DOE F 241.

3
(2-01) p.1 of 2 UNITED STATES DEPARTMENT OF ENERGY (DOE) OMB CONTROL NO.
Announcement of Scientific and Technical Information (STI) 1910-1400
(For Use By Financial Assistance Recipients and Non-M&O/M&I Contractors)
PART I: STI PRODUCT DESCRIPTION
(To be completed by Recipient/Contractor)

A. STI Product Identifiers H. Sponsoring DOE Program Office

1. REPORT/PRODUCT NUMBER(s) Office of Industrial Technologies (OIT)(EE20)
None
2. DOE AWARD/CONTRACT NUMBER(s) I. Subject Categories (list primary one first)
DE-FC36-97ID13554 32 Energy Conservation, Consumption and Utilization
3. OTHER IDENTIFYING NUMBER(s) Keywords: Steel, Hot Strip Mill, Computer Modeling
None J. Description/Abstract
The Hot Strip Mill Model (HSMM) is an off-line PC based software
B. Recipient/Contractor model originally developed by the University of British Columbia (UBC)
INTEG process group, 11279 Perry Highway, Suite 107, Wexford, PA and the National Institute of Standards and Technology (NIST) under
15090 the AISI/DOE Advanced Process Control Program. The HSMM was
C. STI Product Title developed to predict the temperatures, deformations, microstructure
Validation of the Hot Strip Mill Model evolution and mechanical properties of steel strip or plate rolled in a
D. Author(s) hot mill. INTEG process group undertook the current task of
Richard Shulkosky enhancing and validating the technology. With the support of 5 North
David Rosberg American steel producers, INTEG process group tested and validated
Jerrid Chapman the model using actual operating data from the steel plants and
E-mail Address(es): enhanced the model to improve prediction results.
[email protected] K. Intellectual Property/Distribution Limitations
(must select at least one; if uncertain contact your
Contracting Officer (CO))
X 1. UNLIMITED ANNOUNCEMENT (available to U.S. and
non-U.S. public; the Government assumes no liability
E. STI Product Issue Date/Date of Publication for disclosure of such data)
03 30 2005 2. COPYRIGHTED MATERIAL: Are there any restrictions
MM DD YYYY based on copyright? Yes X No.
F. STI Product Type (Select only one) If yes, list the restrictions as contained in your award document

X 1. TECHNICAL REPORT
3. PATENTABLE MATERIAL: THERE IS PATENTABLE
X Final Other (specify) MATERIAL IN THE DOCUMENT.
INENTION DISCLOSURE SUBMITTED TO DOE:
2. CONFERENCE PAPER/PROCEEDINGS DOE Docket Number: S-
(Sections are marked as restricted distribution pursuant
to 35 USC 205)
4. PROTECTED DATA: CRADA Other, specify
Conference Information (title, location, dates)
Release date (required) no more than
5 years from date listed in Part I.E. above MM DD YYYY
3. JOURNAL ARTICLE
5. SMALL BUSINESS INNOVATION RESEARCH (SBIR) DATA
a. TYPE: Announcement Citation Only Release date (required) no more than 4
Preprint Postprint years from date listed in Part I.E. above
MM DD YYYY
b. JOURNAL NAME 6. SMALL BUSINESS TECHNOLOGY TRANSFER RESEARCH
(STTR) DATA
Release date (required) no more than 4
c. VOLUME d. ISSUE
years from date listed in Part I.E. above
MM DD YYYY
e. SERIAL IDENTIFIER (e.g. ISSN or CODEN)
7. OFFICE OF NUCLEAR ENERGY APPLICED TECHNOLOGY
L. Recipient/Contract Point of Contact Contact for
additional information (contact or organization name
4. OTHER, SPECIFY To be included in published citations and who would
Receive any external questions about the content of
the STI Product or the research contained herein)
Richard Shulkosky, Vice President, INTEG process group, inc.
G. STI Product Reporting Period Name and/or Position
11 01 2001 Thru 03 30 2005 [email protected] (724) 933-9350
MM DD YYYY MM DD YYYY E-mail Phone
INTEG process group, inc. 11279 Perry Hwy., Ste 107, Wexford,
PA 15090
DOE F 241.3
(2-01) p.2 of 2 UNITED STATES DEPARTMENT OF ENERGY (DOE) OMB CONTROL NO.
Announcement of Scientific and Technical Information (STI) 1910-1400
(For Use By Financial Assistance Recipients and Non-M&O/M&I Contractors)

PART II: STI PRODUCT MEDIA/FORMAT and PART III: STI PRODUCT REVIEW/RELEASE
LOCATION/TRANSMISSION INFORMATION
(To be completed by Recipient/Contractor) (To be completed by DOE)

A. Media/Format Information: A. STI Product Reporting Requirements Review.

1. MEDIUM OF STI PRODUCT IS:
X Electronic Document Computer Medium 1. THIS DELIVERABLE COMPLETES ALL REQUIRED
Audiovisual Material Paper No Full-text DELIVERABLES FOR THIS AWARD
2. SIZE OF STI PRODUCT 48 Pages, 721 KB 2. THIS DELIVERABLE FULFILLS A TECHNICAL
3. SPECIFY FILE FORMAT OF ELECTRONIC DOCUMENT INFORMATION REPORTING REQUIREMENT, BUT
BEING TRANSMITTED, INDICATE: SHOULD NOT BE DISSEMINATED BEYOND DOE.
SGML HTML XML X PDF Normal
PDF Image TIFFG4 B. Award Office Is the Source of STI Product Availability
WP-indicate Version (5.0 or greater) THE STI PRODUCT IS NOT AVAILABLE IN AN
Platform/operation system ELECTRONIC MEDIUM. THE AWARDING OFFICE WILL
MS-indicate Version (5.0 or greater) SERVE AS THE INTERIM SOURCE OF AVAILABILITY.
Platform/operation system
Postscript C. DOE Releasing Official
4. IF COMPUTER MEDIUM OR AUDIOVISUAL MATERIAL: 1. I VERIFY THAT ALL NECESSARY REVIEWS HAVE BEEN
a. Quantity/type (specify) COMPLETED AS DESCRIBED IN DOE G 241.1-1A,
b. Machine compatibility (specify) PART II, SECTION 3.0 AND THAT THE STI PRODUCT
c. Other information about product format a user needs to know: SHOULD BE RELEASED IN ACCORDANCE WITH THE
INTELLECTUAL PROPERTY/DISTRIBUTION
LIMITATION ABOVE.
B. Transmission Information:
1. STI PRODUCT IS BEING TRANSMITTED: Release by (name)
X a. Electronically via E-Link
b. Via mail or shipment to address indicated in award Date
MM DD YYYY
document (Paper product, CD-ROM, diskettes,
video cassettes, etc.) E- Mail

Phone
2. INFORMATION PRODUCT FILE NAME
X (of transmitted electronic format)
EditedTRP0040FinalReport.
 AISI/DOE Technology Roadmap Program

Final Report

0040 - VALIDATION OF THE HOT STRIP MILL MODEL

by

Richard A. Shulkosky
David L. Rosburg
Jerrid D. Chapman

March 30, 2005

Work performed under Cooperative Agreement

No. DE-FC36-97ID13554

Prepared for
U. S. Department of Energy

Prepared by
American Iron and Steel Institute
Technology Roadmap Program Office
Pittsburgh, PA 15220
 DISCLAIMER

“This report was prepared as an account of work sponsored by an agency of the United States
Government. Neither the United States Government nor any agency thereof, nor any of their
employees, makes any warranty, express or implied, or assumes any legal liability or
responsibility for the accuracy, completeness, or usefulness of any information, apparatus,
product, or process disclosed, or represents that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise, does not necessarily constitute or imply endorsement,
recommendation, or favoring by the United States Government or any agency thereof. The
views and opinions of authors expressed herein do not necessarily state or reflect those of the
United States Government or any age ncy thereof.”

“This report has been reproduced from the best available copy. It is available in paper copy and
electronic format.”

Number of pages in this report: 48

DOE and DOE contractors can obtain copies of this report

FROM: Office of Scientific and Technical Information, P. O.
Box 62, Oak Ridge, TN 37831. (615) 576-8401.

