Professional Documents
Culture Documents
MacNeal-Schwendler
Corporation
815 Colorado Boulevard
Los Angeles, California 90041-1777
NAV70.5ZZZSM-NAS107-NT1
February 1999
DISCLAIMER
The concepts, methods, and examples presented in this text are for educational
purposes only and are not intended to be exhaustive or to apply to any particular
engineering problem or design. The MacNeal-Schwendler Corporation assumes no
liability or responsibility to any person or company for direct or indirect damages
resulting from the use of any information contained herein.
Printed in U.S.A.
1998 by The MacNeal-Schwendler Corporation
All rights reserved.
Seminar
____________________________
Date
Location
____________________________
Instructor ____________________________
____________________________
Title
____________________________
_____________________________________________________________________
Your feedback is very important in helping MSC meet the training needs of its clients. Your specific
written comments are especially helpful. Please use the following scale:
1 = poor
Rating
2 = marginal
3 = average
4 = good
Category
5 = very good
Comments
Effectiveness of instructor in
explaining technical concepts.
Organization of material in a
clear and understandable
manner.
(Continued on back)
Rev 5/93
1 = poor
Rating
2 = marginal
3 = average
Category
4 = good
5 = very good
Comments
Impression of hotel
accommodations. At which
hotel did you stay? Any
problems?
Would you recommend this instructor for future classes? Why or why not? ______________________
__________________________________________________________________________________
Rev 5/93
TABLE OF CONTENTS
TITLE
1.0
SECTION
INTRODUCTION
WHAT IS DESIGN OPTIMIZATION? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BASIC FEATURES IMPLEMENTED IN MSC/NASTRAN . . . . . . . . . . . . . . . . . . . . . . . . . .
MSC/NASTRAN IMPLEMENTATION OF STRUCTURAL OPTIMIZATION. . . . . . . . . . . . .
CONCEPTS PRESENTED IN THIS COURSE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
STRENGTHS OF MSC/NASTRAN STRUCTURAL OPTIMIZATION. . . . . . . . . . . . . . . . . .
GENERAL FUNCTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SUMMARY OF NEW CAPABILITIES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BASIC OPTIMIZATION PROBLEM STATEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
EXAMPLE FORMULATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ONE BAR SUBJECT TO COMPRESSIVE LOAD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.0
NUMERICAL OPTIMIZATION
MINIMIZATION OF FUNCTIONS OF A SINGLE VARIABLE. . . . . . . . . . . . . . . . . . . . . . . .
GRADIENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
UNCONSTRAINED MINIMIZATION METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
KUHN-TUCKER CONDITIONS FOR OPTIMALITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SIMPLE CANTILEVER EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SERIES APPROXIMATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SEQUENTIAL LINEAR PROGRAMMING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
METHOD OF FEASIBLE DIRECTIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MODIFIED METHOD OF FEASIBLE DIRECTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
NUMERICAL IDENTIFICATION OF ACTIVE AND VIOLATED CONSTRAINTS. . . . . . . . .
3.0
1-1
1-2
1-3
1-4
1-5
1-6
1-8
1-10
1-11
1-16
2-1
2-4
2-6
2-10
2-12
2-15
2-19
2-22
2-26
2-32
DESIGN MODELING I
WHAT IS A DESIGN MODEL?. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HOW ARE THE DESIGN AND ANALYSIS MODELS RELATED?. . . . . . . . . . . . . . . . . . . .
DESIGN AND ANALYSIS MODEL COMPARISON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESIGN MODEL DEFINITION PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DEFINING THE ANALYSIS DISCIPLINES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DEFINING THE DESIGN VARIABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESVAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
RELATING DESIGN VARIABLES TO PROPERTIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVPREL1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
IDENTIFYING THE DESIGN RESPONSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DRESP1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DEFINING THE OBJECTIVE FUNCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESOBJ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DEFINING THE DESIGN CONSTRAINTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DCONSTR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DCONADD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESSUB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESGLB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BASIC DESIGN PROCESS CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DOPTPRM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
iii
3-1
3-2
3-3
3-5
3-6
3-7
3-8
3-10
3-11
3-15
3-16
3-24
3-25
3-27
3-28
3-31
3-33
3-34
3-38
3-39
TABLE OF CONTENTS
TITLE
4.0
SECTION
EXAMPLES I
EXAMPLE 3-BAR TRUSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10-BAR TRUSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10-BAR TRUSS RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BEAM VIBRATION FREQUENCY AS A DESIGN CONSTRAINT (M. J. Turner 1970) . . . .
BEAM VIBRATION FREQUENCY AS A DESIGN CONSTRAINT RESULTS . . . . . . . . . . .
5.0
6.0
5-1
5-2
5-5
5-9
5-11
5-12
5-13
5-14
5-15
DESIGN MODELING II
DESIGN MODEL DEFINITION PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
USER-DEFINED DESIGN VARIABLE-TO-PROPERTY RELATIONS. . . . . . . . . . . . . . . . .
DVPREL2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DEQATN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DTABLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
EXAMPLE DEFINING A RECTANGULAR SECTION BAR. . . . . . . . . . . . . . . . . . . . . . . .
USER-DEFINED RESPONSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DRESP2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
EXAMPLE DEFORMATION CONTROL WITH TYPE-2 RESPONSES . . . . . . . . . . . . . .
EXAMPLE BAR AXIAL PLUS BENDING STRESSES . . . . . . . . . . . . . . . . . . . . . . . . . . .
RESTRICTIONS IN FORMING SYNTHETIC RESPONSES . . . . . . . . . . . . . . . . . . . . . . . .
7.0
4-1
4-6
4-10
4-15
4-22
6-1
6-2
6-3
6-5
6-9
6-10
6-11
6-13
6-16
6-19
6-20
EXAMPLES II
STIFFENED PANEL WEIGHT MINIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
25-BAR TRUSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12
CANTILEVER PLATE (REDUCED BASIS FORMULATION). . . . . . . . . . . . . . . . . . . . . . . . 7-26
8.0
iv
8-1
8-2
8-3
8-5
8-6
8-9
8-13
8-16
TABLE OF CONTENTS
TITLE
9.0
SECTION
FUNDAMENTALS OF STRUCTURAL OPTIMIZATION
STRUCTURAL OPTIMIZATION THE DIFFICULTIES. . . . . . . . . . . . . . . . . . . . . . . . . . . .
STRUCTURAL OPTIMIZATION THE SOLUTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESIGN VARIABLE LINKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESIGN VARIABLE LINKING IN SHAPE REDESIGN . . . . . . . . . . . . . . . . . . . . . . . . . . . .
GENERATION OF CONSTRAINTS IN MSC/NASTRAN . . . . . . . . . . . . . . . . . . . . . . . . . . .
CONSTRAINT SCREENING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DSCREEN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FORMAL APPROXIMATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
APRCOD SELECTION OF APPROXIMATE FORM . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SENSITIVITY ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SEMIANALYTIC SENSITIVITY ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DELB FINITE DIFFERENCE STEP SIZE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SENSITIVITY ANALYSIS IMPLEMENTATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ADJOINT SENSITIVITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ADJOINT SENSITIVITY THEORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESIGN SENSITIVITY COEFFICIENT MATRICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DSAPRT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FORMATTED SENSITIVITY PRINT EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MOVE LIMITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MOVE LIMITS IMPOSED ON ANALYSIS MODEL PROPERTIES . . . . . . . . . . . . . . . . . . .
MOVE LIMITS IMPOSED ON DESIGN VARIABLES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AUTOMATIC UPDATES OF MOVE LIMITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MOVE LIMITS UPDATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PARAM OPTEXIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FLOW CHART FOR SOLUTION 200 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CONVERGENCE AT THE DESIGN CYCLE LEVEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CONVERGENCE CRITERIA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SUMMARY OF PARAMETERS FOR DESIGN OPTIMIZATION . . . . . . . . . . . . . . . . . . . . .
10.0
9-1
9-3
9-7
9-11
9-15
9-16
9-18
9-21
9-28
9-30
9-31
9-33
9-35
9-37
9-38
9-47
9-49
9-51
9-52
9-53
9-55
9-56
9-57
9-58
9-59
9-63
9-65
9-67
SHAPE OPTIMIZATION
BASIC EQUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BASIS VECTORS EXAMPLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SHAPE BASIS VECTORS IN THE DESIGN MODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESIGN MODELING INPUT FOR
SHAPE OPTIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MANUAL GRID VARIATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVGRID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AUXILIARY MODELS IN SHAPE OPTIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
EXAMPLE DIRECT INPUT OF SHAPES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVSHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AUXILIARY BOUNDARY MODELS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
EXAMPLE ANALYTIC BOUNDARY SHAPES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AUXCASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AUXMODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
BNDGRID. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVBSHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10-1
10-2
10-6
10-7
10-8
10-10
10-12
10-14
10-19
10-24
10-26
10-30
10-31
10-32
10-33
TABLE OF CONTENTS
TITLE
SECTION
SUMMARY OF SHAPE BASIS VECTOR GENERATION METHODS. . . . . . . . . . . . . . . . . 10-41
GUIDELINES, RECOMMENDATIONS AND LIMITATIONS . . . . . . . . . . . . . . . . . . . . . . . . 10-43
11.0
12.0
11-1
11-2
11-3
11-6
11-18
SUPERELEMENT OPTIMIZATION
GENERAL CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
CASE CONTROL SPECIFICATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-3
EXAMPLE TWENTY FIVE BAR TRUSS, SUPERELEMENT OPTIMIZATION . . . . . . . . . 12-5
13.0
AEROELASTIC OPTIMIZATION
AEROELASTIC OPTIMIZATION MSC BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . .
STATIC AEROELASTICITY DESIGN CONDITIONS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
STATIC AEROELASTICITY SENSITIVITY ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FLUTTER DESIGN. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FLUTTER RESPONSE PROPERTIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
FLUTTER RESPONSE SENSITIVITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AEROELASTIC DESIGN EXAMPLE (HA200A and B) . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ANALYSIS CONDITIONS FOR EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESIGN CONDITIONS FOR EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HA200B OBJECTIVE RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HA200B DESIGN VARIABLE RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HA200B FLUTTER RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HA200A INPUT FILE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DESIGN MODEL AND BULK DATA INPUT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HA200A SELECTED RESULTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HA200B SELECTED RESULTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13-1
13-2
13-3
13-4
13-5
13-6
13-7
13-9
13-10
13-12
13-13
13-14
13-15
13-17
13-23
13-31
vi
A-1
A-3
A-4
A-5
A-6
A-7
A-8
A-9
A-10
A-11
A-12
A-13
A-15
TABLE OF CONTENTS
TITLE
SECTION
DESVAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DLINK. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DOPTPRM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DRESP1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DRESP2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DSAPRT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DSCREEN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DTABLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVBSHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVGRID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVPREL1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVPREL2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
DVSHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MODTRAK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
vii
A-19
A-20
A-21
A-25
A-31
A-34
A-37
A-39
A-40
A-41
A-43
A-46
A-48
A-49
SECTION 1
INTRODUCTION
WHAT IS DESIGN OPTIMIZATION? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
BASIC FEATURES IMPLEMENTED IN MSC/NASTRAN . . . . . . . . . . . . . . . . . . . . . 1-2
MSC/NASTRAN IMPLEMENTATION OF STRUCTURAL OPTIMIZATION. . . . . . . . 1-3
CONCEPTS PRESENTED IN THIS COURSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-4
STRENGTHS OF MSC/NASTRAN STRUCTURAL OPTIMIZATION. . . . . . . . . . . . . 1-5
GENERAL FUNCTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-6
SUMMARY OF NEW CAPABILITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-8
BASIC OPTIMIZATION PROBLEM STATEMENT . . . . . . . . . . . . . . . . . . . . . . . . . . 1-10
EXAMPLE FORMULATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-11
ONE BAR SUBJECT TO COMPRESSIVE LOAD . . . . . . . . . . . . . . . . . . . . . . . . . . 1-16
1-1
1-2
Initial
Design
Improved
Design
The required number of
iterations of the external loop
must be small.
Structural
Response
Analysis
Approximate
Model
Constraint
Screening
Sensitivity
Analysis
Optimizer
Many Times
Finite Element
Analysis
1-3
1-4
1-5
GENERAL FUNCTIONS
Solution Sequence
200 - supports design sensitivity and optimization
Analysis Types
Statics
Normal modes
Buckling
Direct frequency*
Modal frequency*
Modal transient*
Static aeroelastic
Aeroelastic flutter
(*Includes acoustic responses)
Design Variables
Sizing properties(including superelements)
Shape(superelements with manual grid variation only)
1-6
1-7
1-8
1-9
j = 1, 2, . . ., L
Equality constraints:
Hk (X) = 0
k = 1, 2, . . ., M
Side constraints:
x iL x i x iu
i = 1, 2, . . ., N
1-10
EXAMPLE FORMULATIONS
Example 1:
Determination of equilibrium position in a two-spring system
(geometrically nonlinear deformation)
l1 = 10 cm
P2
K1 = 8N/cm
X2
P1
P2 = 5N
P1 = 5N
l2 = 10 cm
X1
K2 = 1N/cm
Undeformed
Deformed
1-11
1-12
t3
t2
t1
h
P
l
--4
l
2t 1 b
t 2 t 2min,
2t 2 h
t 3 t 3min,
t3 b
1-13
l
v = t 3 --- + l [ bh ( h 2t 2 ) ( b 2t 1 ) ]
4
where: { x } = [ b h t 1 t 2 t 3 ]
Design
Variables
Response
Constraints
tube buck 0
2t 1 b 0
2t 2 h 0
Design
Constraints
t3 b 0
t1
min
t1 0
t2
min
t2 0
t3
min
t3 0
1-14
Side
Constraints
1-15
Ainit = 75 mm
, A, E
E = 7.0E4 N/mm
2
3
= 2.7E6 kg/mm
Optimal Design:
P
max = ------------A opt
A opt = 1.0E4 N 200 N mm
= 50 mm
1-16
1-17
+DVP1
1-18
1-19
***************************************************************
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
3
NUMBER OF OPTIMIZATIONS W.R.T. APPROXIMATE MODELS
2
OBJECTIVE AND MAXIMUM CONSTRAINT HISTORY
--------------------------------------------------------------------------------------------------------------OBJECTIVE FROM
OBJECTIVE FROM
FRACTIONAL ERROR
MAXIMUM VALUE
CYCLE
APPROXIMATE
EXACT
OF
OF
NUMBER
OPTIMIZATION
ANALYSIS
APPROXIMATION
CONSTRAINT
--------------------------------------------------------------------------------------------------------------INITIAL
1.012500E-01
-3.333334E-01
1
6.746810E-02
6.747905E-02
-1.623075E-04
3.104401E-04
2
6.747905E-02
6.747905E-02
0.000000E+00
3.104401E-04
--------------------------------------------------------------------------------------------------------------STRUCTURAL OPTIMIZATION TEST PROBLEM (STATIC LOADING)
APRIL 20, 1995 MSC/NASTRAN
8/17/94
PAGE 38
SINGLE BAR, 1 DOF, SUBJECT TO COMPRESSIVE LOAD
1-20
SECTION 2
NUMERICAL OPTIMIZATION
MINIMIZATION OF FUNCTIONS OF A SINGLE VARIABLE. . . . . . . . . . . . . . . . . . . 2-1
GRADIENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4
UNCONSTRAINED MINIMIZATION METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6
KUHN-TUCKER CONDITIONS FOR OPTIMALITY . . . . . . . . . . . . . . . . . . . . . . . . . 2-10
SIMPLE CANTILEVER EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12
SERIES APPROXIMATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-15
SEQUENTIAL LINEAR PROGRAMMING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-19
METHOD OF FEASIBLE DIRECTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-22
MODIFIED METHOD OF FEASIBLE DIRECTIONS. . . . . . . . . . . . . . . . . . . . . . . . . 2-26
NUMERICAL IDENTIFICATION OF ACTIVE AND VIOLATED CONSTRAINTS . . 2-32
dF
extremum is defined for X at which -------- = 0
dX
( X 0 , F 0 ), ( X 0 , F 0 ), ( X 1 , F 1 )
we could use a quadratic polynomial approximation of the form:
2
F ( X ) = a 0 + a 1 X + a 2 X
2-1
F 0 = a 1 + 2a 2 X 0
2
F0 = a0 + a1 X0 + a2 X0
2
F1 = a0 + a1 X1 + a2 X1
From which:
( F1 F0 ) ( ( X1 X0 ) F0 )
a 2 = ----------------------------------------------------------------------( X1 X0 )
a 1 = F 0 2a 2 X 0
2
a0 = F0 a1 X0 a2 X0
Therefore
( X ) = a + 2a X = 0
F
1
2
a1
X ---------2a 2
F ( X ) = 2a 2
Minimum at X if a2 > 0
Maximum at X if a2 < 0
2-2
Fopt
X0
X1
Constrained minimum:
F(X)
XX
Fopt
X0
Side constraint
X
XX
X = X
2-3
X1
GRADIENTS
Multivariate function minimization in MSC/NASTRAN is gradientbased.
Let:
f = f ( X 1 , X 2 , , X n )
i.e., a scalar function of n independent variables.
Gradient of a scalar function is defined as:
f
f
f
f = ---------- e 1 + ---------- e 2 + + ---------- e n
X 1
X 2
X n
where e 1 , e 2 , , e n are unit vectors in the n
th
direction.
= --------e i
X i
2-4
GRADIENTS (Cont.)
Gradient of a scalar vector field
2
Example: f ( X 1 , X 2 ) = X 1 + X 2
(Note that this is the equation for a family of circles in the X1, X2
plane with centers at the origin and radii of
f .)
f
f
f ( X 1 , X 2 ) = ---------- e 1 + ---------- e 2
X 1
X 2
= 2X 1 e 1 + 2X 2 e 2
or
2X 1
f ( X 1 , X 2 ) =
2X 2
X2
f ( X 1, X 2 )
X1
2-5
1
PE = --- K 1
2
X1 + ( l 1 X2 ) l 1
1
+ --- K 2
2
2
2
X1 + ( l 2 X2 ) l 2
P1 X1 P2 X2
minimum occurs at
( PE )
-
--------------X
1
( PE ) =
= 0
(
PE
)
----------------
X
2
PEmin = 41.81 N cm
X1
= 8.631 cm
X2
= 4.533 cm
2-6
100
80
60
X*
+
40
30
20
10
q1
q
+ S
40
10
20 30
= X
2-7
10
12
q
q1
S = F ( X
) =
q1
F
---------X 1
F
---------X 2
.
