,GB,CID
cbush1d, 108, 208, 41, 42
$PBUSH1D,PID,K,C
pbush1d, 208, 1.e+7, 1000.
$
$ $ $ $ $ $ $ $ $ $
$ SOLUTION STRATEGY
$
.
.
.
$
ENDDATA

We investigate two different models of the pendulum in one run. In the first model, we use a Rod element for the stiffness
of the leg, K = 1.e+7 [N/m], and a Visc element for 5% equivalent viscous damping, C = 1,000.0 [Ns/m]. he x- and z-
displacements of the free end are shown in Figure 3-19. The Rod rotates while the viscous damper Visc stays fixed in its initial
horizontal position. The Visc element is linear and does not follow large deformations. The displacements are damped out
because the viscous damper is acting in the x-direction during the whole motion.

Main Index
 670 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Figure 3-19 Displacement of Swinging Pendulum, Model with Rod/Visc Element Versus Model with BUSH1D Element

In the second model, we use a BUSH1D element. The element has a linear stiffness (K) and a viscous damper (C) with the
same values as in the first model. In the BUSH1D element, the spring and damper rotate. The relative axial displacements
and velocities are small because of the high axial stiffness of the element. The rotating damper has no noticeable effect on
the overall motion of the pendulum because it damps only the small relative deformations. The x-and z-displacements of
the free end are not damped, see Figure 3-19.

Main Index
 Appendix C: Nonlinear Analysis 671
General Description of Nonlinear Capabilities

The example demonstrates how the answers can change if dampers rotate with the deformation compared to dampers which
stay fixed in space.

Adaptive Gap Element

The gap element is used in MSC Nastran to model surface contact problems. The adaptivity is based on the subincremental
scheme with the capability to update the stiffness at the proper timing, bisect the increment when it is excessive, and adjust
the penalty stiffness to the proper values. Generally, in an non-adaptive gap element, the axial stiffness has to be properly
chosen in order to reach convergence. However, the main goal of the self-adaptive method is to relieve users from this trial
and error method and to achieve convergence on the first try. Due to the adaptive nature of the algorithm, the adaptive gap
method costs more in CPU time when compared to the nonadaptive gap element. However, the adaptive gap method allows
a user to choose a wider latitude of axial stiffness and yet still achieve convergence. Therefore, with more experience, the
user can decrease CPU time by fine-tuning this axial stiffness.

Usage
The gap element is used in nonlinear static analysis (SOL 106) and nonlinear transient analysis (SOL 129). The gap
element provides point-to-point contact. When the gap element is open, there is no contact and no friction. When the gap
element is closed, there are three different conditions. The first case is when the gap is sliding (no friction). The second case
occurs when the gap element is sticking (static friction). The third case occurs when the gap element is slipping (kinetic
friction). Sometimes, it may be difficult to converge because of the switching among these conditions.

Limitations
One of the contact surfaces should not rotate by a large angle because the gap element orientation is not updated for large
rotations.

Input
How to Use the CGAP and PGAP Bulk Data Entries
Often a mistake is made in the orientation of the gap element, which is defined through G0 or X1, X2, and X3. If the grid
points GA and GB are not coincident, then line AB is the element x-axis and the orientation vector lies in the X-Y plane.
The use of coincident grid points is not recommended.
On the PGAP Bulk Data entry, U0 and F0 are the initial gap opening and the preload, respectively. U0 is the separation of
the gap element. If you are unsure about what preload to use, use the default value of zero. KA, KB, and KT are called the
penalty values. KA is the axial stiffness when the gap is closed, KB is the axial stiffness when the gap is open, and the KT is
the transverse gap stiffness. All you need to input is the KA value. KB is defaulted to be zero, and KT is defaulted as a
function of KA. The selection of KA or a penalty value is critical to the convergence of the problem. The nonadaptive gap
does not update this KA value. The new adaptive gap updates this KA value if you specify a positive value for TMAX. The
default option for the new adaptive gap is NO adaptive stiffness update. This is when the TMAX value is equal to 0.0. If
TMAX is a negative number, then MSC Nastran uses the nonadaptive gap formulation. A good initial KA selection is
roughly 1000 times the Young's modulus of the material that is in the neighborhood of the gap element.

Main Index
 672 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Fields 8 and 9 also require some explanation. For the new adaptive value this is MU1, which is the static coefficient of
friction, and MU2, which is the dynamic coefficient of friction. The nonadaptive gap elements use MUY and MUZ in the
same fields instead of MU1 and MU2 values. Therefore, to compare the results using adaptive and nonadaptive gap
elements, set both of these entries to the same values. On most problems, you will not know how to estimate the penetration
depth or field 2 on the continuation entry. Therefore, if you skip the continuation entry, the default new adaptive gap is
obtained, which is everything adaptive except the KA update.
The recommended allowable penetration of TMAX is 10 percent of the thickness for the shell element, and for the solid
element it is roughly 1/10,000 of the characteristic length for the problem. A small positive value will turn on the adaptive
stiffness update, which will then converge at a faster rate.
The maximum adjustment ratio or MAR in field 3 of the continuation entry is used only for the penalty value adjustment
of the adaptive gap element. The default value of 100 is sufficient for most problems.

Output
The gap element output is obtained by the STRESS output request in the Case Control. An example of the gap output
format is shown in Listing 3-11. The load step is 1.95 on the top left-hand corner. The digit on the left-hand side of the
decimal indicates that subcase 1 is completed. The digits on the right-hand side of the decimal point indicate that 95
percent of subcase 2 is completed. The far left column lists the gap element numbers. The second column provides the
forces in the axial direction of the gap element. The third and the fourth columns represent forces in the Y and Z direction.
These forces should be zero if the gap is in the open position. Columns 5 to 9 represent the displacements of the gap in its
element coordinate system. The last column indicates the status of the gap. This status can be “open”, “stick”, or “slip”.
Sticking is when the ratio of (shear force)/coefficient of friction * normal force) = SR is such that 0.<SR<1. Slip occurs when
SR = 1.0, open is when the gap is not in contact; hence SR = 0.

Main Index
 Appendix C: Nonlinear Analysis 673
General Description of Nonlinear Capabilities

Listing 3-11 Sample of CGAP Output

NONLINEAR STATIC CONTACT OF A SPHERE ON A RIGID NG6603 NOVEMBER 10, 1992 MSC/NASTRAN 11/ 9/92 PAGE 177
USE TMAX=0.0016,FULLY ADAPTIVE FEATURE
KA=1.0E6 SUBCASE 2
LOAD STEP = 1.92500E+00
S T R E S S E S ( F O R C E S ) I N G A P E L E M E N T S ( C G A P )

ELEMENT - F O R C E S I N E L E M S Y S T - - D I S P L A C E M E N T S I N E L E M S Y S T -
ID COMP-X SHEAR-Y SHEAR-Z AXIAL-U TOTAL-V TOTAL-W SLIP-V SLIP-W STATUS
2001 3.60382E+00 -4.47466E-01 0.0 2.23604E-02 -8.30369E-04 0.0 -6.81213E-04 0.0 STICK
2003 5.04829E+00 -1.10082E+00 0.0 8.85048E-02 -7.80763E-03 0.0 -7.44069E-03 0.0 STICK
2005 0.0 0.0 0.0 1.95640E-01 -1.70540E-02 0.0 -1.70540E-02 0.0 OPEN
2007 0.0 0.0 0.0 2.23109E-01 -9.91112E-03 0.0 -9.91112E-03 0.0 OPEN
2009 0.0 0.0 0.0 2.40853E-01 -4.95583E-03 0.0 -4.95583E-03 0.0 OPEN
NONLINEAR STATIC CONTACT OF A SPHERE ON A RIGID NG6603 NOVEMBER 10, 1992 MSC/NASTRAN 11/ 9/92 PAGE 178
USE TMAX=0.0016,FULLY ADAPTIVE FEATURE
KA=1.0E6 SUBCASE 2
LOAD STEP = 1.95000E+00
S T R E S S E S ( F O R C E S ) I N G A P E L E M E N T S ( C G A P )

ELEMENT - F O R C E S I N E L E M S Y S T - - D I S P L A C E M E N T S I N E L E M S Y S T -
ID COMP-X SHEAR-Y SHEAR-Z AXIAL-U TOTAL-V TOTAL-W SLIP-V SLIP-W STATUS
2001 3.64258E+00 -4.47838E-01 0.0 2.23643E-02 -8.30493E-04 0.0 -6.81213E-04 0.0 STICK
2003 5.15982E+00 -1.12016E+00 0.0 8.85160E-02 -7.81408E-03 0.0 -7.44069E-03 0.0 STICK
2005 8.83100E-02 -5.82848E-03 0.0 1.97000E-01 -1.73532E-02 0.0 -1.73532E-02 0.0 STICK
2007 0.0 0.0 0.0 2.25831E-01 -1.01834E-02 0.0 -1.01834E-02 0.0 OPEN
2009 0.0 0.0 0.0 2.44078E-01 -5.09924E-03 0.0 -5.09924E-03 0.0 OPEN
NONLINEAR STATIC CONTACT OF A SPHERE ON A RIGID NG6603 NOVEMBER 10, 1992 MSC/NASTRAN 11/ 9/92 PAGE 179
USE TMAX=0.0016,FULLY ADAPTIVE FEATURE
KA=1.0E6 SUBCASE 2
LOAD STEP = 1.97500E+00
S T R E S S E S ( F O R C E S ) I N G A P E L E M E N T S ( C G A P )

ELEMENT - F O R C E S I N E L E M S Y S T - - D I S P L A C E M E N T S I N E L E M S Y S T -
ID COMP-X SHEAR-Y SHEAR-Z AXIAL-U TOTAL-V TOTAL-W SLIP-V SLIP-W STATUS
2001 3.66578E+00 -4.47205E-01 0.0 2.23666E-02 -8.30282E-04 0.0 -6.81213E-04 0.0 STICK
2003 5.22790E+00 -1.12097E+00 0.0 8.85228E-02 -7.81435E-03 0.0 -7.44069E-03 0.0 STICK
2005 2.37324E-01 -3.59185E-02 0.0 1.97024E-01 -1.73652E-02 0.0 -1.73532E-02 0.0 STICK
2007 0.0 0.0 0.0 2.27630E-01 -1.03924E-02 0.0 -1.03924E-02 0.0 OPEN
2009 0.0 0.0 0.0 2.46395E-01 -5.21857E-03 0.0 -5.21857E-03 0.0 OPEN
NONLINEAR STATIC CONTACT OF A SPHERE ON A RIGID NG6603 NOVEMBER 10, 1992 MSC/NASTRAN 11/ 9/92 PAGE 180
USE TMAX=0.0016,FULLY ADAPTIVE FEATURE
KA=1.0E6 SUBCASE 2
LOAD STEP = 2.00000E+00
S T R E S S E S ( F O R C E S ) I N G A P E L E M E N T S ( C G A P )

ELEMENT - F O R C E S I N E L E M S Y S T - - D I S P L A C E M E N T S I N E L E M S Y S T -
ID COMP-X SHEAR-Y SHEAR-Z AXIAL-U TOTAL-V TOTAL-W SLIP-V SLIP-W STATUS
2001 3.68888E+00 -4.46565E-01 0.0 2.23689E-02 -8.30068E-04 0.0 -6.81213E-04 0.0 STICK
2003 5.29569E+00 -1.12167E+00 0.0 8.85296E-02 -7.81458E-03 0.0 -7.44069E-03 0.0 STICK
2005 3.86731E-01 -6.51561E-02 0.0 1.97039E-01 -1.73749E-02 0.0 -1.73532E-02 0.0 STICK
2007 0.0 0.0 0.0 2.29423E-01 -1.06010E-02 0.0 -1.06010E-02 0.0 OPEN
2009 0.0 0.0 0.0 2.48707E-01 -5.33770E-03 0.0 -5.33770E-03 0.0 OPEN

Example: Hertz Contact Stress

Hertz contact stress is a well known problem arising in the contact of a wheel and a rail. To guard against the possibility of
surface failure, it is necessary to calculate the stress that results from loading one body against another. The gap element is
used to model the point-to-point contact phenomenon. In this case, the gap element represents the contact surface. The
Hertz contact stress problem considered here is a nonlinear static analysis using SOL 106.
Modeling Techniques
For the static analysis, a symmetric half of a 10-degree sector of the sphere is modeled by CHEXA and CPENTA elements
with axisymmetric boundary conditions as shown in Figure 3-20. Contact with a rigid plane is simulated by adaptive gap
elements.

Main Index
 674 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Figure 3-20 Hertz Contact Stress

Uniform vertical displacement of the symmetric plane is maintained by constraining the vertical movement of all the nodes
on the top plane by MPC relations at the center line. The radius of the sphere is 8 inches, Young's modulus is 1000 psi,
and Poisson's ratio is 0.3. The static analysis is performed in 17 increments up to the maximum loading. The characteristic
length of the problem is the diameter of the sphere, i.e., 16 inches. Therefore, the TMAX value is 16/10000, or 0.0016.
The initial KA or axial gap stiffness is selected to be 1000 times Young's modulus. These values are inserted in the PGAP
entry as shown in Figure 3-21.

1 2 3 4 5 6 7 8 9 10
$ EID PID GA GB X1 X2 X3
CGAP 2001 1 105 301 1. 0. 0.

$ PID U0 F0 KA KB KT MU1 MU2

PGAP 1 .022 1.+6 .3 .3 +
$ TMAX MAR TRMIN
+ 0.0016

Figure 3-21 Partial Input Data

Main Index
 Appendix C: Nonlinear Analysis 675
General Description of Nonlinear Capabilities

Discussion of Results
The objective of this example is to demonstrate that the adaptive features of the gap element allows the user to use a wider
range of KA values, and still get good results. Figure 3-22 shows that the if the adaptive features of gap element are used, good
results are obtained for KA values ranging from 1.0E1 to 1.0E6. If adaptive features are not used, good results are obtained
for KA values between1.0E3 and 1.0E5. For KA values higher than 1.0E5 nonadaptive gap element failed to converge, and
below 1.0E3, the accuracy decreased.

Figure 3-22 Comparison of Axial Stiffness Values

3-D Slideline Contact Capability

Slide line contact allows the user to model interactions between two deformable bodies. One of the deformable bodies is
called the primary and the other is called the secondary. The modeling of interaction requires the user to define contact
regions in terms of slide lines. The user may specify as many slide lines as desired.

Main Index
 676 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

A slide line contact region consists of a primary line and a secondary line as shown in Figure 3-23. A primary line is a list of
grid points in the topological order on the primary body. A secondary line is a list of grid points in the topological order on
the secondary body. The grid points on the primary line are called the primary nodes, and on the secondary line are called
the secondary nodes. A line segment joining two consecutive primary nodes is called a primary segment. Thus a primary
line, in general, consists of number of primary segments. A minimum of one primary segment consisting of two primary
nodes are required for the primary line. Similar to primary line, a secondary line, in general, consist of a number of
secondary segments. However, a secondary line may not have any secondary segments; it may have just one secondary node.

k – 1 Secondary Segment
Secondary Line
k
k+1 k–1 k – 1 Secondary Segment Normal
l–1 l-th Primary Segment Normal
l+1 Primary Line
y Normals Must Face
Each Other
l-th Primary Segment
x

z Slideline Plane Vector Direction

 Arrows show positive direction for ordering nodes. Counterclockwise from primary line to
secondary line.
 X-Y plane is the slideline plane. Unit normal in the Z-direction is the slideline plane vector.
 Secondary and primary segment normals must face each other; i.e., the normals must be in the
opposite directions.

Figure 3-23 Typical Slideline Contact Region

The primary and the secondary nodes in a contact region must lie in a plane called the slide line plane. The slide line plane
is defined by a vector normal to the plane called the slide line plane vector. The primary and secondary can have large
relative motions within the slide line plane. However, relative motion outside the slide line plane are ignored, therefore,
must be small compared to a typical primary segment. Thus, slide line contact is ideal for modelling interactions between
two bodies which may come in contact in specific planes. MSC Nastran checks to make sure that in the initial geometry
all the nodes of a contact region lie in the slide line plane. However, no such check is made during analysis.
In general, contact is determined only for secondary nodes and the primary line. This may result in primary nodes
penetrating the secondary line. However, the error involved depends only on the mesh discretization. At the expense of
increase in CPU cost, a user may wish to determine contact between primary nodes and the secondary line in addition to

Main Index
 Appendix C: Nonlinear Analysis 677
General Description of Nonlinear Capabilities

determining contact between secondary nodes and the primary line. This is called symmetric penetration. Thus, for
symmetric penetration, no distinction is made between secondary and primary.
To determine contact between the secondary nodes and the primary line, a local coordinate system is created for each
primary segment. The tangent direction for the primary segment is given by the vector from its first node to the second
node. Remember that the primary nodes are specified in the topological order. The normal direction to the primary
segment is determined by the cross product between the slide line plane vector and the tangent vector. This normal
direction must point from primary line to the secondary line. This is because MSC Nastran determines the contact between
secondary nodes and the primary line by measuring how close the secondary nodes are to the primary segment in this
normal direction.
For symmetric penetration, similar to the primary segments, tangent and normal directions are also defined for secondary
segments. Now, the normals of primary segments and secondary segments must face each other as shown in Figure 3-23. This
is generally accomplished by traversing form primary line to secondary line in a counter-clockwise fashion, or clockwise
fashion depending on whether the slideline plane vector forms a right hand or a left hand coordinate system with the slide
line plane. It is the user's responsibility to make sure that the normals face each other. MSC Nastran checks indirectly to
make sure that the normals face each other by calculating the initial gap between the two bodies. However, this is not a
foolproof check, especially when there is no initial clearance.
Initial penetration between the two bodies is not allowed. The initial gaps between the secondary nodes and the primary
line are calculated based on the coordinates specified for the secondary and the primary nodes. Thus, if there is no initial
gap, MSC Nastran may calculate initial penetration if the coordinates are not specified accurately. To avoid having the user
to calculate the coordinates very accurately, MSC Nastran adjusts the coordinates automatically and issues a warning
message if the initial penetration of a secondary node is less than ten percent of the length of the primary segment to which
it projects. However, if the initial penetration is larger then ten percent, the analysis is terminated and a fatal message is
issued. Sometimes, the analysis may be terminated with this message even though there is a large initial gap between the
two bodies.This implies that the normals of primary or secondary segments are not pointing in the correct direction.
Once a contact between a node and a body is detected, a three node slide line element as shown in Figure 3-24 is
automatically created. The first node of this element is the secondary node and the last two nodes are the primary nodes of
the primary segment with which the secondary node is in contact. The secondary node is forced to slide on the primary
line after contact and must remain on the primary line until a tensile force develops.

Main Index
 678 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

x2 – x1
6
5
n
m2

2
gn
4
S1
t 3
m1

a x2 – x1
gt

a 0 x2 – x1

 S, m 1 , , m 2 = secondary, primary node 1 and primary node 2, respectively

 a, a 0 = current and previous surface coordinate

 g n = penetration of secondary node into the primary segment

 g t = sliding of the secondary node on the primary segment

Figure 3-24 Geometry and Definition of a Typical Three Node Slide Line Element

In MSC Nastran the contact and traction forces/stresses are always associated with secondary nodes. In order to compute
the contact and traction stresses, an area is associated with each secondary node, based on the contributory length and the
width/thickness from the adjacent secondary segments. Therefore, the user needs to specify widths associated with

Main Index
 Appendix C: Nonlinear Analysis 679
General Description of Nonlinear Capabilities

secondary line in a three dimensional model or thickness associated with secondary line in a two dimensional model. A user
can request to output results for any number of secondary nodes in any number of slide line contact regions. The output
for a secondary node consists of (a) the slide line contact region identification number, (b) the primary segment to which
it projects, (c) the parametric surface coordinate to identify the exact projection of secondary node relative to the two
primary nodes of the primary segment, (d) contact and tangential forces and stresses in the element coordinate system, and
(e) a slip ratio to indicate whether the secondary node is sticking, slipping (or sliding). The slip ratio is defined as the (shear
force)/(coefficient of friction  normal force). Thus, a ratio of 0. indicates no shear or no contact. A slip ratio of 1. indicates
slipping. A value between 0. and 1. is a normalized value which indicates how close the value of the shear force is to
impending slipping.
The slide line contact formulation is based on the penalty method which uses an artificial stiffness value. However, the user
need not provide penalty values. MSC Nastran will automatically calculate the appropriate numbers. The user has the
option to scale the calculated penalty values. It is possible that the automatically calculated penalty values, may sometimes,
cause convergence difficulties. However, a user can overcome the convergence problems by restarting the analysis from any
previous converged solution and scaling down the penalty values. Generally, it is recommended to scale down the penalty
values by one order of magnitude at a time.

Diagnostic Output
Diagnostic output related to 3-D slide line contact can be obtained by setting the executive control statement DIAG to 35.

DIAG 35 Prints current coordinates of contact node and other information associated with secondary nodes to debug
slide line contact algorithm for each iteration for slide line contact.

Limitations
The slide line capability is only available for SOL 106 and SOL 129. For all other solution sequences all the Bulk Data
entries and Case Control commands related to slide line contact are ignored.
Contact stress or force for any secondary node could be plotted using XYPLOT BOUT RESP command. The grid ID is
the secondary node external grid ID. However, the secondary node requested for plotting must also be requested for
printout via the BOUTPUT Case Control command.
The user interface for slideline contact consists of one Case Control command, five Bulk Data entries and one PARAMeter.
The Case Control command is BOUTPUT. The Bulk Data entries are BCONP, BLSEG, BWIDTH, BFRIC, and
BOUTPUT. The parameter is ADPCON.

Example
The purpose of this example is to demonstrate how to use the features of slide line contact capability in MSC Nastran.
An elastic sphere in contact with a rigid plane is pressed down with a uniform pressure at the top. No friction is assumed
between the sphere and the rigid plane. The analytical solution to this problem were first published by Timoshenko, S. P.,
and Goodier, J. N., Theory of Elasticity, McGraw-Hill Book Company, 1970, pgs. 409-420. For demonstration purposes, a
symmetric half of a ten degree sector of the sphere is modeled by CHEXA and CPENTA elements with axisymmetric
boundary conditions (see Figure 3-25). Uniform vertical displacements of the symmetry plane is maintained by tying the

Main Index
 680 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

vertical displacements of all the grid points on the top plane (loading plane) by MPC relations to the one at the center line
(z-axis). A total load of 2160 is applied at the top of sphere. Because only a ten degree sector is modeled and because of the
MPC relationships, a load of only 60 is applied at the top center grid point. This load is applied in three subcases with five
load increments per subcase.

Figure 3-25 Model of a Sphere in Contact with Rigid Plane

To model the contact, two slide line regions are defined via the BCONP Bulk Data entry. The first slide line contact region
models the contact between the outer grid points of the sphere at  = 0 and the rigid plane as shown in Figure 3-26. The
outer grid points of the sphere (148, 116, 114, 112, 110, 108, 105) are used to define the secondary line. To define the
primary line a grid point 313 at  = 0 is defined on the rigid plane. The primary line is defined by the lower most center
grid point of the sphere (grid point 106) and the grid point 313. Notice that the grid points on the primary and secondary
are defined in topological order traversing in counter-clockwise form the primary line to the secondary line (see BCONP
with ID = 10, BLSEG with ID's = 10 and 20 in the example input). This is because the slide line plane is the R-Z plane at

Main Index
 Appendix C: Nonlinear Analysis 681
General Description of Nonlinear Capabilities

 = 0 and the slide line plane vector forms a right hand system with the slide line plane (see CORD2R with ID = 1 in
the example listing).

148

116
ts
(counter-clockwise
114 traverse)
ns

112 (normals face

each other)
110
nm
108
105

106 tm 313
Y

X

Z (slide line plane vector direction)

Figure 3-26 Slide Line Contact Region One

The second slide line contact region models the contact between the outer grid points of the sphere at  = 10 and the
rigid plane as shown in Figure 3-27. Another grid point 314 at  = 10 on the rigid plane is defined to model the primary
line. This time we traverse clockwise form the primary line to the secondary line in topological order (see BCONP with ID
= 20 and BLSEG with ID's = 30 and 40 in the example input) because the slide line plane vector forms a left hand system
with the slide line plane (see CORD2R with ID = 3 in the example input).

Main Index
 682 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

208

180

178

176 (clockwise
traverse)
174
ts
172
170 (normal face
each other)

106 tm 314

Y
Z
(slide line plane vector direction)

X
Figure 3-27 Slide Line Contact Region Two

The most important issue is to make sure that the normals of primary and secondary line point towards each other. The
normal to a segment is obtained by e 3  t . e 3 is the unit vector in the z-direction of the coordinate system used in field
CID of BCONP Bulk Data entry. t is the tangential vector form node 1 to node 2 of the line in the order specified in the
BLSEG Bulk Data entry.

Main Index
 Appendix C: Nonlinear Analysis 683
General Description of Nonlinear Capabilities

Listing 3-12 Slideline Example Problem of Sphere Contacting a Flat Plate

TIME 300 $
SOL 106 $
CEND
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
$ HERTZIAN CONTACT OF A SPHERE ON A FLAT SURFACE $
$ TWO UNSYMMETRICAL SLIDELINES $
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
TITLE=NONLINEAR STATIC CONTACT OF A SPHERE ON A RIGID PLANE WITHOUT FRICTION
SET 333 = 103,117,124,131,138,151,152,
105,108,110,112,114,116,148,149,150,168,167
SET 444 = 106, 301 THRU 314
SET 555 = 9001, 9002
$
$ use BOUTPUT to request output for contact regions
$
BOUTPUT(SORT1) = ALL
$
DISP = 333
SPCF = 444
STRESS = 555
SPC = 4
MPC = 100
SUBCASE 1 $ 20 LBS
NLPARM = 410
LOAD = 100
SUBCASE 2 $ 40 LBS
NLPARM = 420
LOAD = 200
SUBCASE 3 $ 60 LBS
NLPARM = 430
LOAD = 300
OUTPUT(XYPLOT)
XTITLE = LOAD AFCTOR
YTITLE = CONTACT NORMALFORCE
XYPLOT BOUTPUT RESP/108(6)
XYPLOT BOUTPUT RESP/105(6)
YTITILE BOUTPUT ELEMENTS NORMAL STRESS - X
XYPLOT BOUTPUT RESP/9001(6),9002(6)
BEGIN BULK
$
$ SLIDELlNE CONTACT related bulk data entries
$
BCONP,1,10,20,, ,,1,1
BCONP,2,30,40,, ,,1,3
BLSEG,10,148,116,114,112,110,108,105 $
BLSEG,20,106,313 $
BLSEG,30,170,172,174,176,178,180,208 $
BLSEG,40,314,106 $
$
$ COORDINATE SYSTEM FOR SLIDE LINE PLANE VECTOR
$
CORD2R,1,2,0.,.0,.0,1.,-90.,.0,+C2
+C2,1.,0.,0.0
CORD2R,3,2,0.,0.0,.0,1.,100.,.0,+C3
+C3,1.,0.,.0

Main Index
 684 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Output Formats
SORT1 Output
NONLINEAR STATIC CONTACT OF A SPHERE ON A RIGID PLANE WITHOUT FRIC MARCH 1, 1993 MSC/NASTRAN 2/17/93 PAGE 109

SUBCASE 3
LOAD STEP = 3.00000E+00
R E S U L T S F O R S L I D E L I N E E L E M E N T S (IN ELEMENT SYSTEM)

SCND CONTAC PRIMARY SURFACE NORMAL SHEAR NORMAL SHEAR NORMAL SLIP SLIP SLIP
GRID ID GRID1 GRID2 CORDINATE FORCE FORCE STRESS STRESS GAP RATIO CODE
148 1 106 313 9.7604E-01 .0 .0 .0 .0 -1.1240E+00 1.5218E-02 .0 OPEN
116 1 106 313 7.3917E-01 .0 .0 .0 .0 -4.1923E-01 -4.1647E-03 .0 OPEN
114 1 106 313 5.7432E-01 .0 .0 .0 .0 -1.1861E-01 -2.3407E-02 .0 OPEN
112 1 106 313 4.5907E-01 2.4318E+00 .0 4.1069E+00 .0 1.2792E-04 -3.9671E-02 .0 SLIDE
110 1 106 313 3.4298E-01 1.1404E+01 .0 1.8499E+01 .0 5.9989E-04 -4.8093E-02 .0 SLIDE
108 1 106 313 2.1905E-01 1.0069E+01 .0 1.6320E+01 .0 5.2968E-04 -3.5747E-02 .0 SLIDE
105 1 106 313 1.0478E-01 5.2337E+00 .0 1.7661E+01 .0 2.7531E-04 -1.8078E-02 .0 SLIDE
170 2 314 106 8.9522E-01 5.2335E+00 .0 1.7660E+01 .0 2.7530E-04 1.8078E-02 .0 SLIDE
172 2 314 106 7.8095E-01 1.0069E+01 .0 1.6320E+01 .0 5.2968E-04 3.5747E-02 .0 SLIDE
174 2 314 106 6.5702E-01 1.1404E+01 .0 1.8499E+01 .0 5.9989E-04 4.8093E-02 .0 SLIDE
176 2 314 106 5.4093E-01 2.4318E+00 .0 4.1069E+00 .0 1.2792E-04 3.9671E-02 .0 SLIDE
178 2 314 106 4.2568E-01 .0 .0 .0 .0 -1.1861E-01 2.3407E-02 .0 OPEN
180 2 314 106 2.6083E-01 .0 .0 .0 .0 -4.1923E-01 4.1647E-03 .0 OPEN
208 2 314 106 2.3956E-02 .0 .0 .0 .0 -1.1240E+00 -1.5218E-02 .0 OPEN

SORT2 Output
NONLINEAR STATIC CONTACT OF A SPHERE ON A RIGID PLANE WITHOUT FRIC MARCH 1, 1993 MSC/NASTRAN 2/17/93 PAGE 99

SUBCASE 1
SCND GRID-ID = 110
R E S U L T S F O R S L I D E L I N E E L E M E N T S (IN ELEMENT SYSTEM)

STEP CONTAC PRIMARY SURFACE NORMAL SHEAR NORMAL SHEAR NORMAL SLIP SLIP SLIP
ID GRID1 GRID2 CORDINATE FORCE FORCE STRESS STRESS GAP RATIO CODE
2.000E-01 1 106 313 3.5166E-01 .0 .0 .0 .0 -1.0336E-01 -4.6982E-03 .0 OPEN
4.000E-01 1 106 313 3.5067E-01 .0 .0 .0 .0 -7.1160E-02 -9.6436E-03 .0 OPEN
6.000E-01 1 106 313 3.4968E-01 .0 .0 .0 .0 -4.3237E-02 -1.4611E-02 .0 OPEN
8.000E-01 1 106 313 3.4868E-01 .0 .0 .0 .0 -1.5314E-02 -1.9579E-02 .0 OPEN
1.000E+00 1 106 313 3.4787E-01 5.6225E-01 .0 9.1206E-01 .0 2.9576E-05 -2.3653E-02 .0 SLIDE
1.200E+00 1 106 313 3.4727E-01 1.8074E+00 .0 2.9318E+00 .0 9.5072E-05 -2.6643E-02 .0 SLIDE
1.400E+00 1 106 313 3.4667E-01 3.0525E+00 .0 4.9516E+00 .0 1.6057E-04 -2.9632E-02 .0 SLIDE
1.600E+00 1 106 313 3.4608E-01 4.2976E+00 .0 6.9713E+00 .0 2.2606E-04 -3.2622E-02 .0 SLIDE
1.800E+00 1 106 313 3.4548E-01 5.5427E+00 .0 8.9911E+00 .0 2.9156E-04 -3.5612E-02 .0 SLIDE
2.000E+00 1 106 313 3.4488E-01 6.7878E+00 .0 1.1011E+01 .0 3.5706E-04 -3.8601E-02 .0 SLIDE
2.200E+00 1 106 313 3.4428E-01 8.0329E+00 .0 1.3031E+01 .0 4.2255E-04 -4.1591E-02 .0 SLIDE
2.400E+00 1 106 313 3.4368E-01 9.2780E+00 .0 1.5050E+01 .0 4.8805E-04 -4.4580E-02 .0 SLIDE
2.600E+00 1 106 313 3.4322E-01 1.0329E+01 .0 1.6756E+01 .0 5.4336E-04 -4.6913E-02 .0 SLIDE
2.800E+00 1 106 313 3.4310E-01 1.0867E+01 .0 1.7628E+01 .0 5.7162E-04 -4.7503E-02 .0 SLIDE
3.000E+00 1 106 313 3.4298E-01 1.1404E+01 .0 1.8499E+01 .0 5.9989E-04 -4.8093E-02 .0 SLIDE

Nonlinear Static Analysis

User Interface
Nonlinear static analysis can be performed using SOLs 106 or 153. Modeling options are compatible with linear analysis;
that is, a SOL 101 model can be converted to SOL 106 analysis with a few additional data entries relevant to the nonlinear
analysis. The nonlinear properties and/or effects are defined by nonlinear material data (for example, MATS1, MATHP,
CREEP, and TABLES1), gap elements (GAP), slideline contact (BCONP, BLSEG and BFRIC) for nonlinear interface, and
PARAMeter LGDISP for geometric nonlinearity.