This report is publicly available from:

Department of Commerce
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
(703) 487-4650

And at the following website:

www.osti.gov/bridge

TRP 0040 – Final Report ii March 30, 2005

 Report Documentation Page

Title and Subtitle:

Validation of the Hot Strip Mill Model

Authors:

Richard A. Shulkosky, David L. Rosburg, Jerrid D. Chapman

Performing Organization Names, Address:

INTEG process group, inc.

11279 Perry Highway, Suite 107
Wexford, PA 15090
USA

Abstract:

The Hot Strip Mill Model (HSMM) is an off- line PC based software model originally developed
by the University of British Columbia (UBC) and the National Institute of Standards and
Technology (NIST) under the AISI/DOE Advanced Process Control Program from 1993 – 1998.
The HSMM was developed to predict the temperatures, deformations, microstructure evolution,
and mechanical properties of steel strip or plate rolled in a hot mill. In 2001, INTEG process
group, inc. undertook the current task of enhancing and validating the technology developed by
the UBC. With the support of the AISI, DOE and five North American steel companies, INTEG
embarked upon a multi- year plan under a DOE TRP project to upgrade, enhance and validate the
model referred to as the AISI Hot Strip Mill Model (HSMM) version 4. The steel company
participants (Dofasco, IPSCO, Stelco, US Steel, Weirton Steel) formed the HSMM Enhancement
Group to provide input and support to the effort. The goals of this project were twofold: 1) test
and validate the existing HSMM using operating data from the plants; and 2) enhance the
HSMM as required to improve the results.

With the release of HSMM version 6.2, the goals of the project have been successfully
completed. An extensive validation and verification program for the enhanced HSMM was
performed using a multitude of samples from the Enhancement Group steel companies.
Excellent agreement was obtained for tensile strength from a variety of steel chemistries and mill
configurations. Enhancement features incorporated into versions 6.0, 6.1, and now the final
version of the HSMM, 6.2, that have made it more flexible and practical to use include:
• Improved user interface
• Ability to link all models and track the material through the entire mill
• Improved temperature and force modeling
• Ability to calibrate the temperature and force models from plant data
• Ability to view and adjust the microstructure calculation algorithms and coefficients

The supporting steel companies have found outstanding value in the HSMM in saving them time
and money for a variety of practical applications. The HSMM continues to be marketed and sold

TRP 0040 – Final Report iii March 30, 2005

on a global basis as the industry’s leading PC-based off- line model for helping steel producers
and researchers improve the hot rolling process.

TRP 0040 – Final Report iv March 30, 2005

 FORWARD
INTEG process group, inc. would like to thank the American Iron and Steel Institute and the
Department of Energy for their continued financial support over the past four years enabling the
HSMM to be enhanced and validated so that it is a practical tool for the steel industry. The
assistance and support received from Joe Vehec, Director of the AISI Technology Roadmap
Program, over the years was very valuable and appreciated. Larry Kavanagh and the staff at the
AISI in Washington, D.C. are also warmly thanked for their continued support throughout the
project.

We would particularly like to acknowledge the efforts of Keith Barnes, Mark Fenton, Brian Joel,
and Tibor Turi of Stelco; Brian Nelson and J.J. Fitzpatrick of Dofasco; Matt Merwin and Eugene
Nikitenko of U.S. Steel; Cache Folkman, Shaojie Chen, and Steve Yocom of IPSCO; and Rich
Frey of Weirton Steel (now part of ISG) for the invaluable input and leadership they have
provided over the course of the four- year project. We are also grateful to all the HSMM
Enhancement Group companies for providing operating data and for their constructive feedback.

We would like to recognize the strong technical contributions of UBC, in particular Dr. Matthias
Militzer and Dr. Vladan Prodanovic for providing background information on the model, adding
new models for a new dual-phase steel, as well as on- going enhancements for the ROT; the
University of Pittsburgh, BAMPRI, in particular Dr. Anthony DeArdo and Dr. Isaac Garcia, for
conducting an analysis on the model; and the consultants who supported various tasks in the
development of the enhanced version including, Dr. Vladimir Ginzburg, Naum Kaplan, Robert
Ballas, Steven Lechuk and Dr. Daqing Jin (the lead researcher for the original project now an
employee of The Timken Company).

TRP 0040 – Final Report v March 30, 2005

 TABLE OF CONTENTS

FORWARD......................................................................................................................................v
EXECUTIVE SUMMARY ....................................................................................................... viii
1 Introduction............................................................................................................................. 1
2 HSMM Enhancements ............................................................................................................ 1
2.1 Improved Software Engineering ..................................................................................... 1
2.1.1 User’s Interface ....................................................................................................... 1
2.1.2 Fortran Code ........................................................................................................... 3
2.2 Improved Practicality in Thermo-Mechanical Calculations ........................................... 5
2.2.1 Material Tracking.................................................................................................... 6
2.2.2 Force Model............................................................................................................ 7
2.2.3 Motor Power Calculations ...................................................................................... 8
2.2.4 Width Changes ...................................................................................................... 10
2.2.5 Limit Checking ..................................................................................................... 12
2.2.6 Added Crown and Shape Models ......................................................................... 13
2.2.7 Additional Mill Equipment ................................................................................... 14
2.3 Improved Flexibility ..................................................................................................... 14
2.3.1 Added Single-Node Calculations.......................................................................... 14
2.3.2 Added Resistance to Deformation Force Model................................................... 16
2.3.3 Added Other Flow Stress Models ......................................................................... 18
2.3.4 Added Temperature Tuning Coefficients ............................................................. 19
2.3.5 Added Automatic Force Model Calibration ......................................................... 21
2.3.6 Added Plant Database Importing .......................................................................... 23
2.3.7 Handle Low Coiling Temperatures....................................................................... 23
2.4 Improved Microstructure/Mechanical Properties Calculations .................................... 24
2.4.1 Allow Chemistry Adjustments.............................................................................. 24
2.4.2 Added GradeBuilder Module................................................................................ 24
2.4.3 Extended ROT Transformation Model into Coiler............................................... 27
2.4.4 Improved Elongation Calculation......................................................................... 27
2.4.5 Improved Vanadium Precipitation Strengthening Calculation............................. 28
2.4.6 Added Models for Dual Phase Steel ..................................................................... 29
3 HSMM Validation ................................................................................................................ 31
3.1 Overview ....................................................................................................................... 31
3.2 Plant Data ...................................................................................................................... 31
3.3 Results ........................................................................................................................... 32
3.4 Validation Summary..................................................................................................... 34
4 HSMM User Documentation................................................................................................ 35
4.1 User’s Manual............................................................................................................... 35
4.2 Getting Started .............................................................................................................. 35
4.3 Calibration Guide .......................................................................................................... 35
4.4 Client Database Link Instructions................................................................................. 35
4.5 Microstructure Guide .................................................................................................... 35
4.6 Technical Manual.......................................................................................................... 35
5 Conclusion ............................................................................................................................ 36
Appendix A – HSMM User Documentation ................................................................................ 37
Appendix B – UBC Report on Dual Phase-Mo Steel................................................................... 37