.
.
F
---------X 1
2-8
= F ( X
q1
iteration (q > 1)
= F ( X
q1 2
) + S
q1
F ( X
)
where = ---------------------------------q2 2
F ( X
)
q1
100
80
60
6
0
X
1
S
40
30
1
20
2
S
10
2S
10
20 30
X*
40
10
12
1.
X is feasible
j = 1, , m
k = 1, , l
gj ( X ) 0
hk ( X ) = 0
2.
j g ( X ) = 0
j
j = 1, , m
j 0
3.
F ( X ) +
j gj ( X
j=1
)+
k + m hk ( X
) = 0
k=1
2-10
X2
F ( X ) = constant
F ( X )
g1 ( X ) = 0
X1
g 2 ( X )
g 1 ( X )
g2 ( X ) = 0
F ( X )
1 g 1 ( X )
2 g 2 ( X )
F ( X )
2-11
L = 500 cm
Section A-A
H
B
6 N
E = 1 10 -----------2
cm
2-12
Bending Stress
PL
= ----------- 2.54
3El
Tip Deflection
H
---- 12
B
Aspect Ratio
1 B 20
Gauge Requirements
20 H 50
2-13
65
60
V = 200000
175000
100000
Height H (cm)
55
150000
125000
H = 50
50
Optimum
45
= 2.54
40
75000
50000
b = 700
35
2.5
5
Width B (cm)
2-14
7.5
SERIES APPROXIMATIONS
Function gradient information can be used to construct first-order
Taylor Series approximations
2
2
x
df
d
f ( x + x ) = f ( x ) + ------ x + ---------f ---------- +
o
2!
2
dx x
dx x o
o
2
f ( x o + x ) = f ( x o ) + df
------- o x + 0 ( x )
dx x
where
2
2-15
4PL
PL
= ----------- = ---------------- 2.54 cm
3
3EI
BH E
2-16
V
V
+ ------ o o B + ------- o o H
B B , H
H B , H
o
o
o o
( B + B, H + H, L ) = ( B , H , L )
+ ------ o o B + ------- o o H
B B , H
H B , H
o
o
o o
( B + B, H + H, L ) = ( B , H , L )
+ ------ o o B + ------- o o H
H B , H
B B , H
o
At ( B , H ) = ( 6, 45 )
o
2-17
65
60
175000
Height H (cm)
55
100000
150000
125000
H = 50
50
Optimum
Approximate
Optimum
X
45
75000
40
35
2.5
7.5
Width B (cm)
2-18
x = x x
j J, (J set of
active constraints)
2-19
F ( x )
x
x
x1
2-20
move limits
X2
o
F ( X )
g(X) = 0
X1
Move limits are required because higher order terms have been
ignored
2-21
F ( X )
g j ( X )
F = Constant
gj ( X )
X1
Feasible
Usable/Feasible
S
Usable
gj ( X )
2-22
X1
(1)
(2)
2-23
60
Height H (cm)
55
H = 50
50
Optimum
45
= 2.54
40
= 700
35
3
5
6
Width B (cm)
2-24
g1 ( X )
X
X
g2 ( X )
F ( X ) = Constant
3
4
X X
X1
In structural applications, unnecessary function evaluations must
be kept to a minimum.
Faster convergence can usually be obtained if the constraint
surfaces are followed to the optimum.
2-25
F s
Subject to:
g j ( X ) s 0
jJ
ss1
2-26
j = 1, ..., m
Direction of steepest descent
Use conjugate direction if unconstrained
for more than one iteration.
If some gj = 0
for j J
2-27
= X
q1
+ S
T 1
X = A [ AA ]
2-28
2-29
60
Height H (cm)
55
H = 50
50
Optimum
45
= 2.54
40
= 700
35
5
6
Width B (cm)
2-30
F = Constant
F
Minimize Constraint
Violation
G
x1
Minimize Objective
Function
Compromise Direction
x2
No Feasible Design
2-31
x1
Feasible Region
gj ( x ) < 0
g j ( x ) = CT
gj ( x ) = 0
Infeasible Region
gj ( x ) > 0
g j ( x ) = CTMIN
x1
2-32
Gj(X)
Feasible
Violated
Active
CTMIN
X
Numerical Constraint
Boundary
CT
Exact Constraint
Boundary
CT
= 0.03
CTMIN = 0.003
2-33
SECTION 3
DESIGN MODELING I
WHAT IS A DESIGN MODEL? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1
HOW ARE THE DESIGN AND ANALYSIS MODELS RELATED? . . . . . . . . . . . . . . 3-2
DESIGN AND ANALYSIS MODEL COMPARISON . . . . . . . . . . . . . . . . . . . . . . . . . . 3-3
DESIGN MODEL DEFINITION PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
DEFINING THE ANALYSIS DISCIPLINES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
DEFINING THE DESIGN VARIABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-7
DESVAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
RELATING DESIGN VARIABLES TO PROPERTIES . . . . . . . . . . . . . . . . . . . . . . . 3-10
DVPREL1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
IDENTIFYING THE DESIGN RESPONSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
DRESP1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-16
DEFINING THE OBJECTIVE FUNCTION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-24
DESOBJ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-25
DEFINING THE DESIGN CONSTRAINTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-27
DCONSTR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-28
DCONADD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-31
DESSUB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-33
DESGLB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34
BASIC DESIGN PROCESS CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-38
DOPTPRM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-39
3-1
DESIGN
IMPROVEMENT
DESIGN
VARIABLES
OBJECTIVE,
CONSTRAINTS
CAD MODEL,
PROTOTYPE
GEOMETRY,
PROPERTIES
LOADS,
BOUNDARY
CONDITIONS
ANALYSIS MODEL
3-2
RESPONSES
tw
tf
...
PBAR
21
3-3
...
...
...
21
...
101
...
CBAR
...
2.36 ...
Analysis
Cross-sectional properties
Design
Cross-section dimensions
3-4
3-5
Analysis =
MODES
normal modes
BUCK
buckling
DFREQ
direct frequency*
MFREQ
modal frequency*
MTRAN
modal transient*
SAERO
static aeroelasticity
FLUTTER flutter
*Includes acoustics
3-6
3-7
DESVAR
Design Variable
Design Variable
Format:
1
DESVAR
ID
LABEL
XINIT
XLB
XUB
DELXV
BARA1
35.0
10.
100.
0.2
10
Example:
DESVAR
Field
Contents
ID
LABEL
XINIT
XLB
XUB
DELXV
Remarks:
1.
DELXV can be used to control the change in the design variable during one optimization
cycle.
2.
If DELXV is blank, the default is taken from the specification of the DELX parameter on
the DOPTPRM entry. If DELX is not specified, then the default is 1.0.
3-8
$ESVAR, ID,
$
DESVAR, 10,
DESVAR, 11,
$
LABEL,
XINIT,
AREA1, 0.05,
THICK1, 0.03,
XLB,
XUB,
0.01,
0.01,
0.1
0.08
DELXV
0.01 x 10 0.1
0.01 x 11 0.08
The entire set of DESVAR entries defines a vector of design variables:
X = [ x 1, x 2, , x n ]
with corresponding bounds,
I
xi xi xi
3-9
pj = Co +
Ci xi
i
3-10
DVPREL1
DVPREL1 the relation between an analysis model property and design variables.
Design Variable to Property Relation
Defines
Format:
1
DVPREL1
ID
TYPE
PID
FID
PMIN
PMAX
C0
DVID1
COEF1
DVID2
COEF2
DVID3
-etc.-
12
PBAR
612
0.2
3.0
0.25
20
20.0
0.3
10
Example:
DVPREL1
Field
Contents
ID
TYPE
PID
FlD
Field position of the property entry, or word position in the element property table
of the analysis model. (Integer 0)
PMIN
Minimum value allowed for this property. If FID references a stress recovery
location, then the default value for PMIN is 1.0+35. PMIN must be explicitly set
to a negative number for properties that may be less than zero (for example, field
ZO on the PCOMP entry). (Real; Default = 1.0E20)
PMAX
C0
DVIDi
COEFi
Remarks:
1.
The relationship between the analysis model property and design variables is given by:
P i = C0 + COEFi DVIDi
i
2.
3-11
DVPREL1
3.
PTYPE = PBEND is not supported, either directly through FIDs or indirectly via word positions in the element property table.
4.
FID may be either a positive or a negative number. If FID > 0, it identifies the field position
on a property entry. If FID < 0, it identifies the word position of an entry in the element
property table. For example, to specify the area of a PBAR, either FID = +4 or FID = 3
can be used. However, if PTYPE = PBEAM, FID must be negative. See the following
element property table for the word positions for PBEAM.
Word
Type
Item
1
2
3
4
5
6
7
8 through 13
14 through 21
22
23
24 through 29
30 through 37
38
39
40 through 45
46 through 53
54
55
56 through 61
62 through 69
70
71
72 through 77
78 through 85
86
87
88 through 93
94 through 101
I
I
I
I
I
I
R
R
R
I
R
R
R
I
R
R
R
I
R
R
R
I
R
R
R
I
R
R
R
Property ID
Material ID
Number of segments
Constant cross section flag (1 = yes, 0 = no)
Unused
Stress output request flag, SO (1 = yes, 0 = no)
X/XB ratio; at end A, X/XB = 0.0
A, I1, I2, I12, J, NSM
C1, C2, D1, D2, E1, E2, F1, F2
Repeat of words 6 through 21 for the 1st intermediate
station
(Continued)
3-12
DVPREL1
Word
Type
102
103
104 through 109
110 through 117
I
R
R
R
118
119
120 through 125
126 through 133
134
135
136 through 141
142 through 149
150
151
152 through 157
158 through 165
166
167
168 through 173
174 through 181
182 through 189
190 through 197
I
R
R
R
I
R
R
R
I
R
R
R
I
R
R
R
R
R
Item
End B
K1, K2, S1, S2, NSI(A), NSI(B), CW(A), CW(B)
M1(A), M2(A), M1(B), M2(B), N1(A), N2(A), N1(B),
N2(B)
3-13
Web Cap
A i = 1.0 x 3
Plate
t i = 1.0 x 1
for the base plate thickness,
PSHELL, 1,
1,
0.15,
1
$
$...Define the design variables:
$
$DESVAR,ID,
LABEL, XINIT, XLB,
XUB,
DELXV
DESVAR, 1,
T-PLATE,0.15,
0.001, 10.0
$
$...Relate the design variables to analysis model properties
$
(linear relations, so use DVPREL1)
$
$DVPREL1,ID,
TYPE,
PID,
FID,
PMIN,
PMAX,
C0,
$+,
DVID1, COEF1, DVID2, COEF2, ...
DVPREL1,1,
PSHELL, 1,
4,
0.01,
,
,
+DP1,
1,
1.0
3-14
+DP1
3-15
DRESP1
DRESP1
Sensitivity
Responseor
Quantities
Defines a set of structural responses that is used in the designDesign
either
as constraints
as an
objective.
Format:
1
DRESP1
ID
LABEL
RTYPE
PTYPE
REGION
ATTA
ATTB
ATT1
ATT2
-etc.-
DX1
STRESS
PROD
10
Example:
DRESP1
102
103
Field
Contents
ID
LABEL
RTYPE
PTYPE
Element flag (PTYPE = ELEM) or property entry name. Used with element type
responses (stress, strain, force, etc.) to identify the property type, since property
entry IDs are not unique across property types. (Character: ELEM, PBAR,
PSHELL, etc.)
REGION
ATTA, ATTB,
ATTi
(Continued)
3-16
DRESP1
Response
Type
(RTYPE)
Response Attributes
ATTA (Integer > 0)
WEIGHT
Blank
Blank
SEIDi or ALL
VOLUME
Blank
Blank
SEIDi or ALL
EIGN
Approximation Code.
See Remark 19.
Blank
FREQ
Approximation Code.
See Remark 19.
Blank
LAMA
Approximation Code.
See Remark 19.
Blank
DISP
Displacement Component
Blank
Grid ID
STRAIN
Blank
STRESS
Blank
FORCE
Blank
CSTRAIN
LAMINA Number
(Integer; Default = 1)
CSTRESS
LAMINA Number
(Integer; Default = 1)
CFAILURE
LAMINA Number
(Integer; Default = 1)
FRDISP
Displacement Component
Grid ID
FRVELO
Velocity Component
Grid ID
FRACCL
Acceleration Component
Grid ID
FRSPCF
Grid ID
FRSTRE
(Continued)
3-17
DRESP1
Response
Type
(RTYPE)
Response Attributes
ATTA (Integer > 0)
FRFORC
TDISP
Displacement Component
Grid ID
TVELO
Velocity Component
Grid ID
TACCL
Acceleration Component
Grid ID
TSPCF
Grid ID
TSTRE
TFORC
TRIM
AESTAT or AESURF
Entry ID
Blank
Blank
STABDER
AESTAT or AESURF
Entry ID
Component
FLUTTER
Blank
Remarks:
1.
Stress, strain, and force item codes can be found in the MSC/NASTRAN Quick Reference
Guide, Appendix A. For stress or strain item codes that have dual meanings, such as von
Mises or maximum shear, the option specified in the Case Control Section will be used;
i.e., STRESS(VONM) or STRESS(MAXS).
2.
RTYPE=CSTRESS, CSTRAIN, and CFAILURE are used only with the PCOMP entry.
CSTRESS and CSTRAIN item codes are described under Table 1. (Element
Stress/Strain Item Codes) in the MSC/NASTRAN Quick Reference Guide, Appendix A.
CFAILURE item codes are described under Table Table 2. (Element Force Item Codes)
in the MSC/NASTRAN Quick Reference Guide, Appendix A. Only force item codes that
refer to failure indices of direct stress and interlaminar shear stress are valid.
(Continued)
3-18
DRESP1
The CFAILURE response type requires the following specifications on the applicable
entries:
a.
b.
c.
Stress limits in the ST, SC, and SS fields on all MATi entries.
3.
ATTB is used only for responses of composite laminae, dynamics, and stability derivatives. For other responses, this field must be blank.
4.
All grids associated with a DRESP1 entry are considered to be in the same region for
screening purposes. Only up to NSTR displacement constraints (see DSCREEN entry)
per group per load case will be retained in the design optimization phase.
5.
6.
7.
If RTYPE = DISP, TDISP, TVELO, TACCL or TSPCF, multiple component numbers (any unique combination of the digits 1 through 6 with no embedded blanks) may be
specified on a single entry. Multiple response components may not be used on any other
response types.
8.
9.
Real/imaginary representation is the default for complex response types. Magnitude/phase representation must be requested by the corresponding Case Control
command; e.g., DlSP(PHASE) = ALL.
10.
REGION is used for constraint screening. The NSTR field on DSCREEN entries gives the
maximum number of constraints retained for each region per load case.
IF RTYPE = WEIGHT, VOLUME, LAMA, EIGN or FREQ, no REGION identification
number should be specified. For all other responses, if the REGION field is left blank, the
default specified in Table 2 is used. Usually, the default value is appropriate.
If the REGION field is not blank, all the responses on this entry as well as all responses on
other DRESP1 entries that have the same RTYPE and REGION identification number will
be grouped into the same region.
(Continued)
3-19
DRESP1
Response Type
Default Region
WEIGHT
No region
VOLUME
No region
LAMA
No region
EIGN
No region
FREQ
No region
DISP
FRDISP
FRVELO
FRACCL
FRSPCF
TDISP
TVELO
TACCL
TSPCF
FLUTTER
OTHER
11.
REGION is valid only among the same type of responses. Responses of different types
will never be grouped into the same region, even if they are assigned the same REGION
identification number by the user.
12.
If RTYPE = WEIGHT or VOLUME, field ATTi = ALL implies total weight/volume of all
superelements except external superelements.
13.
RTYPE = STABDER identifies a stability derivative response. ATTB is the restraint flag
for the stability derivative. ATTB = 0 means unrestrained, and ATTB = 1 means restrained. For example, ATTA = 4000, ATTB = 0, and ATT1 = 3 reference the unrestrained
Cz derivative for the AESTAT (or AESURF) entry ID = 4000.
(Continued)
3-20
DRESP1
RTYPE = FLUTTER identifies a set of damping responses. The set is specified by ATTi:
ATT1 = Identification number of a SET1 entry that specifies a set of modes.
ATT2 = Identification number of an FLFACT entry that specifies a list of densities.
ATT3 = Identification number of an FLFACT entry that specifies a list of Mach numbers.
ATT4 = Identification number of an FLFACT entry that specifies a list of velocities.
15.
16.
For RTYPE = TDISP, TVELO, TACCL, TSPCF, TFORC, and TSTRE, ATTB
specifies a time value. If ATTB is specified, then the responses are evaluated at the closest time selected by the OTIME command. The default for ATTB is all time steps selected
by the OTIME command.
17.
Intermediate station responses on CBAR elements due to PLOAD1 and/or CBARAO entries may not be defined on the DRESP1 entry.
18.
RTYPE = EIGN refers to normal modes response in terms of eigenvalue (radian/time)2 while RTYPE = FREQ refers to normal modes response in terms of natural
frequency or units of cycles per unit time.
19.
For RTYPE = LAMA, EIGN or FREQ, the response approximation used for optimization
can be individually selected. (Approximation Code = 1 = direct linearization, = 2 = Inverse
Linearization).
3-21
ATTB,
ATT1,
501,
3-22
PID1
PID2
DRESP1
.
.
.
PIDm
3-23
EID1
EID2
.
.
.
EIDn
DRESP1
(Case Control
command)
3-24
(or DRESP2)
DESOBJ
Design Objective
DESOBJ the DRESP1 or DRESP2 entry to be used as the design objective.
Selects
Design Objective
Format:
DESOBJ (
MAX
) = N
MIN
Examples:
DESOBJ = 10
DESO = 25
Describer
Meaning
MIN
MAX
Remarks:
1.
A DESOBJ command is required for a design optimization task and is optional for a sensitivity task.