Main Index
 Appendix C: Nonlinear Analysis 685
General Description of Nonlinear Capabilities

Case Control
Each subcase defines a set of loading and boundary conditions which can be subdivided into a number of increments.
Loading conditions are specified using the same loads as in linear static analysis; that is, FORCEi, MOMENTi, PLOADi,
GRAV, TEMP, etc. Constraints can be specified using SPCs and MPCs, which can be changed from subcase to subcase.
The enforced motion can be applied using SPCD or nonzero value in SPC. The specified value for SPCD or SPC is a
measure of displacement with reference to the undeformed position. The enforced motion is also subdivided in the subcase
like the incremental loads. The unique data required for SOL 106 is supplied on the NLPARM entry, which controls the
incremental and iterative solution processes. The snap-through analysis or post-buckling analysis can be performed using
arc-length methods which are controlled by parameters in the NLPCI Bulk Data entry.
The load step is labeled by the cumulative load factor. The load factor varies from 0 to 1 in each subcase. Then the load
step will end with 1, 2, 3, etc. for the first, the second and the third subcase, respectively. There are advantages to divide the
entire loading history into many subcases so that each subcase does not have excessive number of increments; for example,
not to exceed 20 steps. The data blocks containing solutions can be generated at each increment or the end of each subcase
depending on the intermediate output option specified in the INTOUT field of NLPARM data entry. Data blocks are
stored in the database for output process and restarts. As such, the database size or the output quantities are better controlled
with multiple subcases. Furthermore, the nonproportional loading can only be achieved by using multiple subcases.
Output requests for each subcase are processed independently. Requested output quantities for all the subcases are appended
after the computational process for actual output operation. Available outputs are DISPLACEMENT, OLOAD, STRESS,
FORCE, GPSTRESS, and SPCFORCE. However, FORCE and GPSTRESS output are available only for linear elements.
It should be noted that the STRESS or STRESS(PLOT) command will cause the nonlinear stress to be printed. In other
words, the STRESS(PLOT) command suppresses the combined linear/nonlinear stress output only.
Restarts are controlled by parameters (LOOPID and SUBID) which can be specified either in the Case Control Section or
Bulk Data Section. Some optional parameters (ADPCON, BUCKLE, LANGLE, MAXLP, NMLOOP) are provided for
additional control or capabilities. All the superelement model generation options and matrix reduction options are allowed
for the linear portion of the structure.

Restriction and Limitations in SOL 106

 Only CROD, CONROD, CTUBE, CBEAM, CGAP, CQUAD4, CQUAD8, CQUADX, CTRIA3, CTRIA6,
CTRIAX, CHEXA, CPENTA, and CTETRA elements may be defined with material or geometric nonlinear
properties. All other elements will be treated as having small displacements and linear materials.
 The CBEAM is assumed to have its nonlinear region concentrated at the ends. Taper is allowed but pin joints are
illegal for material nonlinearity. However, pin joints can be used if the only nonlinearity is geometric.
 Constraints apply only to the nonrotated displacements at a grid point. In particular, multipoint constraints and
rigid elements may cause problems if the connected grid points undergo large motions. However, also note that
replacement of the constraints with overly stiff elements can result in convergence problems.
 Large deformations of the elements may cause nonequilibrium loading effects. The elements are assumed to have
constant length, area, and volume, except for hyperelastic elements.
 All restarts with grid point or element changes must start over at the first subcase.

Main Index
 686 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

 For large displacement analysis, there are two different approaches for the angular motions: gimbal angle and
rotation vector approach. In Gimbal angle approach angular motions are treated as three ordered rotations about
the z, y, and x axes. In rotation vector approach the three angular motions are treated as a vector. The rotation is
about the rotation axis and the magnitude of rotation is equal to the amplitude of rotation vector.
 The CGAP element contact planes do not rotate as a function of displacement. The actual shapes of the contacting
surfaces would be required for such calculations.
 Since the solution to a particular load involves a nonlinear search procedure, solution is not guaranteed. are must
be used in selecting the search procedures on the NLPARM data. Nearly all iteration control restrictions may be
overridden by the user.
 Follower force effects are calculated for loads which change direction with grid point motion.
 All cold-start runs are assumed to start in a stress-free condition with zero displacements and rotations. Prestressed
and preloaded cases must be simulated by running them as the initial load case, followed by the desired solution
subcase.
 Every subcase must define a new total load or enforced boundary displacement. This is necessary because error
ratios are measured with respect to the load change.
 All superelements, except for the residual (SE = 0), are assumed to be linear. All existing superelement restrictions
also apply here.
 Any changes to a nonresidual superelement requires the use of the appropriate SEMG, SEMA, SELA, etc., case
control request for efficient restarts (see also Nonlinear Analysis With Superelements, 732).
 Only the residual superelement has automatic output data processing. Upstream superelements may be loaded
only via LSEQ and CLOAD data.
 In postbuckling analysis with arc-length methods (NLPCI Bulk Data entry), the iteration algorithm may jump
between different buckling modes when multiple bifurcation points are present in the structure. It is
recommended to check all possible buckling modes with the linear buckling solution sequence SOL 105 before
making a run with SOL 106. The user has to introduce initial imperfections (loads or enforced displacements) to
keep the structure in the intended buckling mode.

Iteration Control: NLPARM Entry

The most crucial data for successful nonlinear static solutions are contained in the NLPARM entry, which defines strategies
for the incremental and iterative solution processes. It is difficult to choose the optimal combination of all the options for
a specific problem. However, based on a considerable number of numerical experiments, the default option was intended
to provide the best workable method for a general class of problems. Therefore, users with little insight or experience in a
specific application should start with the default option. However, users with some experience and insight in a specific
problem may change the default values based on the following observations:
 Computing cost for each line search is comparable to that of an iteration.
 The SEMI method usually provides better convergence than the AUTO method at the expense of higher
computing cost.
 Default tolerances for the convergence criteria may be somewhat conservative. However, loose tolerances may
cause difficulties in the subsequent steps.

Main Index
 Appendix C: Nonlinear Analysis 687
General Description of Nonlinear Capabilities

 The quasi-Newton method is effective in most problems. However, it seems to have adverse effects in some
problems; for example, creep analysis.
 The line search method is effective to cope with difficulties in convergence in some problems. More extensive line
searches may be exercised by a large value of MAXLS and/or a smaller value of LSTOL. On the other hand, line
searches may have adverse effects in some problems; for example, plane stress plasticity.
 The arc-length method should be used if the problem involves snap-through or postbuckling deformation. Then
the Bulk Data entry NLPCI must be attached.
The NLPARM entry format is shown below with default values wherever applicable:

1 2 3 4 5 6 7 8 9 10
$NLPARM ID NINC DT KMETHO KSTEP MAXITER CONV INTOUT
D
NLPARM 0.0 AUTO 5 25 PW NO

$ EPSU EPSP EPSW MAXDIV MAXQN MAXLS FSTRESS LSTOL

$ MAXBIS MAXR RTOLB

5 20. 20.

1.0E-3 1.0E-3 1.0E-7 3 MAXITER 4 0.2 0.5

The ID field is referenced by the NLPARM Case Control command. The NINC field is an integer which specifies the
number of increments to be processed in the subcase. The total load specified in the subcase minus the load specified in the
preceding subcase is equally divided by this integer (NINC) to obtain the incremental load for the current subcase. Another
subcase should be defined to change constraints or loading paths. However, multiple subcases may be required in the
absence of any changes in constraints or loads to use a moderate value (for example, not to exceed 20) for NINC. Use of a
moderate value has advantages in controlling database size, output size, and restarts.
The DT field requires a real number specifying the time increment for each load step in the case of creep analysis. The time
unit must be consistent with the unit used in the CREEP data to define the creep characteristics. The creep time increment
should be determined properly for efficiency and accuracy based on the creep rate represented by the creep law to be used.
The larger the creep rate, the smaller the time increment should be. No creep deformation is considered with the default
value of 0. It is noted that a creep analysis should be preceded by a static analysis with at least one subcase, because the creep
deformation requires stresses in the material.
Stiffness matrix update strategies are determined by a combination of the data specified in the two fields KMETHOD and
KSTEP. Options for KMETHOD are AUTO, SEMI, or ITER. The KSTEP field, which is an auxiliary to the KMETHOD
field, should have an integer equal to or greater than 1. With the AUTO option, the program automatically determines
when to reevaluate the stiffness matrix based on the rate of convergence. At each iteration, the computing time for
convergence without the stiffness matrix update is estimated and compared with the computing time for the matrix update
in order to determine whether the update is more efficient. This decision is deferred in the first two iterations after a new
stiffness is obtained. If the solution tends to diverge, however, the update decision will be made effective immediately. The
stiffness matrix will be updated upon convergence if the number of iterations required for convergence is greater than

Main Index
 688 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

KSTEP. The SEMI option is identical to the AUTO option except for one additional stiffness update after the first iteration
which always occurs unless the solution converges in a single iteration. With the ITER option, the stiffness matrix is
updated at every KSTEP iterations. Thus, the full Newton-Raphson iteration is exercised if KSTEP is 1. If KSTEP >
MAXITER, the stiffness will never be updated.
The MAXITER field is an integer representing the number of iterations allowed for each load increment. If the number of
iterations reaches MAXITER without convergence, the load increment is bisected and the analysis is repeated. If the load
increment cannot be bisected (that is, MAXBIS is reached or MAXBIS = 0) and MAXDIV is positive, the best attainable
solution is computed and the analysis is continued to the next load increment. If MAXDIV is negative, the analysis is
terminated.
The convergence test is performed at every iteration with the criteria specified in the CONV field. Any combination of U
(for displacement), P (for load), and W (for work) may be specified. All the specified criteria must be satisfied to achieve
convergence, except for an absolute convergence condition, under which the solution is converged regardless of criteria.
This subject will be elaborated in the following section. The convergence tolerances are specified in the fields EPSU, EPSP,
and EPSW for U, P, and W criteria, respectively.
The defaults for the nonlinear convergence tolerances have been modified to improve run times for most large nonlinear
static analyses without appreciably sacrificing accuracy. The new default tolerances also reflect values similar to those used
in other popular nonlinear FEA programs such as Marc. Run times of large nonlinear static models with plasticity and/or
large deformation models will usually improve significantly. Run times for models with gaps or line contact may also
improve, but not as significantly. At this time, it appears that static nonlinear heat transfer models and most nonlinear
dynamic analysis models continue to require the present tight default tolerances.
Default nonlinear tolerances in Solutions 106 and 153 will vary according to the model, analysis type, and user-defined
precision. Although default tolerances can still be overridden by entering values of CONV, EPSU, EPSP, and EPSW on the
NLPARM entry, default tolerances can now be easily invoked with a single parameter, leaving the above fields on the
NLPARM entry blank. Depending on the desired level of accuracy and run times, various default tolerance sets may be
selected. These tolerances are designated as “very high,” “high,” “engineering design,” and “preliminary design” and can now
be controlled with the new parameter
PARAM,NLTOL,ITOL
where ITOL is an integer value ranging from 0 (very high) to 3 (preliminary design). The default is 2 (engineering accuracy)
for nonlinear static analysis and 0 (very high) for nonlinear heat transfer and dynamic analyses. See Parameter Descriptions
(p. 794) in the MSC Nastran Quick Reference Guide, for a complete description.
Tables 3-10, 3-11, and 3-12, illustrate the tolerances that will be set by the program if the CONV, EPSU, EPSP, and EPSW
fields on the NLPARM entry are left blank.

Table 3-10 Default Tolerances for Static Nonlinear SOL 106 Models Without Gaps, Contact or Heat Transfer
NLTOL Designation CONV EPSU EPSP EPSW
0 Very high PW _______ 1.0E-3 1.0E-7
1 High PW _______ 1.0E-2 1.0E-3

Main Index
 Appendix C: Nonlinear Analysis 689
General Description of Nonlinear Capabilities

Table 3-10 Default Tolerances for Static Nonlinear SOL 106 Models Without Gaps, Contact or Heat Transfer
NLTOL Designation CONV EPSU EPSP EPSW
2 Engineering PW _______ 1.0E-2 1.0E-2
3 Prelim Design PW _______ 1.0E-1 1.0E-1
none Engineering PW _______ 1.0E-2 1.0E-2

Table 3-11 Default Tolerances for Static Nonlinear SOL 106 Models With Gaps or Contact (Enter NLTOL
Values of 0 or 2 Only or Omit the Parameter)
NLTOL Designation CONV EPSU EPSP EPSW
0 Very high PW _______ 1.0E-3 1.0E-7
2 Engineering PW _______ 1.0E-3 1.0E-5
None Engineering PW _______ 1.0E-3 1.0E-5

Table 3-12 Default Tolerances for Static Nonlinear SOL 106 or 153 Models With Heat Transfer (Enter NLTOL
Value of 0 Only or Omit the Parameter)
NLTOL Designation CONV EPSU EPSP EPSW
0 Very high PW _______ 1.0E-3 1.0E-7
None Very high PW _______ 1.0E-3 1.0E-7

To use the new default tolerances, the CONV, EPSU, EPSP, and EPSW fields of the NLPARM entry must be left blank.
In addition, the NINC field must also either be blank.
As in the past, if the user has a good idea of what to use for the default tolerances and/or number of increments, then these
values should be used. If the problem is nearly linear, a value of NINC=1 sometimes yields acceptable results.
Because nonlinear tolerances now depend on NLTOL, NINC (on the NLPARM entry), and other possible NLPARM
overrides, convergence tolerances actually used by the code are now echoed in the output file (right after the Bulk Data
echo, if any) as shown in the following example:

DEFAULT NONLINEAR TOLERANCES

NLPARM ID 10
NINC 10
EPSU 1.00000E-03 (not normally used)
EPSP 1.00000E-03
EPSW 1.00000E-05
BASED ON
MODEL HAS GAPS AND/OR LINE CONTACT
(VALUES YOU ENTER OVERRIDE THE ABOVE DEFAULTS)

Main Index
 690 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

Examples of Convergence Tolerance Changes

A very large nonlinear model consisting of approximately one million degrees of freedom was studied extensively to
determine the effect of various tolerances on run time and accuracy. This particular model did not employ heat transfer,
gaps, or contact. Results of the study showed that compared to the “very high accuracy” solution, acceptable accuracy can
be achieved for both the “high accuracy” and “engineering accuracy” designations. The “preliminary design” accuracy
tolerances produces results with the same general trends as other reasonable tolerances, but the answers were considered too
inexact for analyses other than preliminary design trade off studies. The following table summarizes the effect of changing
the tolerances for this large model.

Tolerances
Run time CPU
Accuracy Type EPSP EPSW hours Accuracy for the Large Model
Very high 0.001 1.0E-7 163.2 Very Good
High 0.01 0.001 10.3 5% or less accurate
Engineering Design 0.01 0.01 7.0 10% or less accurate
Preliminary Design 0.1 0.1 2.9 10% to 20% less accurate

The INTOUT field requires a specification of YES, NO or ALL. If YES is specified, the output requests
(DISPLACEMENT, FORCE, STRESS, etc.), specified in the Case Control Data, will be processed for every computed
converged solution in the subcase. If NO is specified, the output requests will be processed only for the last load step of the
subcase. If ALL is specified, the output requests are processed for every computed and user specified load increment. For
Newton's iteration methods (that is, without NLPCI), the option ALL is equivalent to the option YES, since the computed
load increment is equal to the user specified load increment. For arc-length methods (that is, NLPCI command is specified),
the computed load increment in general is not the same as the user specified load increment, and is not known in advance.
The option ALL allows the user to obtain solutions at the desired intermediate load increments. With a small database
option, the database will retain only those data blocks corresponding to the incremental load steps for which output is
requested by an INTOUT field. Then, the restarts will be restricted to those starting from these steps. However, if the job
is terminated with an incomplete subcase for some reason (for example, diverging, time expiration, etc.), data blocks for all
the interim steps of that subcase are stored in the database to allow flexible restarts.
The MAXDIV field requires an integer to specify a limit on the probable divergence conditions allowed for each iteration
to continue. There are two classifications in the divergence condition: probable and absolute. The absolute divergence is
treated as two occurrences of the probable divergence. When the probable divergence occurs MAXDIV times or more and
if MAXDIV is positive, the current solution base (displacements) is retracted and the stiffness matrix is updated in reference
to the preceding iteration step. If the solution diverges again in the same load increment while MAXDIV is positive, the
best attainable solution is computed and the analysis is continued to the next load increment. If MAXDIV is negative, the
analysis is terminated on second divergence.
The MAXQN field requires an integer to specify the maximum number of quasi-Newton vectors to be saved in the
database. Quasi-Newton vectors are accumulated, if desired, until MAXQN is reached. If the number of accumulated QN
vectors is greater than or close to MAXQN upon convergence, the stiffness matrix will be updated before the next increment

Main Index
 Appendix C: Nonlinear Analysis 691
General Description of Nonlinear Capabilities

is processed. All the QN vectors already accumulated are purged upon a stiffness update and the accumulation is reinitiated.
The quasi-Newton update may be suppressed with a value of 0 for MAXQN. Details of the quasi-Newton method will be
presented in the following section.
The line search is controlled by the fields MAXLS and LSTOL. The maximum number of line searches allowed for each
iteration is specified in the MAXLS field. The line search process may be suppressed with a value of 0 for MAXLS. The
LSTOL field requires a real number between 0.01 and 0.9 to specify a tolerance for the line search operation. The line search
operation will be conducted if the error defining the divergence rate is greater than LSTOL. If the line search convergence
is not achieved, the line search continues until the number of line searches reaches MAXLS.

The FSTRESS field requires a real number (0.0 < FSTRESS < 1.0), representing a fraction of the effective stress (  ) used
to limit the subincrement size in the material routines. The number of subincrements in the material routines for elasto-
plastic or creep deformation processes is determined such that a subincrement in the effective stress is approximately
FSTRESS   . This data is also used to establish an error tolerance in the yield function to alleviate accumulation and
propagation of the error. If the error in the yield function exceeds FSTRESS   at the converging state, the program will
terminate the job with a fatal error message “ERROR EXCEEDED 20% OF CURRENT YIELD STRESS.” If the
bisection option is selected, bisection will be activated under this condition.
The MAXBIS field requires an integer to specify the number of bisections allowed for a load or arc-length increment. When
the solution diverges, different actions are taken depending on the sign of MAXBIS. If MAXBIS is positive, the stiffness
matrix is updated on the first divergence and the load increment is bisected on the second divergence. If MAXBIS is
negative, the load increment is bisected every time the solution diverges until the limit on bisection is reached. If the
solution does not converge after |MAXBIS| bisections, the analysis is continued or terminated depending on the sign of
MAXDIV.
The MAXR field requires a real number to specify the bounds on the ratio of the adjusted arc-length increment relative to
the initial value. In the adaptive load/arc-length increment, the overall upper and lower bounds on the load/arc-length
increment in the subcase are defined as

1 - l n
----------------  --------  MAXR
MAXR l 0

where l n is the arc-length at step n and l 0 is the original arc-length. The arc-length method for load increments is
selected by the NLPCI Bulk Data entry, which must have the same ID as the NLPARM Bulk Data entry.
The RTOLB field requires a real number to specify the incremental rotation (in degrees) allowed per iteration. The
bisection is activated if the incremental rotation for any degree of freedom (  x ,  y or  z ) exceeds the value specified
for RTOLB. This bisection strategy based on the incremental rotation is controlled by the MAXBIS field.

User Interface for Arc-Length Methods: NLPCI Entry

Most of the parameters in the NLPARM Bulk Data entry are used to control the iteration strategy for arc-length methods.
However, MAXLS is not applicable because the line search procedure is not coupled with arc-length methods currently.
The parameters which are applicable only to the arc-length methods are specified in the Bulk Data entry NLPCI, in

Main Index
 692 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

connection with the NLPARM data by the same ID. The NLPCI entry is shown below with default values, followed by
brief description of the fields:

1 2 3 4 5 6 7 8 9 10
NLPCI ID TYPE MINALR MAXALR SCALE DESITER MXINC
NLPCI CRIS 0.25 4. 0.0 12

ID Must be identical to ID of NLPARM to be active.

TYPE Constraint type (CRIS for Crisfield, RIKS for Riks, or MRIKS for modified
Riks method)
MINALR, MAXALR For variable arc-length method, the arc-length is bounded by
l new
MINALR  --------------  MAXALR
l old
where MINALR  1. and MAXALR  1 . The automatic adjustment is
not activated if MAXALR = MINALR = 1.
SCALE Scaling factor for the magnitude of the load term in arc length. Real > 0 Default
= 0. The constraint equation has a disparity in the dimension by mixing the
displacements with the load factor. The scaling factor ( w ) is introduced as user
input so that the user can make constraint equation unit-independent by a
proper scaling of the load factor,  . As the value of w is increased, the
constraint equation is gradually dominated by the load term. In the limiting
case of infinite w , the arc-length method would be degenerated to the
conventional Newton's method.
DESITER Desired number of iterations for convergence to be used for adaptive arc-length
adjustment.

l new = l old Id
----------
I max
MXINC Maximum number of controlled increments allowed within the specified load
step. There can be cases where the structure never snaps back and collapses to
failure. In these cases, the analysis cannot be completed without bounding the
maximum number of increments. For such cases, this value should be confined
to a reasonable limit, e.g., less than 100, to prevent inadvertent depletion of the
computing resources.

Iteration Related Output Data

At the end of every iteration, the relevant data from the iteration process are printed under the following heading. The
section numbers in parentheses refer to sections in the MSC Nastran Handbook for Nonlinear Analysis.

Main Index
 Appendix C: Nonlinear Analysis 693
General Description of Nonlinear Capabilities

ITERATION iteration count i.

EUI relative error in terms of displacements (see Section 3.6.3).
EPI relative error in terms of loads (see Section 3.6.3).
EWI relative error in terms of work (see Section 3.6.3).
LAMBDA rate of convergence (defined in Section 3.3.1).
DLMAG absolute norm of the load error vector.
FACTOR final value of the line search parameter (see Section 3.4).
E-FIRST divergence rate, initial error before line search (see Section 3.4.3).
E-FINAL error at the end of line search (see Section 3.4.3).
N-QNV number of quasi-Newton vectors appended (see Section 3.5.3).
N-LS number of line searches performed during the iteration.
ENIC expected number of iterations for convergence (see Section 3.3.1).
NDV number of occurrences of probable divergence during the iteration (see Section 3.3.2).
MDV number of occurrences of bisection conditions in the material routine (excessive stress
increment) or in the rotation angle (excessive rotation) during the iteration.

Loads Overview
The static loads in nonlinear analysis consist of a subset of the total MSC Nastran static load set. Most of the relevant loads
data applicable to the linear static analysis are also applicable to nonlinear static analysis except for the Bulk Data entry
DEFORM. The enforced displacements can be applied as loads using the Bulk Data entry SPCD, which defines an
enforced motion of a grid in a specific fixed direction. Each grid point with an enforced displacement must also appear on
an SPC or SPC1 entry.
The loads are selected in each subcase by a Case Control command LOAD. The load for a subcase is subdivided into the
number of increments specified for the subcase. The solution strategy in nonlinear is to apply the loads in an incremental
fashion until the desired load level is reached. The algorithms “remember” the loads from one subcase to the next. If the
load reaches the desired level in a subcase and if the load description is left out in the subsequent subcase inadvertently, the
nonlinear solution algorithm will begin in an incremental fashion to remove the load. The loads described below make up
the valid subset applicable for nonlinear static analysis.

Concentrated Loads
The concentrated loads consist of two categories: those which are stationary in direction through out the analysis and those
which follow the grid motion. The first category consists of the following: FORCE, MOMENT, and SLOAD.
The second category consists of the following: FORCE1, FORCE2, MOMENT1, and MOMENT2. The three FORCE entries
differ only in the manner in which the direction of the force is specified. FORCE uses the components of a vector. FORCE1
uses two grid points, not necessarily the same as the loaded grid points. FORCE2 defines the direction of the force as a

Main Index
 694 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

vector which is the vector product of two other vectors. The distinctions between the three MOMENT entries are similar.
These loads are follower loads in nonlinear and produce follower stiffness terms.

Distributed Loads
The distributed loads consist of two categories: those which are stationary in direction through out the analysis and those
which follow the element motion. The first category consists of the PLOAD1.
The second category consists of the following: PLOAD, PLOAD2, and PLOAD4.
Loads may be applied to the interiors of finite elements. The PLOAD1 entry defines concentrated and linearly distributed
forces and moments to the CBEAM element. However, the stress calculation for nonlinear CBEAM element ignores the
distributed loads, but lumps them onto the nodal points. The PLOAD entry is exceptional in that it references three or four
arbitrarily located grids rather than an element. The PLOAD2 entry defines a uniform static pressure load applied to the
CQUAD4 or CTRIA3 elements. The PLOAD4 defines a pressure load on surfaces of CHEXA, CPENTA, CTRIA3, and
CQUAD4 elements which need not be normal to the surface. When normal to the surface, the PLOAD4 is a follower load.

Mass Related Loads

The mass related loads are gravity and centrifugal loads, which require specification of mass data. The gravity loads which
can be specified in the GRAV Bulk Data entry are stationary in direction through out the analysis. It is used to define the
direction and magnitude of the gravity vector in some identified coordinate system. The components of the gravity vector
are multiplied by the mass matrix to obtain the components of the gravity force at each grid point.
The centrifugal loads to be specified in the RFORCE Bulk Data entry is used to define the components of a spin vector
which is used internally to compute centrifugal forces. Each component of the spin vector is multiplied by a scale factor.

Thermal Loads
Thermal loads are selected in each subcase by the Case Control command TEMP(LOAD). Thermal loading which is a
follower loading is accomplished through the following entries: TEMP, TEMPD, TEMPP1, TEMPP3, and TEMPRB.
The TEMP and TEMPD provide for grid point temperature specification which are then interpolated to points within
elements. For the nonlinear CQUAD4 and CTRIA3 elements, the TEMPP1 provides for an average temperature and
thermal gradient in the thickness direction and the TEMPP3 provides for temperature at points through the thickness. The
TEMPRB provides for average temperature at ends A and B for the CROD and CBEAM element. Average temperature
specified directly for an element will take precedence over the temperature interpolated from the elements connected grid
points.
The Case Control TEMP(INIT) is required above the subcase level. For nonlinear static analysis TEMP(MAT) or
TEMP(BOTH) are not applicable. The use of TEMP(MAT) or TEMP(BOTH) will cause a fatal error. Subcases that do
not contain a TEMP(LOAD) will default to TEMP(INIT), which implies no thermal load. Thermal loads are often used
to provide initial strains.