TRP 0040 – Final Report vi March 30, 2005

 LIST OF FIGURES
Figure 1 – User Interface for HSMM version 4.0 ..................................................................................2
Figure 2 – User Interface for HSMM version 6.2 ..................................................................................3
Figure 3 – Modularity of Software Modules for Easy Modification or Replacement................................4
Figure 4 – Sample original code without error checking, comments, or descriptive names.......................4
Figure 5 – Sample Fortran 95 code with error checking, comments, & descriptive names........................5
Figure 6 – Tracking Calculation Points ................................................................................................6
Figure 7 – Speed Profile and Transfer Times between Mill Stands .........................................................7
Figure 8 – Force Geometric Factor with Peening Effect.........................................................................7
Figure 9 – Mill Stand Drive showing Motor Power and Torque Calculations ..........................................8
Figure 10 – Roll Bite illustrating contact length L’, bite angle ?, roll force P, and tensions S ...................9
Figure 11 – Lever Arm Coefficient ‘m’ as a Function of Roll Bite Geometry........................................10
Figure 12 – Increased Edger Efficiency with Grooved Edger Rolls ......................................................12
Figure 13 – Error and Warning Messages...........................................................................................13
Figure 14 – Shape Envelope and Calculated Curve .............................................................................14
Figure 15 – Single vs. Multiple Node Calculations ..............................................................................15
Figure 16 – Resistance to Deformation Geometric Factor ....................................................................17
Figure 17 – Resistance to Deformation Temperature Factor.................................................................17
Figure 18 – Shida Flow Stress as Function of %Carbon Content ..........................................................18
Figure 19 – Medina Flow Stress for HSLA-50....................................................................................19
Figure 20 – Single Node Thermal Model Tuning Coefficients .............................................................20
Figure 21 – Multiple Node Thermal Model Tuning Coefficients ..........................................................20
Figure 22 – Chart of Calculated (lines) vs. Measured (dots) Temperatures for Tuning ...........................21
Figure 23 – Flow Stress Multiplier Regression....................................................................................22
Figure 24 – Flow Stress Calibration Screen........................................................................................22
Figure 25 – Plant Database Link ........................................................................................................23
Figure 26 – User Chemistry Field ......................................................................................................24
Figure 27 – GradeBuilder Screen.......................................................................................................25
Figure 28 – Thermal Property Selection by Phase (UBC Method) ........................................................26
Figure 29 – BISRA Thermal Property Selection (BISRA Method) .......................................................27
Figure 30 – New Elongation Curve ....................................................................................................28
Figure 31 – Improvement in HSLA-Vanadium Grade Yield Strength Predictions..................................29
Figure 32 – Cooling Path on the Runout Table for Dual Phase Steels ...................................................30
Figure 33 – Finishing Temperature Comparison Figure 34 –Coiling Temperature Comparison .33
Figure 35 – Yield Strength Comparison Figure 36 – Tensile Strength Comparison ...............33
Figure 37 – Ferrite Grain Size Comparison.........................................................................................34

LIST OF TABLES
Table 1 – Mill Configurations of Supporting Steel Companies.............................................................31
Table 2 – Processing Parameter Ranges .............................................................................................31
Table 3 – Chemistry Range ...............................................................................................................32
Table 4 – Statistical Analysis of Comparison between Actual and Calculated .......................................34

Please note that the Appendices noted in the Table of Contents are not included in the DOE
submission due to Intellectual Property and Confidentiality issues.

TRP 0040 – Final Report vii March 30, 2005

 EXECUTIVE SUMMARY
The Hot Strip Mill Model (HSMM) is an inventive off- line PC based software model originally
developed by the University of British Columbia (UBC) and the National Institute of Standards
and Technology (NIST) under the AISI/DOE Advanced Process Control Program from 1993 –
1998. The HSMM was developed to predict the temperatures, deformations, microstructure
evolution, and mechanical properties of steel strip or plate rolled in a hot mill. In 2001, INTEG
process group, inc. undertook the current task of enhancing and validating the technology
developed by UBC. The objective was to test, upgrade and validate the core models used for
predicting the temperature, forces, microstructure evolution and final mechanical properties of
steel produced on a hot strip mill. The scope of work includes validating and/or replacing
various sub- models, adding practical application functions, updating the users interface to
facilitate the ease of use of the model and to provide adequate documentation

With the support of the AISI, DOE and five North American steel companies, INTEG embarked
upon a multi- year plan under a DOE TRP project to upgrade, enhance and validate the model
referred to as the AISI Hot Strip Mill Model (HSMM) version 4. The steel company participants
(Dofasco, IPSCO, Stelco, US Steel, Weirton Steel) formed the HSMM Enhancement Group to
provide input and support to the effort.

The project included a detailed review of each sub- module of the model and a validation and/or
replacement of each sub- module. Practical application functions, an updated user’s interface to
facilitate the ease of use of the model and adequate documentation was to be provided. A five-
phase plan was developed to validate the Hot Strip Mill Model. Phases 1, 2 and 3 of the
extended work plan were to conduct a technical audit of the model and to develop a plan to
improve the model for practical applications. Phases 4 and 5 were to develop, validate and
calibrate an enhanced version of the model with proper documentation, advanced modules, etc.

Phase 1, which undertook several tasks to bring the HSMM to a certain level of usability, was
completed during the 3rd Quarter of 2001. INTEG then released to the participants on August 4,
2001 an updated version of the HSMM.

Phase 2 was to flow chart, document and identify the inputs and outputs of each module (or sub
module) for the cur rent version of the HSMM. Although some areas of the model were difficult
to document due to limited information, this phase was completed as much as practical during
the 1st Quarter of 2002 and was to be completed during phases 4 and 5 when additional
information was available.

Phase 3 was to validate each sub module, but validation of each sub module using alternate
models or plant data was not possible due to the design of the original model. Instead, based
upon previous tests and published results of the model by the steel companies and UBC, an
evaluation of the modules as a whole was completed as much as practical during the 1st Quarter
of 2002.

Phase 4 involved the integration of the existing and new modules to make a cohesive model
capable of covering all the needed functions to properly predict the temperature evolution,

TRP 0040 – Final Report viii March 30, 2005

forces, microstructure evolution and final mechanical properties. This task was completed and
validated with an initial set of data in the 4th Quarter of 2002.

Phase 5 involved the validation of the model and was completed in the 4th Quarter of 2003.
Excellent agreement was obtained between the actual and calculated values for tensile strength
and yield strength. Additional work under Phase 5 was completed in the 1st Quarter of 2004 and
resulted in the addition of GradeBuilder, which allows the user to develop and add new grades of
steel by selecting or adding new algorithms and coefficients. Additional work under Phase 5 was
completed in the 4th Quarter of 2004 that included an upgrade to the ROT tracking and thermal
models, the addition of “soft” coupling of mill equipment, and the implementation of basic
equations for dual phase steels.

The successful result of this project was the final release of the Hot Strip Mill Model (HSMM) as
version 6.2. This version allows users to easily set-up their mill configuration, simulate a rolling
mill schedule and calibrate the model for a variety of grades of steel. The enhanced HSMM was
validated using a multitude of samples from the Enhancement Group steel companies. Excellent
agreement was obtained for comparisons between measured mechanical properties and those
calculated by the HSMM.

Enhancement features incorporated into version 6.2 of the HSMM that have made it more
flexible and practical to use include:
• Improved user interface
• Ability to link all models and track the material through the entire mill
• Improved temperature and force modeling
• Ability to calibrate temperature and force models with plant data
• Ability to adjust microstructure calculation algorithms and coefficients

The supporting steel companies have found outstanding value in the HSMM in saving them time
and money for a variety of practical applications. The HSMM continues to be marketed and sold
on a global basis as the industry’s leading PC-based off- line model for helping steel producers
and researchers improve the rolling process.

TRP 0040 – Final Report ix March 30, 2005

1 Introduction
This report provides a summary of the Enhancements (Section 2), Validation (Section 3),
and Documentation effort (Section 4) that INTEG performed over the life of the entire
“Validation of the Hot Strip Mill Model” project. Detailed user’s manuals and technical
documentation that are provided in Appendices A and B are confidential "Protected Metals
Initiative Data" and are available only to the project participants.

2 HSMM Enhancements
The enhancements that were identified as necessary improvements to the HSMM were
related to four main categories:
• Software Engineering (section 2.1)
• Practicality in Thermo- mechanical Calculations (Section 2.2)
• Improved Flexibility (Section 2.3)
• Microstructure/Mechanical Property Calculations (Section 2.4)

2.1 Improved Software Engineering

The Hot Strip Mill Model version 4.0 as delivered by UBC was a stand-alone Windows 95
application. It was a composition of a graphical User’s Interface and about eight Fortran
executables programs and numerous text data files. Each of the Fortran modules
represented a particular process area such as the Roughing Mill, Finishing Mill, Runout
Table, etc. The interface was an aid for the preparation of input files before launching
control to one of the Fortran modules. The graphical user interface was designed in
Microsoft’s Visual Basic 5.0.