2.
If the DESOBJ command is specified within a SUBCASE, the identified DRESPi Bulk Data
entry uses a response only from that subcase. If DESOBJ appears above all SUBCASE
commands and there are multiple subcases, it uses a global response.
3.
3-25
$
$
$
$
DESOBJ(MIN) = 10
$
$ IN BULK DATA:
$
$DRESP1, ID,
LABEL, RTYPE, PTYPE, REGION, ATTA,
ATTB,
ATT1,
$+,
ATT2,
...
$
DRESP1, 10,
WEIGHT, WEIGHT, ,
,
,
,
ALL
$
^^^
$
(DEFAULT IS 0: JUST THE RESIDUAL STRUCTURE)
$
3-26
rj rj ( x ) rj
3-27
DCONSTR
DCONSTR
Define design
Design Constraints
Design Constraints
constraints.
Format:
1
DCONSTR
DCID
RID
10
10
LALLOW UALLOW
Example:
DCONSTR
1.25
Field
Contents
DCID
RID
LALLOW
UALLOW
Remarks:
1.
The DCONSTR entry may be selected in the Case Control Section by the DESSUB or
DESGLB command.
2.
3.
For a given DCID, the associated RID can be referenced only once.
4.
The units of LALLOW and UALLOW must be consistent with the referenced response defined on the DRESPi entry. If RID refers to an EIGN response, then the imposed bounds
2
must be expressed in units of eigenvalue, (radian/time) . If RID refers to a FREQ response, then the imposed bounds must be expressed in cycles/time.
5.
LALLOW and UALLOW are unrelated to the stress limits specified on the MATi entry.
6.
3-28
DCONSTR
Design Constraints
> GSCAL
As Remark 6 indicates, small values of UALLOW and LALLOW require special processing
and should be avoided. Bounds of exactly zero are particularly troublesome. This can be
avoided by using a DRESP2 entry that offsets the constrained response from zero.
3-29
3-30
DCONADD
Format:
1
DCONADD
DCID
DC1
DC2
DC3
DC4
DC5
DC6
DC7
DC8
-etc.-
10
10
Example:
DCONADD
12
Field
Contents
DCID
DCi
Remarks:
1.
2.
3-31
Subcase-dependent constraints
DESGLB
3-32
DESSUB
Format:
DESSUB = n
Examples:
DESSUB = 10
DESS = 25
Describer
Meaning
Remark:
1.
A DESSUB command is required for every subcase for which constraints are to be
applied.
3-33
DESGLB
Format:
DESGLB = n
Examples:
DESGLB = 10
DESG = 25
Describer
Meaning
Remarks:
1.
A DESGLB command is optional and invokes constraints that are to be applied independent of a particular subcase. These constraints could be based on responses that are
independent of subcases (e.g., WEIGHT or VOLUME).
2.
The DESGLB command can be used to invoke constraints that are not a function of
DRESP1 entries; e.g., DRESP2 responses that are not functions of DRESP1 responses
are subcase independent.
3-34
$
$ IN CASE CONTROL:
$
SUBCASE 20
ANALYSIS = STATICS
DESSUB = 100
(ETC.)
$
$ IN BULK DATA:
$
$DRESP1, ID,
LABEL, RTYPE, PTYPE,
$+,
ATT2,
...
$
$ MAXIMUM STRESS AT END A:
DRESP1, 110,
SAMAX, STRESS, PBAR,
+
2020,
2030
$
$ MAXIMUM STRESS AT END B:
DRESP1, 120,
SBMAX, STRESS, PBAR,
+
2020,
2030
$
$ BOUNDS APPLIED TO BOTH ENDS A & B:
$DCONSTR,DCID,
RID,
LALLOW, UALLOW
DCONSTR,100,
110,
-85.,
+85. $
DCONSTR,100,
120,
-85.,
+85. $
$
REGION, ATTA,
ATTB,
ATT1,
7,
2010,
14,
2010,
3-35
PID1
PID2
DCONSTR
DRESP1
.
.
.
PIDm
3-36
EID1
EID2
.
.
.
EIDn
rj rj ( x )
g 2j 1 ( x ) = ------------------------- 0
L
rj
U
rj ( x ) rj
g 2j ( x ) = -------------------------- 0
U
rj
Normalization provides:
Expression of all constraints in standard form (basic optimization
problem statement)
Ranking of constraints regardless of response magnitude
It is best to use reasonable bounds, avoiding 0.0 if possible.
3-37
3-38
DOPTPRM
Format:
1
DOPTPRM PARAM1
VAL1
PARAM2
VAL2
PARAM3
VAL3
PARAM4
VAL4
PARAM5
VAL5
-etc.-
DESMAX
10
Example:
DOPTPRM IPRINT
10
Field
Contents
PARAMi
VALi
Remarks:
1.
Name
APRCOD
CONV1
CONV2
CONVDV
Relative convergence criterion on design variables. (Real > 0.0; Default = 0.001)
CONVPR
(Continued)
3-39
DOPTPRM
Name
CT
CTMIN
DABOBJ
DELB
DELOBJ
DELP
Fractional change allowed in each property during any optimization design cycle.
This provides constraints on property moves. (Real > 0.0; Default = 0.2)
DELX
Fractional change allowed in each design variable during any optimization cycle.
(Real > 0.0; Default = 1.0)
DESMAX
DOBJ1
DOBJ2
DPMIN
DX1
DX2
DXMIN
Minimum design variable move limit (Real > 0.0; Default = 0.05).
GMAX
Maximum constraint violation allowed at the converged optimum. (Real > 0.0;
Default = 0.005)
GSCAL
(Continued)
3-40
DOPTPRM
Name
IGMAX
IPRINT
IPRNT1
IPRNT2
ISCAL
Design variables are rescaled every ISCAL iterations. Set ISCAL= 1 to turn off
scaling. (Integer; Default = NDV (number of design variables))
ITMAX
Maximum number of iterations allowed at optimizer level during each design cycle.
(Integer; Default = 40)
ITRMOP
ITRMST
JTMAX
Maximum number of iterations allowed at the optimizer level for the Sequential
Linear Programming Method. This is the number of linearized subproblems
solved. (Integer 0; Default = 20)
JPRINT
JWRITE
If JWRITE > 0, file number on which iteration history will be written. (Integer > 0;
Default = 0)
(Continued)
3-41
DOPTPRM
Name
METHOD
P1
Print control items specified for P2. (Integer 0; Default = 0) Initial results are
always printed prior to the first approximate optimization. If an optimization task
is performed, final results are always printed for the final analysis unless
PARAM,SOFTEXIT,YES is specified. These two sets of print are not controllable.
n: Print at every n-th design cycle.
P2
PTOL
STPSCL
Scaling factor for shape finite difference step sizes, to be applied to all shape
design variables. (Real > 0.0; Default = 1.0)
3-42
P1
P2
IPRINT
METHOD
no data recovery
n=0
n>0
3-43
never
n=0
n>0
3-44
SECTION 4
EXAMPLES I
EXAMPLE 3-BAR TRUSS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1
10-BAR TRUSS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
10-BAR TRUSS RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-10
BEAM VIBRATION FREQUENCY AS A DESIGN CONSTRAINT
(M. J. Turner 1970) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15
BEAM VIBRATION FREQUENCY AS A DESIGN CONSTRAINT RESULTS . . . . . 4-22
20 ksi tension
15 ksi compression
Displacements at grid 4:
x direction 0.2 in
y direction 0.2 in
4-1
y
10"
10"
x
A2
A1
1
3
10"
A1
4
Subcase 2
20,000 lbs
x: 16,000 lbs
y: 12,000 lbs
Subcase 1
20,000 lbs
x: 16,000 lbs
y: 12,000 lbs
4-2
4-3
4-4
4-5
------------------------------------------------------------------------------------------------OBJECTIVE FROM
OBJECTIVE FROM
FRACTIONAL ERROR
MAXIMUM VALUE
CYCLE
APPROXIMATE
EXACT
OF
OF
NUMBER
OPTIMIZATION
ANALYSIS
APPROXIMATION
CONSTRAINT
-------------------------------------------------------------------------------------------------INITIAL
4.828427E+00
-3.234952E-01
1
3.007897E+00
3.008492E+00
-1.977251E-04
-3.737402E-03
2
2.821953E+00
2.821638E+00
1.118734E-04
-1.967246E-02
3
2.734469E+00
2.734299E+00
6.217039E-05
-7.241016E-03
4
2.708915E+00
2.708921E+00
-2.024285E-06
-2.369141E-04
5
2.702065E+00
2.702063E+00
7.941219E-07
-2.666992E-04
-------------------------------------------------------------------------------------------------DESIGN VARIABLE HISTORY
------------------------------------------------------------------------------------------------------------------------INTERNAL | EXTERNAL |
|
DV. ID. |
DV. ID. |
LABEL
| INITIAL
:
1
:
2
:
3
:
4
:
5
:
------------------------------------------------------------------------------------------------------------------------1 |
1
| A1
| 1.0000E+00 : 7.1020E-01 : 7.8436E-01 : 8.1374E-01 : 8.1739E-01 :
8.3569E-01 :
2 |
2
| A2
| 2.0000E+00 : 9.9976E-01 : 6.0313E-01 : 4.3271E-01 : 3.9699E-01 :
3.3838E-01 :
------------------------------------------------------------------------------------------------------------------------*** USER INFORMATION MESSAGE 6464 (DOM12E)
RUN TERMINATED DUE TO HARD CONVERGENCE TO AN OPTIMUM AT CYCLE NUMBER =
5.
10-BAR TRUSS
Design model description
Design variables 10 cross-sectional areas
Objective structural weight
Constraints:
Tensile stress 25,000 psi
Compressive stress 25,000 psi
360"
360"
1
5
1
2
10
5
8
9
4
3
6
4
100,000 lbs
Note:
360"
2
100,000 lbs
Several variations of this problem appear in the structural optimization literature. Multiple load paths are available to support the loads at grids 2 and 4.
4-6
4-7
4-8
4-9
4-10
11
10
4-11
INITIAL
2.098234E+03
6.370801E-01
1.891479E+03
1.891557E+03
-4.091475E-05
2.185570E-02
1.727591E+03
1.727633E+03
-2.430620E-05
2.338859E-02
1.686985E+03
1.686986E+03
-3.618000E-07
2.917891E-03
1.649626E+03
1.649633E+03
-4.291912E-06
4.533672E-03
1.633149E+03
1.633161E+03
-7.474479E-06
1.653125E-03
1.622287E+03
1.622258E+03
1.745734E-05
3.786953E-03
1.610578E+03
1.610584E+03
-3.259081E-06
9.830468E-04
1.605705E+03
1.605713E+03
-5.017484E-06
5.797110E-03
1.597221E+03
1.597220E+03
2.292802E-07
2.474219E-04
10
1.593996E+03
1.593996E+03
-2.297439E-07
4.404531E-03
---------------------------------------------------------------------------------------------------------------
4-12
Objective Function
4-13
4-14
6"
A2
A3
t1
t2
t3
20"
20"
20"
Objective
Structural weight minimization
Constraint
Fundamental transverse vibration frequency 20 Hz
Design variables
A1, A2, A3 (bar cross-sectional areas)
t1, t2, t3 (web thickness)
4-15
4-16
4-17
4-18
4-19
7
6
4-20
1.920000E+01
N/A
9.599569E+00
9.599999E+00
-4.480283E-05
9.583256E-02
7.769873E+00
7.771017E+00
-1.472050E-04
1.124312E-02
7.153115E+00
7.153544E+00
-5.999172E-05
-8.152450E-03
7.007634E+00
7.007762E+00
-1.830387E-05
6.831794E-04
6.964279E+00
6.964198E+00
1.157139E-05
2.233088E-03
6
6.962721E+00
6.962722E+00
-2.054529E-07
2.271554E-03
---------------------------------------------------------------------------------------------------------------
4-21
---------------------------------------------------------------------------------------------------------------------------------INTERNAL |
EXTERNAL
|
|
DV. ID. |
DV. ID.
|
LABEL
|
6
:
7
:
8
:
9
:
10
:
11
:
---------------------------------------------------------------------------------------------------------------------------------1 |
1
| A1
|
8.2998E-01 :
2 |
2
| A2
|
4.3973E-01 :
3 |
3
| A3
|
1.1371E-01 :
4 |
4
| T1
|
4.9550E-02 :
5 |
5
| T2
|
4.2818E-02 :
6 |
6
| T3
|
2.6718E-02 :
*** USER INFORMATION MESSAGE 6464 (DOM12E)
RUN TERMINATED DUE TO HARD CONVERGENCE TO AN OPTIMUM AT CYCLE NUMBER =
6.
4-22
4-23
SECTION 5
SPECIAL MODELING TOPICS I
DESIGN VARIABLE LINKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1
DLINK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
REDUCED BASIS FORMULATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
DESIGN RESPONSES AND CASE CONTROL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9
MODE TRACKING FEATURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-11
MODE TRACKING REQUEST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-12
MODE TRACKING PARAMETERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13
EQUALITY CONSTRAINTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-14
FREQUENCY MATCHING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-15
5-1
DLINK
DLINK
Relates
one design variable to one or more other design variables.
Format:
1
DLINK
ID
DDVID
C0
CMULT
IDV1
C1
IDV2
C2
IDV3
C3
-etc.-
10
0.1
0.33
2.0
1.0
7.0
10
Example:
DLINK
Field
Contents
ID
DDVID
C0
CMULT
IDVi
Ci
Remarks:
1.
2.
This capability provides a means of linking physical design variables such as element
thicknesses to nonphysical design variables such as the coefficients of interpolating
functions.
3.
CMULT provides a simple means of scaling the Ci. For example if Ci = 1/7, 2/7, 4/7, etc.
is desired, then CMULT = 1/7 and Ci = 1, 2, 4, etc., may be input.
4.
An independent IDVi must not occur on the same DLINK entry more than once.
5.
5-2
t1
0
t2
10.
t3
20.
t4
30. 40.
t4 t1
t i = ---------------- x i + t 1
30
so,
2
1
t 2 = --- t 1 + --- t 4
3
3
1
2
t 3 = --- t 1 + --- t 4
3
3
5-3
DESVAR, 1,
T1,
1.,
0.01,
5.
DESVAR, 1,
T2,
1.,
0.01,
5.
DESVAR, 1,
T3,
1.,
0.01,
5.
DESVAR, 1,
T4,
1.,
0.01,
5.
$
DLINK, 11,
2,
0.,
0.333, 1,
2.,
4,
1.
DLINK, 12,
3,
0.,
0.333, 1,
1.,
4,
2.
$
DVPREL1,21,
PSHELL, 101,
4,
0.01,
5.,
,
,
+
+,
1,
1.0
DVPREL1,22,
PSHELL, 102,
4,
0.01,
5.,
,
,
+
+,
2,
1.0
DVPREL1,23,
PSHELL, 103,
4,
0.01,
5.,
,
,
+
+,
3,
1.0
DVPREL1,24,
PSHELL, 104,
4,
0.01,
5.,
,
,
+
+,
4,
1.0
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......0
The only independent variables are t1 and t4. t2 and t3 are the
dependent quantities. The design space is now two-dimensional,
rather than four.
Reducing the number of independent design variables offers greater
advantages for large design problems.
5-4
t1
t2
t10
x=L
5-5
F1
F2
F3
0.5
1.0
x
F 1 --- = C 1
L
x
x
F 2 --- = C 1 1 ---
L
L
x
x 2
F 3 --- = C 1 1 ---
L
L
C 1 = 1.0
5-6
x/L
t2
1.0
t
1.0
3
t
1.0
4
t
1.0
5
=
1.0
t6
1.0
t7
1.0
t8
t
1.0
9
t
1.0
10
1.0
1.0
0.9
0.81
0.8
0.64
0.7
0.49
0.4
x1
0.36
x2
0.25
x3
0.16
0.3
0.09
0.2
0.04
0.1
0.01
0.6
0.5
5-7
$PSHELL,PID,
PSHELL, 101,
PSHELL, 102,
PSHELL, 103,
PSHELL, 104,
PSHELL, 105,
PSHELL, 106,
PSHELL, 107,
PSHELL, 108,
PSHELL, 109,
PSHELL, 110,
$
$DESVAR,ID,
DESVAR, 1,
DESVAR, 2,
DESVAR, 3,
$
$DVPREL1,ID,
$+,
DVID1,
DVPREL1,201,
+,
1,
DVPREL1,202,
+,
1,
DVPREL1,203,
+,
1,
DVPREL1,204,
+,
1,
DVPREL1,205,
+,
1,
DVPREL1,206,
+,
1,
DVPREL1,207,
+,
1,
DVPREL1,208,
+,
1,
DVPREL1,209,
+,
1,
DVPREL1,210,
+,
1,
MID1,
100,
100,
100,
100,
100,
100,
100,
100,
100,
100,
T,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
1.0,
MID2,
100,
100,
100,
100,
100,
100,
100,
100,
100,
100,
12I/T3
,
,
,
,
,
,
,
,
,
,
MID3
100
100
100
100
100
100
100
100
100
100
LABEL,
X1,
X2,
X3,
XINIT,
0.33,
0.33,
0.33,
XLB,
-1.0,
-1.0,
-1.0,
XUB,
+1.0
+1.0
+1.0
DELXV
TYPE,
COEF1,
PSHELL,
1.0,
PSHELL,
1.0,
PSHELL,
1.0,
PSHELL,
1.0,
PSHELL,
1.0,
PSHELL,
1.0,
PSHELL,
1.0,
PSHELL,
1.0,
PSHELL,
1.0,
PSHELL,
1.0,
PID,
DVID2,
101,
2,
102,
2,
103,
2,
104,
2,
105,
2,
106,
2,
107,
2,
108,
2,
109,
2,
110,
2,
FID,
COEF2,
4,
1.0,
4,
0.9,
4,
0.8,
4,
0.7,
4,
0.6,
4,
0.5,
4,
0.4,
4,
0.3,
4,
0.2,
4,
0.1,
PMIN,
...