Loads Combination
Various load specifications can be combined using LOAD and (CLOAD, LSEQ).
The LOAD provides for the linear combination of static load sets. The static loads (LOAD, FORCE, etc.) applied to the
upstream superelements cannot be referenced by a Case Control command LOAD. The CLOAD entry is designed to apply

Main Index
 Appendix C: Nonlinear Analysis 695
General Description of Nonlinear Capabilities

static loads to upstream superelements by combining loads defined in LSEQ entries. The (CLOAD, LSEQ) combination
allows the nonlinear algorithms to apply in an incremental fashion on the upstream superelement loads to the boundary of
the residual structure.
The CLOAD defines a static load combination for superelement loads acting on residual boundary and the LSEQ defines
a static load combination for superelement upstream loads. The LSEQ assigns load vectors to the superelements and
numerically labels them via the DAREA field entry value. The LID and TID field entries point to the appropriate load and
temperature distribution entries. The CLOAD defines a static load as a linear combination of previously calculated
superelement loads defined by the LSEQ. The IDVi fields correspond directly to the numeric labels defined via the DAREA
fields of the LSEQ entries.
The CLOAD must be selected in the residual solution subcase via the case control CLOAD entry. Only one LOADSET
may appear in Case Control. The hierarchy of the loads data is shown schematically in Figure 3-28. An example Case Control
Section is shown below:
.
.
.
SEALL=ALL $ Required in SOL 106 only

LOADSET = 1000 $ Selects LSEQ 1000 for upstream loads

SUPER = ALL $ Identify SEs to process
.
.
.
DISP = ALL
.
.
.
.
.
.
.
.
.
$ NONLINEAR SOLUTIONS FOR RESIDUAL SUPERELEMENTS
SUBCASE 10
CLOAD = 1001 $ Refers to CLOAD Bulk Data
NLPARM = 12 $ Iteration control
.
.
.
SUBCASE 20
CLOAD = 1002
LOAD = 10 $ Residual SE load
NLPARM = 22
.
.
.

Main Index
 696 MSC Nastran Reference Guide
General Description of Nonlinear Capabilities

SE 0 Upstream SE

LOAD CLOAD LOADSET

Bulk CLOAD

Defined
on LSEQ
DAREA1 DAREA1
IDV1 IDV2
Correlated
to CLOAD

LSEQ

Static Loads

Figure 3-28 Hierarchy of Load Data

Restarts for Nonlinear Static Analysis

Any restart in SOL 106 must have the FMS command RESTART and the two parameters PARAM,SUBID,value and
PARAM,LOOPID,value.
The SUBID is the sequential number of a subcase. It is recommended that the SUBID value always be incremented by one
and point to a new subcase. The new Subcase should contain any standard subcase commands required.
The LOOPID is the identifier of the converged solution from which the restart is to take place.
Any converged and saved solution always gives the messages:
*** USER INFORMATION MESSAGE 6186,
*** SOLUTION HAS CONVERGED ***

Main Index
 Appendix C: Nonlinear Analysis 697
General Description of Nonlinear Capabilities

SUBID 7 LOADINC 5 LOOPID 67 LOAD STEP 7.000 LOAD FACTOR 1.00000

^^^ DMAP INFORMATION MESSAGE 9005 (NLSTATIC) - THE SOLUTION FOR LOOPID= 67
IS SAVED FOR RESTART
To restart from the above run, the restart might look like:
RESTART
TIME 10000
SOL 106
DIAG 50
CEND
TITLE = SAMPLE RESTART
PARAM,SUBID,8
PARAM,LOOPID,67
SPC = 200
MPC = 250
SET 1 = 1118
DISPLACEMENT = 1
SPCF=1
SUBCASE 1
LOAD = 101
NLPARM = 51
.
.
.
SUBCASE 7
LOAD = 107
NLPARM = 57
SUBCASE 8
LOAD = 108
NLPARM = 58
BEGIN BULK
NLPARM 58 5 AUTO 20 20 PW YES
1-3 1-3 1-7 -3
SPCD 108 1118 3 0.025
FORCE 108 1118 0. 0. 0. 1.
ENDDATA
For a full data recovery restart after the completion of a complete analysis proceed as above but do not actually add the new
Subcase in Case Control. For example, assume the analysis has ended in subcase 8 above with the following messages:
*** USER INFORMATION MESSAGE 6186,
*** SOLUTION HAS CONVERGED ***

SUBID 8 LOADINC 5 LOOPID 72 LOAD STEP 8.000 LOAD FACTOR 1.00000

^^^ DMAP INFORMATION MESSAGE 9005 (NLSTATIC) - THE SOLUTION FOR LOOPID= 72
IS SAVED FOR RESTART
Perform a full data recovery restart as follows:
RESTART
TIME 10000
SOL 106
DIAG 50
CEND
TITLE = FULL DATA RECOVERY RESTART
MAXLINES = 10000000
$ POINT TO A NEW SUBCASE BUT DO NOT ACTUALLY ADD A NEW SUBCASE

Main Index
 698 MSC Nastran Reference Guide
Nonlinear Modal Analysis

PARAM,SUBID,9
PARAM,LOOPID,72
SPC = 200
MPC = 250
SET 1 = 1118
$DISPLACEMENT = 1
$ ADD DISPLACEMENT AND STRESS REQUESTS FOR DATA RECOVERY
DISPLACEMENT=ALL
STRESS=ALL
SPCF=1
SUBCASE 1
LOAD = 101
NLPARM = 51
.
.
.
SUBCASE 7
LOAD = 107
NLPARM = 57
SUBCASE 8
LOAD = 108
NLPARM = 58
BEGIN BULK
ENDDATA
Note that in the above, there is no SUBCASE 9 added.
The result of this last restart will be the complete displacement and stress results output for each converged solution saved
on the data base starting with the first converged solution and ending with the results of the converged solution at LOOPID
= 72.

Nonlinear Modal Analysis

User Interface
Prestressed normal modes can be requested at the end of each subcase in SOL 106 by adding the following items to the
input file.
 Add METHOD=SID in the subcase of interest. A prestressed normal mode analysis will then be performed at the
end of this subcase. The METHOD command points to an EIGRL or EIGR Bulk Data entry which then selects
an eigenvalue method.
 For multiple normal mode analyses in SOL 106, the METHOD command may appear in more than one subcase
or above all subcases. In the latter case, a normal mode analysis is performed at the end of each subcase. The
stiffness used for modal analysis corresponds to the last step of the subcase. Modal analysis cannot be performed at
intermediate solution step.
 A PARAM,NMLOOP,loopid command must appear in the Case Control or Bulk Data Section to request a
normal mode analysis at the end of those subcases with a METHOD command. The actual value of loopid is
unimportant as long as it is a positive integer. This alleviates the cumbersome task of figuring out the exact loopid.
 Add the appropriate EIGRL or EIGR entry in the Bulk Data Section of the MSC Nastran Quick Reference Guide.

Main Index
 Appendix C: Nonlinear Analysis 699
Nonlinear Modal Analysis

Example (TURBINE1.DAT)
The following example is a normal mode analysis of a prestressed turbine blade (see Figure 3-29). The blade is modeled with
CHEXA elements and loaded with centrifugal forces (RFORCE). The default iteration strategy is selected on the NLPARM
Bulk Data entry. MSC Nastran automatically updates the stiffness matrix at the end of the subcase if a METHOD Case
Control command appears in the subcase.

Figure 3-29 Prestressed Modes of a Turbine Blade

An excerpt of the input file is shown.

Main Index
 700 MSC Nastran Reference Guide
Nonlinear Modal Analysis

SOL106
DIAG8
CEND
TITLE = Turbine Blade
ECHO = SORT,PUNCH
SPC= 200
LOAD= 300
NLPAR= 400
METHO= 500
DISP(PLOT) = ALL
ELST(PLOT) = ALL
NLSTRESS = NONE
BEGIN BULK
PSOLID100100
MAT1100 1.+7 .3 1.
RFORCE 300 .5 1. 2
NLPARM400 1 200 +N
+N 1.-6 1.-12 +B
+B 0
EIGRL500 6
PARAMCOUPMASS1
PARAMLGDISP1
PARAMNMLOOP1
PARAMPOST0
$
CHEXA 1 100 10000 10001 10101 10100 11000 11001 +M00000
+M00000 11101 11100
GRID 19324 44. 4.5 -.2 456
SPC1 200 123 10000 10100 10200 10300 11000 11100 +M00648
+M00648 11200 11300 12000 12100 12200 12300 13000 13100 +M00649
+M00649 13200 13300 14000 14100 14200 14300 15000 15100 +M00650
+M00650 15200 15300 16000 16100 16200 16300 17000 17100 +M00651
+M00651 17200 17300 18000 18100 18200 18300 19000 19100 +M00652
+M00652 19200 19300
$
$ PLUS THE REST OF THE MODEL
$
ENDDATA

Restart from SOL 106 into SOL 103 or into Another Linear Solution Sequence
For the purpose of a prestressed normal modes analysis, the old way of restarting from SOL 106 into SOL 103 is no longer
necessary; the user can, instead, restart from a SOL 106 run into another SOL 106 run to perform the prestressed normal
modes calculation.
Restarts from SOL 106 into linear solution sequences are not recommended to the novice user because of several
limitations. The results of the linear restart are incremental values with respect to the preload, not total values. However,
some experienced users restart from SOL 106 into SOL 103 to perform prestressed modal analysis with changing boundary
conditions, or restart into another linear solution sequence to perform a perturbed linear solution on a preloaded structure.
In the SOL 106 cold start run, KMETHOD=’ITER’ or ‘AUTO’ and KSTEP=1 must be specified on the NLPARM Bulk
Data entry in order to restart from the end of the subcase. With these parameters, all nonlinear information is stored on
the database so that the nonlinear tangent stiffness can be recovered in the linear restart run. The following command must
be included in the linear restart run:
 PARAM,NMLOOP,loopid in the Case Control or Bulk Data Section, where loopid is the load step id at the end
of a subcase from the SOL 106 cold start run.
Restarts are also allowed at intermediate steps within a subcase. The parameter INTOUT on the NLPARM entry must be
set to ‘YES’ or ‘ALL’ in the SOL 106 cold start run to store intermediate steps on the database for restart purposes. However,
the follower force stiffness may not be updated at intermediate steps. Therefore, the user should restart from the end of a
subcase if the effect of follower force stiffness is important.

Main Index
 Appendix C: Nonlinear Analysis 701
Nonlinear Modal Analysis

Differential Stiffness and Follower Force Stiffness in Linear and Nonlinear Solution
Sequences
For prestressed normal mode analysis, it is useful to know which stiffness contributions are included. The table below
summarizes the action of the parameters LGDISP and FOLLOWK in SOL 106 as compared to the action of FOLLOWK
in linear solution sequences. In some cases, the default settings for linear and nonlinear solution sequences are not the same.
In nonlinear static analysis (SOL 106), for the default of LGDISP=-1 (geometric linear), the parameter FOLLOWK
(follower force stiffness) is ignored. For LGDISP=1, FOLLOWK is set to YES by default. In the linear solution sequences,
FOLLOWK=YES is the default, see Case No. 3 in Table 3-13.
In theory, the follower force stiffness is unsymmetric. The linear solution sequences always symmetrize the follower force
stiffness. By default, SOL 106 uses the symmetrized follower force stiffness to improve performance. If the user wants to
use the unsymmetric follower force stiffness, the parameter FKSYMFAC can be specified for this purpose. See FKSYMFAC
(p. 834) in the MSC Nastran Quick Reference Guide for further details. The use of the symmetrized follower force stiffness
is recommended because it gives sufficient accurate results with substantial performance improvement.
In SOL 129, follower forces are included, however, the follower force stiffness is never calculated.

Table 3-13 Differential Stiffness and Follower Force Stiffness Contribution

SOL 103 Cold
SOL 106 Restart from SOL Start Using
Parameters Cold Start or Restart 106 into SOL 103 STATSUB
Nonlinear
LGDISP FOLLOWK Element Force Stiffness Stiffness Stiffness
1 1 YES L+M+D+F L+M+D+F L+M+D+F L+D’+F’
2 1 NO L+M+D+F L+M+D L+M+D L+D’
3 -1 YES L+M L+M L+M+D’+F’ L+D’+F’
4 -1 NO L+M L+M L+M+D’ L+D’
5 2 YES L+M+D L+M+D+F L+M+D+F L+D’+F’
6 2 NO L+M+D L+M+D L+M+D L+D’

where:

L = linear part of element force or stiffness

M = material nonlinear part of element force or stiffness
D = differential or geometric nonlinear part of element force or stiffness
D’ = differential part of stiffness for small deformations
F = follower force part of element force or stiffness
F’ = follower force part of stiffness for small deformations

Main Index
 702 MSC Nastran Reference Guide
Nonlinear Modal Analysis

All other linear solution sequences (for example, SOL 105, 108, etc.) have the same default as SOL 103.

Buckling Analysis Enhancement

An accurate solution to a buckling problem requires more meticulous efforts than just following a numerical procedure.
There are a number of factors to consider before a buckling solution can be accepted with confidence. Such points along
with a few examples are discussed by comparing solutions obtained from linear buckling analysis with preloads and
nonlinear buckling with/without unsymmetric stiffness matrix.
A starting step for a nonlinear buckling analysis should be a linear buckling analysis.

Theory
Let us start with a linear equation of motion for a preloaded structure, i.e.,
d
Mu·· + Cu· + Ku + K u = P  t 
d
where M, C, K and K are mass, viscous damping, material and differential stiffness matrices, and P  t  is a forcing
function in time domain. The differential stiffness is created by the initial stress due to preloads, and it may include the
follower stiffness if applicable. Ignoring the damping term to avoid complex arithmetic, an eigenvalue problem may be
formulated as
d 2
K + K  –  M = 0
which is a governing equation for normal mode analysis with a preload, but it could be recast for a dynamic buckling
analysis at a constant frequency. For a static buckling, that constant frequency of vibration is zero and the inertia term drops
out, i.e.,
d
K + K  = 0

in which    represents virtual displacements, or buckling modes shapes. A non-trivial solution exists for an eigenvalue
d
that makes the determinant of  K + K  vanish, which leads to an eigenvalue problem
d
 K + K     =  0 

where  is an eigenvalue which is a multiplier to the applied load to attain a critical buckling load. If there exists constant
preloads other than the buckling load in question, the above equation should include additional differential stiffness; that is,
d
 K + K preload + K buckle     =  0 

Main Index
 Appendix C: Nonlinear Analysis 703
Nonlinear Modal Analysis

in which differential stiffness is distinguished for constant preload and variable buckling load. Notice that no eigenvalue
d
solutions are meaningful if the preload makes the structure buckle, i.e.,  K + K preload  should be positive definite. The
buckling analysis with an excessive preload can render wrong solutions, unless the positive definiteness is ensured.

Example - prld_buck.dat
An example of input data for a linear buckling analysis with a preload is shown below:
ID plate105, v2001 $ SHL 2/23/01
DIAG 8, 15 $
SOL 105
TIME 60
CEND
TITLE = MSC Nastran job created on 20-Feb-01
SPC = 2 $ clamp the bottom edge
DISP=ALL
SUBCASE 10
SUBTITLE= preload tensile edge force 300 lb/in
LOAD = 2
SUBCASE 20
Subtitle= top pressure 100 lb/in.
LOAD = 7
SUBCASE 30
SUBTITLE=buckling analysis
STATSUB(preload)=10
STATSUB(buckle)=20
METHOD=1
BEGIN BULK
PARAM POST 0
EIGRL 1 0. 6. 3 0 1.
.
.
.
ENDDATA
This is an example of a plate model (5 in. by 10 in., 0.1 in thick, steel, clamped on the bottom edge) with lateral tension
by an edge load (in x) of 300 lb/in for a preload and vertical edge load (in y) of 100 lb/in for a buckling load. Alternatively,
the preload and the buckling load may be switched to perform a similar analysis. The first buckling mode is a bending mode
at vertical load of -64.55 lb/in, as shown in Figure 3-30. The second buckling mode in bending at -593 lb/in is also shown
in Figure 3-30, but it is of no physical significance and shown here for verification purposes only. If the lateral load is in
compression instead of tension, the plate could buckle in a twist mode at -123.29 lb/in in the absence of a vertical load, as
shown in Figure 3-31. A parametric study, conducted with various combinations of lateral and vertical loads, resulted in a
buckling envelope (or locus) on a loading plane as shown in Figure 3-32. The buckling shape is a bending mode above the
curve and a twist mode in the left side of the curve. Bifurcation behavior can be observed in the transitional region near
the corner point, where either the twist or bending mode could appear as the first and the second mode. Lower-right side
of the curve is the safety zone.

Main Index
 704 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Figure 3-30 Panel Buckling Shapes for the First Two Bending Modes

Figure 3-31 Panel Buckling in Twist Mode Under Lateral Compression

Main Index
 Appendix C: Nonlinear Analysis 705
Nonlinear Modal Analysis

Figure 3-32 Cantilevered Panel Buckling Envelope

Nonlinear Buckling Analysis with Symmetric Stiffness Matrix

Nonlinear buckling analysis capability has been available in SOL 106 by restarts using PARAM, BUCKLE, 1. While serving
the purpose of a nonlinear buckling analysis following a static nonlinear analysis, this buckling analysis procedure is
cumbersome to the user because it requires a restart. To provide a more convenient and versatile capability for nonlinear
buckling analysis, an option for a buckling analysis in a cold-start run with PARAM, BUCKLE, 2 was introduced in MSC
Nastran Version 70.7. This option allows a buckling analysis in any Subcase that has the METHOD command specified
for an eigenvalue analysis.
There are two more ways in nonlinear analysis to estimate the critical buckling load. One method is the arc-length method
that can provide solutions past the critical buckling load into the post-buckling state. Using the arc-length method, the
nodal displacement of a point with the most noticeable motion should be traced to illustrate the peak point. The applied
load at this peak resembles the critical buckling load. Another method is to use Newton's method until the solution cannot
be obtained due to divergence, in which case the adaptive bisection method is activated in the vicinity of the critical
buckling load and stops at the limit load, very close to the critical buckling point. A similar method was applied to the
nonlinear dynamic stability analysis using adaptive bisection algorithm in time domain for buckling or bifurcation
prediction in the implicit direct-time integration.
The nonlinear buckling analysis procedure in MSC Nastran allows buckling analysis with preloads. However, the nonlinear
procedure follows a different numerical procedure. Therefore, it could provide a means for validation of linear solutions. A
parametric study using nonlinear buckling analysis rendered data points for the panel buckling problem, labeled “Nonlinear
1 and 2” in Figure 3-32, using the lateral force and vertical force as preloads, respectively. Although two different methods
have been used in the nonlinear buckling prediction: namely, (1) eigenvalue estimation by extrapolation and (2) detection
of the critical load by consecutive bisections with no recourse due to divergence, all the nonlinear solution points fell, by

Main Index
 706 MSC Nastran Reference Guide
Nonlinear Modal Analysis

and large, on the same curve. This figure also shows that the problem is basically linear, because the “Large Displacement”
option introduced no visible effects.

Nonlinear Buckling Analysis with Unsymmetric Stiffness Matrix

Follower forces usually create unsymmetric follower stiffness unless the follower forces are completely enclosed, exerting
zero resultant forces. This asymmetry condition should be addressed with caution. The follower stiffness in SOL 106 was
introduced in Version 70.7, where symmetrization was applied conservatively so that it is activated in rare occasions. The
logic was modified in MSC Nastran 2001 to always symmetrize the follower stiffness by default. Some problems exhibited
convergence difficulties as a drawback of symmetrizing follower stiffness. The symmetrization does not show adverse effects
until the follower stiffness effect becomes dominant under large displacements. In MSC Nastran 2004, the symmetrization
is modified to be activated adaptively, detecting the severity of the asymmetry.
For adaptive symmetrization of the follower stiffness, we need a criterion to determine the extent of asymmetry. This can
be achieved by normalizing the norm of unsymmetric part of the follower stiffness by a stiffness matrix that is independent
from follower stiffness. The differential stiffness was chosen for this purpose after some numerical experiments. Then, the
ratio of asymmetry is compared with a tolerance, which is represented by a user parameter FKSYMFAC so that the user can
control the process. The default value for FKSYMFAC is determined to be 0.024 by a parametric study, allowing
symmetrization of the follower stiffness during the stiffness update process if the iterative process can render converged
solutions. If the asymmetry ratio is greater than FKSYMFAC, symmetrization process will not be activated. The asymmetry
ratio increases as the geometric nonlinearity intensifies, which is confirmed by numerous test runs.
When the follower stiffness becomes significantly asymmetric, the real eigenvalue analysis would be invalid or erroneous.
The eigenvalue analysis with unsymmetric matrices requires complex arithmetic with left- and right-handed eigenvectors.
To use the complex eigenvalue analysis, you must specify CMETHOD instead of METHOD in the Subcase and the
CMETHOD should reference EIGC Bulk Data entry. The eigenvalues obtained from the complex eigenvalue analysis for
buckling with asymmetric stiffness are complex numbers, of which imaginary numbers become zeros. All the eigenvectors
are also real numbers. By comparing results from both real and complex eigenvalue analysis, it is confirmed that the real
eigenvalue analysis renders wrong solutions to unsymmetric stiffness for nonlinear buckling. Particularly, the mode shapes
are erroneous with real eigenvalue analysis from unsymmetric problem, whereas the discrepancy in the critical buckling load
is not pronounced.
The buckling shape should be examined graphically to fully understand the physical phenomenon. Fortunately, the
complex eigenvalue analysis produces real eigenvectors from real and unsymmetric matrices. However, the buckling shapes
displayed with eigenvectors currently are with reference to the undeformed shape, and do not visualize the real buckling
shape if the deformed shape at the onset of the buckling has very large displacements.

Example
An example problem is a cantilevered beam model subject to an end force that is a follower force; that is, FORCE1. The
eigenvalue analysis is performed at load factor 4, where the asymmetry ratio is 0.069, which produced buckling shapes
shown in Figure 3-33 (obtained by complex method) and Figure 3-34 (obtained by real eigenvalue method). They are different
and the real eigenvalue method is likely to be in error. The display tool should be improved so that the buckling shapes be
plotted with an eigenvector as a perturbation with reference to the deformed shape to fully perceive the physics.

Main Index
 Appendix C: Nonlinear Analysis 707
Nonlinear Modal Analysis

Figure 3-33 Buckling Shape by Complex Method at L.F.=4

Figure 3-34 Buckling Shape by Real Eigenvalue Analysis at L.F.=4

Main Index
 708 MSC Nastran Reference Guide
Nonlinear Modal Analysis

An Example of Cylinder Buckling

A hollow cylinder is used to demonstrate an interesting behavior in buckling. Its geometric attributes include: 50 in. in
diameter, 100 in. in height and 0.125 in. wall thickness with closed ends of 0.5 in. in thickness. The cylinder, made of steel,
is subjected to a combined load of external pressure on the cylindrical surface and vertical edge force on the top, while the
bottom is fixed in z.
The cylinder could buckle in two different modes depending on the combination of loads as shown in Figure 3-35 and
Figure 3-36, where the first mode with vertical creases appear when the external pressure is predominant (showing a star-
shaped cross-section at the mid-span) and the second mode with horizontal creases appear when the vertical edge force is
more dominant. A failure curve (or buckling envelope) can be constructed by a parametric study using various magnitudes
of the preloads with external pressure or vertical edge load, as shown in Figure 3-37. The first buckling shape appears above
the curve and the second buckling in the right side of the curve. A bifurcation occurs at the corner, where either buckling
mode may occur. The cylinder is free from buckling in the lower-left region of the curve. It is noted that the buckling
resistance against the vertical load does not increase even if the pressure direction is reversed to an internal pressure (creating
a tensile hoop stress), although the buckling load increases against the external pressure as the vertical compression
decreases.
Nonlinear buckling solution points are added to the curve rendered by linear buckling solutions, where three categories of
nonlinear solutions are presented. Solution points in square symbol are from the eigenvalue analysis using an extrapolation
scheme, and those in diamond symbol are from the limit process using consecutive bisections upon divergence during the
incremental/iterative processes. Although the second buckling loads in the right side of the curve predicted by nonlinear
analyses are almost identical to the linear solutions, nonlinear solutions for the first buckling shape above the curve are
somewhat dispersed. In the absence of an exact solution for a reference, these nonlinear solutions seem plausible to indicate
that the geometric nonlinearity has significant effects on the buckling of the first kind with vertical creases, while the
geometric nonlinear effect is negligible in the second kind. These solutions, however, were obtained with symmetrized
stiffness matrix. The solution points in triangular symbol show tremendously different trend in the upper side of the curve,
where the buckling is caused by the external pressure.
The pressure load creates follower stiffness, which is unsymmetric in this problem. The solutions in triangular symbol were
obtained with unsymmetric matrices. By preserving asymmetric stiffness matrices, the solutions converged further and the
predicted buckling loads were higher. The buckling analysis with asymmetric matrices (but real-valued) requires complex
eigenvalue analysis, although the solution turns out to be real-valued. Furthermore, an ideal cylinder without the end
condition does not buckle under pressure because of the follower stiffness effect. That is why the solutions with
asymmetric stiffness converge to much higher external pressure, and the complex eigenvalue analyses with extrapolation at
the last converged steps render solution points shown by triangles. Solutions may continue for increased pressure and it has
been observed that the predicted buckling point increases as the eigenvalue analysis is performed at higher pressure. Two of
the buckling modes of these solutions are shown in Figures 3-38 and 3-39, which are rather spurious and fortuitous.
However, the deformed shape of the second kind of buckling as shown in Figure 3-40 was obtained by a limit process using
adaptive bisections upon divergence, which depicts a genuine buckling shape. It also signifies that the nonlinear eigenvalue
analysis based on extrapolation is an approximation and it can vary appreciably depending on where the extrapolation is
applied.

Main Index
 Appendix C: Nonlinear Analysis 709
Nonlinear Modal Analysis

Figure 3-35 Cylinder Buckling Shape 1 With Vertical Creases

Figure 3-36 Cylinder Buckling Shape 2 with Horizontal Creases

Main Index
 710 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Figure 3-37 Cylinder Buckling Envelope (With Nonlinear Solutions)

Main Index
 Appendix C: Nonlinear Analysis 711
Nonlinear Modal Analysis

Figure 3-38 Buckling Shape at Paxial=2000 and External Pressure=73.44

Figure 3-39 Buckling Shape at Paxial=10000 and External Pressure=72.02

Main Index
 712 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Figure 3-40 Deformed Shape on Buckling at Paxial=12243.4 and

Internal Pressure=20

Nonlinear Transient Response Analysis

User Interface
The input data required for SOL 129 is a combination of direct time integration control data, similar to SOL 109 (for
direct linear transient with superelements), and nonlinear modeling data similar to SOL 106 (for nonlinear statics). The
nonlinear properties and/or effects are defined by nonlinear material data (MATS1, MATHP, and TABLES1), gap elements
(GAP) or slideline contact (BCONP, BLSEG, and BFRIC) for nonlinear interface, and PARAMeter LGDISP for geometric
nonlinearity. The transient effects are produced by time-dependent loading functions (TLOADi, DAREA, LSEQ, etc.),
damping (parameters, elements and material data), and mass properties. Massless degree of freedom should be avoided for
numerical stability and the small damping is also recommended.
The unique data required for SOL 129 is supplied on the TSTEPNL Bulk Data entry. The TSTEPNL entry in itself is a
combination of the TSTEP entry for direct time integration and the NLPARM entry for nonlinear iteration control.
Restarts are controlled by parameters (LOOPID, STIME, SLOOPID, and SDATA) which can be specified either in the
Case Control Section or Bulk Data Section. Some optional parameters (TSTATIC, NDAMP) are provided for additional
control or capabilities.

Main Index
 Appendix C: Nonlinear Analysis 713
Nonlinear Modal Analysis

Case Control
Each subcase defines a time interval starting from the last time step of the previous subcase and the interval is subdivided
into small time steps. The output time is labeled by the cumulative time, including all previous subcases. There are
advantages to divide the total duration of analysis into many subcases so that each subcase does not have excessive number
of time steps; for example, not to exceed 200 steps. The data blocks containing solutions are generated at the end of each
subcase to store in the database for output process and restarts. As such, converged solutions are apt to be saved at many
intermediate steps in case of divergence and more flexible control becomes possible with multiple subcases.
The input loading functions may be changed for each subcase or continued by repeating the same DLOAD command.
However, it is recommended to use the same TLOADi Bulk Data for all the subcases in order to maintain the continuity
between subcases, because TLOADi entries define the loading history as a function of cumulative time. Static loads
(PLOADi, FORCEi, MOMENTi) may be associated with time-dependent functions by the LSEQ Bulk Data entry that can
be selected by a Case Control command LOADSET. However, thermal loads or enforced displacements (SPCD) are not
allowed in the nonlinear transient analysis. Nonlinear forces as function of displacements or velocities (NOLINi) may be
selected and printed by Case Control commands NONLINEAR and NLLOAD, respectively. Each subcase may have a
different time step size, time interval, and iteration control selected by the TSTEPNL request. The Case Control commands
which may not be changed after the first subcase are: SPC, MPC, DMIG, and TF.
Output requests for STRESS, SDISPLACEMENT, SVELOCITY, SACCELERATION, and NONLINEAR may be
specified within a subcase. DISPLACEMENT, VELOCITY, ACCELERATION, OLOAD, FORCE, and SPCFORCE must
be specified above the subcase level. However, FORCE output and GPSTRESS output are not available for nonlinear
elements.
Initial conditions (displacement or velocity) can be specified by the Bulk Data TIC selectable by the Case Control
command IC. If initial conditions were given, all of the nonlinear element forces and stresses must be computed to satisfy
equilibrium with the prescribed initial conditions. On the other hand, initial conditions could be a part of the nonlinear
analysis by applying static analysis for the preload using PARAM, TSTATIC in the first subcase. Then the transient analysis
can be performed in the ensuing subcases. Associated with the adaptive time stepping method, the PARAMeter NDAMP
is used to control the stability in the ADAPT method. The parameter NDAMP represents the numerical damping
(recommended value for usual case is 0.01) which often required to improve the stability and convergence in the contact
problems.
All the superelement model generation options and matrix reduction options are allowed for the linear portion of the
structure. The generalized dynamic reduction, component mode synthesis, and Guyan reduction may be performed for
upstream superelements. The residual superelement may contain scalar degree of freedom representing linear modal
formulations.