2.1.1 User’s Interface

The User’s Interface consists of Microsoft-compatible Windows screens, menu selections,
buttons, data entry and display fields, charts, etc. that the user interacts with for program
control and exchanging data with the Fortran calculation software. Because Microsoft was
encouraging its Visual Studio customers to migrate up to its new .NET Framework
environment, it was an obvious decision to keep up with the current technology and
completely redesign the User’s Interface screens using .NET. Some of the changes that
went into the redesigned User’s Interface software were:
• Divide the Interface screens into five main functional areas:
o Mill Configuration
o Grade Calibration
o Rolling Schedules
o Data Exporting / Reporting
o GradeBuilder
• Switch from saving data in text files to Microsoft Access database files
• Add the ability to import rolling schedule data from the plant’s database
• Exchange data with the Fortran dynamic link library via well-designed large data
structures
Figure 1 provides a picture of the main User’s Screen for HSMM version 4.0.

TRP 0040 – Final Report 1 March 30, 2005

 Figure 1 – User Interface for HSMM version 4.0

The HSMM version 6.2 utilizes a user- friendly interface (see Figure 2) allowing each mill
to be accurately configured, each rolling schedule to be set-up in detail, each grade of steel
to be accurately characterized and the final results to be viewed, charted, reported and
exported, as needed. The user interface is divided into the following main areas:
• The Mill Configuration Screen allows the user to set-up the rolling mill to be
used and includes the furnace area, roughing area (mills, edgers, sprays), heat
retention area (coil box, heat panels), finishing area (mills, edgers, sprays), run out
table and mill exit area.
• The Calibration Screen allows the user to calibrate the model for each grade of
steel being simulated. During the overall project set-up, the user selects a specific
set of coefficients to be used for the grade of steel being processed via a specific
rolling mill schedule.
• The Rolling Schedule Screen is used to enter the processing parameters of the
piece being modeled and to view the results of the single node and multiple node
calculations. The screen allows the user to view and configure the Initial Data,
Pass Data, Speed/Time, Shape/Crown, Temperature Data, Rolling Parameters,
Microstructure, Run Out Table, Charts and Summary Results.
• The Data Exporting Screen allows the user to export data easily from the model
to data files that can be easily read by Microsoft Excel or similar software
packages for further analysis.
• The Reporting Screen is used for printing reports containing Mill Configuration,
Calibration, and Rolling Schedule data
• The GradeBuilder Screens allows the user to “build” his/her own grade in
addition to the nine sample grades characterized for the HSMM

TRP 0040 – Final Report 2 March 30, 2005

 Figure 2 – User Interface for HSMM version 6.2

2.1.2 Fortran Code

The version 4.0 executable programs were built from eight Fortran source files. To make
this software more understandable and maintainable several enhancements were made to
bring the software up to modern software engineering standards:
• Sub-divide the eight Fortran source files into smaller individual modules
• Eliminate duplication of functionality between the original source files
• Use longer, more descriptive variable names
• Add program block separators and descriptive comments
• Add reasonability checking to module input parameters
• Add calculation error checking to avoid crashes (divide by zero, square root of
negative value, exponent over- or underflows)
• Update the code to Fortran 95 standard
Figure 3 illustrates how the software was divided into individual modules to make
maintenance of the software easier. A comparison of the version 4.0 Fortran code in
Figure 4 and the enhanced Fortran 95 code in Figure 5 shows the improvements that were
made in readability and error checking.

TRP 0040 – Final Report 3 March 30, 2005

 User Interface
Mill Config Rolling Sched Coefficients Tools Graphs Reports
Microstructure Equations
Recryst Precipt Heat of Trans Grain Type Hardening Mech Prop
Tracking
Temps Speed Location Nodes Time Data
Temperature
Radiation Conduction Convection Boiling Flow Sprays
Rolling
Power Forces Strain Strain Rate Stresses Deformation

Modify/Replace
Figure 3 – Modularity of Software Modules for Easy Modification or Replacement

C
SIG1=SIG/9.81
REDF=REDC(NROLS3)/100.0
DH=H1-H2
100 FRD = DSQRT (RD / H2)
DQRT=DSQRT(REDF/(1.0D0-REDF))
PHI = DTAN(PI*DLOG(1.0D0-REDF)/(8.0D0*FRD)+0.5D0*
1 DATAN (DQRT)) / FRD
HNUET = 2.0D0 * RD * (1.0D0 - DCOS (PHI)) + H2
QP = PI / 2.0D0 / DQRT * DATAN (DQRT) - PI / 4.0D0 -
1 FRD / DQRT * (DLOG (HNUET / H2) + 0.5D0 *
2 DLOG (1.0D0 - REDF))
P = SIG1 * DSQRT(RD * DH) * QP
ABD = DABS ((P - P0) / P)
IF (ABD .GT. 1.0D-3) THEN
RD = R * (1.0D0 + C * P / DH)
P0 = P
GOTO 100
ENDIF

Figure 4 – Sample original code without error checking, comments, or descriptive names

TRP 0040 – Final Report 4 March 30, 2005

 ! -----------------------------------------------------------------------------
!
! BLOCK 200 - initialize varaibles and calculate Sim's factor
!
! -----------------------------------------------------------------------------

!! Calculate the roll deformation force.

SQRT_PARAM = DEFORM_RADIUS / EXIT_DIM
! Check if SQRT factor is out of range
IF (SQRT_PARAM < 0.) THEN
! Exp function out of range
ERROR_CODE%DESCRIPTION = ERROR_INVALID_CALC
ERROR_CODE%AREA = MOD_SIMS_GEO_FACTOR
RETURN
END IF
ROLL_DEFORM_FORCE = SQRT(SQRT_PARAM)

!! Calculate the frictional force experienced by the roll.

SQRT_PARAM = DECIMAL_REDUCTION / (1.0 - DECIMAL_REDUCTION)
! Check if SQRT factor is out of range
IF (SQRT_PARAM < 0.) THEN
! Exp function out of range
ERROR_CODE%DESCRIPTION = ERROR_INVALID_CALC
ERROR_CODE%AREA = MOD_SIMS_GEO_FACTOR
RETURN
END IF
FRICT_FORCE = SQRT(SQRT_PARAM)

!! Calculate the angle of contact of the strip at a neutral point.

!! See equation 2.55 in the steckel mill model theoretical manual.
CONTACT_ANGLE = ATAN(PI * LOG(1.0 - DECIMAL_REDUCTION) / &
(8.0 * ROLL_DEFORM_FORCE) + 0.5 * ATAN(FRICT_FORCE)) / ROLL_DEFORM_FORCE

!! Find the thickness at the neutral point.

THICK_NEUT = 2.0E0 * DEFORM_RADIUS * (1.0D0 - COS(CONTACT_ANGLE)) + EXIT_DIM

!! Calculate Sim's geometrical factor

!! See equation 2.54 in the steckel mill model theoretical manual.
QP = PI / 2.0 / FRICT_FORCE * ATAN(FRICT_FORCE) - PI / 4.0 - &
ROLL_DEFORM_FORCE / FRICT_FORCE * (LOG(THICK_NEUT / EXIT_DIM) + &
0.5 * LOG(1.0 - DECIMAL_REDUCTION))

Figure 5 – Sample Fortran 95 code with error checking, comments, & descriptive names

After the Fortran code was sub-divided into smaller calculation modules, flow charts were
developed as a permanent record to better understand the program’s logic flow. Instead of
building a series of executable programs, the Fortran source code was built into a single
dynamic link library (dll) file of individual modules that could be called from the User’s
Interface.

2.2 Improved Practicality in Thermo-Mechanical Calculations

The HSMM version 4.0 ran as seven separate models for the various hot mill areas and
mill configuration types: Roughing Mill Model, Reversing Roughing Mill Model, Coil
Box Model, Finishing Mill Model, Runout Table Model, Deformation Model, Down
Coiler Model, and Steckel Mill Model. The results of each model were not linked to the
input of the next successive model. It was discovered that the HSMM version 4.0 lacked
the ability to simulate certain hot mill equipment and normal processing conditions. It was
also possible for the HSMM to simulate impossible, overly-aggressive reduction

TRP 0040 – Final Report 5 March 30, 2005

conditions. To improve the practicality of the HSMM, several enhancements were
incorporated relating to linking all the models together for the entire hot mill, adding
various limit checking, and simulating additional pieces of mill equipment.