0.01,
3,
0.01,
3,
0.01,
3,
0.01,
3,
0.01,
3,
0.01,
3,
0.01,
3,
0.01,
3,
0.01,
3,
0.01,
3,
PMAX,
C0,
,
1.0
,
0.81
,
0.64
,
0.49
,
0.36
,
0.25
,
0.16
,
0.09
,
0.04
,
0.01
5-8
5-9
5-10
i Mi i 1 = ti
Updated DRESP1 data available in the .pch file
$
$ Mode Tracking has been performed successfully. Updated
DRESP1 entries are:
$
DRESP1 301LFREQ FREQ 2
DRESP1 303HFREQ FREQ 4
5-11
MODTRAK
Format:
MODTRAK = n
Example:
MODTRAK = 100
Describer
Meaning
Remark:
1.
Selection of a MODTRAK Bulk Data entry with the MODTRAK Case Control command
activates mode tracking for the current subcase. This request is limited to normal modes
subcases (ANALYSIS = MODES) in design optimization (SOL 200).
5-12
MODTRAK
Format:
1
MODTRAK
SID
10
Example:
MODTRAK
100
26
0.80
Field
Contents
SID
Sets identification number that is selected in the Case Control Section with the
MODTRAK command. (Integer; No Default) See Remark 1.
LOWRNG
HIGHRNG
MTFILTER
(Integer 0,
Remarks:
1.
Only the designed modes for the subcase will be tracked. A designed mode is one that is
used in the design model (in connection with either objective or constraints) and, therefore, identified on a DRESP1 entry.
2.
The range of modes LOWRNG through HIGHRNG, inclusive, will be used to track the designed modes. If LOWRNG and HIGHRNG are both blank, then all computed modes will
be used to search for the designed modes. Since large numbers of computed modes will
result in higher computational costs, limiting the search range with LOWRNG and HIGHRNG is recommended.
3.
Modes are considered to correlate if their mass normalized cross orthogonalities are
greater than MTFILTER.
5-13
EQUALITY CONSTRAINTS
Equality constraints are not provided for directly, although they can
be formulated in MSC/NASTRAN.
The simplest way is to provide lower and upper bounds that are
equivalent:
r r(x) r
This can be written using a single DCONSTR entry
$
$ EIGENVALUE FOR MODE 2:
DRESP1, 15,
LAMA1, EIGN,
,
,
$
$ BOUNDS:
DCONSTR,100,
15,
3.5531E4,3.5531E4
$
$ 30 Hz
5-14
FREQUENCY MATCHING
Design objective:
Assume that the eigenvalues of a natural vibration problem
i, i + 1, i + 2, are found from an analysis of the current model.
Since the results of the shake tests indicate that they should be
i, i + 1, i + 2, we wish to make adjustments to a set of
structural parameters to match both the analysis and test results
Approach 1
Use equality constraints or two narrowly defined inequality
constraints
i i i +
Approach 2
Form a new synthetic response as the sum of frequency errors
squared and minimize it
ERR = ( 1.0 i i )2 + ( 1.0 ( i + 1 i + 1 ) )2 +
5-15
5-16
SECTION 6
DESIGN MODELING II
DESIGN MODEL DEFINITION PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1
USER-DEFINED DESIGN VARIABLE-TO-PROPERTY RELATIONS. . . . . . . . . . . . 6-2
DVPREL2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
DEQATN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5
DTABLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-9
EXAMPLE DEFINING A RECTANGULAR SECTION BAR . . . . . . . . . . . . . . . . . 6-10
USER-DEFINED RESPONSES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-11
DRESP2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-13
EXAMPLE DEFORMATION CONTROL WITH TYPE-2 RESPONSES. . . . . . . . . 6-16
EXAMPLE BAR AXIAL PLUS BENDING STRESSES . . . . . . . . . . . . . . . . . . . . . 6-19
RESTRICTIONS IN FORMING SYNTHETIC RESPONSES. . . . . . . . . . . . . . . . . . . 6-20
6-1
Analysis
Model
Parameters
PL
Po
f(X,C)
C2
X DVPREL1
=
PNL
DVPREL2
6-2
DVPREL2
DVPREL2
Variable
Relation
Defines the relation between an analysis model property and design Design
variables
withtoaProperty
user-supplied
equation.
Format:
1
DVPREL2
ID
TYPE
PID
FID
PMIN
PMAX
EQID
DESVAR
DVID1
DVID2
DVID3
-etc.-
DTABLE
LABL1
LABL2
LABL3
-etc.-
13
PBAR
712
0.2
DESVAR
11
13
DTABLE
PI
YM
10
Example:
DVPREL2
Field
Contents
ID
TYPE
PID
FID
PMIN
Minimum value allowed for this property. If FID references a stress recovery
location field, then the default value for PMIN is 1.0+35. PMIN must be
explicitly set to a negative number for properties that may be less than zero (for
example, field ZO on the PCOMP entry). (Real; Default = 1.E20)
PMAX
EQID
DESVAR
DESVAR flag. Indicates that the IDs of DESVAR entries follow. (Character)
DVIDi
DTABLE
DTABLE flag. Indicates that the IDs for the constants in a DTABLE entry follow.
This field may be omitted if there are no constants involved in this relation.
(Character)
LABLi
6-3
DVPREL2
Remarks:
1.
The variables identified by DVIDi and LABLi correspond to variable names (x1, x2, etc.)
listed in the left-hand side of the first equation on the DEQATN entry identified by EQID.
The variable names x1 through xN (where N = m + n) are assigned in the order DVID1,
DVID2, ..., DVIDn, LABL1, LABL2, ..., LABLm.
2.
If both DESVAR and DTABLE are specified in field 2, DESVAR must appear first.
3.
FID may be either a positive or a negative number. If FID > 0, it identifies the field position
on a property entry. If FID < 0, it identifies the word position of an entry in EPT. For
example, to specify the area of a PBAR, either FID = +4 or FID = 3 may be used. However, if PTYPE = PBEAM, FID must be negative. See Remark 4 on the DVPREL1 entry
description for specification of the PBEAM element property tables.
4.
PTYPE = PBEND is not supported, either directly through FIDs or indirectly via word positions in the element property table.
6-4
DEQATN
DEQATN one or more equations for use in design sensitivity or p-element analysis.
Design Equation Definition
Defines
Format:
1
DEQATN
EQID
10
EQUATION
EQUATION (Cont.)
Example:
DEQATN
14
Field
Contents
EQID
EQUATION
Remarks:
1.
6-5
DEQATN
A
3
F 1 = A + B C ( D + 10 ) + sin ( PI ( 1 ) R ) + -------------BC
F = A + B F1 D
where SIN and PI are intrinsic functions. See Remark 4.
2.
EQUATION may contain embedded blanks. EQUATION must contain less than 12,500
nonblank characters. This is equivalent to approximately 195 continuation entries.
3.
The syntax of the expressions follows FORTRAN language standards. The allowable
arithmetic operations are shown in Table 1 in the order of execution precedence. Parenthesis are used to change the order of precedence. Operations within parentheses are
performed first with the usual order of precedence being maintained within the
parentheses.
Operator
Operation
Sample Expressions
Interpreted As
, +
XY
X(Y)
Exponentiation
XY
(XY)
, +
Negative or Positive
XY
(X)Y
, /
Multiplication or Division
XY+Z
(XY)+Z
+,
Addition or Subtraction
X+Y
X+Y
4.
The expressions may contain intrinsic functions. Table 2 contains the format and descriptions of functions that may appear in the expressions. The use of functions that may be
discontinuous must be used with caution because they can cause discontinuous derivatives. These are ABS, DIM, MAX, MIN, and MOD. For examples and further details see
the MSC/NASTRAN DMAP Module Dictionary.
(Continued)
6-6
DEQATN
Format
Description
Mathematical Expression
ABS(x)
absolute value
|x|
ACOS(x)
arccosine
ACOSH(x)
hyperbolic arccosine
ASIN(x)
arcsine
ASINH(x)
hyperbolic arcsine
ATAN(x)
arctangent
ATAN2(x,y)
arctangent of quotient
ATANH(x)
hyperbolic arctangent
ATANH2(x,y)
COS(x)
cosine
cos x
COSH(x)
hyperbolic cosine
cosh x
DIM(x,y)
positive difference
x MIN(x,y)
EXP(x)
exponential
LOG(x)
natural logarithm
loge x
LOG10(x)
common logarithm
log10 x
LOGX(x,y)
base x logarithm
logx y
maximum
minimum
MOD(x,y)
remainder (modulo)
x y (INT(x/y))
PI(x)
multiples of pi ()
SIN(x)
sine
sin x
SINH(x)
hyperbolic sine
sinh x
SQRT(x)
square root
TAN(x)
tangent
tan x
TANH(x)
hyperbolic tangent
tanh x
-1
cos
cosh
-1
-1
sin
sinh
tan
tan
x
x
-1
-1
-1
x
(x/y)
-1
tanh
-1
tanh
x
(x/y)
6-7
DEQATN
5.
b.
DRESP2 entry, then xi represents the DVIDj, LABLk, NRm, and Gp fields in that
order.
c.
d.
e.
6.
7.
The DMAP logical operators NOT, AND, OR < XOR, and XQV cannot be used as Xi
names.
8.
Input errors on the DEQATN entry often result in poor messages. Substituting a [ for a
parenthesis or violating the restriction against large field format are examples. Known
messages are UFM 215, SFM 233 and UFM 5199. If any of these messages are encountered then review the DEQATN entry input.
6-8
DTABLE
Table Constants
DTABLE a table of real constants that are used in equations (see DEQATN entry).
Defines
Table Constants
Format:
1
DTABLE
LABL1
VALU1
LABL2
VALU2
LABL3
VALU3
LABL4
VALU4
LABL5
VALU5
LABL6
VALU6
-etc.-
PI
3.142
10.1
1.0E6
0.1
5.5E5
100.
Example:
DTABLE
Field
Contents
LABLi
VALUi
Remarks:
1.
Only one DTABLE entry may be specified in the Bulk Data Section.
2.
6-9
10
h = 0.4
A = bh
h
1
C 1 = --- b
2
1
C 2 = --- h
2
bh
I 1 = ---------12
hb
I 2 = ---------12
H1,
.4,
.1,
$
$...BAR PROPERTY RELATIONS, A, I1, I2:
$DVPREL2,ID,
TYPE,
PID,
FID,
$+,
DESVAR,DVID1, DVID2, ...,
$+,
DTABLE,CID1,
CID2,
...
DVPREL2,250,
PBAR,
120,
4,
+,
DESVAR, 10,
11
DVPREL2,251,
PBAR,
120,
5,
+,
DESVAR, 10,
11
DVPREL2,252,
PBAR,
120,
6,
+,
DESVAR, 10,
11
$
$...EQUATIONS:
DEQATN 501
AREA(B,H) = B*H
DEQATN 502
I1(B,H) = B*H**3/12.
DEQATN 503
I2(B,H) = H*B**3/12.
$
$...STRESS RECOVERY POINT LOCATIONS:
$DVPREL1,ID,
TYPE,
PID,
FID,
$+,
DVID1, COEF1, DVID2, COEF2,
DVPREL1,260,
PBAR,
120,
12,
+,
10,
-.5
DVPREL1,261,
PBAR,
120,
13,
+,
11,
.5
6-10
1.0
PMIN,
,
PMAX,
,
EQID,
,
,
,
+
+
501,
1.0E-5, ,
502,
1.0E-5, ,
503,
PMIN,
...
-.5,
PMAX,
C0,
.05,
USER-DEFINED RESPONSES
Design responses can be classified based on how they are defined:
DRESP1
Design
=
Responses
Simple
Responses
(Type 1)
Synthetic
Responses
(Type 2)
DRESP2
Responses written using equations
6-11
U =
displacement
responses
such
as
ux + uy + uz
6-12
DRESP2
Format:
1
DRESP2
ID
LABEL
EQID
REGION
DESVAR
DVID1
DVID2
DVID3
DVID4
DVID5
DVID6
DVID7
DVID8
-etc.-
LABL1
LABL2
LABL3
LABL4
LABL5
LABL6
LABL7
LABL8
-etc.-
NR1
NR2
NR3
NR4
NR5
NR6
NR7
NR8
-etc.-
G1
C1
G2
C2
G3
C3
G4
C4
-etc.-
DPIP1
DPIP2
DPIP3
DPIP4
DPIP5
DPIP6
DPIP7
DPIP8
DPIP9
-etc.-
LBUCK
DESVAR
101
205
209
DTABLE
DRESP1
DNODE
DVPREL1
Example:
DRESP2
201
DTABLE
PI
YM
DRESP1
14
22
33
DNODE
14
22
43
101
102
DVPREL1
Field
Contents
ID
LABEL
6-13
10
DRESP2
EQID
REGION
DESVAR
DVIDi
DTABLE
Flag indicating that the labels for the constants in a DTABLE entry follow.
(Character)
LABLj
DRESP1
NRk
DNODE
Flag signifying that the following fields are designed grid points. See Remark 7.
(Character)
Gm
Cm
DVPREL1
DPIPi
Remarks:
1.
DRESP2 entries may only reference DESVAR, DTABLE, DRESP1, and DVPREL1 entries. They may not reference other DRESP2 entries.
2.
3.
DRESP2 entries must have unique identification numbers with respect to DRESP1
entries.
4.
The DESVAR, DTABLE, DRESP1, DNODE, and DVPREL1 flags in field 2 must
appear in the order given above. Any of these words, along with the identification numbers associated with them, may be omitted if they are not involved in this DRESP2
relationship. However, at least one of these four types of arguments must exist.
5.
The REGION field follows the same rules as for the DRESP1 entries. DRESP1 and
DRESP2 responses will never be contained in the same region, even if they are assigned
the same REGION identification number. The default is to put all responses referenced
by one DRESP2 entry in the same region.
(Continued)
6-14
DRESP2
The variables identified by DVIDi, LABLj, NRk, the Gm, Cm pairs and DPIPi are assigned
(in that order) to the variable names (x1, x2, x3, etc.) specified in the left-hand side of the
first equation on the DEQATN entry referenced by EQID. The variable names x1 through
xN (N = m + n + p + q) are assigned in the order DVID1, ..., DVIDm, LABL1, ..., LABLn,
NR1, ..., NRp, G1, ..., Gq, DPIP1, ..., DPIPr. In the example below,
DESVARs 101 and 3 are assigned to arguments A and B.
DTABLEs PI and YM are assigned to arguments C and D.
Grid 14, Component 1 is assigned to argument R.
DRESP2
LBUCK
DESVAR
101
DTABLE
PI
YM
DNODE
14
DEQATN
7.
(Gm, Cm) refer to a designed grid component. Depending on the scheme used in generating basis vectors, a designed grid component can be one of the following:
a.
b.
A grid component that is free to move in a full auxiliary model when the external displacement fields are used.
c.
When the auxiliary model boundary shape method is used, a designed grid component can be either a loaded grid component in the auxiliary model or a grid
component with motion that is obtained from interpolation of boundary shapes.
d.
When the interface using geometric boundary shapes is used, a designed grid component is either a grid component defined on a DVGRID entry or one with motion that
is obtained from interpolation of boundary shapes.
6-15
desired shape
deformed shape
u1
1
u2
2
u3
u4
4
original shape
Design objective:
Bar displaces but does not bend
Target displacements become:
2u 1 + u 4
u 2 = ----------------------3
and
u 1 + 2u 4
u 3 = ----------------------3
6-16
DEF =
2
2
( u 2 u 2 ) + ( u 3 u 3 )
$ IN CASE CONTROL:
$
DESOBJ(MIN) = 21
$
$ IN BULK DATA:
$
$...IDENTIFY THE FIRST-LEVEL DISPLACEMENT RESPONSES:
$
$DRESP1,ID,
LABEL, RTYPE, PTYPE, REGION, ATTA,
ATTB,
ATT1,
+
$+,
ATT2,
...
DRESP1, 11,
UY1,
DISP,
,
,
2,
,
1
DRESP1, 12,
UY2,
DISP,
,
,
2,
,
2
DRESP1, 13,
UY3,
DISP,
,
,
2,
,
3
DRESP1, 14,
UY4,
DISP,
,
,
2,
,
4
$
$DRESP2,ID,
LABEL, EQID,
REGION, ,
,
,
,
+
$+,
DRESP1, NR1,
NR2,
...
DRESP2, 21,
DEF2,
100,
,
,
,
,
,
+
+,
DRESP1, 11,
12,
13,
14
$
DEQATN 100
F1(A,B,C,D) = (2.*A + D)/3.0 ;
+
+
F2
= (A + 2.*D)/3.0 ;
+
+
F
= SQRT( (B-F1)**2 + (C-F2)**2 )
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......0
6-17
PID1
PID2
DRESP2
.
.
.
DRESP1
PIDm
6-18
EID1
EID2
.
.
.
EIDn
$ ANALYSIS:
$
CBAR,
10,
CBAR,
11,
CBAR,
12,
CBAR,
13,
$
PBAR,
100,
+,
12.5,
$
$ DESIGN:
$
DRESP1, 110,
DRESP1, 111,
DRESP1, 112,
DRESP1, 113,
DRESP1, 114,
$
DRESP2, 210,
+,
DRESP1,
DRESP2, 211,
+,
DRESP1,
DRESP2, 212,
+,
DRESP1,
DRESP2, 213,
+,
DRESP1,
$
DEQATN 303
100,
100,
100,
100,
1,
2,
3,
4,
2,
3,
4,
5,
1.,
1.,
1.,
1.,
101,
25.,
1.25E3, 2.6042E5,6.5104E4,,
12.5,
-25.,
-12.5, -25.,
,
-12.5,
,
25.
SA1,
SA2,
SA3,
SA4,
AXIAL,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
PBAR,
PBAR,
PBAR,
PBAR,
PBAR,
,
,
,
,
,
2,
3,
4,
5,
6,
,
,
,
,
,
100
100
100
100
100
SR1,
110,
SR2,
111,
SR3,
112,
SR4,
113,
303,
114
303,
114
303,
114
303,
114
SRI(POINT,AXIAL) = POINT+AXIAL
6-19
1.,
1.,
1.,
1.,
0.
0.
0.
0.
+
6-20
SECTION 7
EXAMPLES II
STIFFENED PANEL WEIGHT MINIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1
25-BAR TRUSS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12
CANTILEVER PLATE (REDUCED BASIS FORMULATION) . . . . . . . . . . . . . . . . . 7-26
Fz
Web Cap
10404
20204
Fx
10204
10203
Y
20200
10004
10400
10200
Web
Plate
10000
Load cases:
1.