Limitations
 No constraint set charges are allowed between subcases or in restarts.
 No initial preloads such as closed gaps or terminal loads are allowed in a cold start. See PARAM,LOOPID under
“Restarts” in this section.
 Overall structure damping (PARAM,G) is not provided in nonlinear elements. Use the damping field in the
material (MATi) data entry and the PARAMeter W4 to supply damping based on the linear properties.
 Element forces are not computed for nonlinear elements. Output requests will be ignored.

Main Index
 714 MSC Nastran Reference Guide
Nonlinear Modal Analysis

 When multiple subcases are specified, it is recommended that a single time function (see the TLOADi Bulk Data
entry description) be assigned to an applied load throughout the entire time span of the solution. This time
function can be used for all subcases. If different time functions are to be used between subcases, ensure that the
time functions are continuous between subcases.
 Element stresses for nonlinear elements are computed and printed during the nonlinear iteration phase only. They
are not merged with those of the linear elements.

Direct Time Integration

If the effects of inertia and damping forces are to be included in the nonlinear analysis, the nonlinear transient response is
analyzed by a step-by-step integration of an equation of motion, i.e.,
Mu·· + Cu· + F  u  = P  t  (3-1)

By direct numerical integration, the equilibrium of equation (3-1) is satisfied at discrete time steps with an interval of t .
The equilibrium is ensured by iterations until the solution converges to the preset error tolerance.
Procedures, such as the quasi-Newton update and line search process, provide users with additional options to adjust
incremental and iterative processes in search of more efficient and effective strategy for obtaining solutions. A self-adaptive
time stepping method that adjusts time step automatically during analysis is available.
Newmark's direct time integration method is implemented using the two-point recurrence (or one-step) formula as a
foundation of this self-adaptive time stepping algorithm. The optimal time step size, which is required for accuracy and
efficiency, changes continuously in the transient dynamic environment. The primary concept of automatic time step
adjustment is that the proper size of the time step can be predicted based on the dominant frequency in the incremental
deformation pattern at the previous time step. This concept presents a deficiency of time lag involved in the prediction
process. Furthermore, changes in nonlinearity cannot be predicted from the deformation pattern at the previous time step.
This deficiency is overcome by the bisection process, which is activated when any difficulties arise in terms of convergence
during the iteration.
Frequent decomposition of the dynamic stiffness matrix is inevitable in adaptive time stepping using the implicit
integration method. Efforts have been made to minimize the thrashing in the algorithm (for example, too frequent or
repetitive adjustment of the time step size) and yet to provide a solution reliably to any type of problem. The iteration
process for the equilibrium employs expeditious methods such as BFGS updates and line searches as well as an adaptive
stiffness matrix update strategy. The present self-adaptive time stepping algorithm for the nonlinear transient analysis is
proving to be a robust and practical method, aimed at efficiency, effectiveness, and user friendliness.
The transient response analysis capability is provided in the self-contained solution sequences, SOL 129. Although this
solution sequences are developed for nonlinear transient response analysis, linear transient response analysis can be
performed in these solution sequences in order to take advantage of the automatic time stepping and restart capabilities.
The core parts of SOL 129 are the nonlinear modules (NLTRD2 and NLTRD) which perform the incremental and iterative
processes for implicit direct time integration.

Main Index
 Appendix C: Nonlinear Analysis 715
Nonlinear Modal Analysis

Implicit Integration Control: TSTEPNL Data Entry

The input fields of the TSTEPNL Bulk Data entry specify the time step size, the number of steps, and the output interval
as well as the nonlinear iteration options. The TSTEPNL Bulk Data is selectable by the Case Control command
TSTEPNL. Although the same TSTEPNL Bulk Data may be selected by more than one subcase, it is recommended to
select different TSTEPNL entry for each subcase in preparation for changes in the restarts.
The choice of time step size is determined primarily by the frequency content of the input load and the highest frequency
mode-of-interest. A general guideline is that seven or more steps per cycle be provided for reasonable accuracy. Modes with
shorter periods (higher frequency) will be attenuated by the numerical process. Highly nonlinear problems may require
smaller step size. However, the adaptive time stepping capability is the vital part of 129 and is recommended for any
problem (linear or nonlinear). Nevertheless, the initial time step size should be estimated by the user according to the
aforementioned principles.
A caution is necessary in using the automatic time step adjustment if the forcing function is a short duration pulse. Since
the automatic time step adjustment algorithm does not consider the loading history, short pulses could be skipped if the
time step is automatically adjusted to a larger value than the pulse duration. It is advised that a drastic change in the time
step size between subcases be avoided. The drastic change (for example, a ratio exceeding 1000) could cause a loss of
accuracy upon subcase switch. In such case, an intermediate subcase should be provided for a transition period of short
interval to reduce the ratio.
The parameters for controlling the accuracy and stability of the incremental and iterative process can be specified in the
TSTEPNL Bulk Data entry. The controls are applicable to the automatic time step adjustment and bisection process in
addition to stiffness matrix updates, BFGS updates and line searches similar to those on the NLPARM Bulk Data entry.
Since default values are resulted from numerous test runs, the analysis should be started with the default setting and changed
if necessary. The TSTEPNL data format is shown below with default values:
Format:

1 2 3 4 5 6 7 8 9 10
TSTEPNL ID NDT DT NO KSTEP MAXITER CONV
EPSU EPSP EPSW MAXDIV MAXQN MAXLS FSTRESS
MAXBIS ADJUST MSTEP RB MAXR UTOL RTOLB

Example With Defaults:

1 2 3 4 5 6 7 8 9 10
TSTEPNL 250 100 .01 1 2 10 PW
1.E-2 1.E-3 1.E-6 2 10 2 0.2
5 5 0.75 16.0 0.1 20.

The TSTEPNL Bulk Data entry is selected using ID by the Case Control command TSTEPNL. Each subcase (residual
superelement solutions only) requires a TSTEPNL entry. Multiple subcases are assumed to occur sequentially in time.
Therefore, the initial conditions of each subcase are defined by the end conditions of the previous subcase.

Main Index
 716 MSC Nastran Reference Guide
Nonlinear Modal Analysis

The NDT field specifies the number of time steps with DT as the size of each time step. The total duration for the subcase
can be assessed by multiplying NDT with DT (that is, NDT  DT ). The time increment ( t ) remains constant during
the analysis in AUTO and TSTEP options, and is equal to DT. However, the time increment ( t ) changes during the
analysis in the ADAPT option and the actual number of time steps will not be equal to NDT. In the ADAPT option, DT
is used as an initial value for t .
The NO field specifies the time step interval for output, i.e., every NOth step solution is saved for output. The data will
be output at steps 0, NO, ..., etc., and the last converged step for printing and plotting purposes. The Case Control
command OTIME may also be used to control the output points.
The program automatically adjusts the incremental time and uses the bisection algorithm in case of divergence. During the
bisection process the stiffness is updated at every KSTEPth successful bisection. The stiffness matrix is always updated for
a new subcase or restart, irrespective of the option selected.
The number of iterations for a time step is limited to MAXITER. If the solution does not converge in MAXITER iterations,
the process is treated as a divergent process; that is, either a bisection or stiffness matrix update takes place based on the
value of MAXBIS. The sign of MAXITER provides a control over the ultimate recourse (reiteration) in case of failure in
convergence or bisection. If the MAXITER is negative, the analysis is terminated when the divergence condition is
encountered twice during the same time step or the solution diverges for five consecutive time steps. If MAXITER is
positive, the program computes the best attainable solution and continues the analysis on second divergence.
The convergence test is controlled by convergence test flags (U for displacement error test, P for load equilibrium error test,
W for work error test) and the error tolerances (EPSU, EPSP and EPSW) which define the convergence criteria. All
requested criteria (combination of U, P, and/or W) are satisfied upon convergence. It should be noted that at least two
iterations are necessary to check the displacement convergence criterion.
The MAXDIV field provides control over diverging solutions. Depending on the rate of divergence, the number of
diverging solutions (NDIV) is incremented by 1 or 2. The solution is assumed to be divergent when NDIV reaches
MAXDIV during the iteration. If the bisection option is used, the time step is bisected upon divergence. Otherwise, the
solution for the time step is repeated with a new stiffness based on the converged state at the beginning of the time step. If
NDIV reaches MAXDIV twice within the same time step, the analysis is terminated with a fatal message.
The BEGS quasi-Newton updates and the line search process work in the way as in static analysis except for the default
setting. The MAXQN field defines the maximum number of quasi-Newton vectors to be saved on the database, and the
MAXLS defines the number of line searches allowed per iteration. Nonzero values of MAXQN and MAXLS activate the
quasi-Newton update and the line search process, respectively.

The FSTRESS field defines a fraction of the effective stress (  ) which is used to limit the subincrement size in the material
routine. The number of subincrements in the material routines is determined such that the subincrement size is
approximately FSTRESS   (equivalent stress). FSTRESS is also used to establish a tolerance for error correction in the
elasto-plastic material, i.e.,

Error in yield function  FSTRESS  

Main Index
 Appendix C: Nonlinear Analysis 717
Nonlinear Modal Analysis

If the limit is exceeded at the converging state, the program will exit with a fatal error message. Otherwise, the stress state
is adjusted to the current yield surface.
MAXBIS is the maximum number of bisections allowed for each time step ( – 9  MAXBIS  9 ). The bisection process
is activated when divergence occurs and MAXBIS  0 . The number of bisections for a time increment is limited to
|MAXBIS|. If MAXBIS is positive and the solution does not converge after MAXBIS bisections, the best solution is
computed and the analysis is continued to the next time step. If MAXBIS is negative and the solution does not converge
in |MAXBIS| bisection, the analysis is terminated.
The parameter ADJUST allows the user to control the automatic time stepping in the ADAPT option. A value of zero for
ADJUST turns off the automatic adjustment completely. If ADJUST is positive, the time increment is continually adjusted
for the first few steps until a good value of t is obtained. After this initial adjustment, the time increment is adjusted every
ADJUST time steps only. A value of ADJUST an order greater than NDT will turn off adjustment after the initial
adjustment. Since the automatic time step adjustment is based on the mode of response and not on the loading pattern, it
may be necessary to limit the adjustable step size when the period of the forcing function is much shorter than the period
of dominant response frequency of the structure. It is the user's responsibility to ensure that the loading history is properly
traced with ADJUST option. The ADJUST option should be suppressed for the duration of short pulse loading. If unsure,
start with DT which is much smaller than the pulse duration in order to properly represent the loading pattern.
MSTEP defines the desired number of time steps to obtain the dominant period response accurately
( 10  Integer  200 ). RB defines bounds for maintaining the same time step for the stepping function in the automatic
time step adjustment method ( 0.1  Real  1.0 ). Parameters MSTEP and RB are used to adjust the time increment
during the analysis in the ADAPT option. The adjustment is based on the number of time steps desired to capture the
dominant frequency response accurately. The time increment is adjusted as follows:
t n + 1 = f  r t n

where:

r = 1 -  2 1 -
------------------ ------  -------
MSTEP w n t n   

with:
f = 0.25 for r < 0.5  RB
f = 0.5 for 0.5  RB < r < RB
f = 1.0 for RB < r < 2
f = 2.0 for 2.  r < 3./RB
f = 4.0 for r > 3./RB

Main Index
 718 MSC Nastran Reference Guide
Nonlinear Modal Analysis

The recommended value of MSTEP for nearly linear problems is 20. A larger value (for example, 40) is required for highly
nonlinear problems. In the default option, the program automatically computes the value of MSTEP based on the changes
in the stiffness.
The MAXR field defines the maximum ratio for the adjusted incremental time relative to DT allowed for time step
adjustment (1.0  Real  32.0). MAXR is used to define the upper and lower bounds for adjusted time step size; that is,

DT DT
MIN  --------------------- , -----------------  T  MAXR  DT
 2 MAXBIS MAXR

The UTOL field defines the tolerance on displacement increment below which there is no time step adjustment (0.001 <
Real  1.0). UTOL is used to filter undesirable time step adjustment; that is, no time step adjustment is performed if

U· n
----------
-  UTOL
U· n max

The RTOLB field defines the maximum value of incremental rotation (in degrees) allowed per iteration to activate bisection
(Real > 2.0). The bisection is activated if the incremental rotation for any degree of freedom (  x ,  y or  z ) exceeds
the value specified for RTOLB. This bisection strategy based on the incremental rotation is controlled by MAXBIS field.

Iteration Related Output Data

During the incremental and iterative computation, the process information consisting of iteration data is printed at the end
of each iteration or time step. They are printed under the following heading:

TIME Cumulative time for the duration of the analysis.

ITER Iteration count for each time step.
DISP Relative error in terms of displacements defined as

ui – ui – 1 -
i =  i ------------------------------
EU for ADAPT method
 1 –  i u max

ui – ui – 1 -
i = ------------------------------
EU for AUTO or TSTEP
 1 –  i u max

where u max = max  u 1 , u 2 , , u n 

Main Index
 Appendix C: Nonlinear Analysis 719
Nonlinear Modal Analysis

LOAD Relative error in terms of load vectors defined as

Ri
E pi = ----------------------------------------- for ADAPT method
max  F n , P n 

Ri
E pi = ------------------------------------ for AUTO or TSTEP
max  F , P
where F and P are three-point average values for internal and external forces,
respectively.
WORK Relative error in terms of work defined as

T
 ui – ui – 1  Ri (3-2)
E wi = ----------------------------------------------------------------------------
T T
max   u n   F n ,  u n   P n  
LAMBDA(I) Rate of Convergence in iteration where

E pi
 i = ------------ (3-3)
E pi – 1
LAMBDA(T) Ratio of the load error for two consecutive time steps computed only for AUTO or
TSTEP method

E pn
 nn = -------------- (3-4)
E pn – 1
LAMBDA-BAR Average of LAMBDA(T) over the last 3 steps, computed only for AUTO or TSTEP
method.
DLMAG Absolute norm of the load vector, R . The absolute convergence is defined using
DLMAG by R  10 – 12 .
FACTOR Final value of the line search parameter.
E-FIRST Divergence rate, error before line search.
E-FINAL Error at the end of line search.
NQNV Number of quasi-Newton vectors appended.
NLS Number of line searches performed during the iteration.
ITR DIV Number of occurrences of divergence detected during the adaptive iteration by the
iteration module NLTRD2.

Main Index
 720 MSC Nastran Reference Guide
Nonlinear Modal Analysis

MAT DIV Number of occurrences of bisection conditions in the material routine (excessive
stress increment) or in the rotation angle (excessive rotation) during the iteration
using the ADAPT method.
NO. BIS Number of bisections executed for the current time interval.
ADJUST Ratio of time step adjustment relative to DT within a subcase.

Restarts
Since SOL 106 and SOL 129 share the same database storage for nonlinear tables and matrices, the restart system for
transient analysis can use either a previous static or transient nonlinear analysis as its initial conditions. A restart from
SOL 129 into SOL 106 is not available.

Restarting from SOL 106 into SOL 129

The options for a restart from SOL 106 into SOL 129 are static to static and static to dynamic analysis. For a restart from
a previous static analysis, only the first subcase is affected. Simply provide a database created in SOL 106 (or 129) and
specify the parameter PARAM,SLOOPID,N where N is the printed value of LOOPID for the desired static solution. The
initial transient load should be identical to static loads at the restart state. Constraint sets, direct input matrices, mass, and
damping may be changed.

Restarting within SOL 129

Restarting within SOL 129 allows static to static, static to dynamic, and dynamic to dynamic analysis. Restarts from a
previous nonlinear transient execution are available for a number of cases. If the same model is to be re-executed, only the
residual superelement needs to be reassembled. If the final results from the previous transient run are to be used as the initial
conditions at t = 0, add N dummy SUBCASE commands to start the residual Case Control execution and STIME = 0.
The normal restart for a transient run is to be continued from the last step of a previous subcase with different loads and/or
TSTEPNL data. For the normal restart provide the following parameters:
LOOPID = N: Start from the Nth subcase
STIME = t: Start from time t
Note that constraint sets should not be changed to avoid incompatible matrix sizes. The values of LOOPID and STIME,
which are printed with the iteration information for each subcase, can be directly read from the printout of the previous
run. If the adaptive time stepping algorithm is not activated, the value of t may also be calculated by the following equation:

 N 
t =   NDT i  DT i
 
i = 1 

where NDTi and DTi are the number of time steps and the time increment of the i-th subcase, respectively.
If a SOL 129 run is terminated abnormally in the middle of a subcase, it may or may not be restartable depending upon
the cause of the abnormal exit. If the job is stopped due to a diverging solution, it can be restarted either from the end of a
previous subcase or from the last saved solution step. The restart procedure for the former is identical to that for the normal

Main Index
 Appendix C: Nonlinear Analysis 721
Nonlinear Modal Analysis

restart as described in the preceding paragraph. The latter case also requires parameters LOOPID and STIME, however,
the input value for STIME differs depending on the value of METHOD specified on the TSTEPNL entry. If METHOD
= AUTO or TSTEP (the NLTRD module), STIME is the time corresponding to the last output step which may be
calculated based on the output skip factor (that is, the NO on the TSTEPNL entry). If METHOD = ADAPT (NLTRD2
module) the last converged solution is treated as an output step and is always saved for the restart so that STIME can be
the time of the last converged step. The values of STIME and LOOPID can also be found in the printout, if the ADAPT
method is used. Once STIME and LOOPID are known, determine the number of remaining time steps in the subcase and
create a new TSTEPNL entry for the remaining time. Insert a new subcase which references the new TSTEPNL entry prior
to the remaining subcases in the Case Control Section.
A solution may be terminated in the middle of a subcase due to insufficient CPU time: (1) the CPU time specified in the
Executive Control Section is insufficient so the run is forced to exit by the program, or (2) the CPU time exceeds the limit
specified in the computer system which leads to a sudden job abortion by the system. In the first case, MSC Nastran is able
to detect the specified CPU time in the Executive Control Section and automatically activate the wrapping-up procedure
before time expiration. When completed, the solution can be restarted from the termination point as in the solution
diverging case. In the second case the solution can only be restarted from the end of a subcase.
Restarts may also be performed solely for data recovery by providing the following parameters:
SDATA = –1 : Recover data without running the solution module LOOPID = N : from the 1st through the N-th subcases.
Note that solution sets DISP, VELO, ACCE, OLOAD, SPCF (printout and plotting) and NLLOAD (plotting only) are
recoverable while STRESS, SDISP, SVELO and SACCE sets are not.

Example
A simply supported elastic beam is subjected to a single cycle of a sinusoidal forcing function at the quarter span as
illustrated in Figure 3-41. The beam is modeled using 20 linear CBEAM elements. A stopper (snubber) is present underneath
the center of the beam with a clearance of 0.02 inches. This gap is simulated by a displacement dependent force which is
active only when the beam is in contact with the snubber. The problem is linear except for the effect of the stopper. The
input data for this problem is given in Listing 3-13.
One subcase with 100 time steps and an initial time step size of 0.1 msec was analyzed in the cold start run. Since the
adaptive time stepping algorithm was activated, the job ran to completion at 9.7 msec instead of the specified time span,
100 x 0.1 = 10 msec. A new subcase with 600 time steps was added in the first restart run to extend the time span for another
30 msec. The adaptive time stepping method was turned off and the execution time was specified to be 0.8 min. As
expected, the run was aborted at 36 msec due to insufficient time. In the second restart run, the execution time was changed
to 5.0 min. and another subcase with 164 time steps and a time increment of 0.1 msec was added to extend the analysis
time span up to 40 msec. The third restart was simply to recover the DISP, VELO, ACCE, OLOAD, NLLOAD and SPC
output. The response at the loading point of the beam is shown in Figure 3-42.

Main Index
 722 MSC Nastran Reference Guide
Nonlinear Modal Analysis

z
A
x
0.02 in
A = 0.314 in 2
25 in F(t) Stopper
I = 0.157 in 4

50 in 50 in  = 0.3 lb/in 3

20 Beam Elements

F F n  U 10010 

47.2 0.02 in
(lbs)  U 10010 
0.001 (sec)
990

Forcing Function NOLIN1 Representing GAP

Figure 3-41 Simply Supported Beam With a Restrained Motion

Main Index
 Appendix C: Nonlinear Analysis 723
Nonlinear Modal Analysis

Figure 3-42 Simply Supported Beam: Loading Point Response

Main Index
 724 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Listing 3-13 Input Data Listing for Simply Supported Beam

$RESTART$ ***added for the restart runs
ID TNOLIN, V66$ SSH 7/10/91
TIME $$
SOL 129$ NONLINEAR TRANSIENT ANALYSIS
DIAG 8,50$ PRINT MATRIX TRAILERS AND ITERATIONS
CEND
TITLE = DYNAMICS HANDBOOK DEMONSTRATION PROBLEM D9918
SUBTITLE = DIRECT TRANSIENT RESPONSE, NONLINEAR FORCE
LABEL = NOLIN IN SOL 129
SEALL = ALL
ECHO = UNSORT
SET 1 = 10005
SET 2 = 10010
SET 3 = 10000
SPC = 1002
$----------------------
$PARAM,LOOPID,1*** used for the 1st restart
$PARAM,STIME,0.0093
$----------------------
$PARAM,LOOPID,2*** used for the 2nd restart
$PARAM,STIME,0.0236
$----------------------
$ DISPL=2
$ VELO=2*** used for the 3rd restart
$ ACCEL=2
$ OLOAD=1
$ NLLOAD=2
$ SPCF=3
$PARAM,LOOPID,3
$PARAM,SDATA,-1
$----------------------
SUBCASE 1
DLOAD= 30
TSTEPNL= 10
NONLINEAR= 13$ SELECT NONLINEARFORCE
SUBCASE 2
DLOAD= 30
TSTEPNL= 20
NONLINEAR= 13$ SELECT NONLINEARFORCE
SUBCASE 3
DLOAD= 30
TSTEPNL= 30
NONLINEAR= 13$ SELECT NONLINEARFORCE
BEGIN BULK
$ MODELING INFORMATION FOR BEAM ONLY
CBAR10110010000100010.0.1.
CBAR10210010001100020.0.1.
CBAR10310010002100030.0.1.
CBAR10410010003100040.0.1.
CBAR10510010004100050.0.1:
CBAR10610010005100060.0.1.
CBAR10710010006100070.0.1.
CBAR10810010007100080.0.1.
CBAR10910010008100090.0.1.
CBAR11010010009100100.0.1.
CBAR11110010010100110.0.1.
CBAR11210010011100120.0.1.
CBAR11310010012100130.0.1.
CBAR11410010013100140.0.1.
CBAR11510010014100150.0.1.

Main Index
 Appendix C: Nonlinear Analysis 725
Nonlinear Modal Analysis

Listing 3-13 Input Data Listing for Simply Supported Beam (continued)
CBAR11610010015100160.0.1.
CBAR11710010016100170.0.1.
CBAR11810010017100180.0.1.
CBAR11910010018100190.0.1.
CBAR12010010019100200.0.1.
CONM21210010.1
GRID1050.-1.
GRID100000.0.0.1246
GRID100015.0.0.1246
GRID1000210.0.0.1246
GRID1000315.0.0.1246
GRID1000420.0.0.1246
GRID1000525.0.0.1246
GRID1000630.0.0.1246
GRID1000735.0.0.1246
GRID1000840.0.0.1246
GRID1000945.0.0.1246
GRID1001050.0.0.1246
GRID1001155.0.0.1246
GRID1001260.0.0.1246
GRID1001365.0.0.1246
GRID1001470.0.0.1246
GRID1001575.0.0.1246
GRID1001680.0.0.1246
GRID1001785.0.0.1246
GRID1001890.0.0.1246
GRID1001995.0.0.1246
GRID10020100.0.0.1246
MATl10003.+7.3.3
PARAMGRDPNT10010
PARAMWTMASS.002588
PBAR1001000.31416.15708 +PBl
+PBl 1. 0.
SPC100210123456
SPC1002100203100003
$
$ MODELING INFORMATION FOR CENTER SPRING
CROD10101010010
MATl1010.0.
PROD10101.
MATSl10PLASTIC 0.113.+8
$ LOADING AND SOLUTION INFORMATION
TLOAD230330..01145187.33-90.
DAREA3310005347.2
TSTEPNL10100.00011ADAPT
TSTEPNL20600.000051ADAPT+T521
+T521+T522
+T5220
TSTEPNL30164.000101ADAPT+T531
+T531+T532
+T5320
$
$ MODELING INFORMATION FOR NONLINEAR SPRING
NOLINl 131001031.10010313
TABLEDl 13+TAB
+TAB -2.5-24.95-2.0-20.0.0.ENDT
$
ENDDATA

Performance in SOL 129

Nonlinear transient response analysis in SOL 129 has become more robust and reliable. It can be used for linear analysis to
take advantage of the automatic time stepping. If the problem is truly linear (that is, without NOLIN) SOL 129 should
converge in one iteration at every time step. The computing speed is comparable to SOL 109 in such cases if the same
number of time steps are used. The automatic time stepping is based on the dominant response frequency which is based
on the system characteristics. This strategy is not effective if the loading changes faster than the system response. In case of
an impact loading, an indiscriminate use of automatic time stepping would cause the impulse represented inadequately or
unrecognized by the algorithm. This could be avoided by turning off the automatic time stepping. But it relies on the
user's intervention and therefore is not user-friendly.

Main Index
 726 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Benefits
SOL 129 will activate time step adjustment unless it is suppressed. As a result, the solution will be more accurate and the
solution time is reduced, especially the engineering time will be significantly saved because accurate solution may be
attained in the first trial. That is the goal of the automatic and adaptive time stepping method. However, if the model
includes NOLIN entries, SOL 129 is likely to activate iteration for equilibrium, while SOL 109 takes one-step solution for
each time step. That is because SOL 109 uses explicit approach for the NOLIN process, taking NOLIN forces at the
previous time step for the equilibrium of the current time step. The iteration makes SOL 129 slower than SOL 109 when
NOLIN entries are present, but the solution is more accurate. The iteration could be avoided by using large tolerances for
convergence criteria.
Saving engineering time is by far more efficient even if the computation takes more time. The automatic time stepping
strives for a one-shot solution for this reason, and the algorithm takes a rather conservative approach; that is, taking a smaller
stride to avoid divergence or convergence difficulties. Particularly, bisection algorithms are devised for accuracy and
effectiveness, not for efficiency. If the accuracy can be compromised for efficiency, the adaptive bisections should be turned
off. Also larger strides can be achieved by specifying a smaller value for MSTEP field (for example, 10) in the TSTEPNL
entry.
The following performance comparison of the analyses of a model with a lot of NOLIN entries shows the contrast in
efficiency versus strategy:

Table 3-14
Version method tsteps usertime systime elapsed Remarks
2001 fixed 3200 1787.69 1208.59 50:52 no bisections
2001 auto 1662.74 1187.90 49:20 no bisections
2004 fixed 3200 1520.04 1147.16 45:20 nolin=0,
tzeromax=0
2004 fixed 37039 17567.10 10535.30 9:48:34 bisections allowed
2004 auto 4540 1434.06 1228.18 47:42 nolin=0,
tzeromax=0
2004 auto 14652 11854.37 5724.78 5:02:58 nolin=1.,
tzeromax=1
2004 auto 1320 829.84 570.50 23:38 nolin=0.,
tzeromax=1,
mstep=10

Theory
Follower Force Generation

First, the load generation should be performed in the iteration module for efficiency. Direct time integration (in NLTRD2)
is to solve the equation
Mu·· + Cu· + F  u  = P  t 

Main Index
 Appendix C: Nonlinear Analysis 727
Nonlinear Modal Analysis

at every incremental time step. This process involves an integration in time domain combined with an iterative process at
each time step to satisfy equilibrium. The iteration is based on an incremental process using

4- 21 –  i+1  i 
------- M + -------------------- C + K̃  U  =  Rn + 1 
t
2 t  
with

 i   i  41 – 
M  U· n  +  1 – 2   P n – F n 
i
 R n + 1  =  P n + 1 – F n + 1  + --------------------
t
   

4 21 –   i 
– -------- M + -------------------- C  U n + 1 – U n 
t
2 t  

where  denotes numerical damping (PARAM, NDAMP).

The current design in MSC Nastran is to generate load vectors in NLTRD2 one column at each time step, so that no update
or interpolation is necessary; that is,

i  i i
P t  u  = f nf  t    P nf  + f f  t    P f  + N t
 
where the first term represents non-follower forces, which is to be generated in NLTRD2. Notice that the nonfollower
force vectors are formed once per time step, but the follower force vector (the second and the third terms) must be computed
at every iteration. It improves not only load-dependent adaptive time stepping but also the memory usage and reduces the
I/O operations.

Load-Dependent Time Step Adjustment

The second step is to make the time step adjustment algorithm adaptive to the loading history based on the TLOADi data.
The adaptive time step algorithm for TLOAD1 is implemented by considering time steps specified in the TABLEDi entries.
A new time step size is compared against time interval in the TABLED(1-4) at every time step. When adjusted step exceeds
the user-specified load increment interval or causes an abrupt changes in load magnitude, the adjusted value is bisected to
be consistent with the binary-based time adjustment algorithm. This process will continue sweeping through the time
duration for these subcase.
The transient loads specified by TLOAD2 are analytical functions, continuous and smooth, but the excitation frequency
should be considered for adaptive time stepping so that the period (field F on TLOAD2) is less than 8 times the incremental
time step; that is, bisect if the time step size is larger than 1/8 of a period.
This algorithm activates an adaptive bisection upon load spikes to account for impulsive loads with adequate number of
time steps. This modifications work for both structural analysis and heat transfer analysis.