2.2.1 Material Tracking

As mentioned above, version 4.0 ran each mill area (rougher, finisher, runout table, coiler)
as separate models. In version 6.2, the entire hot mill is simulated sequentially from drop
out of the reheat or tunnel furnace to exiting into the coiler or cooling bed. Not only did
this improve the efficiency of running the model for the user, but also improved the
accuracy of the temperature and microstructure calculations by continuously tracking the
material’s process parameters such as temperature, grain size, precipitation in austenite,
retained strain, etc. through all areas of the mill. Three calculations points along the
material length were chosen for tracking: the headend, the middle point, and the tailend as
shown in Figure 6.

Calculation Points
Tail Middle Head

Figure 6 – Tracking Calculation Points

Accurately tracking the timing of the three points through the entire hot mill requires user
input of the threading and top speeds during rolling and tables speeds during transfer
between stands as well as the stopping distances and delay times for passes at reversing
stands. An example speed profile between two individual rolling stands is shown in Figure
7. Applying the actual acceleration and deceleration rates when changing speeds improves
the timing and temperature calculations. With more accurate temperature predictions
being calculated and provided to the microstructure calculations, more accurate
microstructure and mechanical property calculations were also achieved.

TRP 0040 – Final Report 6 March 30, 2005

 Stand 1 Top Table Speed Stand 2
Rolling Speed Rolling Speed

Head Middle Tail Head Middle Tail

Out of out of out of into into into
Stand 1 Stand 1 Stand 1 Stand 2 Stand 2 Stand 2

Figure 7 – Speed Profile and Transfer Times between Mill Stands

2.2.2 Force Model

HSMM version 4.0 used the NIST developed equations and coefficients to calculate flow
stress ‘s’ and the traditional Sim’s geometric factor ‘Qp’ to calculate the rolling force.

2
F= σ Q p R' ∆h W (2.1)
3

However, when rolling thick product in the early roughing passes, deformation beyond the
arc of roll contact (also known as the peening effect) occurs that results in higher rolling
forces. To compensate for this effect, an adjustment was made to the Sim’s geometric
factor as shown in Figure 8. This adjustment is a function of the roll bite aspect ratio ‘a’
(contact length L' divided by average thickness).

Qp = 0.7924 + 1.778 * exp(-2.148*a) for a < 1.0 (2.2)

Qp = Force Geometric Factor

3

2.5
Geometric factor

Peening
2 Effect
1.5
Sim's Factor
1

0.5

0
0 2 4 6 8 10
Aspect Ratio

Figure 8 – Force Geometric Factor with Peening Effect

TRP 0040 – Final Report 7 March 30, 2005

2.2.3 Motor Power Calculations
In addition to roll separating force limitations, other concerns in a hot mill are the limits
from the mill stands’ motor power and torque. It is a futile exercise to develop a new
rolling practice for a product that achieves the target mechanical properties, but the mill
doesn’t have the power to produce it. HSMM version 6.2 was enhanced with the addition
of optional motor power and torque calculations as shown in Figure 9. These calculations
are optional to the HSMM user because power calculations require a number of motor data
parameters (rated power, RPMs, maximum load ratio, gear ratio of the gear box, etc) be
input for each rolling stand.

Power:
Electrical Mechanical

Spindle Work Roll

Gear Material
Motor Shaft
Box

Spindle Work Roll

Torque: at Shaft at Spindles

Figure 9 – Mill Stand Drive showing Motor Power and Torque Calculations

Before calculating motor power, the total rolling torque ‘M’ is calculated from the rolling
force ‘P’ in kN and the lever arm ‘a’ in mm and multiplied by 2 to consider both work
rolls.

2* P*a
M = [ kN − m ] (2.3)
1000

The total rolling torque is affected when entry and/or exit tension on the material is
present. In this calculation, tension not only lowers the rolling force and therefore the
rolling torque, but entry tension S1 increases the rolling torque while exit tension S2
decreases it.

2 * a * W * L' ( K − s avg ) R (S − S 2 )
M= * + * 1 [ kN − m] (2.4)
1000 1000 1000 1000

savg is the average specific tension in MPa.

s avg = β * s1 + (1 − β ) * s 2 (2.5)

TRP 0040 – Final Report 8 March 30, 2005

 Relative Edging Efficiency
2.00

(eff. grooved rolls)/(eff.flat rolls) 1.95

1.90

1.85

1.80

1.75

1.70

1.65
0 20 40 60 80 100 120 140

width draft, mm

Figure 12 – Increased Edger Efficiency with Grooved Edger Rolls

2.2.5 Limit Checking

In addition to the power checking that was described in section 2.2.3, HSMM version 6.2
has incorporated a system of checking both user-entered and calculated values against
maximum and/or minimum limit values. These limits are configured by the user and many
are of these limits are optional. A list of any limit violations is displayed to the user.

Entered parameters that are limit checked:

• Slab temperatures
• Slab dimensions
• Roll diameters
• Work roll speeds
• Table speeds

Calculated parameters that are limit checked:

• Material lengths and widths
• Bite angles
• Rolling forces
• Rolling torques
• Motor output powers
• Edger buckling

Error and limit warning messages are displayed to the user as shown in Figure 13.

TRP 0040 – Final Report 12 March 30, 2005

 Figure 13 – Error and Warning Messages

2.2.6 Added Crown and Shape Models

Even though the hot mill can roll a particular product within its own limits, the product
may not be salable if its shape (flatness) is unacceptable. Another level of practicality was
added to version 6.2 with the incorporation of the crown and shape models. These models
can be turned on by the user to calculate the crown (profile) on the work piece after each
reduction. The exit crown of the work piece is calculated from the deflection of the roll
stack due to the rolling load, the crowns on the work and backup rolls, and any applied
mechanical bending forces. To maintain a flat product with good shape, its relative (%)
crown can change only so much until the internal stresses either cause buckles down the
center of the strip or waves down the edges. The amount of allowed crown change has
been defined by an upper and lower limit that produces a shape “envelope” as shown in
Figure 14.

c1 c 2
∆c = − (2.11)
h1 h 2
where c1 = entry crown, c2 = exit crown

1 .86 1. 86

(Edge Wave) − 80 

h h
> ∆c > 40  (Center Buckle) (2.12)
 w  w

TRP 0040 – Final Report 13 March 30, 2005

 Shape Envelope
20

10

Relative Crown Change

Center Buckle
0
(X1000) F1 F2 F3 F4 F5 F6 F7
-10
Edge Wave
-20

-30

-40
FM Stand
Center Buckle Edge Wave Strip Shape

Figure 14 – Shape Envelope and Calculated Curve

By adjusting the reductions in the finishing mill and the bending forces (if available) the
material shape can be made flat or at least improved.

2.2.7 Additional Mill Equipment

Enhancing the HSMM for version 6.2 included expanding the temperature models to
simulate some common mill equipment items such as heat covers (also known as thermal
covers, thermal panels, table covers, etc) and cooling beds for plate products.

Heat Covers
Heat covers are modeled by applying an elevated ambient temperature input for the
headend and a calculated ambient temperature for the tailend based on the headend
temperature that pre- heats the covers. Both top-and-bottom and top-only heat covers can
be modeled.

Cooling Bed
A cooling bed is available for plate products to be sent after the Runout table for
simulation of radiation and convection cooling at the mill ambient temperature.
Simulation of forced convection, however, was not included.

2.3 Improved Flexibility

Several of the enhancements to HSMM version 6.2 originated from the need to provide
flexibility to the user in making choices and adjustments to help improve the model’s
results and handle more processing conditions.