2.
7-1
7-2
1
--2
2
A
A
I 12 = I 12 ------- , C j = C j -------
0 A 0
0 A 0
7-3
7-4
10402,
10403,
10404,
20100,
20101,
20102,
20103,
20104,
20200,
20201,
20202,
20203,
20204,
,
,
,
,
,
,
,
,
,
,
,
,
,
5.0,
7.5,
10.0,
0.0,
2.5,
5.0,
7.5,
10.0,
0.0,
2.5,
5.0,
7.5,
10.0,
10.0,
10.0,
10.0,
5.0,
5.0,
5.0,
5.0,
5.0,
5.0,
5.0,
5.0,
5.0,
5.0,
0.0
0.0
0.0
1.0
1.0
1.0
1.0
1.0
2.0
2.0
2.0
2.0
2.0
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15,
16,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
1,
10000,
10001,
10002,
10003,
10100,
10101,
10102,
10103,
10200,
10201,
10202,
10203,
10300,
10301,
10302,
10303,
10001,
10002,
10003,
10004,
10101,
10102,
10103,
10104,
10201,
10202,
10203,
10204,
10301,
10302,
10303,
10304,
10101,
10102,
10103,
10104,
10201,
10202,
10203,
10204,
10301,
10302,
10303,
10304,
10401,
10402,
10403,
10404,
10100
10101
10102
10103
10200
10201
10202
10203
10300
10301
10302
10303
10400
10401
10402
10403
21,
22,
23,
24,
25,
26,
27,
28,
2,
2,
2,
2,
2,
2,
2,
2,
10200,
10201,
10202,
10203,
20100,
20101,
20102,
20103,
10201,
10202,
10203,
10204,
20101,
20102,
20103,
20104,
20101,
20102,
20103,
20104,
20201,
20202,
20203,
20204,
20100
20101
20102
20103
20200
20201
20202
20203
31,
32,
33,
34,
3,
3,
3,
3,
20200,
20201,
20202,
20203,
20201,
20202,
20203,
20204,
0.0,
0.0,
0.0,
0.0,
1.0,
1.0,
1.0,
1.0,
1,
2,
3,
0.06,
1,
1,
1,
0.6,
0.15,
0.2,
0.144,
0.06,
1
1
1.728-4,1.728-2,1.745-2,,
-0.6,
-0.06, -0.6,
-0.06,
7-5
0.0
0.0
0.0
0.0
,
0.6
+PB3
7-6
7-7
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
7-8
7-9
8
7
7-10
--------------------------------------------------------------------------------------------------------------OBJECTIVE FROM
OBJECTIVE FROM
FRACTIONAL ERROR
MAXIMUM VALUE
CYCLE
APPROXIMATE
EXACT
OF
OF
NUMBER
OPTIMIZATION
ANALYSIS
APPROXIMATION
CONSTRAINT
--------------------------------------------------------------------------------------------------------------INITIAL
5.784520E+00
1.311878E+00
1
1.122267E+01
1.122355E+01
-7.868297E-05
4.089278E-01
2
1.636887E+01
1.636941E+01
-3.285838E-05
-1.610547E-03
3
1.294620E+01
1.294636E+01
-1.259646E-05
-8.604820E-02
4
8.748927E+00
8.751337E+00
-2.753791E-04
1.988992E-02
5
7.379158E+00
7.379156E+00
3.230973E-07
9.885500E-02
6
7.954716E+00
7.954452E+00
3.315011E-05
4.145312E-04
7
7.954473E+00
7.954452E+00
2.637622E-06
4.145312E-04
---------------------------------------------------------------------------------------------------------------
7-11
25-BAR TRUSS
This design problem calls for a minimum weight structure subject to
member stress, simple Euler buckling, and joint displacement
constraints, under static loading conditions.
Each structural element is a hollow tube with a thickness-to-diameter
ratio of 1:10. Diameters will be the design variables.
1
2
6
3
4
5
10
7
Y
8
X
7-12
7-13
7-14
EI
cr = -----------2
AL
With:
A = 2Rt
2
= ------D
10
3
I = R t
4
= ------D
80
(2)
(2)
are:
2
8L F s
= --------- = --------------------- 1
2
cr
( D ) E
7-15
7-16
7-17
7-18
+
+
+
+
+
+
+
+
+
+
7-19
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
7-20
7-21
13
12
6.493687E+02
2.031843E-00
7-22
8.805606E+02
8.805025E+02
6.599126E-05
1.628765E-01
8.741666E+02
8.741613E+02
6.074461E-06
-2.911091E-04
8.605269E+02
8.605145E+02
1.439852E-05
1.001358E-04
8.496406E+02
8.496353E+02
6.321647E-06
2.777576E-04
8.410117E+02
8.410120E+02
-2.902939E-07
3.701448E-04
8.340327E+02
8.340328E+02
-2.195423E-07
6.405115E-04
8.280418E+02
8.280425E+02
-8.108119E-07
7.916689E-04
8.236416E+02
8.236408E+02
9.633533E-07
6.154776E-04
8.201868E+02
8.201854E+02
1.711575E-06
7.835627E-04
10
8.175861E+02
8.175862E+02
-7.465287E-08
8.921623E-04
11
8.163202E+02
8.163205E+02
-3.738431E-07
7.225275E-04
12
8.155923E+02
8.155926E+02
-2.993414E-07
8.862019E-04
7-23
7-24
7-25
p2
X
t1 t2 t3
t4
t5
40"
t6 t7
t8
10"
Design variables 1, 2, 3
These three parameters determine the longitudinal thickness
distribution of the plate
Initial design 1 = 2 = 3 = 1.0
Design objective
Weight minimization
7-26
t1
1.0
1.0
1.0
t2
1.0
0.875
0.7656
7-27
7-28
7-29
7-30
6,
4,
+DP6
7,
4,
+DP7
8,
4,
+DP8
responses
RTYPE, PTYPE,
REGION, ATTA,
ATTB,
ATT1,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
STRESS,
DISP ,
DISP ,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
1
1
1
1
2
2
2
2
3
3
3
3
4
4
4
4
5
5
5
5
6
6
6
6
7
7
7
7
8
8
8
8
9
29
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
PSHELL,
,
,
7,
9,
16,
17,
7,
9,
16,
17,
7,
9,
16,
17,
7,
9,
16,
17,
7,
9,
16,
17,
7,
9,
16,
17,
7,
9,
16,
17,
7,
9,
16,
17,
3,
3,
7-31
7-32
2.0,
7-33
9
8
7-34
--------------------------------------------------------------------------------------------------------------OBJECTIVE FROM
OBJECTIVE FROM
FRACTIONAL ERROR
MAXIMUM VALUE
CYCLE
APPROXIMATE
EXACT
OF
OF
NUMBER
OPTIMIZATION
ANALYSIS
APPROXIMATION
CONSTRAINT
--------------------------------------------------------------------------------------------------------------INITIAL
7.843750E+01
1.072155E+01
1
1.170497E+02
1.170496E+02
6.518085E-07
4.135007E+00
1.616799E+02
1.616796E+02
1.604404E-06
1.589539E+00
2.065549E+02
2.065548E+02
2.954913E-07
5.076001E-01
2.363098E+02
2.363113E+02
-6.392500E-06
1.115919E-01
2.423183E+02
2.423193E+02
-4.093034E-06
3.467807E-02
2.447022E+02
2.447021E+02
4.988528E-07
1.603751E-02
2.461020E+02
2.461019E+02
1.240038E-07
7.689251E-04
8
2.461019E+02
2.461019E+02
0.000000E+00
7.689251E-04
---------------------------------------------------------------------------------------------------------------
7-35
SECTION 8
SPECIAL MODELING TOPICS II
BEAM CROSS-SECTION LIBRARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1
BEAM CROSS-SECTION LIBRAY OPTIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . 8-2
COMPARISON BETWEEN DESIGNING PBAR AND PBARL. . . . . . . . . . . . . . . . . . 8-3
DVPREL1 INCLUDED ON DRESP2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
DESIGN VARIABLE LINKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-6
CONSTRAINT EQUATION SCALING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-9
BEAM DESIGN GENERAL ISSUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-13
BEAM DESIGN SIMPLE EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-16
May specify cross section dimensions (e.g., height and width for a
rectangular section) rather than properties (e.g., I, A, etc.)
Cross sectional properties are automatically calculated.
Two new Bulk Data entries: PBARL and PBEAML
8-1
8-2
1
1
--- B, --- H
2 2
E
B
*************************************
* Input Data for Designing PBAR
*
*************************************
$
PBAR
9
1
5.0
10.4167 0.4167
-0.5
2.5
0.5
2.5
0.5
-2.5
-0.5
-2.5
C
$
ID
LABEL
INIT
MIN
MAX
DESVAR 1
HEIGHT 5.0
0.01
10.
DESVAR 2
WIDTH
1.0
0.01
10.
$
$ Explict expression for cross-section area
$
ID
TYPE
PID
FID
PMIN
PMAX
EQID
DVPREL2 1
PBAR
9
4
0.01
100
DESVAR 1
2
DEQATN 100
A(H,B)=H*B
$
$ Explict expression for first moment intertia
DVPREL2 2
PBAR
9
5
0.01
200
DESVAR 1
2
DEQATN 200
A(H,B)=B*H**3/12.
$
$ Explict expression for stress data recovery point D (0.5B, 0.5H)
$
ID
TYPE
PID
FID
PMIN
PMAX
C0
DVPREL1 11
PBAR
9
14
0.01
1
0.5
DVPREL1 12
PBAR
9
15
0.01
2
0.5
8-3
9,
1.0
ID
1
2
1,
5.0
LABEL
HEIGHT
WIDTH
BAR
$
INIT
MIN
MAX
DESVAR
5.0
0.01
10.
DESVAR
1.0
0.01
10.
$
$ Dimension Height references to design variable 1
$
ID
TYPE
PID
FID
PMIN
PMAX
DVPREL1 11
BARL
9
13
0.01
1
1
$
$ Dimension Width references to design variable 2
DVPREL1 12
PBARL
9
12
0.01
1
2
8-4
C0
DRESP2
ID
LABEL
DESVAR
DRESP1
DNODE
DVPREL1
EQID REGION
DTABLE
-etc.-
-etc.-
NR1
NR2
NR8
-etc.-
G1
NR3
NR4
NR5
NR6
C1
G2
C2
G3
C3
G4
C4
-etc.-
DPIP1
DPIP2
DPIP3
DPIP4
DPIP5
DPIP6
DPIP8
DPIP9
-etc.-
NEW!
8-5
NR7
DPIP7
10
4
bh
I 1 = ---------- = 1.042E-10 m
12
4
hb
I 2 = ---------- = 4.167E-10 m
12
$PBAR, PID,
MID,
A,
I1,
I2
PBAR,
101,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
102,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
103,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
104,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
105,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
106,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
107,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
108,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
109,
100,
5.E-5, 1.042-10,4.167-10
PBAR,
110,
100,
5.E-5, 1.042-10,4.167-10
$
$...INDEPENDENT DESIGN VARIABLE SET:
$DESVAR,ID,
LABEL, XINIT, XLB,
XUB,
DELXV
DESVAR, 1,
X1,
1.0,
-1.0,
+1.0
DESVAR, 2,
X2,
0.0,
-1.0,
+1.0
DESVAR, 3,
X3,
0.0,
-1.0,
+1.0
$
$...DEPENDENT DESIGN VARIABLE SET:
$DESVAR,ID,
LABEL, XINIT, XLB,
XUB,
DELXV
DESVAR, 11,
H1,
.01,
1.E-4, 1.0
DESVAR, 12,
H2,
.01,
1.E-4, 1.0
DESVAR, 13,
H3,
.01,
1.E-4, 1.0
8-6
8-7
C2,
1.0,
0.9,
0.8,
0.7,
0.6,
0.5,
0.4,
0.3,
0.2,
0.1,
,
,
+
+
PBAR,
16
PBAR,
17
PBAR,
18
PBAR,
19
PBAR,
20
106,
6,
1.0E-12,,
40,
107,
6,
1.0E-12,,
40,
108,
6,
1.0E-12,,
40,
109,
6,
1.0E-12,,
40,
110,
6,
1.0E-12,,
40,
8-8
2
8L
E 2
c + ---------- ( D + T ) 0 ------------------------------------ c 1.0
2
2
2
2
8L
E(D + T )
Synthetic Response
0.4ET
D
c + ---------------- 0 ---------------- c 1.0
D
0.4ET
Synthetic Response
8-9
g u = ------------------------------------------------------ 0
Then
g u
1
---------- = ----- 1
2
g
1
-----2
CTMIN
1
CT
Active
8-10
g
1
--------------------2 max
CTMIN
1
1
CT
Active
8-11
8-12
End A
Words
6-21
End B
Words
166-181
8-13
End A Intermediate
Words
Words*
6-21
22-37
End B
Words
166-181
8-14
End A Station C
Words Words
6-21
22-37
Intermediate
Words*
38-53
End B
Words
166-181
8-15
b1
A
h1
h2
B
b2
With initial design:
b
1
1.0
h
1
2.0
{x} = =
b2
0.5
h2
1.0
8-16
8-17
8-18
8-19
COMPARISON BETWEEN INPUT PROPERTY VALUES FROM ANALYSIS AND DESIGN MODELS -----
8-20
--------------------------------------------------------------------------------------------------------------PROPERTY
PROPERTY
FIELD
ANALYSIS
DESIGN
LOWER
UPPER
DIFFERENCE
TYPE
ID
ID
VALUE
VALUE
BOUND
BOUND
FLAG
--------------------------------------------------------------------------------------------------------------PBEAM
100
-177
-5.100000E-01
-5.000000E-01
-1.000000E+35
1.000000E+20
WARNING
PBEAM
100
-175
5.100000E-01
5.000000E-01
-1.000000E+35
1.000000E+20
WARNING
PBEAM
100
-170
4.200000E-02
4.166667E-02
1.000000E-03
1.000000E+20
WARNING
PBEAM
100
-169
1.040000E-02
1.041667E-02
1.000000E-03
1.000000E+20
WARNING
PBEAM
100
-168
5.100000E-01
5.000000E-01
1.000000E-03
1.000000E+20
WARNING
PBEAM
100
-33
-5.100000E-01
-5.000000E-01
-1.000000E+35
1.000000E+20
WARNING
PBEAM
100
-31
5.100000E-01
5.000000E-01
-1.000000E+35
1.000000E+20
WARNING
PBEAM
100
-26
4.200000E-02
4.166667E-02
1.000000E-03
1.000000E+20
WARNING
PBEAM
100
-25
1.040000E-02
1.041667E-02
1.000000E-03
1.000000E+20
WARNING
PBEAM
100
-24
5.100000E-01
5.000000E-01
1.000000E-03
1.000000E+20
WARNING
PBEAM
100
-17
-1.010000E+00
-1.000000E+00
-1.000000E+35
1.000000E+20
WARNING
PBEAM
100
-15
1.010000E+00
1.000000E+00
-1.000000E+35
1.000000E+20
WARNING
PBEAM
100
-10
6.670000E-01
6.666667E-01
1.000000E-03
1.000000E+20
WARNING
PBEAM
100
-9
1.670000E-01
1.666667E-01
1.000000E-03
1.000000E+20
WARNING
PBEAM
100
-8
2.010000E+00
2.000000E+00
1.000000E-03
1.000000E+20
WARNING
1.
IF FIELD ID IS LESS THAN ZERO, IT IDENTIFIES THE WORD POSITION OF AN ENTRY IN EPT.
2.
IF FIELD ID IS GREATER THAN ZERO, IT IDENTIFIES THE FIELD POSITION ON A PROPERTY BULK DATA ENTRY.
3.
THE DIFFERENCE
IF THE FLAG IS
IF THE FLAG IS
IF THE FLAG IS
FLAG IS USED TO CHARACTERIZE DIFFERENCES BETWEEN ANALYSIS AND DESIGN MODEL PROPERTIES:
NONE, THEN THERE IS NO SIGNIFICANT DIFFERENCE BETWEEN THE TWO VALUES.
WARNING, THEN THE USER IS ADVISED THAT DIFFERENCES EXIST.
FATAL, THEN THE DIFFERENCES ARE GREATER THAN 1.00000E+35 AND THE RUN WILL BE TERMINATED.
SECTION 9
FUNDAMENTALS OF STRUCTURAL OPTIMIZATION
STRUCTURAL OPTIMIZATION THE DIFFICULTIES . . . . . . . . . . . . . . . . . . . . . . . 9-1
STRUCTURAL OPTIMIZATION THE SOLUTION. . . . . . . . . . . . . . . . . . . . . . . . . . 9-3
DESIGN VARIABLE LINKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-7
DESIGN VARIABLE LINKING IN SHAPE REDESIGN . . . . . . . . . . . . . . . . . . . . . . 9-11
GENERATION OF CONSTRAINTS IN MSC/NASTRAN . . . . . . . . . . . . . . . . . . . . . 9-15
CONSTRAINT SCREENING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-16
DSCREEN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-18
FORMAL APPROXIMATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-21
APRCOD SELECTION OF APPROXIMATE FORM . . . . . . . . . . . . . . . . . . . . . . . 9-28
SENSITIVITY ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-30
SEMIANALYTIC SENSITIVITY ANALYSIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-31
DELB FINITE DIFFERENCE STEP SIZE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-33
SENSITIVITY ANALYSIS IMPLEMENTATION . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-35
ADJOINT SENSITIVITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-37
ADJOINT SENSITIVITY THEORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-38
DESIGN SENSITIVITY COEFFICIENT MATRICES . . . . . . . . . . . . . . . . . . . . . . . . . 9-47
DSAPRT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-49
FORMATTED SENSITIVITY PRINT EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-51
MOVE LIMITS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-52
SECTION 9
FUNDAMENTALS OF STRUCTURAL
OPTIMIZATION (Cont.)