Main Index
 728 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Adaptive Time Stepping for NOLIN and NLRGAP

The third phase makes the time step adjustment algorithm sensitive to the NOLIN data. The NOLIN forces, denoted by
N t  u u·  , are implicit functions of time (except for NLRGAP that could be explicit function of time), which may vary
abruptly as a function of time. Then the solution may not render a correct solution with a large time step. The NOLIN
entries consists of
 NOLIN1 with TABLEDi: N1 i = s  T j  u u· 

 NOLIN2: N2 i = s  X j  X k

A
 NOLIN3: N3 i = s  X j

A
 NOLIN4: N4 i = – s   – X j 
 NLRGAP: simulates a radial forces on a journal bearing joint
The criteria for the abruptness can be as follows:
–4
t n + 1 = 0.5  t if N t + t – N t  F n  Max  N t  10 

where the factor F n is determined based on the numerical experiments (defaulted to 1000). This factor F n is controlled
by system cell 386 (=nolin), i.e.,
F n = 1000.  nolin

where nolin is the alias name of the real-valued sys(386) as follows:

nolin = 0 bisection is suppressed (same as MSC Nastran 2001)

nolin = 0.001 increase accuracy slightly
nolin = 0.01 increase accuracy a little more
nolin = 1.0 allow full adaptive bisection (default)

Consecutive bisections are allowed (KSTEP+1) times, in which KSTEP is defaulted to 2 in the TSTENL entry.
In order to make a full implementation of NOLIN forces as internal forces, they have to be subtracted from element forces,
and thus NOLINs are elements having stiffness and damping. To remedy the automatic time stepping without full
implementation as elements, NOLINs need to be considered like an internal force in computing a dominant frequency for
a time stepping criterion. The dominant frequency should then be computed by

Main Index
 Appendix C: Nonlinear Analysis 729
Nonlinear Modal Analysis

T
2  u n   F n – N n 
n = -----------------------------------------------------
T
 u n   M   u n 

where:

N n =  N n  –  N n – 1 
The effectiveness for NOLIN-dependent time step adjustment is demonstrated in the two examples below: a vibrating
beam with a stopper simulated by NOLIN1 and a thermostat model with a NOLIN1. The result is the accuracy
improvement, not efficiency, as shown in the figures. The vibrating beam experiences abrupt changes in the NOLIN1 force
(contact) when the GAP opens or closes. The adaptive bisection algorithm detects the transition from open to closed or
vice versa. The thermostat model has similar characteristics as a vibrating beam.

Figure 3-43 Vibrating with a Stopper

Main Index
 730 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Figure 3-44 Thermal System with a Thermostat

Initial Time Step Adjustment

The adaptive time stepping algorithm in the NLTRD2 module starts with a user-specified value (DT field in the TSTEPNL
entry). The algorithm estimates a dominant frequency in the system, from which an appropriate time step size is computed.
However, this process requires two converged time steps. If the process continues without reverting back to time zero, the
solution could be erroneous for the first two steps with a large DT. This is particularly true if you specify excessively (two
orders of magnitude) large DT as demonstrated in some heat transfer problems.
An initial quadri-section algorithm has been implemented in MSC Nastran to remedy this deficiency. This initial quadri-
section is repeated until a right time step size is attained, which is controlled by an integer system cell 373 called
TZEROMAX, i.e.,

TZEROMAX > 0 : maximum number of times to return to time zero

TZEROMAX = 0 : no initial time step adjustment
TZEROMAX < 0 : no limit on DT adjustment

with a default value of TZEROMAX=4. When the initial time step adjustment is activated, a UWM (7600) is issued as:
INITIAL TIME STEP SIZE IS REDUCED TO X.XXXE-X.
This algorithm is applicable to both structural and heat transfer analyses, but especially effective for heat transfer analysis.
It allows the adjustment of user-specified value of DT by almost three orders of magnitude by default. This initial time step
adjustment algorithm accommodates the existing automatic time stepping capability as it currently works. However, the

Main Index
 Appendix C: Nonlinear Analysis 731
Nonlinear Modal Analysis

parameter MAXR (in TSTEPNL entry) has been modified to accept any real number without a bound, with a default value
of 32.
An example of a thermal conduction process is analyzed with/without initial time step adjustment starting with DT=1.0e-
4. The solution without initial time step adjustment shows gross error at the beginning as shown in Figure 3-45. The initial
time step adjustment algorithm adjust DT to 1.56e-6 to render the second solution.

Figure 3-45 Temperature History

Figure 3-46 Phase Transformation

Main Index
 732 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Nonlinear Analysis With Superelements

In nonlinear analysis with superelements, the nonlinear model must be in the residual model (Superelement = 0). All basic
superelement rules apply with the exception that a nonlinear element can not have any of its degrees of freedom in the “o-
set”.
The unique feature pertaining to load application for superelements are listed in 3-D Slideline Contact Capability, 675 under
“Load Combination”.

Grid Point Force Balance and Element Strain Energy in Nonlinear Static Analysis
In nonlinear analysis, the strain energy, E , for an element is defined by integrating the specific energy rate, the inner
product of strain rate and stress, over element volume and time
t
·T
E =   dV d (3-5)
0V
where:

 = stress tensor
·
 = strain rate
V = element volume
t = actual time in the load history

The integration over time leads to the following recursive formula using the trapezoidal rule
1 T
E n + 1 = E n + E n + 1 = E n + ---   n + 1   n +  n + 1  dV (3-6)
2
where:

n+1 = current load step

n = previous load step
E = strain energy increment
 = strain-increment

The steps n and n + 1 are converged solution steps.

By integrating equation (3-6) over the element volume, we get

Main Index
 Appendix C: Nonlinear Analysis 733
Nonlinear Modal Analysis

1 T
E n + 1 = E n + --- u n + 1  f n + f n + 1  (3-7)
2
where:

f = internal element forces

u = displacement increment from n to n + 1

For computational convenience, MSC Nastran uses equation (3-7) to calculate the element strain energy. The internal
element forces are readily available in every step because they are needed for the force equilibrium. Note that temperature
effects are included in the internal element forces.
When loads from temperature differences or element deformation are present, the default definition of element strain
energy for linear elements differ from the definition for nonlinear elements. For linear elements, the element strain energy
is defined as
1 T T
E = --- u K e u – u P et (3-8)
2
where P et is the element load vector for temperature loads and element deformation. Equation (3-8) assumes that the
temperatures are constant within a subcase. For nonlinear elements, the definition of equation (3-5) is used. In the case of
linear material and geometry, equation (3-5) becomes
1 T 1 T
E = --- u K e u – --- u P et (3-9)
2 2
Equation (3-9) assumes that the temperature varies linearly within a subcase. The user may request the definition of equation
(3-9) to be applied to linear elements by adding PARAM,XFLAG,2 to the input file.

Example
The following MSC Nastran input file represents a simplified model of a mechanical clutch that consists of springs, beams,
rigid elements, and gap elements. A geometric nonlinear analysis is performed. Both GPFORCE and ESE output requests
are applied above all subcases. This is a good example to show the grid point force balance with both linear and nonlinear
elements, that includes:
 applied loads
 element forces
 SPC forces
 MPC forces.
The model is shown in Figure 3-47. For clarity, only the elements are displayed.

Main Index
 734 MSC Nastran Reference Guide
Nonlinear Modal Analysis

Figure 3-47

The dashed lines represent rigid elements, the rest are gap and beam elements. Spring elements are located at the
intersection point of all dashed lines and are not visible in the figure because their connection points are coincident.

Grid Point Force Output

1 GPF005B - FORD ELETRO-MECHANICAL CLUTCH SEPTEMBER 26, 2000 MSC Nastran 9/25/00 PAGE 248
DISPLACEMENT IN MM., FORCE IN NEWTON, STRESS IN MPA
0 SUBCASE 2
LOAD STEP = 2.00000E+00
G R I D P O I N T F O R C E B A L A N C E

POINT-ID ELEMENT-ID SOURCE T1 T2 T3 R1 R2 R3

1 APP-LOAD .0 .0 .0 .0 .0 -2.400000E+02
1 F-OF-SPC .0 .0 .0 .0 .0 2.400000E+02
1 10001 BEAM .0 .0 .0 .0 .0 .0
1 10002 BEAM .0 .0 .0 .0 .0 .0
1 *TOTALS* .0 .0 .0 .0 .0 .0
0 100 F-OF-SPC -5.366200E-25 -4.514879E-24 .0 4.535559E-23 -6.400176E-24 .0
100 F-OF-MPC 5.366200E-25 4.514879E-24 .0 -4.535559E-23 6.400176E-24 1.292178E-24
100 3 ELAS2 .0 .0 .0 .0 .0 1.070983E-24
100 *TOTALS* .0 .0 .0 .0 .0 2.363161E-24
0 101 F-OF-SPC .0 .0 -3.078859E-28 -3.549490E-26 2.535672E-24 .0
101 101 BEAM -3.701448E-08 2.907432E-07 .0 .0 .0 -5.474456E-09
101 2004 BEAM -3.888788E-07 -1.144153E-07 3.078859E-28 3.549490E-26 -2.535672E-24 -1.777338E-23
101 1 GAP 1.270000E-10 1.438901E-27 .0 .0 .0 .0
101 *TOTALS* -4.257663E-07 1.763280E-07 .0 .0 .0 -5.474456E-09
0 102 101 BEAM -2.715682E-08 -2.918291E-07 .0 .0 .0 -2.549396E-09
102 102 BEAM -1.637746E-08 1.273060E-07 .0 .0 .0 -1.374512E-25
102 2 GAP 1.270000E-10 -9.144001E-27 .0 .0 .0 .0

Main Index
 Appendix C: Nonlinear Analysis 735
Nonlinear Modal Analysis

Element Strain Energy Output

1 GPF005B - FORD ELETRO-MECHANICAL CLUTCH SEPTEMBER 26, 2000 MSC Nastran 9/25/00 PAGE 231
DISPLACEMENT IN MM., FORCE IN NEWTON, STRESS IN MPA
0 SUBCASE 1
LOAD STEP = 1.00000E+00
E L E M E N T S T R A I N E N E R G I E S

ELEMENT-TYPE = BEAM * TOTAL ENERGY OF ALL ELEMENTS IN PROBLEM = 4.422530E+02

SUBCASE 1 * TOTAL ENERGY OF ALL ELEMENTS IN SET -1 = 4.422530E+02
0
ELEMENT-ID STRAIN-ENERGY PERCENT OF TOTAL STRAIN-ENERGY-DENSITY
101 1.211380E+01 2.6928 5.680606E-02
102 1.049874E+00 .2334 4.923244E-03
103 1.323984E-01 .0294 6.208644E-04
104 1.323428E-01 .0294 6.206040E-04
105 1.374425E-01 .0306 6.445181E-04
106 1.407006E-01 .0313 6.597968E-04
107 1.478413E-01 .0329 6.932822E-04
108 1.483829E-01 .0330 6.958218E-04
109 1.032130E-01 .0229 1.936014E-04
110 3.163312E-01 .0703 7.020313E-04
111 4.605607E-01 .1024 1.927717E-03
112 7.067609E-01 .1571 2.958209E-03
113 4.186914E-01 .0931 1.752469E-03
121 3.068841E+00 .6822 2.905759E-02

Remarks
1. Linear and nonlinear elements can be mixed. The output for grid point force and element strain energy in nonlinear
analysis follow the same format as in linear analysis.
2. The reference system for the grid point force output is the grid point global coordinate system. The user cannot
specify another output coordinate system. In nonlinear, the grid point forces are not aligned with element edges.
Therefore, PARAM,NOELOF and PARAM,NOELOP are ignored in nonlinear analysis.
3. In nonlinear analysis, the element strain energy must be calculated for each intermediate load step even if the output
is requested only in the last load step. To save computations, the element strain energy is only calculated upon user
request. The Case Control commands, GPFORCE or ESE must be present to activate grid point force output or
element strain energy calculations and output. If the user wants to save the grid point forces and element strain
energy on the database and wants an output later in an eventual restart run, the Case Control commands,
GPFORCE(PLOT) or ESE(PLOT), must be specified in the cold start run.
4. By default, if GPFORCE or ESE is present, then the output of grid point forces or element strain energy is only
available at the end of each subcase. For output of intermediate steps, the parameter INTOUT on the NLPARM
Bulk Data entry must be set to YES or ALL.
5. For a model with superelements, the grid point force balance is not in equilibrium along the boundary of the
residual and upstream superelements. This is because the applied load in the residual does not include the load
contributed from upstream superelements when the grid point force balance is computed. The same situation also
occurs in each upstream superelement whose applied load does not include the contribution from other
superelements and the residual. When the total of the grid force balances of all superelements is added together,
equilibrium is reached.
For example, grid point 123 is located at the boundary between the residual (superelement ID = 0) and an
upstream superelement (ID = 10) and is connected to two QUAD4 elements (ID 22 and 21). The output of the
grid point force balance is shown below.

Main Index
 736 MSC Nastran Reference Guide
Nonlinear Modal Analysis

1 GPF006A - NONLINEAR GP FORCE TEST PROBLEM SEPTEMBER 26, 2000 MSC Nastran 9/26/00 PAGE 26
Q4 ELEMENTS + MULTIPLE S.E.'S SUPERELEMENT 0
0 NONLINEARITY ANALYSIS - AUTO/NLPCI + INTOUT=YES SUBCASE 1
LOAD STEP = 1.00000E+00
G R I D P O I N T F O R C E B A L A N C E

POINT-ID ELEMENT-ID SOURCE T1 T2 T3 R1 R2 R3

0 123 APP-LOAD -3.287216E-01 1.006007E-02 -5.592173E-03 .0 .0 .0
123 F-OF-SPC .0 .0 5.127304E+00 -2.287795E-04 -1.193804E-02 .0
123 21 QUAD4 -6.765498E-01 8.791827E+00 -2.111628E+00 8.897618E-04 1.142736E-02 -3.166363E-02
123 *TOTALS* -1.005271E+00 8.801888E+00 3.010083E+00 6.609823E-04 -5.106755E-04 -3.166363E-02

1 GPF006A - NONLINEAR GP FORCE TEST PROBLEM SEPTEMBER 26, 2000 MSC Nastran 9/26/00 PAGE 33
Q4 ELEMENTS + MULTIPLE S.E.'S SUPERELEMENT 10
0 NONLINEARITY ANALYSIS - AUTO/NLPCI + INTOUT=YES SUBCASE 1
LOAD STEP = 1.00000E+00
G R I D P O I N T F O R C E B A L A N C E

POINT-ID ELEMENT-ID SOURCE T1 T2 T3 R1 R2 R3

123 22 QUAD4 1.005275E+00 -8.801861E+00 -3.010083E+00 -6.609823E-04 5.106755E-04 3.166376E-02
123 *TOTALS* 1.005275E+00 -8.801861E+00 -3.010083E+00 -6.609823E-04 5.106755E-04 3.166376E-02

Obviously, neither the residual nor the upstream superelement has the balanced grid point force. However, if both
sides are added together, the balance will be reached.
6. Computational time and memory requirement increase when grid point force balance or element strain energy
output is requested. For a large nonlinear model, allocation of sufficient memory is important. Spill logic is built in
the code, but it may increase the overall computational time.

Main Index
 Appendix D: Acoustic Cavity Modeling
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

D Acoustic Cavity Modeling


Introduction 738
 Assumptions and Limitations 739

Acoustic Cavity Example Problem 740

Main Index
 738 MSC Nastran Reference Guide
Introduction

Introduction
The Bulk Data entries used specifically for acoustic cavity analysis are described below, and the entry formats are exhibited
in the MSC Nastran Quick Reference Guide. Their purposes are analogous to the use of structural data entries. A mesh is defined
over the longitudinal cross section of an acoustic cavity, and finite elements are connected between these points to define
the enclosed volume.
The points are defined by GRlDF entries for the axisymmetric central fluid cavity and by GRIDS for the radial slots. The
GRlDF points are interconnected by finite elements via the CAXlF2, CAXlF3, and CAXlF4 entries to define a cross-
sectional area of the body of rotation. The CAXIF2 element defines the area of the cross section between the axis and two
points off the axis (the GRlDF points may not have a zero radius). The CAXIF3 and CAXlF4 entries define triangular or
quadrilateral cross sections and connect three or four GRIDF points, respectively. The density and/or bulk modulus at each
location of the enclosed fluid may also be defined on these entries.
The GRIDS points in the slot region are interconnected by finite elements via the CSLOT3 and CSLOT4 entries. These
define finite elements with triangular and quadrilateral cross-sectional shapes, respectively. The width of the slot and the
number of slots may be defined by default values on the AXSLOT entry. If the width of the slots is a variable, the value is
specified on the GRIDS entries at each point. The number of slots, the density, and/or the bulk modulus of the fluid may
also be defined individually for each element on the CSLOT3 and CSLOT4 entries.
The AXSLOT Bulk Data entry is used to define the overall parameters for the system. Some of these parameters are called
the “default” values and may be selectively changed at particular cross sections of the structure. The values specified on the
AXSLOT entry will be used if the corresponding fields on the GRIDS, CAXlFi, or CSLOTi entries are left blank. The fields
RHOD and RHO (density) and BD and B (bulk modulus) are properties of the fluid. If the value given for bulk modulus
is zero, the fluid is considered incompressible by the program. The fields MD and M (number of slots) and WD (slot width)
are properties of the geometry. The fields MD and M define the number of equally spaced slots around the circumference,
with the first slot located at  = 0 . The field N (harmonic number) is selected by the user to analyze a particular set of
acoustic modes. The pressure is assumed to have the following distribution:
p  r,z,  =  r,z  cos N
If N = 0 , the breathing and longitudinal modes will result. If N = 1 , the pressure at  = 180 will be the negative
of the pressure at  = 0 . If N = 2 , the pressures at  = 90 and  = 270 will be the negative of that at
 = 0 . Values of N larger than M/2 have no significance.
The interface between the central cavity and the slots is defined with the SLBDY Bulk Data entries. The data for each entry
consist of the density of the fluid at the interface, the number of radial slots around the circumference, and a list of GRIDS
points that are listed in the sequence in which they occur as the boundary is traversed. In order to ensure continuity between
GRlDF and GRIDS points at the interface, the GRIDF points on the boundary between the cylindrical cavity and the slots
are identified on the corresponding GRIDS entries rather than on GRlDF entries. Thus, the locations of the GRlDF points
will be exactly the same as the locations of the corresponding GRIDS points.
Various standard bulk data entries may be used for special purposes in acoustic analysis. The SPCi Bulk Data entry may be
used to constrain the pressures to zero at specified points, such as at a free boundary. The formats for these entries are
included in the MSC Nastran Quick Reference Guide. Dynamic load entries, direct input matrices, and scalar elements may be
introduced to account for special effects.

Main Index
 Appendix D: Acoustic Cavity Modeling 739
Assumptions and Limitations

Assumptions and Limitations

The accuracy of the acoustic model will be dependent on the selection of the mesh of finite elements. The assumption for
each element is that the pressure field has a linear variation over the cross section and a sinusoidal variation around the axis
in the circumferential direction. In areas where the pressure gradient changes are large, such as near a sharp corner, the
points in the mesh should be placed closer together so that large changes in flow may be defined accurately by the finite
elements.
The shape of the finite elements play an important part in the accuracy of the results. It has been observed that long, narrow
elements produce disproportionate errors. Cutting a large square into two rectangles will not improve the results, whereas
dividing the square into four smaller squares may decrease the local error by as much as a factor of ten.
The slot portion of the cavity is limited to certain shapes because of basic assumptions in the algorithms. The cross section
of the cavity normal to the axis must have a shape that is reasonably well defined by a central circular cavity having equally
spaced, narrow slots. Various shapes are shown in Figure 4-1 in the order of increasing expected error.
It is recommended that shapes such as the cloverleaf and square cross section be analyzed with a full three-dimensional
technique. The assumption of negligible pressure gradient in the circumferential direction within a slot is not valid in these
cases.
The harmonic orders of the solutions are also limited by the width of the slots. The harmonic number, N, should be no
greater than the number of slots divided by two. The response of the higher harmonics is approximated by the slot width
correction terms discussed in Section 17.1 of The MSC Nastran Theoretical Manual.

SMALL
ERRORS

LARGE
ERRORS

Figure 4-1 Modeling Errors for Various Shapes

Main Index
 740 MSC Nastran Reference Guide
Acoustic Cavity Example Problem

The output data for the acoustic analysis consist of the values of pressure in the displacement vector selected via the Case
Control command PRESSURE = n. The velocity vector components corresponding to each mode may be optionally
requested by the Case Control command STRESS = n where n is the set number indicating the element numbers to be
used for output, or by the words STRESS = ALL.
Plots of the finite element model and/or of the pressure field may be requested with the MSC Nastran plot commands. The
central cavity cross section will be positioned in the XY plane of the basic coordinate system. The slot elements are offset
from the XY plane by the width of the slot in the +Z direction. The radial direction corresponds to X and the axial direction
corresponds to the Y direction. Pressures will be plotted in the Z direction for both the slot points and the central cavity
points. The PLOTEL elements are used for plotting the acoustic cavity shape. The plot request entry SET n INCLUDE
PLOTEL must be used where n is a set number.

Acoustic Cavity Example Problem

Table 4-1 contains a listing of the data entries used as a simple example of acoustic cavity analysis. The problem to be solved
is illustrated in Figure 4-2. The model was subdivided into only ten finite elements in order to limit the number of data
entries. For reasonable engineering accuracy, this model should be subdivided into at least four times that number of
elements.

Main Index
 Appendix D: Acoustic Cavity Modeling 741
Acoustic Cavity Example Problem

1” 8” 4”
Rigid
Boundaries
2” 4” Slot
Nozzle

4” Central Cavity
Symmetry P=0
Z
A
Symmetry 14” 2”
Section A-A

Parameters:Density: RHOD= 1.2  10 – 7 lb-sec 2 /in 4

Bulk Modulus: BD= ra 2 = gRT = 21.0 lb/in 2

Harmonic: N=1
Number of Slots: MD=4

Finite Element Model:

r

GRIDS points 6 9
3
GRIDF points
1 2
1 Element IDs 5 8 3
2 12
1 11
4 7 7
8
4 5 6 13
10 9
z
Figure 4-2 Description of Acoustic Cavity Example Problem

Main Index
 742 MSC Nastran Reference Guide
Acoustic Cavity Example Problem

Each data entry in Table 4-1 is given a number on the left side. The format for each type of Bulk Data entry is given in
parentheses above the group of that type. The following is a brief description of each part of the input data:

Lines
1-4 Each statement in the Executive Control Section has the format of a request word and a selection separated
by blanks or a comma. The CEND statement is last, but the statements above may appear in any order.
Solution Sequence 103, normal modes analysis, was chosen. A limit of 2 minutes CPU time was set
(TIME2).
5-7 The TITLE, SUBTITLE, and LABEL commands may contain any list of letters and numbers following the
(=) sign. This list will appear on the first three lines of each output page.
8 The method of eigenvalue extraction is selected with the METHOD command. The number 11 refers to
the identification number of an EIGRL Bulk Data entry which appears below as line 32.
9-11 A simple output request is illustrated with these commands. PRES = ALL will result in the printout of all
pressures at the GRIDF and GRIDS points. STRESS = ALL will result in the printout of all velocities in the
elements. This printout will occur for all extracted eigenvectors. Selected points or elements can be printed
via the SET command.
12 The BEGIN BULK command denotes the beginning of the Bulk Data Section. The Bulk Data entries may
occur in any order. Putting these entries in alphabetic sort will save sorting time in large problems, however.
13 In this problem, all the parameters except slot width WD are constant throughout the volume. The values
on AXSLOT will be used whenever a corresponding field in the following entries is blank.
14-20 The location of points on the slot are defined with these entries. Entries 14, 16, 18, and 20 serve a dual
purpose by defining a GRIDS point identification number in field 2 and a GRlDF point identification
number in field 6. The two types of points thereby are forced to have the same locations at the interface.
21-22 The location of points within the axisymmetric fluid cavity are described by the GRIDF entry. No points are
allowed to have a zero or negative radius.
23-31 These entries describe the elements shown in Figure 4-2. Each element is given a unique identification
number and a list of the connected GRIDS or GRlDF points. Since the parameters p and B are constants,
these fields are left blank, so the values on the AXSLOT entry will be used.
32 The ElGRL entry is used to define parameters for eigenvalue extraction (resonant frequencies). More than
one of these entries may appear. The method to be used is selected with the METHOD
command (command 8). With this particular command, the Lanczos method was selected with three (ND
= 3) output mode shapes specified.
33 The SLBDY entry defines the boundary between the slot and the central cavity. Both the density (RHO and
the number of radial slots (M) are blank, so the AXSLOT defaults are used; i.e., RHOD = 1.2 x 10-7 and M
= 4. Only four GRIDS points are on the boundary, so a continuation entry is not necessary. Field 8 is left
blank to signify the last entry.
34 The ENDDATA entry is required to denote the end of the Bulk Data Section. Any following entries will be
ignored.

Main Index
 Appendix D: Acoustic Cavity Modeling 743
Acoustic Cavity Example Problem

Table 4-1 Acoustic Cavity Example Problem Data

Line
No.
1 $DEMO PROBLEM D03D92 WITH SOL 103 AND LANCZOS
2 SOL 103 Executive Control statements
3 TIME 2
4 CEND
5 TITLE = ACOUSTIC CAVITY EXAMPLE PROBLEM
6 SUBTITLE = FIRST HARMONIC
7 LABEL = NORMAL MODES (LANCZOS)
8 METHOD = 11
9 OUTPUT Case Control commands
10 PRES = ALL
11 STRESS = ALL
12 BEGIN BULK

Main Index
 744 MSC Nastran Reference Guide
Acoustic Cavity Example Problem

Table 4-1 Acoustic Cavity Example Problem Data (continued)

Line
No.
(AXSLOT, (RHOD, BD, N, WD, MD)

13 AXSLOT,1.2-7,21.0,1,,4

(GRIDS,ID,R,Z,W,ID

14 GRIDS,2,4.0,0.0,0.23 01,1
15 GRIDS,3,8.0,.0,1.0
16 GRIDS,5,4.0,4.0,2.0,4
17 GRIDS,6,8.0,4.0,1.0
18 GRIDS,8,4.0,8.0,2.0,7
19 GRIDS,9,8.0,8.0,1.0
20 GRIDS,12,4.0,1.2+1,2.0,11

(GRIDF,ID,R,Z )

21 GRIDF,10,2.0,12.0
22 GRIDF,13,2.0,1.4E1

(CSLOT4,1D,P1,P2,P3,P4,RHO,,M )

23 CSLOT4,1 ,2,3,6,5
24 CSLOT4,2,5,6,9,8
Bulk Data entries
(CSLOT3,1D,P1,P2,P3,,RHO,, )
25 CSLOT3,3,8,9,12

(CAXIF2,1D,P1,P2,,,RHO )
26 CAXIF2,4,1 ,4
27 CAXIF2,5,4,7
28 CAXIF2,6,7,10
29 CAXIF2,9,10,13

(CAXIF3,ID,P1,P2,P3,,RHO,B )
30 CAXIF3,7,7,10,11
31 CAXIF3,8,10,11,13

(EIGRL,SID,V1,V2,ND )
32 EIGRL,11,,,3
(SLBDY,RHO,M,ID1,ID2,ID4,etc. )
33 SLBDY,,,12,8,5,2
34 ENDDATA

Main Index
 Appendix E: Cyclic Symmetry
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

E Cyclic Symmetry


An Overview 746
 Symmetry 746

Theory 751

Main Index
 746 MSC Nastran Reference Guide
An Overview

An Overview
Symmetry has long been a subject on which philosophers and mathematicians have focused their attention. Their inquiries
into the concepts of symmetry have led to powerful results that have, in turn, led to the formulation of analysis procedures
of general applicability in applied mechanics. A familiar example is provided by the routine application of reflective or
bilateral symmetry in many engineering disciplines. The principles of reflective symmetry allow the analyst to reduce the
size of his analytical model by one-half for each plane of reflective symmetry in the system under investigation.
Reflective symmetry is only one of the types of symmetry that exist in nature. Among the other types is cyclic or rotational
symmetry in which the properties of a body are repeated at equal intervals about an axis of symmetry. Mathematicians have
developed a general theory of cyclic groups to deal with such problems, which electrical engineers have applied, in the
method of symmetrical components originated by Fortescue in 1918, to the analysis of polyphase electrical networks.
An original method for the application of the concepts of cyclic symmetry to structural analysis is available in MSC Nastran.
The method introduces no sacrifice in theoretical rigor or generality and provides substantial reductions in computer time
for the analysis of linear systems that possess rotational and/or reflective planes of symmetry. These techniques are currently
available in the following solution sequences:

SOLution 114 Superelement Static Analysis

SOLution 115 Superelement Normal Modes Analysis
SOLution 116 Cyclic Buckling
SOLution 118 Superelement Direct Frequency Response

In addition to significant savings in computer time, the analyst who grasps the concept of cyclic symmetry can realize a
considerable reduction in problem preparation tasks. For example, if a body consists of N identical segments in the sense
that the entire body can be generated through reflections and/or rotations of one (the fundamental region) of these N
segments, the analyst is only required to define a finite element model for the fundamental region and to provide a list of
the boundary grid points for this fundamental region. The finite element model may include any members of the MSC
Nastran element library other than axisymmetric elements. All boundary conditions are automatically handled internally
by MSC Nastran.
Complete generality in the definition of loading conditions is achieved by allowing the user to independently specify loads,
enforced motions, and/or temperatures on each of the N segments, even though the fundamental region is the only segment
explicitly represented in the analysis. Output requests are handled with the same degree of generality in that solutions are
provided for each of the N segments.

Symmetry
Prior to specific consideration of the cyclic symmetry capability in MSC Nastran, it is useful to briefly review the aspects
of symmetry that concerns the analyst. Specifically, the concern is with those symmetry operations that involve the
permutations of a fundamental region of a body that leave the body unchanged. Several examples are presented to illustrate
the points of interest.

Main Index
 Appendix E: Cyclic Symmetry 747
Symmetry

A-Frame (Dihedral Symmetry)

An A-frame has a plane of reflective symmetry as indicted by the dashed line in the figure below.

A reflection of the left half of the symmetric frame through the plane of symmetry results in the right half of the frame.
This familiar reflective type of symmetry is known as dihedral (DIH) symmetry. Either symmetric half of the frame may
be taken as the fundamental region.

Pinwheel (Rotational Symmetry)

An example of a rotationally symmetric body is provided by the following example in which the “leaves” of a pinwheel have
curved boundaries.