2.3.1 Added Single-Node Calculations

The HSMM version 4.0 used the implicit finite difference method for calculating
temperatures at multitude nodes through the thickness down the center of the work piece.
This calculation method has been preserved in version 6.2 and calculates 101 nodes

TRP 0040 – Final Report 14 March 30, 2005

 k G = b1 * α + c1;. for.α <= α c (2.20)

k G = a 2 *α 2 + b 2 * α + c2;. for.α > α c (2.21)

Resistance to Deformation
Geometric Factor
4
3.5
Geometric Factor

3
2.5
2
1.5
1
0.5 αc
0
0 1 2 3 4 5 6
Arithmetic Average Aspect Ratio

Figure 16 – Resistance to Deformation Geometric Factor

The temperature factor (kT ) is a function of the temperature difference between the selected
Normalized Temperature (TN) and the material temperature. The graph of the following
two equations for the temperature factor is shown in Figure 17.

k T = 1 + b1 * (TN − T ) ; for T >= TN (2.17)

k T = 1 + b 2 * (TN − T ) ; for T < TN (2.18)

Resistance to Deformation
Temperature Factor
1.6
Temperature Factor

1.4
1.2
1
0.8
0.6
0.4
0.2 TN
0
800 900 1000 1100 1200 1300
Material Temperature [ºC]

Figure 17 – Resistance to Deformation Temperature Factor

TRP 0040 – Final Report 17 March 30, 2005

The normalized resistance to deformation (K N) is the value of resistance to deformation of
the rolled material at the selected normalized temperature and at a normalized aspect ratio
which has a value of 1.

This method is semi-empirical, but allows the model to accurately calculate the force
predictions by using plant data from previously processed coils for any grade of steel.
Once the model is calibrated using this method, new rolling schedules for the same grade
can be accurately simulated for conducting what- if analysis.

The user has the choice of which rolling force model to use, either flow stress or resistance
to deformation.

2.3.3 Added Other Flow Stress Models

Both the Shida and Medina flow stress calculation methods were added for using grades of
steel not characterized in the lab for the NIST developed equations nor were previously
rolled in the user’s mill to provide data for the resistance to deformation calibration. Like
the NIST flow stress method, these methods define the flow stress of steels during hot
plastic deformation as a function of temperature, strain, strain rate, and austenite grain size.
However, what distinguishes these two flow stress models and makes them useful is that
they calculate flow stress also as a function of the steel’s chemical composition.

The Shida flow stress model was developed by S. Shida of Hitachi Research in 1974. This
model is applicable to C-Mn steel grades that may contain a small amount of
microalloying elements. Figure 18 is a graph illustrating the effect of changing carbon
content of the Shida flow stress.

Shida's Flow Stress

from Carbon Content
400
Flow Stress [MPa]

350
300 %C

250 0.038
200 0.1
150 0.5
100
50
700 800 900 1000 1100 1200 1300
Temperature [C]

Figure 18 – Shida Flow Stress as Function of %Carbon Content

The Medina flow stress model was developed by S.F. Medina and C.A. Hernandez in
1996. This model can be applied to C-Mn steels as well as those containing microalloys

TRP 0040 – Final Report 18 March 30, 2005

such as Vanadium (V), Titanium (Ti), and Niobium (Nb). A graph of the Medina flow
stress for a HSLA-50 grades is shown in Figure 19.

Medina's Flow Stress

HSLA - 50
250
Flow Stress [MPa]

200

150

100

50

0
700 800 900 1000 1100 1200 1300
Temperture [C]

Figure 19 – Medina Flow Stress for HSLA-50

2.3.4 Added Temperature Tuning Coefficients

The ability of the HSMM to accurately simulate rolling loads and final mechanical
properties is directly dependent on its ability to accurately simulate the correct material
temperature evolution through the hot mill. Temperature evolution in the material mainly
involves the effects of radiation, conduction to the work rolls, conduction to water sprays,
and heating from deformation. Although the HSMM requires the user to input a number of
parameters that characterize the mill equipment and operating conditions, there will always
be a set of unaccountable factors that are difficult, if not impossible or impractical, to
include in any thermal model (e.g. the cooling effect of roll cooling water that reaches the
strip or the effect of a water spray that remains partially plugged). Because these variable
factors cannot be modeled, other thermal models need to be adjusted with tuning
coefficients to compensate for these factors and produce reliable results. Separate thermal
model tuning coefficients and multipliers were added for the single- node and multiple-
node models as shown if Figures 20 and 21, respectively.

TRP 0040 – Final Report 19 March 30, 2005

 Figure 20 – Single Node Thermal Model Tuning Coefficients

Figure 21 – Multiple Node Thermal Model Tuning Coefficients

The HSMM can plot calculated temperatures as well as entered measured temperatures
entered by the user. Measured temperatures may be recorded in Engineering Logs or
stored in the plant’s database. The source of these temperatures may be from pyrometer
readings at various locations in the mill or they may be calculated by the plant’s on-line
Level 2 computer at each stand or pass. From the HSMM temperature chart of calculated

TRP 0040 – Final Report 20 March 30, 2005

vs. measured values, the user can adjust the thermal model tuning coefficients in an
iterative process of running the model and adjusting the tuning value until the calculated
values match the measured ones as shown in Figure 22.

Figure 22 – Chart of Calculated (lines) vs. Measured (dots) Temperatures for Tuning

2.3.5 Added Automatic Force Model Calibration

Like the thermal models need for a tuning method to be more accurate, the three flow
stress methods needed a tuning method to make the force model more accurate. All three
methods consider the temperature, strain, strain rate, and austenite grain size for their
calculation of flow stress. If the force model is using one of the flow stress methods and it
is not providing accurate force predictions, the difficulty is determining which flow stress
coefficients to adjust.

To simplify the calibration procedure, it was decided to only adjust the flow stress based
on temperature and to let the HSMM calculate its own calibration coefficients for each
grade. By entering measured roll bite entry temperatures and rolling forces into the
HSMM for one or more rolling schedules of the same grade, the flow stress calibration
procedure can be initiated by the click of a button. This procedure calculates the ratios of
the measured to the calculated rolling forces and then performs a second order polynomial
regression on this set of ratios vs. temperatures to determine the A, B, and C coefficients
for the flow stress tuning multiplier. An example regression calculation and graph is
provided in Figure 23 and the flow stress calibration screen is shown in Figure 24.

TRP 0040 – Final Report 21 March 30, 2005

 Flow Stress Multiplier = A * T 2 + B * T + C (2.18)

Temp Fmeas Fcalc Ratio = m/c

1225 1505 1500 1.0033 1.02 2
1.01 y = 7E-08x + 0.0001x + 0.7721
1220 1750 1735 1.0086 2
1.00 R = 0.9569
1210 2135 2164 0.9866
0.99
1200 2080 2099 0.9909
0.98
1180 2165 2206 0.9814
0.97
1150 2215 2243 0.9875 0.96
1050 2630 2769 0.9498 0.95
1030 2590 2711 0.9554 0.94
1005 2410 2549 0.9455 0.93
990 2095 2228 0.9403 0.92
970 1690 1802 0.9378 900 950 1000 1050 1100 1150 1200 1250
950 1450 1562 0.9283

Figure 23 – Flow Stress Multiplier Regression

Figure 24 – Flow Stress Calibration Screen

Once a grade is calibrated with its own set of A, B, and C calibration coefficients, the flow
stress multiplier function produces multiplier values within a range around 1.0 that adjust
the flow stress and force calculations to better match the measured forces. New rolling
schedules created for the grade can be expected to have improved force predictions.

TRP 0040 – Final Report 22 March 30, 2005

2.3.6 Added Plant Database Importing
The ability to import data from a plant database and automatically create a Rolling
Schedule was added in version 6.2. This enhancement allows the user to simulate a
previously rolled coil without having to manually enter the rolling parameters, such as
thicknesses, speeds, measured forces, etc. To implement this feature the user must set up
an ODBC connection to the plant database via Administrator Tools in his Windows
Control Panel. Then by invoking a query into the plant database, three tables of data must
be generated in a format required by the HSMM. Once the tables are created, the HSMM
Database Link utility screens are used to import data and create new rolling schedules as
shown in Figure 25.

Figure 25 – Plant Database Link

2.3.7 Handle Low Coiling Temperatures

The runout table model for HSMM version 4.0 was developed and tested for normal
coiling temperatures down to 550° C. For simulating certain advanced high strength steels
(AHSS), much lower coiling temperatures are required to produce the desired bainite and
martensite phases. It was discovered during HSMM simulations that the runout table
models for both the multiple node and single node models could not be tuned to
simultaneously produce the intermediate temperatures and low coiling temperatures that
were actually observed in plant trials. The plant data showed there was very rapid cooling
of the material at temperatures below 450° C.