MOVE LIMITS IMPOSED ON ANALYSIS MODEL PROPERTIES . . . . . . . . . . . . . 9-53
MOVE LIMITS IMPOSED ON DESIGN VARIABLES . . . . . . . . . . . . . . . . . . . . . . . . 9-55
AUTOMATIC UPDATES OF MOVE LIMITS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-56
MOVE LIMITS UPDATE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-57
PARAM OPTEXIT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-58
FLOW CHART FOR SOLUTION 200 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-59
CONVERGENCE AT THE DESIGN CYCLE LEVEL . . . . . . . . . . . . . . . . . . . . . . . . 9-63
CONVERGENCE CRITERIA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-65
SUMMARY OF PARAMETERS FOR DESIGN OPTIMIZATION . . . . . . . . . . . . . . . 9-67
9-1
STRUCTURAL OPTIMIZATION
THE DIFFICULTIES (Cont.)
A Brute Force coupling of an analysis program and an
optimization program suffers from:
Too many design variables
Too many design constraints
Too many detailed analyses
Finite
Element
Analysis
Program
9-2
Numerical
Optimization
Program
Approximation
Concepts
Design
Variables
Constraints
Temporary deletion of
nonactive constraints and
dynamic updates of
retained constraints.
Small number of
detailed structural
analyses and tests
Explicit approximations of
a small number of the
retained constraints.
Experience and
intuition
Taylor Series
Regression Analysis
Analysis
Results
Notes:
1.
2.
Constraint screening and formal approximations are performed automatically by the code.
9-3
STRUCTURAL OPTIMIZATION
THE SOLUTION (Cont.)
Program flow using approximation techniques
Perform detailed finite element analysis
Calculate all constraints. Delete those constraints that are not
potentially critical (a common sense approach)
Calculate the gradients of all retained constraints
Create a high-quality approximation of the responses with
respect to the design variables
Solve this approximate problem
Update the analysis variables
Perform a detailed analysis of the proposed design
Calculate all constraints
Check for convergence to an optimum
Repeat if necessary
9-4
STRUCTURAL OPTIMIZATION
THE SOLUTION (Cont.)
Approximate Design Model
Improved Designs
Finite
Element
Analysis
Program
Numerical
Optimization
Program
Approximate
Design
Model
9-5
STRUCTURAL OPTIMIZATION
THE SOLUTION (Cont.)
Program Structure with Approximate Model
Initial
Design
Improved
Design
The required number of
iterations of the external loop
must be small.
Structural
Response
Analysis
Approximate
Model
Constraint
Screening
Sensitivity
Analysis
Many Times
Finite Element
Analysis
9-6
Optimizer
360"
360"
1
5
1
2
10
5
8
9
4
3
6
360"
4
100,000 lbs
9-7
2
100,000 lbs
A2
1.0
A
1.0
3
A
1.0
4
A
0.0
5
=
0.0
A6
0.0
A7
0.0
A8
A
0.0
9
A
0.0
10
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1.0
x1
0.0
x2
0.0
x3
1.0
0.0
1.0
0.0
1.0
0.0
1.0
1.0
1.0
9-8
360"
360"
1
5
1
2
10
5
8
360"
9
4
3
4
100,000 lbs
Proposal 1: A 1, A 2, A 3, , A 10
Proposal 2: A 1, A 2, A 3, , A 10
Proposal 3: A 1, A 2, A 3, , A 10
9-9
2
100,000 lbs
1
A1
A1
A
1
2
A2
1
A3
A3
1
A4
A4
A
1
A5
5
=
1
A
A
6
6
1
A7
A7
A8
A8
1
A
A9
9
1
A 10
A
10
A1
2
A2
2
A3
2
A1
3
A2
3
A3
3
A4
A4
2
A5
2
A6
2
A7
2
A8
2
A9
2
A 10
3
A5
4
A6
3
A7
4
A8
5
A9
6
A 10
x
1
x2
x3
9-10
h1
h2
5
{ x }D = [ x1 x3 x4 x5 ]
9-11
9-12
R
Circular Arc
Straight Line
1 Grid Point
1 Parameter (R)
2 Grid Points
Cubic Polynomial
3 Grid Points
1 Slope
9-13
F1 ( )
F2 ( )
F3 ( )
.
.
.
Basis Functions
9-14
(DRESP1, 2)
(DCONSTR)
9-15
CONSTRAINT SCREENING
Purpose:
The number of constraints for structural design problems could be
extremely large, but most of them do not directly influence the
design process. Since the design sensitivity analysis becomes
very expensive for large numbers of constraints, and the amount of
data transferred to approximate optimization is proportional to the
number of constraints in that phase, it is necessary to delete all the
unnecessary constraints before moving into the sensitivity
analysis.
Constraint deletion consists of two steps:
Step I - Deletion:
Based on constraint value. A constraint G(X) is deleted temporarily
if G(X) < TRS, where TRS is the truncation threshold value given by
a DSCREEN entry for each constraint type.
Gj(X)
x x x x x x x x
x
0
TRS
1.0
List of Constraints
9-16
x x x x x x x x
0
TRS
1.0
Retained
Responses
Region 1
1
Region2
0
9-17
Region 3
2
DSCREEN
Format:
1
DSCREEN
RTYPE
TRS
NSTR
0.7
10
Example:
DSCREEN STRESS
Field
Contents
RTYPE
Response type for which the screening criteria apply. See Table 1. (Character)
TRS
NSTR
Maximum number of constraints to be retained per region per load case. See
Remark 3. (Integer > 0; Default = 20)
Remarks:
1.
Displacement and displacement derivative constraints associated with one particular load
case are grouped by the specification of DRESP1 entries. From each group, a maximum
of NSTR constraints are retained per load case.
RTYPE
DISP
STRESS
STRAIN
FORCE
EQUA
CSTRESS
CSTRAIN
CFAILURE
FLUTTER
FRDISP
Default Region
Specification
RTYPE
DRESP1
Property ID
Property ID
Property ID
DRESP2
Property ID
Property ID
Property ID
DRESP1
DRESP1
FRVELO
FRACCL
FRSTRE
FRFORC
FRSPC
TACCL
TDISP
TVELO
TSTRE
TFORC
TSPC
(Continued)
9-18
Default Region
Specification
DRESP1
DRESP1
Property ID
Property ID
DRESP1
DRESP1
DRESP1
DRESP1
Property ID
Property ID
DRESP1
DSCREEN
Stress-strain constraints are grouped by the property; i.e., all elements belonging to the
set of PIDs specified under ATTi on a DRESPi entry are regarded as belonging to the
same region. In superelement sensitivity analysis, if the property (PID) is defined in more
than one superelement, then separate regions are defined. A particular stress constraint
specification may be applied to many elements in a region generating many stress constraints, but only up to NSTR constraints per load case will be retained.
3.
4.
If a certain type of constraint exists but no corresponding DSCREEN entry is specified, all
the screening criteria used for this type of constraint will be furnished by the default
values.
5.
Constraints can be retained only if they are greater than TRS. See the Remarks under
the DCONSTR entry for a definition of constraint value.
6.
9-19
9-20
FORMAL APPROXIMATIONS
These form a key ingredient in the approximate model in that it is
necessary to obtain high-quality and explicit approximations of the
implicit structural responses computed by the finite element
analyses
High-quality Maintain reasonable accuracy in predicting
responses for large variations of the independent design variables
Reasonable Direct the design toward a practical optimum design
Large At least 10 to 20% changes in the design variables
9-21
x
f ( x + x ) = f ( x ) + f ( x )x + f ( x ) ---------- +
2!
First-order, approximations, directly in terms of the design
variables:
n
f(x) f(x ) +
(x x )
------i
x i x o i
i=1
9-22
f(x) f(x ) +
- ( Y ( x ) Y i ( x i ) )
------Y i x o i i
o
i=1
1
Y i = ---xi
now, let
Y i
1
-------- = -----2
x i
xi
then
and
fR ( x ) = f ( x )
o 2 1
f
1
------(
x
)
---
-----
x x o i x
o
i
i x
i=1
i
= f(x )
i=1
f x i o
------------ ( x x i )
x x o x i
9-23
Stress = P/A
DISPL = PL/AE
Both are proportional to 1/A
9-24
1/A
1/A
9-25
j = 1,m
xi xi xi
i = 1,n
o
Minimize: F ( x ) = F ( x ) +
- ( Y i ( x i ) Y i ( x i ) )
------Y i x o
o
i=1
subject to:
n
g j ( x ) = g j ( x ) +
i=1
g j
o
-------- Y i ( x i ) Y i ( x i )
Y i x o
j = 1,m
Move limits:
l
xi xi xi
where Yi
Yi
9-26
i = 1,n
Answer:
g ( X ) = g ( X o ) +
------x i
i=1
( x i x io )
x = xo
9-28
use g D ( X )
if g D ( X ) g R ( X ) < 0,
use g R ( X )
Thus:
N
g
g(X ) +
------( x i x io )
o
x i
i=1
x = xo
g ( X ) =
N
2 1
g
1
g(X )
x io ---- --------
------o
x x
x
i
io
i=1 i x=x
if
g
x i ------x i
if
g
x i ------x i
9-29
x = xo
< 0
x = xo
SENSITIVITY ANALYSIS
Formal Approximations are based on the assumption that gradient
information is available.
Sensitivity Analysis yields this
information.
Static Displacement Sensitivity Analysis
KU = P
Differentiate with respect to a single design variable:
U
K
P
------- U + K ------- = ------x i
x i
x i
U
Solve for ------x i
U
P K
K ------- = ------- ------- U
x i
x i x i
The equation above is exact. (The right-hand side is often referred
to as a pseudo-load vector.)
This first-order displacement sensitivity information allows the approximations of all other static response gradients to be made:
dr j r j
r j ( X + x j , U + u ) r j ( X, U )
-------- -------- = --------------------------------------------------------------------------dx i x i
x i
U
where U = ------- x i
x i
9-30
e K ( x + x ) K ( x )
K
i
---------- -----------------------------------------------------------x i
x i
e K ( x + x ) K ( x x )
K
i
i
---------- --------------------------------------------------------------------------x i
2x i
9-31
------- = --------------------------------------------T
x i
j M j
Buckling sensitivity
K k + k K d k = 0
Differentiating with respect to a design variable xi:
K d
T K
k ------- + k ---------- k
x i
k
x i
--------- = -----------------------------------------------------T
x i
( k Kd k )
9-32
e
k ( p j + DELB P j ) k ( p j ) p j
k
--------- ------------------------------------------------------------------------------ --------
o
x i
x i x o
DELB p j
9-33
------- -------------- =
p j x i
x i
o
( u + u, p j + DELB p j ) ( u, p j ) p j
--------------------------------------------------------------------------------------------------- --------
o
x i x o
DELB p j
9-34
K
------- U =
x i
M
------- U =
x i
m=1
m=1
k m
e
----------- U
x i
m m
e
------------- U
x i
9-35
9-36
ADJOINT SENSITIVITY
Motivation
Consider a modal frequency analysis with
Four subcases
Two hundred frequencies
100,000 DOF
Solution vectors require 1.28 gigabytes
Now do a sensitivity analysis with 100 design variables
Disk space requirement goes up by at least 100
An alternative algorithm is available if:
Only grid responses are active
Design model is in the residual
nresp < ndv (nsub nfreq)
9-37
[ M + iB + K ] { u } = { P }
Grid response (r) is extracted from the global displacement vector (u)
r = f(u)
Sensitivity of the response with respect to design variables is
T
dr dx = f u u x
In the direct method, u x is obtained from:
2
2
[ M + iB + K ] { du dx } = [ dM dx + idB dx + dK dx ] { u }
The adjoint method solves for a solution vector of the form:
T
2
[ M + iB + K ] { } = { f u }
Defining
2
[ FAC ] = [ M + iB + K ]
9-38
[ FAC ] = [ I ]
9-39
ADJOINT SENSITIVITY
WHAT IS SUPPORTED
Statics
Static Aeroelasticity
Frequency Response
Transient Response
Gravity Loads
Grid Responses
DISP
FRDISP
FRVELO FRACCL
p-elements
Element Responses
STRESS
STRAIN FORCE
CSTRESS CSTRAIN CFAILURE
FRSTRE
FRFORC FRSPCF
9-40
Model
Design Task
Minimize sensor jitter
nfreq
u si
i=1
Constraints
Weight 8.5
f 1 and f 2 100 Hz
9-41
2335
Number of elements:
2129
Number of subcases:
101
13
Number of responses:
91
Performance Results
Parameter
V70
V69.1
CPU Time
1913.4 secs
4709.9
Scratch Space
90.5 MB
608.4 MB
70.4 MB
356.6 MB
Final Objective
1.559
1.3862
9-42
9-43
9-44
2615
Number of elements:
2647
Number of subcases:
1075
Number of responses:
Performance Results
Parameter
V70
V69.1
CPU Time
48.1 secs
410.2
Scratch Space
24.4 MB
554.0 MB
7.7 MB
259.1 MB
9-45
14990
Number of elements:
16174
Number of subcases:
Number of frequencies:
Number of modes:
67
2514
Number of responses:
Performance Results
Parameter
V70
V69.1
CPU Time
3504.2 secs
6787.7 secs
Scratch Space
541.0MB
5238. MB
487.4MB
1953. MB
Final Objective
0.6379
0.6386
-0.4218
-.4151
9-46
Response
Design Variable
r j p k
---------- ---------p k x i
Sensitivity
Analysis
DVPREL1,
DVPREL2
where pk = properties
9-47
.
.
.
.
.
.
x2
.
.
.
.
.
.
.
.
.
r j
----------
x n k
xn
9-48
.
.
.
r j
----------
x 1 1
rj
Subcase 2
.
.
.
r 1j
-----------
x 1 1
r1
Subcase 2
.
.
.
x1
rj
Subcase 1
.
.
.
r1
Subcase 1
DSAPRT
Format:
ALL
FORMATTED
NOEXPORT
DSAPRT ( UNFORMATTED ,
, [ START = i ], [ BY = j ], [ END = k ]) = n
EXPORT
NOPRINT
NONE
Examples:
DSAPRT(FORMATTED,EXPORT)
DSAPRT(FORMATTED,START = FIRST,BY = 3,END = LAST) = 101
DSAPRT(UNFORMATTED,START = FIRST)
DSAPRT(UNFORMATTED,EXPORT)
DSAPRT(FORMATTED,END = 4) = ALL
DSAPRT(UNFORMATTED,END = SENS) = ALL
DSAPRT(NOPRINT,EXPORT)
Describers
Meaning
FORMATTED
UNFORMATTED Output will be printed as a matrix print (see description of the MATPRN module
in the MSC/NASTRAN DMAP Module Dictionary).
NOPRINT
EXPORT
NOEXPORT
START = i
Specifies the first design cycle for output. (Integer > 0 or Character: FIRST or
LAST; Default = 1 or FIRST)
BY = j
Specifies the design cycle interval for output. (Integer 1, or > 0; Default = 0)
See Remark 2.
END = k
Specifies the last design cycle for output. (Integer > 0 or Character: FIRST,
LAST, or SENS; Default = LAST)
ALL
All design responses (defined in DRESP1 and DRESP2 entries) will be output.
9-49
DSAPRT
Remarks:
1.
Only one DSAPRT may appear in the Case Control Section and should appear above all
SUBCASE commands.
2.
Sensitivity data will be output at design cycles i, i + j, i + 2j, ..., k. Note that the BY = 0
default implies no sensitivity analysis at the intermediate design cycles.
3.
END = SENS requests design sensitivity analysis, and no optimization will be performed.
4.
DSAPRT(NOPRINT,EXPORT,END = SENS)
DSAPRT(UNFORMATTED,START = LAST)
9-50
9-51
----------------------------------------------------------------------------------------------------------------------DRESP1 ID=
20
RESPONSE TYPE= WEIGHT
SEID=
0
RESP VALUE
DESIGN VARIABLE COEFFICIENT
DESIGN VARIABLE COEFFICIENT
----------------------------------------------------------------------------------------------------------------------4.8284e+00
1 A1
2.8286e+00
2 A2
1.0006e+00
SYMMETRIC THREE BAR TRUSS DESIGN OPTIMIZATION D200X1
JANUARY 18, 1996 MSC/NASTRAN
1/17/96
PAGE
24
BASELINE - 2 CROSS SECTIONAL AREAS AS DESIGN VARIABLES
----------------------------------------------------------------------------------------------------------------------DRESP1 ID=
23
RESPONSE TYPE= STRESS
ELEM ID=
1
COMP NO=
2 SEID=
0
SUBCASE RESP VALUE
DESIGN VARIABLE COEFFICIENT
DESIGN VARIABLE COEFFICIENT
----------------------------------------------------------------------------------------------------------------------1
1.3530E+04
1 A1
-1.1895E+04
2 A2
-8.1543E+02
2 -9.0973E+03
1 A1
1.0742E+04
2 A2
-8.2031E+02
----------------------------------------------------------------------------------------------------------------------DRESP1 ID=
25
RESPONSE TYPE= STRESS
ELEM ID=
3
COMP NO=
2 SEID=
0
SUBCASE RESP VALUE
DESIGN VARIABLE COEFFICIENT
DESIGN VARIABLE COEFFICIENT
----------------------------------------------------------------------------------------------------------------------1 -9.0973E+03
1 A1
1.0742E+04
2 A2
-8.2031E+02
2
1.3530E+04
1 A1
-1.1895E+04
2 A2
-8.1543E+02
MOVE LIMITS
Move limits are directly applied to the design variables as side
constraints
Limits are also imposed as upper and lower bounds on each
analysis property related to the design model. These bounds are
applied to the optimization of the current approximate model and
are treated as constraints by the optimizer, not side constraints.
Allowable move limits are problem dependent, and there is no
general a priori method of estimation
true function
F(X)
linear approximation
X = current design
allowable
move limits
For sizing problems, use relatively large move limits (such as 0.5
1.0 initially), reducing them if convergence becomes a problem.
Smaller move limits are recommended for dynamics problems
since such responses are often highly nonlinear.
9-52
pj pj ( x ) pj
where
p j p j DELP
u =
p j + p j DELP
pj
pj
pj
pj
9-53
pj
l
l
o
p j = max min [ p j ( DELP ), ( p j DPMIN ) ], PMIN
u
u
o
p j = min max [ p j ( DELP ), ( p j + DPMIN ) ], PMAX
pj
pj
pjo
pj
pj
Note that PMIN and PMAX from the DVPREL1 and 2 entries have
been accounted for as well.