In this case, there are no planes of dihedral symmetry, but it is observed that each of the six “leaves” of the figure are identical
and that rotations of the fundamental region (for example, segment 1) through the effectively distinct multiples of
360  6 , will “generate” the complete pinwheel. This type of symmetry is known as rotational symmetry. In general,
rotational symmetry includes N symmetry operations corresponding to the effectively distinct multiples 360  N where
N is the number of identical segments in the body.

Main Index
 748 MSC Nastran Reference Guide
Symmetry

N-Gons
In more general cases, a body may be susceptible to both DIH and ROT symmetry operations. For example, consider the
three regular N-gons: the equilateral triangle, the square, and the pentagon.

D
D c C d c
C
E C
0
c b d b
0 0 e b

A a B A a B A a B

N=3 N=4 N=5

In these figures, the internal solid lines in conjunction with the edges define fundamental regions (for example, A0B) that
may be permuted through the ROT symmetry operations. The axis of rotation is normal to the plane of the page and passes
through the point 0.
In addition, the dotted internal lines such as a0 may be used to, for example, reflect A0a into a0B. Also, a0B may be reflected
through B0 to obtain B0b, and so on around each N-gon until the identity of A0a is reached. Clearly, the fundamental
region A0a for the DIH symmetry operations is smaller than the fundamental region A0B for ROT symmetry operations,
and both of these fundamental regions are substantially smaller than the fundamental region required by traditional
reflective symmetry techniques.

Three-Dimensional Bodies
The reader has undoubtedly noted that the previous examples involved planar bodies. However, the previous discussions
would not in any way be altered if these examples were assumed to be three-dimensional bodies; that is, if the illustrations
represent a cross-section of a uniform body perpendicular to the plane of the page. In fact, there are no restrictions with
regard to changes in cross-sectional geometry as a function of distance along the axis of rotational symmetry.
In the example, the fundamental region is indicated by the shaded volume.

Main Index
 Appendix E: Cyclic Symmetry 749
Symmetry

The symmetry operations are of the type DIH.

Cyclic Symmetry versus Traditional Reflective Symmetry Techniques

Engineers are generally familiar with the traditional techniques of reflective symmetry that could be employed in the
analysis of any of the previous examples other than the pinwheel. The following list compares the major steps in the analysis
procedures required with traditional reflective techniques and with cyclic symmetry techniques. The advantages of the
cyclic symmetry techniques are most vivid.
 Analytical Model
a. Traditional Techniques
The size of the analytical model may be reduced by one-half for each plane of reflective symmetry.
b. Cyclic Symmetry
The size of the analytical model is at most equal to the size of the model required by traditional analysis
techniques (A-frame). In general, the analytical model required for cyclic symmetry analysis techniques is
substantially smaller.
 Boundary Conditions
a. Traditional Techniques
In the general case of p planes of reflective symmetry, the analyst must explicitly define 2 p sets of boundary
conditions. These boundary conditions are defined through SPCi Bulk Data entries.
b. Cyclic Symmetry
The user supplies a list of the grid points on the boundary of the fundamental region. Boundary conditions are
automatically imposed by the program.
 Applied Loads
a. Traditional Techniques

Main Index
 750 MSC Nastran Reference Guide
Symmetry

The analyst must decompose the various loading conditions into symmetric and antisymmetric components.
b. Cyclic Symmetry
Applied loads are defined to act at arbitrary locations throughout the entire structure respective of whether any
or all of these locations are in the explicitly defined fundamental region.
 Computer Solution
a. Traditional Techniques
For each set of defined boundary conditions, a lengthy matrix decomposition and equation solution are
required. Static condensations are generally not economically feasible in static analysis.
b. Cyclic Symmetry
A static condensation is performed once to reduce the size of the analysis set. In static analysis, all but the
boundary degrees of freedom should be omitted. In real eigenvalue analysis, buckling analysis, and dynamic
analysis, the boundary degrees of freedom and selected internal points are kept in the analysis set. The static
condensation involves a lengthy decomposition which, however, should be less lengthy than any one of the
decompositions required by traditional techniques because the fundamental region for cyclic symmetry is
generally substantially smaller than the region required for other analysis techniques. Several additional
decompositions of relatively trivial magnitude are required after the static condensation.
 Data Recovery
a. Traditional Techniques
The results for each boundary condition must be combined to obtain the complete response of the structure.
The user must define such combinations through SUBCOM or SYMCOM Case Control commands.
b. Cyclic Symmetry
The response for the entire structure is automatically provided without any special user action. SUBCOMs and
SYMCOMs are not necessary and must not be used.
In summary, cyclic symmetry provides a uniform treatment of symmetry with a potential for substantial savings in data
preparation and solution cost.

Summary
The examples that have been discussed provide heuristic evidence that cyclic symmetry may be of significant value in
structural analysis. One advantage is that the analyst will be required to provide an analytical model for only the
fundamental regional of a structure, and this fundamental region is generally substantially smaller than the analytical model
required by more traditional analysis techniques. In addition, the comparison between cyclic symmetry and more
traditional analysis procedures provided in the previous section indicates that both user convenience and reductions in
computer time requirements are uniquely available through cyclic symmetry procedures.
The validity of these statements relies upon the development of transformation relations that allow the analyst to
manipulate the mathematical representation of the fundamental region in accordance with the appropriate symmetry
operations. In graphical terms, the following operations are performed.

Main Index
 Appendix E: Cyclic Symmetry 751
Theory

MSC Nastran input file

Step 1
for fundamental region

Mathematical representation
of the model in terms of
Step 2
physical quantities, e.g.,
Ku = P

Transformation of mathematical
representation to cyclic
Step 3 quantities, e.g.,
Ku = P

Problem solution in terms of

Step 4
cyclic quantities

Transformation of results from

Step 5 cyclic quantities to physical
quantities
Steps 3 and 5 are of particular interest because they depend on the existence of the necessary transformations. Step 4 is also
of interest in that it must be demonstrated that problem solution procedures in terms of cyclic quantities are valid and
require less computer time than more traditional solution procedures. These matters are addressed in the next section.

Theory
The theory of cyclic symmetry is presented in some detail for the ROT case, and the corresponding results for the DIH case
are summarized. This form of presentation will provide the interested user with sufficient information to independently

Main Index
 752 MSC Nastran Reference Guide
Theory

investigate many of the theoretical aspects of cyclic symmetry that cannot be adequately covered within the limits of this
discussion.

Symmetrical Components
Consider an arrangement of N identical objects symmetrically disposed about a normal. The angle between these objects is

2
a = ------ (5-1)
N
and, in addition, let U 1  r,t  be any variable displacement, force, stress, temperature, etc. in object 1, the fundamental
region. The variables r and t denote distance and time respectively. The corresponding quantity in the n-th object is U n
 r,t  and as a single valued solution is required
Un = Un + N n = 1,2,...,N

The various U n 's are, in general, unsymmetric quantities that are not related in a simple manner.

The work by Fortescue on symmetrical component theory was referenced in the introduction of this section. This theory
states that a general vector field on N-vectors with arbitrary phase relationships may be resolved into N-sets of N-
symmetrically phased vectors of equal magnitude. For N = 3, the vectors may be illustrated by the following diagram.

Main Index
 Appendix E: Cyclic Symmetry 753
Theory

x 10
U2 x 20
x 30
U1

x 31 x 11

U3

x 21

x 32

x 12
x 22
Now, in accordance with the theory of symmetrical components, the following vector sums may be stated.

U 1 = x 10 + x 11 + x 12
U 2 = x 20 + x 21 + x 22 (5-2)

U 3 = x 30 + x 31 + x 32

Here, the U n 's represent physical quantities and the x nk terms represent symmetrical components. The superscripts k
denote the order of the symmetric components. The subscripts n are taken to be the number of a particular vector in the
vector field. In the general context of cyclic symmetry, the subscripts n denote the number applied to each of the N
identical objects under consideration. The object n = 1 is arbitrarily taken as the fundamental region.

Main Index
 754 MSC Nastran Reference Guide
Theory

Transformation from Physical Components to Cyclic Components (ROT)

These transformations for the symmetric components in terms of the general field vectors may be written in summation
form for the general case of N vectors.
N
1
x 1k = ----  U n e i  n – 1 ka (5-3)
N
n=1

where a is given by equation (5-1).

Transformation from Cyclic Components to Physical Components (ROT)

After the solution phase of the problem, it will normally be necessary to transform the results from cyclic components back
into physical components.
Equation (5-2) may be written in the following form.
N–1
Un =  x 1k e – 1  n – 1 ka
k=0
= x 10 +  x 11 e – 1  n – 1 a + x 1 N – 1  e i  n – 1 a 
+  x 12 e – 1  n – 1 a + x 1 N – 2  e i  n – 1 2a  (5-4)

+ x 13 e – 1  n – 1 a + x 1 N – 3  e i  n – 1 3a
+ ,...,
n – 1
+ x 1N  2  – 1 

Here,

e –1  n – 1   N – k a = e – 1  n – 1 ka e – 1  n – 1 Na = e – 1  n – 1 ka

and
n–1
e –1  n – 1   N  2 a =  – 1 

If the exponentials in the bracketed terms in the above summation are replaced by the relations e  ix = cos x  sin x , the
typical bracketed term takes the following form:

  x 1k + x 1 N – k   cos  n – 1 ka – i  x ik – x 1 N – k   sin  n – 1 ka 

Main Index
 Appendix E: Cyclic Symmetry 755
Theory

The purpose for the somewhat arbitrary introduction of equation (5-4) is now recognized, and the typical bracketed term
assumes the following form:

 U kc cos  n – 1 ka + U ks sin  n – 1 ka 

The final form of the transformation from cyclic components to physical components is therefore,
KL
n–1 N2
Un = x 10 +   U kc cos  n – 1 ka + U ks sin  n – 1 ka  +  – 1  X1
(5-5)
k=1
n = 1, 2, 3, , N

In equation (5-5), U n represents physical quantities such as displacement force, stress, temperature in the n-th segment in
terms of the cyclic components in the fundamental region. The following points should be noted:
N – 1
1. K L = ----------------- and x 1N  2 is nonexistent when N is odd.
2
N – 2
2. K L = ----------------- when N is even.
2
3. The notation for cyclic components of order k = 0 and k = N  2 is distinct from the notation for cyclic
components of order 1  k  k L because of the somewhat peculiar nature of these two components.

Boundary Conditions and Problem Solution (ROT)

The transformation equations of Theory, 751 are assumed to allow one to transform a problem that is mathematically defined
in terms of physical components into a set of mathematical statements written in terms of cyclic components. Specifically,
the mathematical description of the n replications of the fundamental region may be written in the form
 Fn  =  Yn   Un  (5-6)

This expression represents the mathematical description of the system in terms of physical coordinates.
The application of previously defined transform equations may now be introduced to obtain the desired mathematical
description of the system in terms of cyclic components. First, however, consider the matrix representation of the transform
equations (5-3) and (5-5), respectively. Equation (5-3) assumes the form

Main Index
 756 MSC Nastran Reference Guide
Theory

1- 1- 1- 1
--- --- ---  ----
N N N N
 U1 
 x 10  2-
--- 2-
--- 2-
--- 2  
  cos a cos 2a  ---- cos  N – 1 a  U2 
N N N N
 U 1c   
  2- 2- 2 U 
 U 1s  0 --- sin a --- sin 2a  ---- sin  N – 1 a  3
N N N
    ..
 U 2c  = 2 2- 2- 2   .
..   ---- --- cos 2a --- cos 4a  ---- cos  N – 1 2a  
  N N N N   ..
. .
 K s   
U L    ..
 N2 2 2 2   .
 x1  0 ---- sin K L a ---- sin 2K L a  ---- sin  N – 1 K L a U 
N N N
 N
1- ---1- 1- 1 N–1
--- ---  ----  – 1 
 U k  =  T xU  NU n–N N N

Equation (5-5) assumes the form

 U1  1 1 0 1  0 1 x 01 
   
 U2  1 cos a sin a cos 2a  sin K L a –1
 U 1c 
   
 U3  1 cos 2a sin 2a cos 4a  sin 2K L a 1  
   1s 
  =  U 
   
   
   K S
  U L 
U  N – 1 
1 cos  N – 1 a sin  N – 1 a cos  N – 1 2a  sin  N – 1 K L a  – 1  N2 
 N  x1 
 Un  =  T Ux   U k 
(5-7)
It is easily demonstrated that
 T xU   T Ux  =  I 

Return now to equation (5-6), the mathematical representation of the system in terms of physical coordinates, and represent
the energy of the system in the form

U =  Un  T  Fn  (5-8)

From equations (5-6), (5-7), and (5-8)

Main Index
 Appendix E: Cyclic Symmetry 757
Theory

  U T  Y   U 
 n n n
U = 
  U k T  T T  Y   T   U k 
 Ux n Ux

The system energy in terms of cyclic components may be written as

T
U =  Uk   Yk   U k  (5-9)

Since the energy must be invariant to the transformations, equations (5-9) and (5-10) may be equated
T
 Y k  =  T Ux   Y n   T Ux  (5-10)

That is, the equations of the system in terms of physical components are transformed into a set of equations in terms of
cyclic components. As there is no coupling between the N identical objects of the system, the matrix Y n may be written
so that it consists of N identical blocks along the diagonal. These blocks represent, in general, the equations of motion of
each of the N identical objects of the system.

As an example, assume N = 4  K L = 1  , and expand equation (5-10). Note that Y 1 = Y 2 = Y 3 = Y 4 .

1 1 1 1 1 1 0 1
 Y k  = Y 1 1 cos a cos 2a cos 3a
I
1 cos a sin a –1
0 sin a sin 2a sin 3a 1 cos 2a sin 2a 1
1 –1 1 –1 1 cos 3a sin 3a –1
40 00
= Y1 0 2 00
00 20
00 04

In general, for N objects the above equation may be written

Y k = Y 1 N for k = 0, N  2
N
Y k = ---- Y 1 for 1  k  K L
2

These equations imply that the equations of motion in terms of cyclic components are either N or N  2 times the
corresponding equations written in terms of physical components. This argument provides an explanation for the presence
of the coefficients that appear in the terms of equation (5-6).

Main Index
 758 MSC Nastran Reference Guide
Theory

To this point, it has been found that except for boundary conditions, the equations for motion in terms of cyclic
components ( k = 0, 1, ..., N  2 ) may be developed from the equations of motion of the fundamental region ( n = 1 ).
In essence, an arbitrarily large physical system may be investigated in terms of a fundamental region which has been
transformed into a number of models defined in terms of cyclic components. The boundary conditions between cyclic
models must now be defined.
The only connection between the n identical objects in question exists along the boundaries between adjacent segments.
Specifically, consider three of the N objects as indicated in the following figure.

j + 1

u 1,j + 1 Side 1

u 2,j Side 2
u 1,j
j

u 2,j – 1 Side 1

Side 2
j – 1

These objects are displaced from one another in order to more clearly define the boundary conditions between the objects.
Note the designation of the two sides of each of the elements as side 1 and side 2 as one proceeds in the counterclockwise
direction. From the requirement of displacement compatibility, it is necessary that the typical boundary degrees of freedom,
u , satisfy the equations
u 2,j – 1 = u 1,j
u 2,j = u 1,j + 1

These physical degrees of freedom may be transformed into cyclic components of order K through equation (5-3).

Main Index
 Appendix E: Cyclic Symmetry 759
Theory

1
 x k  side 1 = ---- u 1,j e i  j – 1 ka
N
(5-11)
1
 x k  side 2 = ---- u 1,j + 1 e ijka
N
A comparison of equation (5-11) provides the result

 x k  side 2 =  x k  side 1 e ika (5-12)

for 1  k  k L . For k = 0 ,

 x k  side 2 =  x k  side 1 (5-13)

N
and for k = ---- ,
2

 x k  side 2 = –  x k  side 1 (5-14)

The cyclic component of equation (5-14) may be written in terms of the variables U kc and U ks in accordance with the
definitions.

 U kc  side 1 =   x k  side 1 +  x N – k  side 1 

 U ks  side 1 = – i   x k  side 1 –  x N – k  side 1 
 U kc  side 2 =   x k  side 2 +  x N – k  side 2 
 U ks  side 2 = – i   x k  side 2 –  x N – k  side 2 
Equation (5-12) may now be introduced to yield the equations

 U kc  side 2 =   x k  side 1 e ika +  x N – k  side 1 e – ika 

 U kc  side 2 = – i   x k  side 1 e ika –  x N – k  side 1 e – ika 
Upon the replacement of the exponentials with Euler's equation, the appropriate boundary conditions may be represented
as follow:
kc kc
  U  side 2  cos ka – sin ka   U  side 1 
 ks  =  ks  (5-15)
  U  side 2  sin ka cos ka   U  side 1 

Main Index
 760 MSC Nastran Reference Guide
Theory

These boundary conditions are of the form of constraint equations in which the variables on side 2 are dependent and the
variables on side 1 are independent. Indeed, these equations may be viewed as MPC equations that couple kc and ks
terms of side 1 to the corresponding terms of side 2.
The equations of motion in terms of cyclic components may be written in the following form.

N  x0   F0 
----    
2  U 1c   F 1c 
N    
 U 1s   F 1s 
N    
N  U 2c   F 2c 
   
 Y1  .  .  =  . 
   
.  .   . 
.  .   . 
   
N  UKLs   FKLs 
   
N
----  x N  2   FN  2 
2    
From the previous discussion on boundary conditions, the above equation may be rewritten in terms of the following
disjoint problems.

k = 0  Y0   X 0  = F 0

Y K 0  U kc   F kc 
1  k  KL  ks  =  ks 
0 Yk  U  F 
N  YN  2   x N  2  =  F N  2 
k = ----
2
The boundary equations (5-13), (5-14), and (5-15) must be introduced to remove the dependent variables that are defined to
be on side 2 of the region. If, in addition, the user has taken advantage of the OMIT feature in MSC Nastran, the only
independent degrees of freedom in the above disjoint problems are the degrees of freedom on side 1 that have not previously
been declared dependent through MPC, SPCi, and/or OMITi Bulk Data entries. Thus, the matrices in the above disjoint
problems, although dense, are of small size relative to, say, the matrices required to perform the same analysis by more
traditional methods.

Summary of Transformations, Boundary Conditions, and Problem Solution (DIH)

Dihedral symmetry (DIH) is an advantageous option in the two general cases indicated below.

Main Index
 Appendix E: Cyclic Symmetry 761
Theory

1. DIH takes advantage of reflective symmetry to reduce the fundamental region by one-half relative to the size of the
fundamental region required in the ROT option.
2. The DIH option can be utilized for analyses that would normally be handled with traditional reflective symmetry
techniques. The A-frame that was discussed above is an example of this situation.
The form of the transformations between cyclic and physical components and the boundary conditions between the
segments of the object in question are, to some extent, dependent on the notation system. Figure 5-1 defines the notation
system used in the DIH option in MSC Nastran.
Segment 2L

Segment 2R

Segment 1L
Side 1
Side 2
Side 2 Segment 1R
Side 1 (Fundamental Region)

Plane of Overall
T2 T2 Reflective Symmetry
T1
Segment NL

Figure 5-1 Example of Dihedral Option

Note that:
 The midplanes or relative planes are always denoted as side 2. The other boundary, which must also be planar, is
always denoted as side 1. In the specific case of a structure with a single overall plane of reflective symmetry (that
is, the complete structure is represented by segments 1R and 1L) the reflective plane may be defined as side 1 or
side 2.
 The two halves of a segment that join at a reflective plane are denoted as the right, “R”, and left, “L”, halves.
 The fundamental region is taken as 1R.
 R halves use a right-handed coordinate system and L halves use a left-handed coordinate system. Therefore, at each
boundary a degree of freedom is associated with its mirror image on the adjacent segment so that compatibility
always involves a left- and a right-hand coordinate system. The degrees of freedom on the boundaries are further
categorized into odd and even components in accordance with the following table.

Component Degrees of Freedom

Even Displacement vectors parallel to the boundary and rotation vectors normal to the boundary.
Odd Displacement vectors normal to the boundary and rotation vectors parallel to the boundary.

Main Index
 762 MSC Nastran Reference Guide
Theory

The several equations of interest for the DIH option are indicated below along with the equation number for the
corresponding expression under the ROT option. In the following equations, the starred (*) terms donate anti-symmetric
motions and unstarred terms denote symmetric motions. The subscripts R and L refer to the right and left halves
respectively.

Transformation from Physical Components to Cyclic Components (DIH)

(See equation (5-5) for the corresponding equations for the ROT option.)

U kc  
 = ----   U n,R cos  n – 1 ka  U n,L cos  N – n ka 
U kc *  N
n

U ks  -
ks
 = ---
N   Un,R sin  n – 1 ka  Un,L sin  N – n  ka 
U * n
Here,

 1  2 for k = 0, N  2
 = 
 1 for 1  k  K L
and the positive sign is associated with the unstarred terms.

Transformation from Cyclic Components to Physical Components (DIH)

(See equation (5-15) for the corresponding relation for the ROT option.)

U n,R =   U kc + U kc *  cos  n – 1 ka +  U ks + U ks *  sin  n – 1 ka

k

U n,L =   U kc + U kc *  cos  N – n ka +  U ks + U ks *  sin  N – n ka

k
N
0  k  ----
2

Boundary Conditions and Problem Solution (DIH)

(See equation (5-5) for the corresponding boundary conditions for the ROT option.)
Side 1, EVEN

Main Index
 Appendix E: Cyclic Symmetry 763
Theory

U kc * = 0
U ks = 0
Side 1, ODD

U kc = 0
U ks * = 0
Side 2, EVEN
ka
U kc sin ------ + U ks cos ka
------ = 0
2 2
ka
U kc * cos ------ – U ks * sin ka
------ = 0
2 2
Side 2, ODD
ka ka
U kc cos ------ – U ks sin ------ = 0
2 2
ka ka- = 0
U kc * sin ------ + cos -----
2 2
In static analysis with cyclic symmetry, the equation
K K K
 K aa   u a  =  Pa 

K
is solved in SSG3 (Static Solution Generator – Part 3) for  u a 

K T K G  + GT K G 
 K aa  =  G ck aa ck sk aa sk

and
K T P KC T P ks
 Pa  =  G ck +  G sk

are formed in CYCLIC3 (Cyclic Symmetry Transformation – Module 3).

kc ks x
The symmetrical components of the loads  P  and  P  are subsets of  P  which is obtained in CYCLIC3 by
multiplying the load vector formed in SSG1 by the GFORE matrix formed in CYCLIC1 module
x
 P  =  P l   GFORE 

Main Index
 764 MSC Nastran Reference Guide
Theory

where  P l  is formed in SSG2 (Static Solution Generator – Part 2) and the transformation matrix [GFORE] is formed
in CYCLIC1 module.
The cyclic component constraint data (CYCD) used in CYCLIC3 for the formulation of the transformation matrices
 G ck  and  G sk  are generated in CYCLIC2 (Cyclic Symmetry – Module 2).

K
The solution  u a  is transformed to symmetric components in CYCLIC4 as follows:

kc K
 ux  =  G ck   u a 
kc K
 ux  =  G sk   u a 

x
The symmetric components for each harmonic index are appended to form  u  in the CYCLIC4 module. The physical
segment displacements are obtained as follows:
x
 u a  =  u   GBACK 

where the transformation matrix is formed in the CYCLIC1 module.

In the case of vibration analysis with cyclic symmetry, the following equation is used in READ to determine the eigenvalues
and eigenvectors,
K
 K aa – M aa   u  = 0

K
where  M aa  is formed in a manner similar to  K aa  in CYCLIC3. The symmetrical components of the eigenvectors
are recovered in CYCLIC4 and the physical segment data is recovered using the GBACK1 matrix created in CYCLIC4
module.
Since all of the transformation matrices are extremely sparse, none of the matrix multiplications indicated above will require
large amounts of computer time. The most significant operation is the triple product associated with the determination of
K
 K aa  . In the usual application this matrix is dense and approximately a-size. The computer time for the triple product
can be estimated as ten matrix packing operations on a full matrix of a-size.

Using Cyclic Symmetry

Many structures, including pressure vessels, rotating machines and antennas for space communications, are made up of
virtually identical segments that are symmetrically arranged with respect to an axis. As shown in Figure 5-2 and Figure 5-3,
there are two types of cyclic symmetry: simple rotational symmetry, in which the segments do not have planes of reflective
symmetry and the boundaries between segments may be general doubly-curved surfaces; and dihedral symmetry, in which
each segment has a plane of reflective symmetry and the boundaries between segments are planar. The use of cyclic

Main Index
 Appendix E: Cyclic Symmetry 765
Theory

symmetry allows the user to model only one of the identical substructures. There will also be a large saving of computer
time for most problems.
The total model consists of N identical segments which are numbered consecutively from 1 to N. The user generates the
model for one segment, using regular elements and standard modeling techniques. All other segments and their coordinate
systems are automatically rotated to equally spaced positions about the polar axis by MSC Nastran. The boundaries must
be conformable; that is, when the segments are put together, the grid points and the displacement coordinate systems of
adjacent segments must coincide. This is easiest to ensure if a cylindrical or spherical coordinate system is used, but such is
not required.
The two boundaries are called Sides 1 and 2. In the case of rotational symmetry, Side 2 of segment n is connected to Side
1 of segment n+1, as shown in Figure 5-1. In the case of dihedral symmetry, Side 1 is on the boundary of the plane of overall
reflective symmetry, and Side 2 is on the midplane or reflective plane of the segment, as shown in Figure 5-3. In either case,
the grid point numbers on Sides 1 and 2 must be specified on the Bulk Data entry, CYJOlN.
The type of symmetry (rotational, dihedral, or axisymmetric) and the number of segments in the model are specified on
the Bulk Data entry CYSYM. Grid points that lie on the axis of symmetry must be listed on the CYAX Bulk Data entry.
The harmonics to be used in the cyclic symmetry analysis are specified on the HARMONICS Case Control command.
The HARMONICS command must be present in Case Control, and it must appear above the subcase level. In buckling
analysis the program limits the static preload to the zero harmonic and the information on the HARMONICS command
is used only for the eigenvalue analysis.
No restrictions are placed on the use of the single-point constraint, the multipoint constraint, or the OMIT feature.
Constraints between segments are automatically applied to the degrees of freedom at grid points specified on CYJOIN Bulk
Data entries which are not otherwise constrained. The SPCD Bulk Data entry may be used to vary the magnitude of
enforced displacements for each of the segments.
In the case of static analysis, the OMIT feature may be used to remove all degrees of freedom at internal grid points without
any loss of accuracy. Since this reduction is applied to a single segment prior to the symmetry transformations, it can greatly
reduce the amount of subsequent calculation. In the case of vibration analysis, the OMIT feature is used in the usual way
to reduce the size of the analysis set and involves the usual approximations. The SUPORTi entries for the free bodies cannot
be used with cyclic symmetry.
Provision is made for the use of rigid body supports with the use of the CYSUP Bulk Data entry. All supported degrees of
freedom for a rigid body must be specified at a single grid point. Additional restrictions on the location of the support point
and the uses of coordinate systems are given on the CYSUP Bulk Data entry.
Special provision is made in DIH-type problems for reflective symmetry about one or two planes. Symmetrical or
antisymmetrical boundary conditions are specified on the DSYM Case Control command. A single plane of reflective
symmetry must be with respect to Side 1, and if a second plane of reflective symmetry exists, it must be perpendicular to
Side 1. In the case of two planes of symmetry, the model must contain an even number of whole segments.
Static loads are applied to the model with a LOADCYH (harmonic load), LOADCYN (physical segment load), or
LOADCYT (AXI option) Bulk Data entry. Only a single subcase is needed for each loading condition, and it must contain
the keyword LOAD, TEMP(LOAD), or DEFORM to select a LOADCYH, LOADCYN, or LOADCYT Bulk Data entry.
The LOADCYH and LOADCYN entries in turn reference FORCE, MOMENT, PLOAD, SLOAD, SPCD, DEFORM, and
TEMP entries. The LOADCYH entry may also reference an RFORCE or GRAV entry in order to specify centrifugal or

Main Index
 766 MSC Nastran Reference Guide
Theory

gravity loading, respectively. The LOADCYT entry is used only with the AXI option to specify loads as a function of
azimuth angle.
The NOUTPUT Case Control command is used to request physical output within a subcase for all output quantities, such
as forces and displacements, requested in the subcase. The HOUTPUT Case Control command is used to request
harmonic output within a subcase for all output quantities requested in the subcase. In the case of thermal or deformation
loading, element force and stress output can only be requested consistent with the form of the loading, i.e., physical output
if the loads are defined with a LOADCYN Bulk Data entry and harmonic output if the loads are defined with a LOADCYH
Bulk Data entry.
The use of a single subcase for each loading condition introduces the need to refer to “coded” subcases for SORT2 output
and plotting purposes. Since a single subcase now refers to NSEG segments (in the case of ROT and AXI options) and two
times NSEG segments (including Right and Left halves, in the case of DIH option), a unique coded subcase ID is necessary
to refer to a particular segment of the structure for plotting purposes in PLOT commands wherever a subcase ID is to be
supplied. This coded ID is obtained as follows:

Coded Subcase ID for a Particular Segment

= SUBCASE ID  10000 + SEGMENT ID  10 + zero (for ROT or AXI option)

 1  if R segment  
= SUBCASE ID  10000 + SEGMENT ID  10 +  
 2  if L segment  

Coded Subcase ID for a Particular Harmonic

1  if component = C  
 
2  if component = S  
= SUBCASE ID  10000 + HARMONIC ID  10 +  
3  if component = C*  
4  if component = S*  


Solution of Axisymmetric Problems

The Cyclic Symmetric Solution Sequences provide a convenient and effective procedure for solving axisymmetric problems
without the need for special axisymmetric elements. The following procedure is recommended for the solution of static load
problems:
1. Select the AXI symmetry option on the CYSYM Bulk Data entry.
2. Select a large number of segments (for example, 180) on the CYSYM Bulk Data entry.
3. Construct a model with two rows of grid points and a single row of elements in the meridional direction.