To increase the heat transfer in this low temperature region, a multiplier was introduced
into version 6.2 that could be tuned to match actual data. This low coiling temperature
multiplier is an equation that is a function of temperature and a tuning factor ‘A’. Separate
tuning factors are applied to the single node and multiple node models.

TRP 0040 – Final Report 23 March 30, 2005

2.4 Improved Microstructure/Mechanical Properties Calculations

2.4.1 Allow Chemistry Adjustments

The microstructure evolution and final mechanical properties models developed by UBC
for version 4.0 were based on eight grades of steel with specific chemistries. These models
contained equations with coefficients determined from lab tests. Many of these equations
were chemistry-dependent, but the chemistry values applied to these equations were fixed
for each of the eight grades that could be selected.

The first step in version 6.2 in making the microstructure evolution models more flexible
was to allow the user to enter the actual chemistry of the grade he/she is simulating. Aft er
entering the actual User Chemistry as shown in Figure 26, the user then selects the grade
with chemistry closest to the entered chemistry. The grades available for selection include
the nine sample grades and any others that have been built using GradeBuilder as described
in Section 2.4.2. By allowing the user to enter a chemistry that differs somewhat from the
selected grade’s chemistry, the microstructure results are generally improved. If the user
enters chemistry values that deviate significantly from the selected grade, a message is
displayed that warns the user that the microstructure results may be suspect.

Figure 26 – User Chemistry Field

2.4.2 Added GradeBuilder Module

The next step in version 6.2 in making the microstructure evolution models more flexible
was to allow the user to “build” his/her own grade in addition to the nine sample grades.
The purpose of adding the GradeBuilder module to the HSMM was to change the user’s
view of the microstructure models from being a rigid “black box” to being an open
configuration panel for building a new grade or modifying an existing grade.
GradeBuilder not only allows the user to see what equations and coefficients are used, but
allows him/her to select which algorithms to use and adjust the coefficients. The user can
even write their own algorithms and select them for use with their own grade.

To build a new grade, the sample grade that is closest to the new grade can be duplicated
and given a new name. Then the equations and coefficients for each microstructure
process during the austenite phase, phase transformation, and final mechanical properties
can be selected to best represent the microstructure characteristics of the new grade. The
austenite process selection screen of the GradeBuilder is shown in Figure 27.

TRP 0040 – Final Report 24 March 30, 2005

 Figure 27 – GradeBuilder Screen

Within GradeBuilder, the user also has the ability to select between two different methods
of determining thermal properties of the grade being built. Method 1 – (UBC) (see Figure
28) has the thermal properties split into three phases (austenite, ferrite and pearlite curves).
This method will accurately calculate the thermal properties based on when phase
transformation occurs.

TRP 0040 – Final Report 25 March 30, 2005

 Figure 28 – Thermal Property Selection by Phase (UBC Method)

Method 2 – (BISRA) uses curves developed by the British Iron and Steel Research
Association (see Figure 29). This method will describe the thermal properties based on
when phase transformation occurred during development of the curves. The advantage of
using these curves is that the range of the model can now be expanded to uses outside of
the scope of the sample grades of steel for thermo- mechanical calculations (i.e.: stainless
steel, Dual Phase, TRIP steels).

TRP 0040 – Final Report 26 March 30, 2005

 Figure 29 – BISRA Thermal Property Selection (BISRA Method)

2.4.3 Extended ROT Transformation Model into Coiler

In HSMM version 4.0, all transformation was expected to occur on the Runout Table. In
cases where transformation did not fully occur, an empirical equation was employed to
predict the final ferrite grain size and final ferrite fraction.

In HSMM version 6.2, this empirical equation was removed, and the transformation
prediction equations used for the Runout Table are extended for use in the coiler. In this
way, the correct cooling path is used to more accurately predict the transformation
conditions in the coiler.

2.4.4 Improved Elongation Calculation

In the HSMM version 4.0 for the eight base grades, the elongatio n was calculated as a
function of the tensile strength defined by two straight lines. Plant data showed that for the
low tensile strength grades the elongation was being under-predicted. A power curve
shown in Figure 30 was fit from plant data and introduced into version 6.2 to improve the
elongation calculations, especially in the lower tensile strength range.

TRP 0040 – Final Report 27 March 30, 2005

 Elongation = 7085*T^-0.88
50

40
Elongation %
30

20

10

0
200 400 600 800 1000
Tensile Strength, MPa
Line 1 Line 2 Curve Fit

Figure 30 – New Elongation Curve

2.4.5 Improved Vanadium Precipitation Strengthening Calculation

For HSMM version 6.2, a chemistry-based Vanadium precipitation strengthening model
was developed. Version 4.0 provided only a constant value for potential precipitation
strengthening that was independent of the Vanadium content. After combining with
Titanium, any free Nitrogen combines with Vanadium in a 4:1 ratio. The maximum
strengthening that is available from precipitation is a function of the VN and excess
vanadium. .

VNeff = Min ([ Nfree] * 4, [V ]) (2.19)

P.S . = a * VNeff + b * ([V ] − VNeff ]) (2.20)

where coefficients a and b were determined

The actual amount of precipitation strengthening is a function of the Shercliff- Ashby aging
curve as before. An example of the improved results is shown in Figure 31.

TRP 0040 – Final Report 28 March 30, 2005

 Vanadium Grade Yield Strengths
490
470
450
430
410
390
370
350
1 2 3 4 5

Meas New Old

Figure 31 – Improvement in HSLA-Vanadium Grade Yield Strength Predictions

2.4.6 Added Models for Dual Phase Steel

For this enhancement, UBC was contracted to perform the necessary lab tests to develop a
microstructure model for hot strip rolling of Dual Phase-Mo 600 steel. The result of this
work produced the following new models for this steel:
• Ferrite Transformation Model
o Enhanced JMAK model
o Enhanced ferrite grain size model
• Bainite Transformation Model
• Martensite Transformation Model
• New Mechanical Properties Model

Figure 32 shows the cooling path required to produce this grade. See Appendix B for UBC
report.

TRP 0040 – Final Report 29 March 30, 2005

 TStart

CR I
α formation
Isothermal (JMAK)
Holding

TFinish ~600°C

BS f (Y)

CR II Cooling time
to Ms

Ms f (Y)

Coiling

time
Figure 32 – Cooling Path on the Runout Table for Dual Phase Steels

TRP 0040 – Final Report 30 March 30, 2005

3 HSMM Validation
3.1 Overview
In 2003, INTEG completed an extensive validation of the HSMM version 6.0 using a
variety of grades of steel, rolled under a variety of processing conditions and from a variety
of rolling mills. Four of the Enhancement Group companies providing data to validate the
HSMM encompassed five rolling mills of various configurations (Table 1).

Steel Mill Type Roughing Heat Finishing Run Out Exit Area
Company Area Retention Area Table
68” Hot Strip 1 Reversing Heat 7 Stand 19 Banks 3 Down
Dofasco Mill Stand Retention Tandem Mill of Coilers
Panels Headers
148” Plate 1 Reversing None 1 Stand 4 Banks 1 Up Coiler
Stelco – Mill Stand Steckel Mill of &1
Hamilton Headers Cooling
Bed
2050mm Hot 1 Reversing Coil box 5 Stand 6 Banks 2 Down
Stelco – Lake
Strip Mill Stand Tandem Mill of Coilers
Erie
Headers
80” Hot Strip 5 Continuous None 6 Stand 20 Banks 2 Down
US Steel –
Mill Stands Tandem Mill of Coilers
Irvin Works
Headers
54” Hot Strip 1 Rev Stand Heat 7 Stand 18 Water 2 Down
Weirton
Mill & 1 Cont. Retention Tandem Mill Walls Coilers
Steel
Stand Panels
Table 1 – Mill Configurations of Supporting Steel Companies

3.2 Plant Data

The data supplied for the HSMM validation covered a variety of thicknesses, speeds,
finishing temperatures, coiling temperatures, tensile strengths and amount of water used on
the run out table. Some variation in the chemistry within the microstructure grade families
was also introduced. The steel companies provided engineering logs, data scan files
(rolling speeds, forces, temperatures, etc.) and laboratory data (yield strength, tensile
strength, elongation, grain size, etc.). Data was obtained for seven of the eight HSMM
microstructure grades. To further improve the accuracy of the model, the actual chemistry
of each piece was used for the microstructure calculations. Table 2 displays the
approximate range of processing parameters and Table 3 the range of key elements of the
steels utilized.