Because property limits are treated as constraints, it is possible
they may become violated during the design process if this helps
satisfy a more seriously violated constraint. In the extreme, a
physically meaningless design may result.
9-54
xi
DELTA x i x i
xi
xi
DELTA
if x i
o
xi
xi
+ DXMIN
9-55
9-56
********************************************************
*
*
*
USER WARNING MESSAGE
*
*
*
* IF YOU WANT TO CONTINUE THE DESIGN AFTER THIS JOB
*
* IS COMPLETED, YOU MUST INCLUDE A REVISED DOPTPRM
*
* BULK DATA ENTRY IN THE BULK DATA SECTION WITH THE
*
* FOLLOWING ITEMS MODIFIED AS SHOWN:
*
*
*
*
DELP =
1.0000E-01
*
*
DPMIN =
5.0000E-03
*
*
DELX =
5.0000E-01
*
*
DXMIN =
2.5000E-02
*
*
*
********************************************************
9-57
********************************************************
*
*
* IF A DELXV IS SPECIFIED ON A DESVAR BULK DATA ENTRY, *
* UPDATED DELXV VALUES ARE PRESENT ON THE DESVAR
*
* ENTRIES CONTAINED IN THE PUNCH FILE.
*
*
*
* (NOTE: THERE MAY BE MORE THAN ONE MESSAGE LIKE
*
*
THIS. THE LAST ONE IN THIS RUN SHOULD
*
*
BE LOCATED AND USED.)
*
*
*
********************************************************
CONVERGENCE NOT ACHIEVED YET
(HARD CONVERGENCE DECISION LOGIC)
---------------------------------------------------------------------------------------------2.8778E-02 MUST BE LESS THAN
1.0000E-03
1.8917E-01 MUST BE LESS THAN
1.0000E-20
--- AND --MAXIMUM CONSTRAINT VALUE
: 3.3610E-02 MUST BE LESS THAN
5.0000E-03
(CONVERGENCE TO A FEASIBLE DESIGN)
--- OR --MAXIMUM OF RELATIVE PROP. CHANGES
: 0.0000E+00 MUST BE LESS THAN
1.0000E-03
AND
MAXIMUM OF RELATIVE D.V. CHANGES
: 4.0000E-01 MUST BE LESS THAN
1.0000E-03
(CONVERGENCE TO A BEST COMPROMISE INFEASIBLE DESIGN)
---------------------------------------------------------------------------------------------OR
:
:
PARAM OPTEXIT
The OPTEXIT parameter provides a means of exiting a design task in
SOL 200.
OPTEXIT
Value
Description
Exit after design sensitivity analysis and write the data blocks
related to sensitivity coefficients (DSCM2 and DSCMCOL) to
an external file using the OUTPUT2 and OUTPUT4 modules.
This is equivalent to the DSAPRT (NOPRINT,EXPORT
END+SENS) Case Control command. See related
parameters ITAPE, IUNIT, and OMAXR.
Exit after the first update of the analysis model based on the
first approximate optimization of the design model.
9-58
NO
IFPL
Does a Design
Model Exist?
YES
PREDOM
EXITOPT
9-59
PHASE0
DESINIT,
SETSOLAP
NO
If AMBS
AMBS
Auxiliary Model
Boundary Shapes
YES
Static Analysis for All Auxiliary Models,
Loop over Auxiliary Models, Boundary
Conditions (FEAOPT = Analysis)
NO
If AMBS
or GMBS
FEA
GMBS
Geometry Model
Boundary Shapes
YES
BNDSHP
NO
FEA
If AMBS
or GMBS
YES
NO
EXIT
YES
Begin Design Sensitivity
and Optimization
Pre-Sensitivity Initialization Operations.
Shape Basis Vector Initialization
and Scaling Operations.
DOM11, DOPR2, DOPR3, DOPR4, DOPR5
9-60
Maximum
Design Cycles
Completed?
YES
EXIT
EXITOPT
NO
3
EXITOPT
If Hard
Convergence
YES
FEA
*OPTEXIT=7 includes computation of sensitivities for the current design cycle (all operations up to DOM9).
9-61
If Soft
Convergence
NO
9-62
UPDATE
IFP
PREDOM
FEA
DESCON
Initialization
DOM12
PSLGDV
FEA
RESPSEN
DOM9
DOM10
DOM12
DOM11
Note:
Print Output
2
Soft Convergence
(Check If the Optimizer Made Progress)
9-63
Internal
Variable
Criterion
(P)
CHGOBJ
ACHOBJ
OBJ
CHGPRP
CHGDV
CONMAX
User
Parameter
0.001
CONV1
1.E-20
CONV2
0.001
CONVPR
0.001
CONVDV
0.005
GMAX
(P 1)
OBJ
OBJ
------------------------------------------------------(P 1)
OBJ
(P)
Default
Limit
OBJ
(P 1)
P (P) P (P 1)
i
i
max
----------------------------------------
(P 1)
1 i NPROP
Pi
x (P) x (P 1)
i
i
max
---------------------------------------
(P 1)
1 i NDV
xi
max { g ( x ) }
k
k
9-64
CONVERGENCE CRITERIA
Soft Convergence Decision Logic
Compares the results of the approximate optimization with the
previous finite element analysis
CHGOBJ CONV1
or
ACHOBJ CONV2
No
YES
CHGPRP CONVPR
No
YES
CONMAX GMAX
or
CHGDV CONVDV
No
YES
SOFTCV = TRUE
SOFTCV = FALSE
9-65
No
Yes
Yes
Yes
Convergence to a
Unique Design
No
CHGPRP<CONVPR
and
CHGDV<CONVDV
Best Compromise,
Infeasible Design
Found
Nonunique If
CHGPRP > CONVPR or
CHGDV > CONVDV
9-66
No
Continue with
Optimization
DESPCH
DSNOKD
NASPRT
NO
No output.
N=0
N>0
0.0
No output.
9-67
SOFTEXIT
UPDTBSH
N>0
YES
if
soft
YES
Note:
Regardless
of
the
value
of
UPDTBSH, shape basis vectors are
still updated for every design cycle
(interpolation to the interior grids).
UPDTBSH only controls updates on
the boundary shapes.
9-68
SECTION 10
SHAPE OPTIMIZATION
BASIC EQUATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1
BASIS VECTORS EXAMPLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-2
SHAPE BASIS VECTORS IN THE DESIGN MODEL. . . . . . . . . . . . . . . . . . . . . . . . 10-6
DESIGN MODELING INPUT FOR SHAPE OPTIMIZATION . . . . . . . . . . . . . . . . . . 10-7
MANUAL GRID VARIATION. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-8
DVGRID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-10
AUXILIARY MODELS IN SHAPE OPTIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . 10-12
EXAMPLE DIRECT INPUT OF SHAPES. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-14
DVSHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-19
AUXILIARY BOUNDARY MODELS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-24
EXAMPLE ANALYTIC BOUNDARY SHAPES . . . . . . . . . . . . . . . . . . . . . . . . . . 10-26
AUXCASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-30
AUXMODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-31
BNDGRID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-32
DVBSHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-33
SUMMARY OF SHAPE BASIS VECTOR GENERATION METHODS . . . . . . . . . . 10-41
GUIDELINES, RECOMMENDATIONS AND LIMITATIONS. . . . . . . . . . . . . . . . . . 10-43
BASIC EQUATION
10
Basis
Vectors
[T]
Design Variable
Changes
{x}
where {G} = { G }
i + 1 { G }i
{x}
= {x}
i + 1 { x }i
= current design
i+1
= updated design
h = 1.0
y
2
1
x
1 2 h
h
h
{ G } = { G } { G } = x { T }
T
{ T } = 0. 1. 0. 0. 1. 0. 0. 0.5 0. 0. 0.5 0. 0. 0. 0. 0. 0. 0.
10-2
5
R
3
R = -1.0
1 1
{ G } = { G } { G } = x { T }
{ T } = 1. 0. 0.
GRID 1
0. 0. 0. 1. 0. 0. 0. 0. 0. 1. 0. 0.
GRID 2
GRID 3
10-3
0. 0. 0.
Tx = 2.0E4 N/m
0.15m
0.15m
Material:
aluminum, 7075T6 sheet
2
E = 7.2E10 N/m
= 0.33
= 2.8E3 kg/m
10-4
x 2
x-direction
variation
y-direction
variation
10-5
a
b
10-6
10-7
{G}i
y
x
G ix
N1
{ G } i = G iy = COEFF N2 x DVID
G iz
N3
DVGRID
DVID
GRID
CID
COEFF
N1
N2
N3
10-8
10
10-9
DVGRID
DVGRID the relationship between design variables and grid point locations.
Design Variable to Grid Point Relation
Defines
Format:
1
DVGRID
DVID
GID
CID
COEFF
N1
N2
N3
108
0.2
0.5
0.3
1.0
10
Example:
DVGRID
Field
Contents
DVID
GID
Grid point (GRID) or geometric point (POINT) identification number. (Integer > 0)
CID
COEFF
Ni
Remarks:
1.
A CID of zero or blank (the default) references the basic coordinate system.
2.
Multiple references to the same grid ID and design variable result in vectorial addition of
the participation vectors defined by CID, COEFF, and Ni. There is no restriction on the
number of DVGRID entries that may reference a given grid (GID) or design variable
(DVID).
3.
{ g }i { g }i =
The vector { N } = [ N x N y N z ] is determined from CID and Ni. Note that it is a change
0
in a design variable from its initial value X , and not the absolute value of the design
0
10-10
DVGRID
The DVGRID entry defines the participation coefficients (basis vectors) of each design
variable for each of the coordinates affected by the design process in the relationship
{ g }i =
{ T }ij X j
j
5.
DVGRID entries that reference grid points on MPCs or RSSCON entries produce incorrect
sensitivities. Often the sensitivities are 0.0 which may result in a warning message indicating zero gradients which may be followed by UFM 6499. Other rigid elements produce
correct results.
10-11
10-12
{ G } = [ T ] { x }
10-13
From the above stress distribution, we can observe that the culvert
interior profile is not optimal.
10-14
10-15
10-16
10-17
10-18
DVSHAP
DVSHAP
Variable
to Basis
Vector(s)
Defines a shape basis vector by relating a design variable identificationDesign
number
(DVID)
to columns
of a displacement matrix.
Format:
1
DVSHAP
DVID
COL1
SF1
COL2
SF2
COL3
SF3
2.0
1.0
10
Example:
DVSHAP
Field
Contents
DVID
COLi
SFi
Scaling factor applied to the COLi-th column of the displacement matrix. (Real;
Default = 1.0)
Remarks:
1.
2.
3.
Multiple references to the same DVID and/or COLi will result in a linear combination of
displacement vectors. In the example above, the shape basis vector is a linear combination of the fourth column and twice the second column.
4.
The displacement matrix must have been created by MSC/NASTRAN and be available on
a database, which is attached via the DBLOCATE FMS statement shown below:
ASSIGN
DBLOCATE
10-19
10-20
3.73200,.00
3.66176,.00
3.46643,.00
3.14600,.00
3.14600,.00
3.14600,.00
3.14600,.00
4.46400,.00
4.42888,.00
4.33122,.00
4.17100,.00
4.17100,.00
4.17100,.00
4.17100,.00
5.19600,.00
5.19600,.00
5.19600,.00
5.19600,.00
5.19600,.00
5.19600,.00
5.19600,.00
1,
2,
2,
3,
3,
4,
5,
6,
6,
7,
7,
8,
9,
10,
10,
11,
11,
12,
13,
14,
14,
15,
15,
16,
20,
21,
21,
22,
22,
23,
27,
28,
28,
29,
29,
30,
16,
17,
17,
18,
18,
19,
23,
24,
24,
25,
25,
26,
30,
31,
31,
32,
32,
33,
0
0
6,
7,
8,
10,
11,
12,
17,
18,
19,
21,
22,
23,
28,
29,
30,
35,
36,
37,
24,
25,
26,
31,
32,
33,
38,
39,
40,
1250.
2500.
5
6
7
9
10
11
16
17
18
20
21
22
27
28
29
34
35
36
23
24
25
30
31
32
37
38
39
10-21
-1.
-1.
10-22
10-23
Bar Elements
Primary Model
10-24
Constant
Linear
Quadratic
P
Cubic
10-25
154
165
143
132
121
110
Z
Y
X
10-26
Z
Y
X
10-27
Z
Y
X
10-28
10-29
AUXCASE
Format:
AUXCASE
Examples:
AUXCAS
AUXC
Remarks:
1.
AUXCASE indicates the beginning of Case Control commands for an auxiliary model.
AUXCASE must follow the primary model Case Control commands.
2.
All Case Control commands following this entry are applicable until the next AUXCASE or
BEGIN BULK command. Commands from preceding Case Control Sections are ignored.
3.
Each auxiliary model Case Control must be delimited with the AUXCASE command.
4.
The AUXMODEL command is used to associate the auxiliary model Case Control with a
particular auxiliary model.
10-30
AUXMODEL
Format:
AUXMODEL = n
Examples:
AUXMODEL = 4
AUXM = 4
Describer
Meaning
Remarks:
1.
AUXMODEL references a particular auxiliary model for analysis and may only be specified
in the auxiliary model Case Control Section.
2.
See the BEGIN BULK command for the Bulk Data definition of an auxiliary model.
10-31
BNDGRID
Format:
1
BNDGRID
GP1
GP2
GP3
GP4
GP5
GP6
GP7
GP8
-etc.-
123
41
42
43
44
45
46
47
10
Example:
BNDGRID
49
GP1
"THRU"
GP2
BNDGRID
123
41
THRU
49
Field
Contents
GPi
Shape boundary grid point identification number. (0 < Integer < 1000000; For
THRU option, GP1< GP2)
Remarks:
1.
C specifies the components for the listed grid points for which boundary motion is
prescribed.
2.
Multiple BNDGRID entries may be used to specify the shape boundary grid point identification numbers.
3.
Both fixed and free shape boundary grid point identification numbers are listed on this
entry.
4.
The degrees of freedom specified on BNDGRID entries must be sufficient to statically constrain the model.
5.
Degrees of freedom specified on this entry form members of the mutually exclusive s-set.
They may not be specified on other entries that define mutually exclusive sets. See the
MSC/NASTRAN Quick Reference Guide, Appendix B for a list of these entries.
10-32
DVBSHAP
DVBSHAP
Design Variable
Boundary Shapes
Associates a design variable identification number to a linear combination
of to
boundary
shape
vectors from a particular auxiliary model.
Format:
1
DVBSHAP
DVID
AUXMOD
COL1
SF1
COL2
SF2
COL3
SF3
1.6
10
Example:
DVBSHAP
Field
Contents
DVID
AUXMOD
COLi
SFi
Scaling factor for load sequence identification number. (Real; Default = 1.0)
Remarks:
1.
2.
Multiple references to the same DVID and/or COLi will result in the vector addition of the
referenced boundary shape vectors.
3.
10-33
10-34
2
2
2
1
1
1
10-35
80
1
1
1
10-36
C2,
GP7,
139
147
155
163
7
15
23
31
40
62
73
95
106
128
89
10-37
963
964
965
966
967
968
969
2
2
2
2
2
2
2
147
148
149
150
151
152
153
158
159
160
161
162
163
164
159
160
161
162
163
164
165
11
2
200
2.1E+5
11
123456
0.8E+5
0.20
1
0.3
11
12
0.00
200
200
220
220
200
12
123456
11
33
3
300
300
300
330
330
300
11
34
3
3
11
123456
12
143
123456 1
143
3
165
3
3
143
0.0
33
165
22
1.0
33
144
154
THRU
-1.0
-1.0
154
0.0
0.0
0.0
0.0
0.0
0.0
0.0
23
22
THRU
1.0
1.0
22
133
148
149
150
151
152
153
154
155
165
132
154
165
10-38
-1.0
10-39
***************************************************************
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
6
NUMBER OF OPTIMIZATIONS W.R.T. APPROXIMATE MODELS
5
OBJECTIVE AND MAXIMUM CONSTRAINT HISTORY
--------------------------------------------------------------------------------------------------------------OBJECTIVE FROM
OBJECTIVE FROM
FRACTIONAL ERROR
MAXIMUM VALUE
CYCLE
APPROXIMATE
EXACT
OF
OF
NUMBER
OPTIMIZATION
ANALYSIS
APPROXIMATION
CONSTRAINT
--------------------------------------------------------------------------------------------------------------INITIAL
8.000000E+06
-3.827773E-01
1
7.401214E+06
7.401214E+06
0.000000E+00
-3.720914E-01
2
6.562918E+06
6.562912E+06
9.142283E-07
-3.570341E-01
3
5.389287E+06
5.389288E+06
-2.783299E-07
-4.672768E-02
4
5.350312E+06
5.350312E+06
0.000000E+00
8.081818E-04
5
5.350312E+06
5.350312E+06
0.000000E+00
8.081055E-04
---------------------------------------------------------------------------------------------------------------------AUXILIARY MODEL 1
MARCH 16, 1994 MSC/NASTRAN
3/15/94
PAGE
226
DESIGN VARIABLE HISTORY
------------------------------------------------------------------------------------------------------------------------INTERNAL | EXTERNAL
|
|
DV. ID. |
DV. ID.
| LABEL
| INITIAL
:
1
:
2
:
3
:
4
:
5 :
------------------------------------------------------------------------------------------------------------------------1 |
1
| UPPER
| 1.0000E+00 : 1.4000E+00 :
1.9600E+00 : 2.7440E+00 : 2.7700E+00 : 2.7700E+00 :
2 |
2
| LOWER
| 1.0000E+00 : 1.4000E+00 :
1.9600E+00 : 2.7440E+00 : 2.7700E+00 : 2.7700E+00 :
------------------------------------------------------------------------------------------------------------------------*** USER INFORMATION MESSAGE 6464 (DOM12E)
RUN TERMINATED DUE TO HARD CONVERGENCE TO AN OPTIMUM AT CYCLE NUMBER =
5.