Main Index
 Appendix E: Cyclic Symmetry 767
Theory

4. Define a set of harmonics on a SET Case Control command. The set of harmonics should be reasonably small for
efficiency, but sufficient in size to yield reasonably accurate results.
5. Select the defined set of harmonics on the HARMONICS Case Control command. Convergence of the harmonic
output towards zero for the higher harmonics will verify that the number of harmonics specified on the
HARMONICS Case Control command and used in the analysis is sufficient.
6. Select a static loading condition with the LOAD, TEMP(LOAD), and DEFORM Case Control commands which
reference LOADCYN, LOADCYH, or LOADCYT Bulk Data entries.
7. Place static loads only on Side 1 or on elements. The boundary conditions will take care of loading on Side2.
8. Prepare the referenced TABLED1 entries if LOADCYT entries are used to specify loads as a function of azimuth
angle. The number of x-y pairs in a given table should be sufficient to produce reasonably accurate results when
linearly interpolated.
9. Select the physical output for a reasonably small number of segments with the NOUTPUT Case Control command
or ALL of the harmonic output with the HOUTPUT Case Control command.
The following procedure is recommended to obtain the vibration modes of an axisymmetric structure:
1. Select the AXI symmetry option on the CYSYM Bulk Data entry.
2. Select a large number of segments (for example, 90) on the CYSYM Bulk Data entry.
3. Construct a model with two rows of grid points and a single row of elements in the meridional direction.
4. Define a set of harmonics on a SET Case Control command where the number of harmonics is reasonably small
for efficiency, but sufficient to cover the frequency range of interest.
5. Select the defined set of harmonics on the HARMONICS Case Control command.
6. Select ALL of the harmonic output with the HOUTPUT Case Control command. This will give the eigenvectors,
in a concise form, which may be interpreted by the formula

U    = U kc cos k + U ks sin k
where  is the azimuth angle measured from Side 1 of the fundamental region and K is the harmonic index.

Main Index
 768 MSC Nastran Reference Guide
Theory

Segment 2

Segment 1
Side 2

Side 1

T2
Segment N
T1

Conformable Interface

1. The user models one segment.

2. Each segment has its own coordinate system which rotates with the segment.
3. Segment boundaries may be curved surfaces. The local displacement coordinate system must conform at the joining
points. The user gives paired list of points on Side 1 and Side 2 which are to be joined.

Figure 5-2 Rotational Symmetry

Segment 2L

Segment 2R

Side 1
Segment 1L

Side 2 Midplane
Side 2 Segment 1R
(Fundamental Region)
Side 1

Plane of Overall
T2 T2 Reflective Symmetry
T1
Segment NL

1. The user models one-half segment (an R segment). The L half segments are mirror images of the R half segments.
2. Each half segment has its own coordinate system which rotates with the segment. The L half segments use left-hand
coordinate systems.
3. Segment boundaries must be planar. Local displacement systems axes, associated with the inter-segment boundaries,
must be in the plane or normal to the plane. The user lists the points on Side 1 and Side 2 which are to be joined.

Main Index
 Appendix E: Cyclic Symmetry 769
Theory

Figure 5-3 Dihedral Symmetry

Main Index
 Appendix F: Local Adaptive Mesh Refinement
MSC Nastran Implicit Nonlinear (SOL 600) User’s GuideMSC Nastran Reference Guide

F Local Adaptive Mesh

Refinement
 Introduction 771
 Adaptive Mesh Refinement Loop 771

Refinement by Regular Subdivision 773

Transference of Analysis Data Between Unrefined and Refined Meshes 798

Main Index
 Appendix F: Local Adaptive Mesh Refinement 771
Introduction

Introduction
The basics of local adaptive meshing is documented in Chapter 18: Adaptive Meshing of the MSC Nastran Linear Static Analysis
User’s Guide. A more detailed explanation of the local meshing refinements are covered in this chapter.

Adaptive Mesh Refinement Loop

During the adaptive mesh refinement process, a sequence of analysis supported over a sequence of different finite element
meshes is sequentially performed within an automatic loop (Figure 6-1).

Figure 6-1 Adaptive Mesh Refinement Loop

This adaptive mesh refinement loop can be summarized as follows:

1. The user inputs an initial, preferably coarse, finite element mesh.
2. An analysis is run and a finite element solution (supported on the current mesh) is computed.
3. Some elements of the previous mesh are scheduled or marked for refinement. These elements are chosen according
to a user specified adaptivity criterion and implicit refinement propagation rules.
4. The elements scheduled for refinement are refined and a new finite element mesh with new elements and grid points
are thus created.
5. Element properties, boundary conditions, constraints and loads are transferred or mapped from the previous mesh
to the new mesh.
6. Steps 2 to 5 are repeated until a termination criterion is met.

Main Index
 772 MSC Nastran Reference Guide
Adaptive Mesh Refinement Loop

Table 6-1 schematically illustrates the first two iterations of the adaptive mesh refinement loop. Elements scheduled for
refinement due to the user specified adaptivity criterion are depicted in green. Notice that there are neighboring elements
that are also refined during the process (yellow elements). Implicit rules to propagate the refinement from elements meeting
the user specified criterion to their neighbors are explained in Propagation of Refinement, 793.

Table 6-1 First Two Iterations of the Adaptive Mesh Refinement Loop
First Iteration Second Iteration
1. Initial Mesh

2. Analysis ... ...

3. Mark for refinement

4. Refine

5. Transfer ...

Main Index
 Appendix F: Local Adaptive Mesh Refinement 773
Refinement by Regular Subdivision

Refinement by Regular Subdivision

Mesh refinement in MSC Nastran is accomplished by the so-called regular or isotropic subdivision of a subset of parent
elements into offspring or children sub-elements. Figure 6-2 illustrates the subdivision rules for individual elements of the
different types:

Line Elements
CTRIA*

CQUAD*

CHEXA

CTETRA

CPENTA

Figure 6-2 Regular (isotropic) subdivision rules for individual elements of different types

These subdivision rules are called regular or isotropic because all edges in the boundary of a refined element are subdivided
into the same number of segments (two) as opposed to, for example, the subdivision of a quadrilateral or a triangular
element by bisection (Figure 6-3).

Main Index
 774 MSC Nastran Reference Guide
Refinement by Regular Subdivision

Figure 6-3 Isotropic vs. Anisotropic Subdivision of a Quadrilateral and a Triangular Element

For a tetrahedron, there are three possible regular subdivision schemes into eight children tetrahedra. These three schemes
are obtained as follows: first, each edge is bisected. This defines four corner tetrahedra and an internal octahedron. Then,
the latter might be subdivided into four additional tetrahedra in three different ways, according to each of its three
diagonals.
In MSC Nastran, the diagonal selected to subdivide the internal octahedron is the one connecting node 7 (mid-node of the
edge 1-3) to node 9 (mid-node of edge 2-4) as depicted in Figure 6-4. Furthermore, corner nodes of the children tetrahedra
are numbered according to the special rule illustrated in Figure 6-4. This special labeling convention of nodes along with the
selection of the diagonal connecting nodes 7 to 9 to subdivide the inner octahedron ensures minimization of element
distortion with successive refinements.

Figure 6-4 Labeling Convention for Corner Nodes of Children Tetrahedra

Main Index
 Appendix F: Local Adaptive Mesh Refinement 775
Refinement by Regular Subdivision

During subdivision, the following properties are preserved:

1. Preservation of element type: Quadrilateral elements are subdivided into quadrilateral elements, triangular elements
into triangular elements, etc., as opposed to, for example, subdivision of a quadrilateral into triangles or subdivision
of a triangle into quadrilaterals (Figure 6-5).

Figure 6-5 Preservation of Element Type during Subdivision

2. Preservation of element orientation (Figure 6-6): The outlining nodes of children elements will be listed in clockwise
(respectively counterclockwise) order when the father element have been defined in clockwise (respectively
counterclockwise) order.

Figure 6-6 Preservation of Orientation during Subdivision

Similarly, linear (3-noded) triangular elements (CTRIA3, CTRIAR) will be subdivided into four linear (3-noded)
triangular elements whereas quadratic (6-noded) triangular elements will be subdivided into four quadratic
triangular elements (Figure 6-7). The linear case requires the creation of three new grid points whereas the quadratic
case involves the creation of nine new grid points, two on each edge and one on each internal edge.

Figure 6-7 Preservation of Interpolation Order in Triangular Elements during Subdivision

Main Index
 776 MSC Nastran Reference Guide
Refinement by Regular Subdivision

The same rule applies to 3-D elements (CTETRA, CPENTA, CTRIA); that is, a linear tetrahedron (4-noded),
pentahedron (6-noded) or hexahedron (8-noded) will be respectively subdivided into eight linear tetrahedra,
pentahedra or hexahedra requiring respectively the creation of 6, 12, and 19 new grid points (Figure 6-8).

Figure 6-8 Preservation of Interpolation Order in 3-D Linear Elements during Subdivision

Location of New Grid Points

In linear elements, new edge nodes are placed at the mid-side of the (straight) edge. Similarly, new face nodes of linear
quadrilateral surface elements (CQUAD4, CQUADR) or quadrilateral faces of linear pentahedral and hexahedral elements (6-
noded CPENTA or 8-noded CHEXA) or new internal nodes of linear tetrahedral or hexahedral elements (4-noded CTETRA
or 8-noded CHEXA) are placed at the baricenter of the surface element, face or 3-D element, that is, at the position obtained
by averaging the position of the corner nodes (Figure 6-9).
Mid-edge nodes are placed at the mid-side of the straight edge, mid-face nodes at the baricenter of the face, and internal
nodes at the baricenter of the element.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 777
Refinement by Regular Subdivision

Figure 6-9 Location of New Mid-edge and Mid-face Nodes in a Linear Element

In quadratic elements new nodes are positioned by making use of the isoparametric mapping. The parametric space of the
element is uniformly bisected, and mid-edge and mid-face nodes are mapped back to the physical space using the element
(isoparametric) shape functions (Figure 6-10).

Figure 6-10 Uniform Subdivision of the Parametric Domain and Resulting Subdivision in Physical Space

No special provisions are taken during refinement of very distorted quadratic elements. In this case, the user should expect
distorted children elements (Figure 6-11).

Main Index
 778 MSC Nastran Reference Guide
Refinement by Regular Subdivision

Figure 6-11 Subdivision of a Distorted Quadratic (Quadrilateral) Element

The default method of placement of mid-edge nodes on mid-side edges might render inaccurate solutions when the initial
mesh provided by the user is very coarse and the boundary of the domain of analysis is therefore poorly approximated.
For example, consider the analysis of a circular planar shell subjected to compression and initially discretized with very few
elements as illustrated in Figure 6-12. If mid-edge nodes are placed on the mid-side of edges, inaccurate results are obtained
because the circular domain remains poorly approximated during all mesh refinement cycles. Default location for mid-edge
nodes is on the mid-side of edges.

Figure 6-12 Compression of a Circular Planar Shell

To address this inaccuracy, the user can request to place mid-edge nodes on a smooth approximation of the analysis domain
boundary interpolated from the initial mesh.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 779
Refinement by Regular Subdivision

Given the initial mesh, a smooth curve is interpolated using the nodes located on the mesh boundary to approximate the
analysis domain boundary. Then, mid-edge nodes are projected onto this smooth approximation. Figure 6-13 depicts this
alternative for the case of the compressed circular shell example shown in Figure 6-12 with the projection of mid-edge nodes
onto a smooth approximation of the geometric boundary interpolated from the initial mesh.

Figure 6-13 Projection of Mid-edge Nodes onto a Smooth Approximation of the Geometric Boundary Interpolated from the
Initial Mesh

It’s important to note that the smooth approximation of the boundary is computed using the boundary nodes of the initial
mesh provided by the user and that the accuracy of this approximation is determined by the coarseness of this initial mesh.
The success and accuracy of this smooth boundary approximation depends also on appropriate detection of corners and
edges. In order to identify corners and edges, the initial mesh is preprocessed using an automatic Geometric Feature Detection
Algorithm, see Detection of Material, and Superelement Interfaces, 811.
The alternative method of projecting edge-nodes onto a smooth approximation of the mesh boundary is available only for
edge nodes belonging to edge boundaries of 2-D and 3-D geometries. No repositioning for edge and face nodes belonging
to 3-D surfaces and face-boundaries of D geometries is supported.

Hanging Nodes and Multipoint Constraints on Hanging Nodes

When an element is refined (subdivided) but its adjacent elements are not refined a non conforming mesh is generated.
Nodes created on the boundary between a refined and a non refined element are referred to as hanging nodes (Figure 6-14).

Main Index
 780 MSC Nastran Reference Guide
Refinement by Regular Subdivision

Figure 6-14 Hanging Node

Displacement on hanging nodes need to be constrained or tied to the displacement of corner nodes to avoid a discontinuity
in the displacement field, as illustrated in Figure 6-15:

Figure 6-15 Displacement Field over an Incompatible Mesh created due to the Presence of an Unconstrained Hanging Node

All degrees of freedom (1 to 6) associated to a hanging node are automatically constrained using internal Multipoint
Constraint (MPC) equations derived from the isoparametric mapping.
Figures 6-16 and 6-17 depicts the MPC equation for hanging nodes laying respectively on a linear and a quadratic edge.
Notice that in the linear case, the MPC equation ties each component of the hanging node displacement U M with those
corresponding to the corner nodes 1 and 2 whereas in the quadratic case each components of the hanging node
displacements U M needs to be tied to the corresponding displacements of both the corner nodes 1, 2 and the mid-edge
node 3.

Figure 6-16 Constraint Equations for Hanging Nodes laying on a Linear Edge

Main Index
 Appendix F: Local Adaptive Mesh Refinement 781
Refinement by Regular Subdivision

Figure 6-17 Constraint Equations for Hanging Nodes laying on a Quadratic Edge

Figures 6-18 and 6-19 show the MPC equations for hanging nodes laying respectively on a linear and a quadratic face. In
the linear case, the MPC equation ties each component of the hanging node displacement with those corresponding to the
corner nodes 1, 2, 3 and 4 whereas in the quadratic case, each component of the hanging node displacements needs to be
tied to the corresponding displacements of both the corner nodes 1, 2, 3, 4 and the mid-edge nodes 5, 6, 7 and 8.

Figure 6-18 Constraint Equation for Hanging Nodes laying on a Linear Face

Figure 6-19 Constraint Equation for Hanging Nodes laying on a Quadratic Face

An example illustrating the need of enforcing constraints on hanging nodes is depicted in Figures 6-20, 6-21, and 6-22
concerning the deformation of a cylindrical shell subjected to a central concentrated load. Figure 6-20 shows the finite
element mesh after the third iteration in the adaptive mesh refinement loop along with the hanging nodes involved in MPC
equations. Figures 6-21 and 6-22 compare the deformed configuration and von Mises stresses obtained when the hanging

Main Index
 782 MSC Nastran Reference Guide
Refinement by Regular Subdivision

node constraints are enforced (left) or not enforced (right). Notice when the hanging node constraints are not enforced, an
incompatible configuration is obtained.

Figure 6-20 Multipoint Constraints on Hanging Nodes

Figure 6-21 Compatible (left) and Incompatible (right) Deformations obtained when Multipoint Constraints are Enforced
(left) or not (right)

Main Index
 Appendix F: Local Adaptive Mesh Refinement 783
Refinement by Regular Subdivision

Figure 6-22 Compatible (left) and Incompatible (right) von Mises Stresses obtained when Multipoint Constraints are
Enforced (left) or not (right)

Selection of Elements to Refine

Elements that will be refined during a given iteration in the adaptive loop are selected in two steps; first, all elements meeting
the user-specified adaptivity criterion are searched for and scheduled for refinement. Second, some of the elements adjacent
to the latter are also scheduled for refinement according to a set of implicit propagation rules. If no elements meeting the
user-specified criterion are found, the adaptivity loop is terminated.
MSC Nastran currently supports four refinement criteria (see Refinement Criteria, 784):
1. Error indicator based criterion
2. Nodes within a spherical spatial region criteria
3. Nodes within an orthogonal spatial region criteria
4. Nodes in contact criteria
The set of implicit refinement propagation rules are the following (see Propagation of Refinement, 793):
1. Horizontal propagation (2-to-1 rule)
2. Horizontal propagation from inner to outer children in triangular and tetrahedral elements
3. Vertical propagation
4. Propagation across superelement boundaries.

Main Index
 784 MSC Nastran Reference Guide
Refinement by Regular Subdivision

Refinement Criteria
The mesh refinement criteria available in MSC Nastran are the following:

Error Indicator Based Criterion

In this case an error indicator E e is computed over each element ‘e’ in the finite element mesh. Then, an element is refined if

2 2
E e  fE

where f is a scalar factor such that 0  f  1 and chosen by the user (as part of the Bulk Data entry HADACRI) and E is
the quadratic mean of the error indicator defined as:
N
2 1 2
E = ----  Ee
N
e=1
with N the total number of elements in the element set where element ‘e’ belongs, Thus, an element is refined if its estimated
error is larger than a fixed percentage of the quadratic mean.
Figure 6-23 schematically depicts a mesh with its corresponding elemental error indicator over a 1-D mesh (shown in blue)
with N elements and corresponding quadratic mean (shown in red). Error indicator distribution over a 1-D mesh is in blue
with N elements and corresponding quadratic mean is shown in red.

Figure 6-23 Mesh with Element Error Indicator and Qadratic Mean

The factor f is specified by the user in the F1 field of the HADACRI Bulk Data entry (HADACRI (p. 2082) in the ).
The error indicator E e is a scalar, elemental magnitude that provides a relative measure of the discretization error; that is,
the error between the finite element solution and the analytical solution of the underlying differential equations of the
problem under analysis. It is computed using the grid point stresses and element stress discontinuity following the
procedure utilized by the ELSDCON Case Control command and described in the MSC Nastran Reference Guide, Section
8.3.

This procedure can be summarized as follows:

Main Index
 Appendix F: Local Adaptive Mesh Refinement 785
Refinement by Regular Subdivision

Na
e e
 Let  aij =  W a  aij be the weighted average over all elements ‘e’ concurrent to a given node ‘a’ of each
e=1
e e
component ‘ij’ of the grid point stresses  aij where W is a weighting factor assigned to element ‘e’ and Na is the
number of elements connected to the given node ‘a’.
 An estimate of the error in a particular component of stress ‘ij’ at a grid point ‘a’ is then be computed as
N
2 e 2
E aij =  W a   aij –  aij 
e=1
Averaging the latter over the different stress components, ‘ij’, over the different shell fibers (for shell elements) and over the
different grid points ‘a’ connected by a given element ‘e’ the elemental, scalar error indicator E e is obtained.

Figure 6-24 shows an example using the error indicator based adaptivity criterion involving the analysis of a 2-D mode-I
fracture specimen. Notice that this criterion tends to cluster the refinement near areas of stress concentration. This is due
to the fact that stress gradients (and therefore element stress discontinuities and error indicators) are considerably higher in
those zones than in the rest of the mesh.

Figure 6-24 Analysis of a 2-D Mode-I Fracture Specimen

This refinement criterion is available for any of the surface or volume elements, namely CTRIA3, CTRIAR, CQUAD4,
CQUAD8, CTETRA, CPENTA, CHEXA. It is not available however for the family of line elements. The latter might be

Main Index
 786 MSC Nastran Reference Guide
Refinement by Regular Subdivision

subdivided either by using any other criteria of the remaining refinement criterion, or because they are connected to the
boundary of a surface or volume element (see Vertical Propagation, 796).

Elements Within a Spatial Spherical Region Criterion

In this criteria, the user defines a spherical region in space by specifying its center in basic coordinate system  X 0 Y 0 Z 0 
and its radius R. Then, all elements with at least one node with basic coordinates  X Y Z  within the spherical region
(i.e., such that   X Y Z  –  X 0 Y 0 Z 0   R  || will be refined.

Figure 6-25 shows the mesh obtained in an example involving a 3-D cylindrical body using this criterion. Figure 6-26 shows
a detail of the mesh obtained after the fourth refinement cycle along with the spatial spherical region selected for refinement
with only the bottom half of the cylinder is shown in the right graphic.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 787
Refinement by Regular Subdivision

Figure 6-25 Sequence of Meshes obtained on a 3-D Cylindrical Body using the “Elements within a Spherical Region”
Adaptivity Criterion

Main Index
 788 MSC Nastran Reference Guide
Refinement by Regular Subdivision

Figure 6-26 Mesh obtained after the Third Refinement Cycle of 3-D Cylindrical Body using the “Elements within a Spherical
Region” Adaptivity Criterion

Elements Within a Spatial Orthogonal Region Criterion

In this refinement criterion, the user defines an hexahedral region in space or box aligned with the basic coordinates system
by specifying the basic coordinates of opposite corners  X 0 Y 0 Z 0  and  X 1 Y 1 Z 1  of the box. Then, all elements
with at least one node with basic coordinates  X Y Z  within the specified hexahedral region (i.e., such that
X 0  X  X 1 , Y 0  Y  Y 1 and Z 0  Z  Z 1 ) will be refined.
Figure 6-27 shows the mesh obtained in the same 3-D cylindrical body used in Figure 6-25 but with the nodes within a box
refinement criterion. A detail of the mesh obtained after the third refinement along with the orthogonal refinement region
is shown in Figure 6-28 with only the bottom half of the cylinder is shown in the right graphic.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 789
Refinement by Regular Subdivision

Figure 6-27 Sequence of Meshes obtained on a 3-D Cylindrical Body using the “Elements within an Orthogonal Region”
Adaptivity Criterion

Main Index
 790 MSC Nastran Reference Guide
Refinement by Regular Subdivision

Figure 6-28 Mesh obtained after the Third Refinement Cycle of 3-D Cylindrical Body using the “Elements within an
Orthogonal (Box) Region” Adaptivity Criterion

Elements in Contact Criterion

In this criterion, all touching elements with at least one node involved in contact and touched elements with at least one
face in contact will be refined. MSC Nastran supports glued contact, rigid-to-deformable body contact, and deformable-
to-deformable body contact situations. Figure 6-29 shows the initial mesh of two 3D deformable bodies composed
exclusively of linear hexahedral elements and brought into contact after a vertical displacement is applied on the top body.

Figure 6-29 Two 3-D Deformable Bodies in Contact

Figure 6-30 shows the sequence of meshes obtained during the mesh refinement process.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 791
Refinement by Regular Subdivision

Figure 6-30 Sequence of Meshes during Adaptive Mesh Refinement Process using the “Nodes in Contact” Criterion on Two
3-D Deformable Bodies In Contact

Figures 6-31 and 6-32 depict the sequence of meshes obtained during mesh refinement using the “Nodes in Contact”
refinement criterion in a situation involving 3-D rigid-to-deformable contact with 8-noded CHEXA elements and 3-D glue
contact between two deformable bodies with 10-noded CTETRA elements, respectively.

Main Index
 792 MSC Nastran Reference Guide
Refinement by Regular Subdivision

Figure 6-31 Sequence of Meshes during Adaptive Mesh Refinement Process using the “Nodes in Contact” Criterion on a
Rigid-to-Deformable Body Contact Setting

Main Index
 Appendix F: Local Adaptive Mesh Refinement 793
Refinement by Regular Subdivision

Figure 6-32 Sequence of Meshes during Adaptive Mesh Refinement Process using the “Nodes in Contact” Criterion on Glue
Contact

Propagation of Refinement
Once elements meeting the refinement criteria are scheduled for refinement, the refinement is propagated to some of their
adjacent elements according to a set of implicit propagation rules. These rules are the following:

Horizontal Propagation (2 to 1 rule)

The first refinement propagation rule is the 2-to-1 rule. The 2-to-1 rule restricts the number of hanging nodes on each edge
to one, see Hanging Nodes and Multipoint Constraints on Hanging Nodes, 779. To this end, all the edge neighbors of elements
scheduled for refinement are selected for refinement as well. This is illustrated in Figure 6-33. When one element scheduled
for refinement (green element) is refined, two hanging nodes are created on its adjacent edges. To restrict the number of
hanging nodes to one, the refinement is propagated to its edge-neighbors (yellow elements).

Main Index
 794 MSC Nastran Reference Guide
Refinement by Regular Subdivision

Figure 6-33 Edge Neighbors of Elements Scheduled for Refinement

Notice that in 3-D meshes, there might be more than one edge-neighbor per edge as illustrated in Figure 6-34.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 795
Refinement by Regular Subdivision

Figure 6-34 Propagation of Refinement from an Element Meeting the User-specified Refinement Criterion (green element)
to its Edge-neighbors (yellow elements) to Enforce the 2-to-1 Propagation Rule

Horizontal Propagation in Triangles and Tetrahedra

Consider a refined triangular or tetrahedral element. If the internal triangle or internal tetrahedra are scheduled for a second
refinement, then the external triangle or tetrahedral are automatically selected for refinement as well. This is to avoid the
creation of redundant degrees of freedom on internal edges or faces that would otherwise be constrained with no net gain
of mesh resolution. As shown in Figure 6-35, if the internal (green) triangle of a refined triangular element is further refined,

Main Index
 796 MSC Nastran Reference Guide
Refinement by Regular Subdivision

no net addition of degrees of freedom is obtained (top row). To avoid this redundancy, the refinement is automatically
propagated to all external (yellow) triangles of the refined triangular element (bottom row).

Figure 6-35 Refined Triangular Element

Vertical Propagation
Consider a line element (CBEAM, CBEND, CBAR, CONROD, CROD, CTUBE, CVISC) attached to an edge of a surface
element (CQUAD4, CQUADR, CTRIA3, CTRIAR) of a surface element attached to the face of a 3-D element (CTETRA,
CPENTA, CHEXA). Then, if the element of higher dimensionality (green) is scheduled for refinement, then the element of
lower dimensionality (yellow) attached to its face is automatically selected for refinement as well as shown in Figure 6-36.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 797
Refinement by Regular Subdivision

Figure 6-36 Automatic Propagated Refinement - Higher to Lower

The same rule applies in the opposite direction: if an element of lower dimensionality attached to the boundary of another
element of higher dimensionality is scheduled for refinement, then the latter is automatically selected for refinement as
shown in Figure 6-37.

Figure 6-37 Automatically Propagated Refinement - Lower to Higher

Propagation of Refinement Across Partitioned Superelement Boundaries

Hanging nodes cannot occur at partitioned superelement boundaries because their corresponding degrees of freedom
cannot belong simultaneously to two different Degree-of-Freedom Sets (see Degree-of-Freedom Sets in the MSC Nastran
Quick Reference Guide). In order to prevent this condition, the refinement is automatically propagated across superelement
boundaries. In this way, hanging nodes are moved from the boundary to the interior of the affected superelements. When

Main Index
 798 MSC Nastran Reference Guide
Transference of Analysis Data Between Unrefined and Refined Meshes

elements on a given superelement are scheduled for refinement (green elements), the refinement is propagated into the
neighboring superelement (yellow elements) to avoid the creation of hanging nodes on the superelement boundaries as
shown in Figure 6-38.

Figure 6-38 Automatically Propagated Refinement - Superelement

Transference of Analysis Data Between Unrefined and Refined

Meshes
Once a refined mesh is obtained by subdividing selected elements from the previous mesh, analysis data must be
communicated or transferred from the old mesh to the new mesh in order set up the next analysis on the new mesh. Analysis
data to transfer includes element properties, shell thicknesses, material orientations, pressure loads and permanent and
single point constraints (displacement boundary conditions). The rules to transfer this data are described in the following
sections.

Transference of Element Properties

Children elements created after refinement inherit their parent’s property ID. In surface elements (CQUAD4, CQUADR,
CTRIA3, CTRIAR) with nonuniform shell thicknesses, the thickness for corner nodes in the children elements are linearly
interpolated from the corner nodes of the parent element as shown in Figure 6-39.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 799
Transference of Analysis Data Between Unrefined and Refined Meshes

Figure 6-39 Transference of Non-uniform Shell Thickness Data from the Parent to the Children Elements

Main Index
 800 MSC Nastran Reference Guide
Transference of Analysis Data Between Unrefined and Refined Meshes

Figure 6-40 Transference of Material Orientation Angle from the Parent of the Children Elements

Furthermore, children elements inherit the material orientation angle (see THETA field on the CQUAD4, CQUADR,
CTRIA3, CTRIAR entries). The orientation angle THETA for children element is computed using the equations given in
Figure 6-40 with the blue arrow pointing in the direction of the material orientation which is conserved during refinement.
Notice that this does not imply the conservation of the THETA angle. This takes into account that THETA is defined as
the angle between the edge joining nodes 1 and 2 of the element and the material direction and, therefore, might not be
uniform within children elements. Notice that uniform material orientation does not imply uniform angle THETA.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 801
Transference of Analysis Data Between Unrefined and Refined Meshes

Transference of Distributed Loads and Concentrated Forces

Pressure loads (PLOAD, PLOAD2, PLOAD4) distributed over parent elements are automatically redistributed over children
elements.
Thus, a uniform pressure load distributed over a quadrilateral surface element (PLOAD, PLOAD2) is copied over children
elements and redistributed with same magnitude and direction over their smaller area.
Similarly, the magnitude of non-uniform pressure loads distributed over surface elements faces of 3-D elements or faces of
surface elements (PLOAD4) are not just copied over but interpolated linearly from corner pressures (Figure 6-41) and applied
to children elements with the same direction. Mid-edge and mid-face pressures are linearly interpolated from corner values.

Figure 6-41 Transference of PLOAD4 applied to the Face of a 3-D (CHEXA) Element

No special provisions are taken regarding transference of concentrated forces or moments. Thus, no concentrated loads
will be created or new grids created for refinement but just carried over existing grid points in the old mesh to the same grid
points in the new mesh.