Finished Thickness (Strip) 2mm to 9mm

Finished Thickness (Plate) 9mm to 16mm
Finishing Temperature Range 800°C to 950°C
Coiling Temperature Range 600°C to 725°C
Yield Strength Range 200MPa to 650MPa
Tensile Strength Range 300MPa to 700MPa
Table 2 – Processing Parameter Ranges

TRP 0040 – Final Report 31 March 30, 2005

 Table 3 – Chemistry Range
Data for approximately 50 coils of steel was evaluated. Since the effort covered five
different rolling mills, differences in data gathering, reporting, terminology and testing
were introduced. Every effort was made to be as consistent as possible for selecting
comparison points between mill data and HSMM calcula ted data. The final analysis
indicated that having the exact temperature reading or the exact force measurement or the
exact whatever was not extremely important to the final mechanical property results.
These measurements could include their own natural margin of error and the HSMM could
still predict, with very acceptable accuracy, the tensile strength of the piece being modeled.
If anything, the variation in data measurement, collection and testing provided a possible
source of error that was not necessarily caused by the models, but observed when
comparing actual versus predicted results.

3.3 Results
The measured parameter that deviated the most from the predicted value was the final
ferrite grain size. On a percentage basis, when comparing actual versus calculated, the
final ferrite grain size comparison varied from as little as a 1% error to as much as a 50%
error. However, even though the final mechanical property calculations are partially grain
size dependent, the results did not consistently show the same relative magnitude of error
for tensile strength comparisons between actual and calculated. This can be primarily
explained by the error that occurs in the “measurement” of the ferrite grain size.

Since no uniform practices were issued any to all of the supporting steel companies prior to
their submission of the grain size measurements, a natural error in measurement can be
expected. However, it is important to point out that the grain sizes calculated by the
HSMM were indicative of the magnitude of the measured grain sizes. For example, one
steel sample had a measured grain size of 7.9 microns while the model predicted a final
ferrite grain size of 4.9 microns. Although this was almost a 40% error, the magnitude of
the grain size prediction was in an acceptable range because the tensile strength prediction
was within 1%.

The ultimate goal of the HSMM is to predict the final mechanical properties of the steel
being rolled in a hot mill. Due to the variations mentioned above, it was decided that the
best or most consistent and reliable parameter that could be used to measure the model’s
performance would be the tensile strength. The tensile strength is viewed as the best
measure of performance because this test is the most repeatable in the lab and thus has the
least deviation (error) built- in on the measurement side. The yield strength calculation, on

TRP 0040 – Final Report 32 March 30, 2005

the other hand, not only introduces a degree of error in the accuracy and repeatability of
the test, but also introduces a variety of methods to report the results such as a Lower Yield
Point, 0.2% Offset or 0.5% Under Load thus creating some potential error in comparison
using data gathered from multiple steel companies.

The following charts (Figures 33, 34, 35, 36, 37) provide a summary of the comparison
between the actual and calculated values for the temperature exiting the finishing mill, the
coiling temperature, the yield strength, the tensile strength, and the ferrite grain size.
Range lines are added to the graphs to show a range of ± 20°C for the temperatures and ±
5%MPa for the yield and tensile strengths. A fixed error range for the temperatures was
used because the relative spread between the lowest and highest temperature was only
about two hundred degrees. For the mechanical property charts, a percentage error range
was used because the range from the lowest to the highest was about 400MPa.

Figure 33 – Finishing Temperature Comparison Figure 34 –Coiling Temperature Comparison

Figure 35 – Yield Strength Comparison Figure 36 – Tensile Strength Comparison

TRP 0040 – Final Report 33 March 30, 2005

 Figure 37 – Ferrite Grain Size Comparison

3.4 Validation Summary

As can be seen from the above charts and the statistical summary in Table 4, the HSMM
has been validated using data from coils produced on a variety of mills and good
agreement has been achieved for the variety of products and processing parameters
covered. When comparing the final results for tensile strength, a very acceptable range of
errors have been achieved with an average percent error (calculated from the average
absolute error) of 3.03% or a ± 3% error. With this type of performance, the HSMM
version 6.2 can be used for conducting a variety of off- line analyses knowing that a proper
trend and/or relative prediction can be achieved.

Parameter Avg. Absolute Error Avg. Percentage Error

Finishing Temperature 9.14°C 1.02%
Coiling Temperature 11.21°C 1.73%
Yield Strength 23.93MPa 7.49%
Tensile Strength 13.27Mpa 3.03%
Ferrite Grain Size 2.65µm 19.67%
Table 4 – Statistical Analysis of Comparison between Actual and Calculated

TRP 0040 – Final Report 34 March 30, 2005

4 HSMM User Documentation
A complete set of documentation including users and technical manuals were generated
during the project and are included in Appendix A.

4.1 User’s Manual

The purpose of this document is to provide an overview of the HSMM, a brief background
on some of the theories used in the HSMM, and a thorough description of the HSMM
User’s Interface and its functionality.

4.2 Getting Started

This document is intended to help the user get quickly oriented with the HSMM,
understand how to utilize the HSMM to study and improve his/her mill operations and to
make him/her aware of some of the advanced features of the HSMM. The sections of this
document are:
• Part I – Quick Tour
• Part II – Working with the HSMM

4.3 Calibration Guide

This guide provides the procedures for properly setting up an HSMM Calibration Module
to accurately simulate a particular grade. These procedures involve using plant data for
tuning the temperature and force model coefficients to get the calculated values to closely
match the measured ones for both the single and multiple node models.

4.4 Client Database Link Instructions

This document contains general information and instructions on how to connect the
HSMM interface to a plant database for importing data into new rolling schedules.

4.5 Microstructure Guide

The purpose of this document is to provide an understanding of the underlying
methodologies used for microstructure modeling in the HSMM, and how the user can best
apply this model to his/her grades of steel using the GradeBuilder Module.

4.6 Technical Manual

The document describes the thermo- mechanical calculations that are performed in the
HSMM and how they are applied in simulating a work piece rolling through an entire hot
strip mill. The equations and numerical methods that are used in these calculations are
also provided.

TRP 0040 – Final Report 35 March 30, 2005

5 Conclusion
With the release of version 6.2, the validation and enhancement goals of this project were
successfully achieved in January, 2005. At that time the HSMM had already been
purchased by three steel producing companies located on three different continents. They
and the supporting steel companies continue to find outstanding value in the HSMM as a
beneficial tool in saving them time and money for a variety of practical applications.

TRP 0040 – Final Report 36 March 30, 2005

Appendix A – HSMM User Documentation
The HSMM User Documentation as described in Section 4 is provided in the following
files in PDF format:
• User Manual.doc
• Getting Started.doc
• Calibration Guide.doc
• Client Database Link Instructions.doc
• Microstructure Guide.doc
• Technical Manual.doc
• Flow Stress Grade Development Procedure.doc
• Microstructure Grade Development Procedure.doc

Appendix B – UBC Report on Dual Phase-Mo 600 Steel

The UBC report titled “Microstructure model for hot strip rolling of DP-Mo steel” referred
to in Section 2.4.6 is provided in the following PDF file:
• UBC-ReportforInteg_Nov2004Mechanical Properties.doc
• UBC-ReportforInteg_Nov2004microstructure.doc

Protected Metals Initiative Data

This Protected Metals Initiative Data was produced under a Cooperative Agreement
identified as DE-FC36-97ID13554 under a DOE Metals Initiative Project and may not be
published, disseminated or disclosed to others until five (5) years from March 30, 2005
unless written authorization is obtained from the American Iron and Steel Institute, Vice
President of Manufacturing and Technology. Upon expiration of the period of protection
set forth in this legend, the Government shall have unlimited rights in this data.

TRP 0040 – Final Report 37 March 30, 2005