Transference of Displacement Coordinate System, Displacement Boundary

Conditions, and Constraints
Displacement coordinate systems, permanent single point constraints, single point constraints and multipoint constraints
on the new grid point c created during refinement are internally enforced according to the following rules:
 If the coordinates systems defined on corner grid points (specified on the CD field in the GRID Bulk Data entry)
are identical, then the same displacement coordinate system is assigned for a new mid-edge or mid-face node
created during refinement. Furthermore, for edges of surface or volume elements, each degree of freedom
associated to the new mid-node created on the edge is either permanently constrained (assigned to degree of

Main Index
 802 MSC Nastran Reference Guide
Transference of Analysis Data Between Unrefined and Refined Meshes

freedom set SG), or explicitly constrained via an internal Single point constrain SPC (assigned to the degree of
freedom set SB) or constrained to corner nodes via an internal Multipoint Constraints M (see Degree-of-Freedom
Sets in the MSC Nastran Quick Reference Guide) depending upon the degree of freedom set where the corner
nodes.
For faces of volume elements, each degree of freedom of a new mid-face node created during refinement is also
either permanently constrained, or explicitly constrained (via internal SPC) or tied to corner nodes (via internal
MPC) according to the permanent constraints, enforced displacement or multipoint constraints defined on the
corner nodes.
The specific constraints to be enforced internally for mid-edge or mid-face nodes are:
• If permanent constraints have been defined on a degree of freedom i at both corner nodes, then a permanent
constraint is also assigned to the degree of freedom i at the mid-edge of mid-face node. For example (as shown
in Figure 6-42), if both corner nodes a and b on an edge have been permanently constrained in all degrees of
freedom (by defining PS a = 123456 and PS b = 123456 on the PS field in the GRID Bulk Data entry),
then node c will be also constrained permanently in all degrees of freedom (by defining PS c = 123456
internally).

Figure 6-42 Coordinate systems ( CD a and CD b )

Coordinate systems ( CD a and CD b ) defined on corner node a and b (using the CD field in the Grid Bulk
Data entry) must match in order for constraints to be enforced on mid-edge node c. Permanent constraints (PS
field on the GRID Bulk Data entry) on each degree of freedom are then carried over the mid-edge node.
• If single point constraints have been defined on a degree of freedom i for all corner nodes, then the degree of
freedom i on the mid-edge or mid-face node will be tied internally to the corner nodes according to the same
multipoint constraint equations used for hanging nodes on straight edges (see Hanging Nodes and Multipoint
Constraints on Hanging Nodes, 779, Figures 6-16 and 6-18).

Main Index
 Appendix F: Local Adaptive Mesh Refinement 803
Transference of Analysis Data Between Unrefined and Refined Meshes

Figure 6-43 If displacement is enforced on a given degree of freedom on corner nodes a and b (using SPC or SPC1 entries),
then a multipoint constraint that ties the mid-edge node c with both corner nodes is internally enforced, provided that
displacement coordinate systems for all corner nodes coincide (CD field in the GRID entry).

Notice that this is equivalent to imposing an SPC on the mid-edge node, degree of freedom i, with a value for
enforced displacement averaged from the value at corner nodes. For example, if a displacement with value U a
along direction 1 is enforced on node a (by defining an SPC on node a, direction 1) and a displacement in the
same direction with value U b is enforced on node b (by defining another SPC on node b, direction 1), then a
multipoint constrain for node c (direction 1) is internally defined such that U c = 1  2  U a + U b  as shown
in Figure 6-43.
• The same multipoint constraint equation is applied on the mid-edge node c (in a given direction i) if one of the
corner nodes has been constrained permanently and the other corner node has been constrained via a Single
Point Constraint (SPC) on the same direction i. For example, if a displacement with value U a along direction
1 is enforced on node a (by defining an SPC on node a, direction 1) and a permanent constraint in the same
direction has been enforced for corner node b (by defining PS b = 1 on the PS field in its GRID Bulk Data
entry), then a multipoint constraint for node c (direction 1) is internally defined such that
U c = 1  2  U a + U b  = 1  2  U a + 0  = 1  2 U a as shown in Figure 6-44.

Figure 6-44 If a displacement is enforced on a given degree of freedom on corner node a (using SPC or SPC1 entries) and
a permanent constraint is enforced on the same degree of freedom on corner node b (PS=1 on its GRID Bulk Data entry), then a
multipoint constraint that ties the mid-edge node c with both corner nodes is internally enforced, provided that displacement
coordinate systems for all corner nodes coincide (CD field in GRID entry).

Main Index
 804 MSC Nastran Reference Guide
Transference of Analysis Data Between Unrefined and Refined Meshes

• Similarly, the same multipoint constraint is applied on the mid-edge or mid-face nodes (in a given degree of
freedom i) if some of the corner nodes are involved in a permanent constraint (PS field in the GRID Bulk Data
entry) or single point constraint (SPC or SPC1 Bulk Data entry) and the others corner nodes are involved in a
multipoint constraint (MPC Bulk Data entry) on the same degree of freedom i.
• If the corner nodes are involved in contact, either as touching nodes or touched element faces or edges, then the
mid-edge nodes are regarded as nodes potentially in contact. Therefore, constraints on any of the degrees of
freedom associated to the latter are determined by the contact detection algorithm.
The set of relations just outlined is summarized in the following table:

Corner node a Corner node b Mid-node c

SG SG SG
SB SB M
SG SB M
SB SG
SG or SB M M
M SG or SB
Node in contact Node in contact Node potentially in contact

 If the displacement coordinate system defined on corner nodes on an edge or face are different, then the
displacement coordinate system for the mid-node on the edge or face is set to the basic coordinate system.
Furthermore, no constraints are enforced on any of its associated degrees of freedom independently of the
constraints that might have been imposed on corner nodes as shown in Figure 6-45.

Figure 6-45 If coordinate systems ( CD a and CD b ) defined on corner nodes a and b (using the CD field in the GRID Bulk
Data entry) are different then the mid-edge node c is left free and its displacement coordinate system is set to the basic.

Detection of Geometric Features

Prior to the initiation of the adaptive mesh refinement loop, the initial mesh provided by the user is preprocessed using an
automatic Geometric Feature and Material interface Detection Algorithm aimed to identify:

Main Index
 Appendix F: Local Adaptive Mesh Refinement 805
Transference of Analysis Data Between Unrefined and Refined Meshes

 Geometric features such as:

• Sharp corners and edges
• Non-manifold edges and vertices, i.e., edges joining more than two surfaces (in 3D) or vertices joining more
than two curves in 2-D or 3-D.
The Geometric Feature Detection Algorithm identifies edges and corners by comparing the angle between each pair of
adjacent elements with the “Feature Angle” VARPHI. The feature angle is a scalar parameter specified by the user
(PARAM,VARPHI) that defines how sharp a mesh edge or vertex should be in order to be considered as a geometric feature.
More precisely, face outward normal vectors N 1 N 2 of each pair of adjacent mesh faces and the edge oriented tangents
T 1 T 2 of each pair of adjacent mesh edges are computed (Figure 6-46) by the geometric feature detection algorithm. If the
angle between N 1 and N 2 for mesh faces, or between T 1 and T 2 for mesh edges is larger than the feature angle VARPHI
then the common edge or vertex will be considered a splitting edge or vertex where surfaces or lines are broken and a
geometric feature is thus defined.

Figure 6-46

Mesh faces and elements are preprocessed to ensure consistent orientation and that the appropriate sign of face normal
vectors and edge tangents will be accounted for during the computation of their mutual angle.
Figures 6-47, 6-48, 6-49, and 6-50 show the edges identified by the Geometric Feature detection algorithm on a surface
mesh (Figure 6-47) and three 3-D volume meshes (Figures 6-47, 6-48, and 6-49).

Main Index
 806 MSC Nastran Reference Guide
Transference of Analysis Data Between Unrefined and Refined Meshes

Figure 6-47 Edges detected by the geometric feature detection algorithm in a surface mesh of triangular elements of a
mechanical part with a non-manifold edge

Figure 6-48 Edges detected by the geometric feature detection algorithm in a 3-D hexahedral mesh of a cylindrical body

Main Index
 Appendix F: Local Adaptive Mesh Refinement 807
Transference of Analysis Data Between Unrefined and Refined Meshes

Figure 6-49 Edges detected by the geometric feature detection algorithm in a 3-D hexahedral mesh of an engine cup

Figure 6-50 Edges detected by the geometric feature detection algorithm in a 3-D hexahedral mesh of an engine cylinder
head

Adequate identification of geometric features (by appropriately adjusting the feature angle VARPHI) is essential to ensure
the convergence of the mesh refinement process to expected results.
Thus, if the error indicator based refinement criterion is selected and sharp edges are not properly identified by the
geometric feature detection algorithm, the refinement might cluster indefinitely in the neighborhood of the undetected
sharp edges as shown in Figure 6-47. This anomaly occurs mainly on sharp intersections between shells due to the fact that
the error indicator indirectly measures membrane stress jumps between adjacent elements and the latter might be very high
due to the abrupt change in shell normal directions at sharp intersections.
Figures 6-51 and 6-52 depict two orthogonal planar shells joined on a common edge and subjected to a vertical
displacement on the top. The feature angle must be chosen smaller than   2 in order for the joining edge to be detected
and the error estimator to ignore (or filter) the large membrane stress discontinuity across this edge as shown in Figure 6-51.

Main Index
 808 MSC Nastran Reference Guide
Transference of Analysis Data Between Unrefined and Refined Meshes

If the feature angle is not appropriately chosen, then the geometric feature algorithm will fail to identify the joining edge
and the mesh refinement will cluster in its neighborhood as shown in Figure 6-52.

Figure 6-51 Clustering during Mesh Refinement

An adequate value of the VARPHI parameter (   4 in this case) will ensure that the sharp edge shared by both planar shells
is detected by the geometric feature detection algorithm and the big membrane stress jumps occurring at the edge are
filtered out.

Figure 6-52 Filtering of Stress Jumps at Edges

The geometric feature detection algorithm fails to detect the sharp edge shared by both planar shells because the VARPHI
parameter is too large. As a consequence, the big membrane stress jumps occurring at the edge are not filtered out and the
refinement clusters in the neighborhood of the sharp edge.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 809
Transference of Analysis Data Between Unrefined and Refined Meshes

The adequate identification of corners is also required to improve the smooth approximation of the analysis domain
boundary constructed by interpolating the mesh boundary nodes and used as a method to place new mid-edge nodes during
refinement (see Location of New Grid Points, 776) alternative to the default location at the mid-side of the edge.
Figures 6-53 and 6-54 compare the two edge-node placement methods (mid-side placement and projection of mid-edge
nodes onto a smooth approximation of the boundary) in an example involving a 2-D treble shaped planar shell subjected
to compression. The boundary of this mesh exhibits three sharp corners located at the intersection of each pair of circular
leaves. The mesh is refined everywhere (uniform refinement). This is accomplished by selecting the “nodes within a spatial
sphere” refinement criteria (see Refinement Criteria, 784) with a spherical refinement region big enough to contain the whole
mesh.

Figure 6-53 Mid-edge Nodes are placed in the Mid-side of Edges

Main Index
 810 MSC Nastran Reference Guide
Transference of Analysis Data Between Unrefined and Refined Meshes

Figure 6-54 Projection of Mid-edge Nodes onto a Smooth Approximation of the Geometric Boundary Interpolated from the
Initial Mesh

Notice that the three sharp corners are appropriately detected by the Geometric Feature Detection Algorithm and kept as
hard points during the mesh refinement process. By contrast, sharp corners might become smeared out if the geometric
feature detection algorithm is not successful due to an inadequate selection of the feature angle (parameter VARPHI)
(Figure 6-55).

Figure 6-55 Corners might be smeared out if they are not appropriately detected by the automatic geometry feature
detection algorithm. Corner detection can be controlled by the user adjusting the VARPHI parameter.

Main Index
 Appendix F: Local Adaptive Mesh Refinement 811
Transference of Analysis Data Between Unrefined and Refined Meshes

Detection of Material, and Superelement Interfaces

Interfaces between mesh regions with different properties IDs or superelement IDs are also automatically detected during
the preprocessing phase prior to the beginning of the adaptive mesh refinement loop.
Different properties or superelement may reference different materials or different shell thicknesses. Therefore, stress of
different order of magnitude are expected in areas with different properties or superelements. This type of discontinuities
must be filtered out by the error indicator (which averages stress jumps across interelement boundaries) in order to be able
to capture the discontinuities introduced by the finite element discretization exclusively.

Main Index
 MSC.Mvision Builder and Evaluator 2002 Installation Guide

Index
MSC Nastran Reference Guide
Index

A BSBRPD subroutine, 87
ACCELERATION, 27, 28 BSBRT subroutine, 87
ACSRCE, 308 BSCON subroutine, 87
Index BSGRQ subroutine, 87
adaptive bisection, 727
ADPCON, 191 BSMSG subroutine, 87
ADUM8, 181 buckling analysis, 19, 26, 702
ADUM9, 181 Bulk Data Section, 11
ADUMi, 46 BUSH1D, 205, 206, 670
aeroelastic response, 20, 27 BWIDTH, 191
AEROF, 25, 27
ANALYSIS, 15, 21, 22 C
angular acceleration, 300 Case Control commands for SOLs
angular velocity, 300 GRID, 15
APRESSURE, 25 METHOD(FLUID), 18
ASETi, 332 SDMAPING(FLUID), 18
ASSIGN, 463 SMETHOD, 15
Augmented Lagrange Multiplier Method, 167 SPC, 15
AUTOASGN, 465 SUPORT1, 15
automatic time stepping, 725 Case Control Section, 11, 14
AUTOSPC, 136, 330 CAXIF2, 738
axial damping element, 205 CAXIF3, 738
axisymmetric elements, 45, 179 CAXIF4, 738
AXSLOT, 738 CAXIFi, 47, 49, 51
CBAR, 47, 49, 51, 53, 65
B CBARAO, 299
B2PP, 19, 30 CBEAM, 47, 49, 51, 53, 58, 60
BAR, 242 CBEAM3, 47
bar element, 65 CBEND, 47, 49, 51, 53, 67
BAROR, 67 CBUSH, 47, 49, 51, 201
BCONP, 191 frequency dependency, 201
BEAM, 242 load-displacement dependency, 201
beam element, 53 CBUSH1D, 47, 49, 51, 205
BEGIN BULK delimiter, 11 CCONEAX, 45, 47, 49, 51
BFRIC, 191 CDAMP1, 179
BLSEG, 191 CDAMP2, 179
BOUTPUT, 191 CDAMP3, 179
BSBRCD subroutine, 87 CDAMP4, 179
BSBRGD subroutine, 87 CDAMP5, 179
BSBRID subroutine, 87 CDAMPi, 47, 49, 51

Main Index
 INDEX 813

CDUMi, 46 constraint elements, 46

CELAS1, 178 contact, 688
CELAS2, 178 convergence, 688
CELAS3, 178 CORD1C, 37
CELAS4, 178 CORD1R, 37
CELASi, 47, 49, 51 CORD1S, 37
CEND delimiter, 11 CORD2C, 37
CFAST, 47 CORD2R, 37
CFLUIDi, 47, 49, 51 CORD2S, 37
CGAP, 48, 49, 51, 186, 609 COUPMASS, 54, 67, 71, 126
CHBDYG, 49 CPENTA, 48, 49, 51, 155, 243
CHBDYi, 48, 51 CQUAD4, 48, 49, 51, 125, 243
CHBDYP, 49 CQUAD8, 48, 49, 51, 125
CHEXA, 48, 49, 51, 155, 243 CQUADR, 48, 50, 52, 125, 243
CLOAD, 24 CQUADX, 45, 179
CMASS1, 179 CRAC2D, 47, 49, 51, 181
CMASS2, 179 CRAC3D, 47, 49, 51, 181
CMASS3, 179 crack tip elements, 181
CMASS4, 179 CROD, 48, 50, 52, 53, 71
CMASSi, 48, 49, 51 CSEAM, 48
CMETHOD, 19, 706 CSHEAR, 48, 50, 52, 124
comments, 11 CSLOT3, 738
complex eigenvalue analysis, 19, 26 CSLOT4, 738
component mode synthesis, 332 CSLOTi, 48, 50, 52
Composite Rotations, 166 CTETRA, 48, 50, 52, 154, 243
Composites CTRIA3, 48, 50, 52, 125, 243
MAT8, 266 CTRIA6, 48, 50, 52, 125
composites CTRIAR, 48, 50, 52, 125, 243
data recovery, 279 CTRIAX, 45, 179
failure criteria, 266 CTRIAX6, 45, 48, 50, 52, 179
failure indices, 279 CTUBE, 48, 50, 52, 53, 71
Hill’s theory, 281 CVISC, 46, 48, 50, 52, 72
Hoffman’s theory, 281 CWELD, 48, 50, 52, 206, 212
lamina, 264 cyclic symmetry, 22
MAT2, 266 dihedral symmetry, 747
material properties, 266 N-gons, 748
maximum strain theory, 283 reflective symmetry, 746, 749
PCOMP, 266 rotational symmetry, 747
PSHELL, 266 theory, 751
Tsai-Wu theory, 281 three-dimensional bodies, 748
concentrated mass elements, 192 CYJOIN, 765
CONM1, 39, 46, 192 cylindrical coordinate system, 37, 38
CONM2, 39, 46, 192 CYSYM, 765
CONMi, 47, 49, 51
CONROD, 47, 49, 51, 53, 71

Main Index
 814 MSC Nastran Reference Guide

D element strain energy, 735

DAREA, 22, 306, 307, 308, 309, 316 ELEMID, 208, 212
database, 462 ELPAT, 209
DBCLEAN, 465 ELSDCON, 442, 455
DBLOAD, 466 ENDDATA delimiter, 11
DBLOCATE, 464 enforced motion, 305
DBSETDEL, 464 EPOINT, 39
DBsets, 462 ESE, 25, 26, 27, 28, 735
DBALL, 462 Euler angles, 164
MASTER, 462 Executive Control Section, 11
SCRATCH, 462 Executive Control statements, 14
USROBJ, 462
USRSOU, 462 F
DBUNLOAD, 466 failure criteria, 284, 286, 287
DDRMM, 72, 179 HILL, 284
DEFORM, 15, 19, 303 HOFFMAN, 286
DEFUSET, 334 TSAI-Wu, 287
Degree of Freedom Sets Failure Index (FI), 284
Bulk Data entries, 327 File Management Section (FMS), 11, 14
definitions, 325 flexibility factors, 69
DELAY, 306, 307, 317 FLUTTER, 19
delivery database, 464 FLUX, 28, 29
design optimization, 22 FMETHOD, 19
differential stiffness, 702, 703 follower force, 727
direct input matrices, 39 follower stiffness, 706
Direct Matrix Abstraction Program, 460 FOLLOWK, 701
DISPLACEMENT, 25, 26, 27, 28 FORCE, 25, 26, 27, 28, 299, 312
displacement coordinate system, 37 FORCE1, 299, 312
DLOAD, 20, 22, 30, 306, 307, 316, 317 FORCE2, 299, 312
DMAP, 460 FORCEAX, 299
DMIG, 39, 196, 332 FREQ, 20, 306
DPHASE, 306 FREQ1, 20, 306
dummy elements, 46 FREQ2, 20, 306
DYNSPCF, 394 FREQUENCY, 20, 30
frequency dependency, 201
E frequency response, 20, 27
EDE, 27, 28
EIGC, 19, 706 G
Eigenvalue problem, 584 GAP, 186
EIGR, 17, 22 gap element, 186
EIGRL, 17, 22 gaps, 688
EKE, 26, 27, 28 GENEL, 46, 192, 196
element deformation, 303 general element, 192
element geometry checks, 241 GEOMCHECK, 242, 244
element offsets, 37 Gimbal angles, 164

Main Index
 INDEX 815

global coordinate system, 37 K

GPFORCE, 25, 26, 28, 735 K2GG, 15, 18, 22, 24
GPSDCON, 442, 455 K2PP, 19, 30
GPSTRAIN, 26 K6ROT, 136
GPSTRESS, 26, 28, 442, 451, 455, 457
GRAV, 300
gravity load, 300
L
Lagrange Elimination Method, 163
GRDPNT, 32
Lagrange elimination method, 169
GRDSET, 39
Lagrange rigid element, 162
GRID, 35, 37
Lanczos method, 590
grid point force, 734
Large mass method, 591
grid point stress, 442, 443, 457
Left Rotation, 166
exception points, 449
LGDISP, 660, 701
geometric method, 447
line elements, 53
output, 452
Linear Elimination Method, 167
topological method, 444
linear steady state heat transfer, 21, 28
grid points, 35, 39
LOAD, 15, 19, 21, 23
GRIDF, 738
LOADCYH, 310, 765
GRIDID, 208, 212
LOADCYN, 310, 765
GRIDS, 738
LOADCYT, 765
Guyan reduction, 332
load-displacement dependency, 201
LOADSET, 20, 21, 24, 306, 307, 308, 314, 316
H local coordinate systems, 35
HARMONICS, 765 LOOPID, 21, 22, 24
HDOT, 29 LSEQ, 20, 21, 24, 306, 308, 310, 313, 316
heat transfer, 688
HEXA, 242
HOUTPUT, 766
M
M2GG, 15, 18, 22, 24, 196
hyperelastic elements, 196
M2PP, 19, 30
hyperelastic materials, 126
MacNeal/Harder test, 139
hyperelasticity
mass properties, 39
curve-fitting algorithm, 628
MAT1, 45, 58, 67
nonlinear stress output, 657
MAT2, 45
MAT3, 45
I MAT8, 45, 265
IC, 21, 22 MAT9, 45
INCLUDE statement, 11 material properties, 45
inertia relief, 331, 377, 402, 408 MATGPR module, 334
INIT, 463 MATHP, 46, 196
INIT MASTER, 464 MATS1, 46, 53
initial conditions, 316 MATT1, 45
INREL, 391 MATT2, 45
MATT3, 45
MATT9, 45
maximum stress (strain), 288

Main Index
 816 MSC Nastran Reference Guide

Mechanism, 588 nonlinear analysis, 311

mechanism, 372 3-D slideline contact, 675
METHOD, 17, 22, 24, 706 elastic-plastic material, 633
MODES, 18 element library, 609
MOMAX, 299 elements, 638
MOMENT, 299, 312 gap element, 671
MOMENT1, 299, 312 geometric nonlinearity, 609, 638
MOMENT2, 299, 312 hyperelastic elements, 655
MPC, 17, 30, 160, 329 hyperelastic material, 609, 625
MPCADD, 329 material nonlinearity, 609, 623
MPCAX, 329 modeling, 621
MPCF, 26, 27, 28 nonlinear analysis with superelements, 732
MSC.Marc, 688 nonlinear elastic material, 624
MSGLIMIT, 242 nonlinear transient response analysis, 712
MSGTYPE, 242 parameters, 618
multipoint constraint, 328 plot codes, 621
recommendations, 623
N restarts, 720
nastran command, 10 small strain elements, 651
NASTRAN statement, 11, 13 SOL 106, 612
NDDL SOL 129, 612
Nastran Data Definition Language, 488 SOL 153, 612
Newton’s method, 705 SOL 159, 612
Newton-Raphson method, 164 static analysis, 684
NLLOAD, 29, 316 subcase structure, 614
NLPARM, 23, 686, 688, 689 temperature dependent material, 637
NLPCI, 691 viscoelastic material, 636
NLRGAP, 317, 728 nonlinear static analysis, 23
NLSTRESS, 26 nonlinear steady state heat transfer, 21
NOCOMPS, 266 nonlinear transient heat transfer, 21, 29
NOLIN, 725, 728 nonlinear transient response analysis, 24
NOLIN1, 307, 728 nonstructural mass, 246
NOLIN2, 308, 728 normal modes analysis, 17, 26
NOLIN3, 308, 728 NOUTPUT, 766
NOLIN4, 308, 728
NOLINEAR, 316 O
NOLINi, 205, 317 Offsets, 120
NONLINEAR, 20, 22, 28 OFREQUENCY, 20, 27, 396
OLOAD, 25, 26, 27, 28, 29
OMITAX, 332
OMITi, 332
OTIME, 396
OUTPUT(PLOT), 26, 27, 28
OUTPUT(POST), 442, 451, 455, 457
OUTPUT(XYPLOT), 27, 28

Main Index
 INDEX 817

P POINTAX, 299
P2G, 15, 22, 24 POST, 482
PARAM, 284 PostScript plot, 479
BUCKLE, 705 PRAC2D, 183
FKSYMFAC, 701, 706 PRESAX, 300
LANGLE, 171 prestressed normal mode, 700, 701
LMFACT, 171 PROD, 71
NDAMP, 727 PROJECT, 465
NLTOL, 688 PSHEAR, 124
NOELOF, 735 PSHELL, 126, 130
NOELOP, 735 PSOLID, 46, 155
ORIGID, 171 PTUBE, 71
PENFN, 171 punch, 490
QRMETH, 142 PVISC, 72
SHLDAMP, 154
SRCOMPS,YES/NO, 284 Q
XFLAG, 733 QBDYi, 21, 22
PARAML module, 334 QHBDY, 21, 22
PARAMs QUAD4, 136, 139, 242
BAILOUT, 589 QUAD8, 138
MAXRATIO, 587 QUADR, 139, 242
PARTPAT, 209 quaternion, 164
PBAR, 66, 72 QVECT, 21, 22
PBARL, 72 QVOL, 21, 22
PBCOMP, 53, 56, 57
PBEAM, 53, 56, 57, 60, 72
PBEAML, 72
R
RANDOM, 20, 490
PBEND, 67
random response, 20, 27, 306
PBUSH1D, 205
RANDPS, 20, 306
PBUSHT, 201
RANDT1, 306
PCOMP, 265
RANDTi, 20
PCOMPG, 265
RBAR, 46, 130, 159, 329
PDAMP, 179
RBAR1, 172
PDUMi, 46
RBE1, 46, 159, 329
PELAS, 178
RBE2, 46, 159, 329
PENTA, 242
RBE3, 46, 160, 329
PGAP, 186
rectangular coordinate system, 37, 38
PLOAD, 312
REPCASE, 15, 17, 18, 21, 30
PLOAD1, 53, 299
RESTART, 465
PLOAD2, 300, 312
restarts, 696
PLOAD4, 300, 312
RFORCE, 300
PLOADX1, 179, 300
Right Rotation, 166
PLOTSUP, 468
RIGID, 170, 172
PLPLANE, 196
rigid elements, 162
PLSOLID, 196
RJOINT, 172, 173

Main Index
 818 MSC Nastran Reference Guide

RLOAD1, 305, 306, 308 SPCF, 25, 26, 27, 28

RLOAD2, 305, 306, 308 SPCFORCE, 28, 29
rod element, 71 spherical coordinate system, 37, 38
Rodrigues parameters, 164 split database operations, 465
RROD, 46, 159, 329 SPOINT, 39
RSPLINE, 46, 160, 329 static analysis, 15, 25
RSSCON, 160, 329 STATSUB, 19
RTRPL1, 172 steady state heat transfer, 29
RTRPLT, 46, 159, 329 STIME, 22, 24
RTRPLT1, 172 STRAIN, 26, 27
strain energy, 733
S strength ratio, 296
SACCELERATION, 27, 28 STRESS, 25, 26, 27, 28, 442
scalar elements, 178 stress discontinuities, 442, 458
SDAMPING, 19, 30 stress intensity factor, 183
SDISP, 29 STRFIELD, 457
SDISPLACEMENT, 27, 28, 29 SUBCASE, 16, 17, 18
SEALL, 25 subcase combinations, 310
SEPLOT, 468 subcases, 15, 17, 19
shear panel, 124 SUBCASID, 21
shear relief, 59 SUBCOM, 16, 30, 301, 303, 310, 750
shell elements, 125 SUBID, 21, 24
single point constraint, 330 subroutines
Slideline contact BSBRGD, 87
guidelines, 189 BSBRID, 87
slideline contact BSBRPD, 87
input, 191 BSBRT, 87
penetration, 189 BSCON, 87
primary line, 188 BSGRQ, 87
secondary line, 188 BSMSG, 87
slideline contact region, 188 SUBSEQ, 30
SLOAD, 299, 312 SUPAX, 331
SLOOPID, 22, 24 SUPER, 25, 31
SOL 106, 685 superelement, 24, 31
SOL 129, 712 superelement analysis, 425
SOL 400, 173 SUPORT, 331
SOL DBTRANS, 466 SUPORT1, 17, 30, 331
solution sequence, 336 SURFACE, 442, 452, 455, 457
rigid format, 336 SVECTOR, 26, 27
SORT1, 25 SVELO, 29
SORT2, 25 SVELOCITY, 27, 28, 29
SPC, 17, 30, 304, 305, 330 SYM, 17, 303
SPC1, 306, 330 SYMCOM, 17, 301, 303, 310, 750
SPCADD, 330 system cell
SPCAX, 304, 330 210, 488
SPCD, 304, 305

Main Index
 INDEX 819

T UPARTN module, 334

TABDMP1, 19 USET, 334
TABLEDi, 306, 307, 317 USET1, 334
TABLEM1, 45
TABLEM2, 45 V
TABLEM3, 45 VEC module, 334
TABLEM4, 45 VECTOR, 26, 27
TABLES1, 46 VELO, 29
TABLEST, 46 VELOCITY, 27, 28, 29
TABRNDi, 306 VOLUME, 442, 452, 455, 457
tapered beam, 61
TEMP, 21, 22, 28, 29, 39, 301, 313
TEMP(LOAD), 302
W
warping, 58, 60
TEMPAX, 39, 301
TEMPD, 21, 22, 39, 301, 313
TEMPERATURE, 15, 19
TEMPERATURE(INIT), 24
TEMPERATURE(INITIAL), 21
TEMPERATURE(LOAD), 23
temperatures, 300
TEMPP1, 300, 313
TEMPP2, 300
TEMPP3, 300, 313
TEMPRB, 300, 313
TETRA, 242
TF, 332
TFL, 30
TFLAG, 151
THERMAL, 28, 29
thermal field, 39, 301
thermal load, 301
TIC, 21
TLOAD1, 305, 306, 307, 308, 316, 727
TLOAD2, 305, 306, 307, 308, 316
TLOADi, 20, 22
transient response, 20, 28
TRIA3, 242
TRIAR, 242
TRIM, 15
TSTEP, 21, 30
TSTEPNL, 21, 24, 206, 715

U
UMERGE module, 334
UMERGE1 module, 334

Main Index