Z
Y
X
Z
Y
X
10-40
10-41
within
the
MSC/NASTRAN
text
10-42
input
10-43
10-44
SECTION 11
DYNAMIC RESPONSE OPTIMIZATION
DYNAMIC RESPONSE OPTIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1
GENERAL CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
BASIC EQUATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
EXAMPLE MODAL FREQUENCY RESPONSE OPTIMIZATION . . . . . . . . . . . . . 11-6
EXAMPLE ACOUSTIC OPTIMIZATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-18
11-1
11
GENERAL CONSIDERATIONS
Design responses are computed just for the output frequencies and
output time steps:
For frequency response:
OFREQ
For transient response:
OTIME
NOi
TSTEP
11-2
BASIC EQUATIONS
Direct frequency response
Equation of motion
2
[ M + iB + K ]u = F
Displacement sensitivities, let
= ------x i
2
[ M + iB + K ]u = F [ M + iB + K ]u
Modal frequency response
Modal transformation,
u =
Differentiate
------x i
and apply
[ u = + = ]
which avoids calculation of eigenvector sensitivities
11-3
[ M + i T B + T K ]
T
= F [ M + iB + K ]U
Modal transient response
Equation of motion
Mu + Bu + Ku = F
with modal transformation u = differentiating, and using
T
+ [ B ] + [ K ]
[ M ]
T
= F ( Mu + Bu + Ku )
Limitations in Dynamic Response Sensitivity
F assumed zero in direct and modal frequency and modal
transient
This assumption is usually good, except for those situations in
which the following may be significant:
Gravity (or other mass-related) loads
11-4
K = K 1 + igK 1 + i
ge ke + K2
e
Limitations:
g, ge, and K2 must be constant (no dependence on design variables).
Similarly, M2 and B2 are constant.
Accurate sensitivities for the modal solutions typically require
retaining more modes (e.g., twice as many) as are required for
accurate modal analysis
11-5
Grid 1110
X
(Uniform Pressure)
11-6
Grid 1110
t1 = x1
Free
Edge
x
Ten independent design variables are used to describe the ten
bands of plate thicknesses.
The mean square displacement minimization will be subject to a
weight budget (minimize the displacement without appreciably
changing the weight):
7.99 vol 8.01
A mean square displacement can be written as:
100
min =
2
i
( u z, 1110 ) + 2
i = 20
200
i = 51
11-7
2i
( u z, 1110 )
11-8
11-9
105,
400,
+
+
1106,
A GROUP
.08,
150
302,
303,
403,
402
*(100), *(100), *(100), *(100)
303,
*(1),
304,
*(1),
.08,
150
404,
*(1),
403
*(1),
403,
404,
504,
503
*(100), *(100), *(100), *(100)
404,
*(1),
405,
*(1),
.08,
150
505,
*(1),
504
*(1),
504,
505,
605,
604
*(100), *(100), *(100), *(100)
505,
*(1),
506,
*(1),
.08,
150
606,
*(1),
605
*(1),
605,
606,
706,
705
*(100), *(100), *(100), *(100)
606,
*(1),
607,
*(1),
.08,
150
707,
*(1),
706
*(1),
706,
707,
807,
806
*(100), *(100), *(100), *(100)
707,
*(1),
708,
*(1),
.08,
150
808,
*(1),
807
*(1),
807,
808,
908,
907
*(100), *(100), *(100), *(100)
808,
*(1),
.08,
809,
*(1),
150
909,
*(1),
908
*(1),
908,
909,
1009,
1008
*(100), *(100), *(100), *(100)
909,
910,
1010,
1009
11-10
11-11
$
DEQATN
201
20
*(1)
VOLUME
g1110L
=
VOLUME
FRDISP
=
3
=
20.0
*(1.0)
1110
=
102
*(2)
G1110H
=
FRDISP
=
3
=
102.0
*(2.0)
1110
=
1
DRESP1
UZ2
20
27
34
41
48
55
62
69
76
83
90
97
108
122
136
150
164
178
192
1
21
28
35
42
49
56
63
70
77
84
91
98
110
124
138
152
166
180
194
22
29
36
43
50
57
64
71
78
85
92
99
112
126
140
154
168
182
196
23
30
37
44
51
58
65
72
79
86
93
100
114
128
142
156
170
184
198
24
31
38
45
52
59
66
73
80
87
94
102
116
130
144
158
172
186
200
25
32
39
46
53
60
67
74
81
88
95
104
118
132
146
160
174
188
26
33
40
47
54
61
68
75
82
89
96
106
120
134
148
162
176
190
UZ2(U20,u21,u22,u23,u24,u25,U26,U27,U28,U29,U30,
U31,U32,U33,U34,U35,U36,U37,U38,U39,U40,
U41,U42,U43,U44,U45,U46,U47,U48,U49,U50,
U51,U52,U53,U54,U55,U56,U57,U58,U59,U60,
U61,U62,U63,U64,U65,U66,U67,U68,U69,U70,
U71,U72,U73,U74,U75,U76,U77,U78,U79,U80,
U81,U82,U83,U84,U85,U86,U87,U88,U89,U90,
U91,U92,U93,U94,U95,U96,U97,U98,U99,U100,
U102,U104,U106,U108,U110,U112,U114,U116,U118,U120,
U122,U124,U126,U128,U130,U132,U134,U136,U138,U140,
U142,U144,U146,U148,U150,U152,U154,U156,U158,U160,
U162,U164,U166,U168,U170,U172,U174,U176,U178,U180,
U182,U184,U186,U188,U190,U192,U194,U196,U198,U200)
= u20**2 + u21**2 + U22**2 + U23**2+ U24**2 +
U25**2 + U26**2 + U27**2 + U28**2 + U29**2 +U30**2 +
U31**2 + U32**2 + U33**2 + U34**2 + U35**2 +
U36**2 + U37**2 + U38**2 + U39**2 + U40**2 +
U41**2 + U42**2 + U43**2 + U44**2 + U45**2 +
U46**2 + U47**2 + U48**2 + U49**2 + U50**2 +
U51**2 + U52**2 + U53**2 + U54**2 + U55**2 +
u56**2 + u57**2 + u58**2 + u59**2 + u60**2 +
U61**2 + U62**2 + U63**2 + U64**2 + U65**2 +
U66**2 + U67**2 + U68**2 + U69**2 + U70**2 +
11-12
11-13
8
7
11-14
2.077626E+02
1.858892E+02
1.176688E-01
-1.247635E-03
1.743351E+02
1.609400E+02
8.323036E-02
-1.242462E-03
1.552931E+02
1.549932E+02
1.935096E-03
-1.248490E-03
1.483332E+02
1.423191E+02
4.225774E-02
-1.247516E-03
1.352799E+02
1.396561E+02
-3.133548E-02
-1.240314E-03
1.376002E+02
1.359294E+02
1.229173E-02
-1.239836E-03
7
1.338581E+02
1.346727E+02
-6.048443E-03
-1.231839E-03
---------------------------------------------------------------------------------------------------------------
11-15
---------------------------------------------------------------------------------------------------------------------------------INTERNAL |
EXTERNAL
|
|
DV. ID. |
DV. ID.
|
LABEL
|
6
:
7
:
8
:
9
:
10
:
11
:
---------------------------------------------------------------------------------------------------------------------------------1 |
1
| T1
|
9.2234E-02 :
1.0245E-01 :
2 |
2
| T2
|
8.9215E-02 :
7.8416E-02 :
3 |
3
| T3
|
6.6226E-02 :
6.4560E-02 :
4 |
4
| T4
|
3.6059E-02 :
3.2658E-02 :
5 |
5
| T5
|
5.5572E-02 :
5.8001E-02 :
6 |
6
| T6
|
7.6552E-02 :
7.8469E-02 :
7 |
7
| T7
|
9.6333E-02 :
9.7656E-02 :
8 |
8
| T8
|
1.3458E-01 :
1.3406E-01 :
9 |
9
| T9
|
1.5716E-01 :
1.5932E-01 :
10 |
10
| T10
|
1.4968E-01 :
1.5128E-01 :
*** USER INFORMATION MESSAGE 6464 (DOM12E)
RUN TERMINATED DUE TO HARD CONVERGENCE TO AN OPTIMUM AT CYCLE NUMBER =
7.
3.622 @ 58 Hz
2.929 @ 52 Hz
11-16
Thickness
11-17
10
11-18
c1
c2
c3
f1
f2
f3
minimize x
subject to P ( f1 ) x 0
P ( f2 ) x 0
P ( f3 ) x 0
One such constraint is written for every response of interest.
can only be minimized if the peak pressures are lowered.
11-19
11-20
11-21
11-22
11-23
10127
10128
10010
10019
10138
.
.
.
12530
.
.
.
12531
.
.
.
12412
.
.
.
12422
.
.
.
12540
12531
12532
12413
12423
12541
11-24
16
15
11-25
11-26
---------------------------------------------------------------------------------------------------------------------------------INTERNAL |
EXTERNAL
|
|
DV. ID. |
DV. ID.
|
LABEL
|
6
:
7
:
8
:
9
:
10
:
11
:
---------------------------------------------------------------------------------------------------------------------------------1 |
1
| P1
|
3.4322E-02 :
2.9154E-02 :
2.4363E-02 :
2.7226E-02 :
2.4670E-02 :
2.4619E-02 :
2 |
2
| P2
|
3.1484E-02 :
2.6989E-02 :
3.1301E-02 :
2.5696E-02 :
2.3196E-02 :
2.3115E-02 :
3 |
4
| P4
|
5.6109E-03 :
6.5327E-03 :
5.8686E-03 :
5.4315E-03 :
4.3781E-03 :
3.8628E-03 :
4 |
5
| P5
|
2.4430E-02 :
3.1121E-02 :
3.3386E-02 :
3.8369E-02 :
4.0770E-02 :
4.0748E-02 :
5 |
6
| P6
|
1.8545E-02 :
1.7878E-02 :
1.7679E-02 :
1.4659E-02 :
1.6339E-02 :
1.6324E-02 :
6 |
7
| P7
|
1.3480E-02 :
1.9997E-02 :
2.1885E-02 :
2.5364E-02 :
2.7816E-02 :
2.7732E-02 :
7 |
8
| BETA
|
2.0764E-01 :
2.1034E-01 :
2.1049E-01 :
2.1332E-01 :
2.4781E-01 :
1.8586E-01 :
---------------------------------------------------------------------------------------------------------------------------------INTERNAL |
EXTERNAL
|
|
DV. ID. |
DV. ID.
|
LABEL
|
12
:
13
:
14
:
15
:
16
:
17
:
---------------------------------------------------------------------------------------------------------------------------------1 |
1
| P1
|
2.4258E-02 :
2.6005E-02 :
2.6021E-02 :
2.6021E-02 :
2 |
2
| P2
|
2.3058E-02 :
2.1841E-02 :
2.1834E-02 :
2.1834E-02 :
3 |
4
| P4
|
3.3257E-03 :
2.7535E-03 :
2.4141E-03 :
2.4141E-03 :
4 |
5
| P5
|
4.3248E-02 :
4.3566E-02 :
4.3587E-02 :
4.3587E-02 :
5 |
6
| P6
|
1.4020E-02 :
1.1520E-02 :
1.1632E-02 :
1.1632E-02 :
6 |
7
| P7
|
2.7829E-02 :
2.9923E-02 :
2.9964E-02 :
2.9964E-02 :
7 |
8
| BETA
|
1.3939E-01 :
1.4246E-01 :
1.0685E-01 :
1.0685E-01 :
*** USER INFORMATION MESSAGE 6464 (DOM12E)
RUN TERMINATED DUE TO HARD CONVERGENCE TO AN OPTIMUM AT CYCLE NUMBER =
15.
140.5 dB @ 100 Hz
115.9 dB @ 100 Hz
11-27
Mode
No.
Structural
Eigenfrequencies
Initial
Final
Fluid
Eigenfrequencies
75.99
7.861680E+01
50.01
95.29
8.777968E+01
100.41
104.16
9.964240E+01
100.41
130.78
1.187981E+02
100.41
133.35
1.262226E+02
112.19
143.74
1.309330E+02
112.19
153.16
1.448834E+02
142.00
173.69
1.457052E+02
142.00
142.00
11-28
SECTION 12
SUPERELEMENT OPTIMIZATION
GENERAL CONSIDERATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
CASE CONTROL SPECIFICATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-3
EXAMPLE TWENTY FIVE BAR TRUSS, SUPERELEMENT OPTIMIZATION . . . 12-5
GENERAL CONSIDERATIONS
12
12-1
responses
can
be
from
any,
or
all,
design
variables
and
12-2
ID MSC, D200SE3 $
SOL 200
$ OPTIMIZATION
CEND
TITLE=STATIC ANALYSIS OF A 25-BAR TRUSS
DISP=ALL
SPC=100
ANALYSIS=STATICS
SUPER=ALL
DESOBJ=15
SUBCASE 1
DESSUB=1
LABEL=LOAD CONDITION 1
LOAD=300
SUBCASE 2
DESSUB=2
LABEL=LOAD CONDITION 2
LOAD=310
BEGIN BULK
12-3
D200SE3
ANALYSIS=STATICS
DESOBJ=15
SUBCASE 10
DESSUB=1
SUPER=1,1 $ SUPERELEMENT 1, LOAD SEQ 1
LOAD=300
SUBCASE 20
DESSUB=2
SUPER=1,2 $ SUPERELEMENT 1, LOAD SEQ 2
LOAD=310
SUBCASE 30
DESSUB=3
LOAD=300
SUBCASE 40
DESSUB=4
LOAD=310
BEGIN BULK
12-4
Superelement 1
3
4
5
10
Superelement 0
X
12-5
to
Constraints:
eight
D
where ---- 100
t
Dt ( D + t )
I ---------------------------------8
12-6
AE ( 100 + 1 )
b = -------- ------------------------------2
8100
L
A feasible buckling condition is:
b
Which can be normalized as:
------ 1
b
12-7
D
where ---- = 100
t
12-8
12-9
12-10
+DP1
+DP2
+DP3
+DP4
+DP5
+DP6
+DP7
+DP8
12-11
12-12
12-13
9
8
12-14
6.614414E+02
1.102773E-01
1
5.985143E+02
5.985219E+02
-1.254311E-05
2.294609E-03
2
5.763787E+02
5.763831E+02
-7.730217E-06
8.119856E-04
3
5.605763E+02
5.605792E+02
-5.117301E-06
3.697191E-04
4
5.513699E+02
5.513676E+02
4.206515E-06
2.951452E-03
5
5.484974E+02
5.484994E+02
-3.560852E-06
3.337860E-05
6
5.465300E+02
5.465295E+02
7.817438E-07
1.285757E-04
7
5.459542E+02
5.459532E+02
1.788729E-06
3.235681E-06
8
5.459244E+02
5.459244E+02
0.000000E+00
3.865957E-04
---------------------------------------------------------------------------------------------------------------
12-15
SECTION 13
AEROELASTIC OPTIMIZATION
AEROELASTIC OPTIMIZATION MSC BACKGROUND . . . . . . . . . . . . . . . . . . . . . 13-1
STATIC AEROELASTICITY DESIGN CONDITIONS . . . . . . . . . . . . . . . . . . . . . . . . 13-2
STATIC AEROELASTICITY SENSITIVITY ANALYSIS . . . . . . . . . . . . . . . . . . . . . . 13-3
FLUTTER DESIGN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-4
FLUTTER RESPONSE PROPERTIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-5
FLUTTER RESPONSE SENSITIVITY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-6
AEROELASTIC DESIGN EXAMPLE (HA200A and B) . . . . . . . . . . . . . . . . . . . . . . 13-7
ANALYSIS CONDITIONS FOR EXAMPLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-9
DESIGN CONDITIONS FOR EXAMPLE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-10
HA200B OBJECTIVE RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-12
HA200B DESIGN VARIABLE RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-13
HA200B FLUTTER RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-14
HA200A INPUT FILE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-15
DESIGN MODEL AND BULK DATA INPUT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-17
HA200A SELECTED RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-23
HA200B SELECTED RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-31
Aeroelasticity
Design Optimization
Aeroelastic
Optimization
13-1
10.0
20.0 e 20.0
13-2
13-3
require
additional
pseudo-
FLUTTER DESIGN
The PK method solves for p at a set of specified velocities
Given p:
= Imag ( p )
= Real ( p )
13-4
13-5
is
conceptually
13-6
straightforward
but
2000
1000
Canard
2100
2007
1007
1131
2103
1100
2115
1119
Aileron
3100
2131
3115
Wing
3103
Rudder
This is a standard MSC/NASTRAN test case (Example HA144E).
13-7
Designed Bars
98
98
90
90
221
97
121
99
99
120
220
122
222
211
111
210
110
212
112
100
311
310
312
13-8
100
13-9
W =
i Ai l i
i=1
1
0.1736
=
X i
I
2.0000
2
0.4630
J i
13-10
( u5 )
( u5 )
100
( u5 )
( u5 )
0.5, M = 1.2 , Level flight, q = 863 psf
220
100
120
g 0.00 at V = 1000, 1300 and 1500 ft/s for M = 0.9 and 1.2
13-11
4000
3500
3000
2500
2000
1500
1000
500
0
0
Iteration Number
13-12
Design Variable
2.0
= DV10
= DV20
= DV30
1.0
0.0
0
Iteration Number
13-13
1.00
0.75
Optimized Design
Initial Design
Damping (g)
0.50
0.25
0.00
0.25
VREQ
0.50
0.75
1.00
6.0
Frequency (Hz)
5.0
4.0
3.0
Optimized Design
2.0
Initial Design
1.0
0.0
1000
1200
1400
Velocity (ft/sec)
13-14
1600
1800
13-15
13-16
13-17
13-18
13-19
13-20
0
0
4
4
XXX
XXX
XXXX
ID_MODE +DR
88
+DR1
+DRES2
+DR2401
13-21
13-22
PAGE
13-23
----COMPARISON BETWEEN INPUT PROPERTY VALUES FROM ANALYSIS AND DESIGN MODELS ------------------------------------------------------------------------------------------------------------------PROPERTY
PROPERTY
FIELD
ANALYSIS
DESIGN
LOWER
UPPER
DIFFERENCE
TYPE
ID
ID
VALUE
VALUE
BOUND
BOUND
FLAG
--------------------------------------------------------------------------------------------------------------PBAR
101
4
1.500000E+00
1.500000E+00
1.000000E-03
1.000000E+20
NONE
PBAR
101
5
1.736110E-01
1.736110E-01
1.000000E-03
1.000000E+20
NONE
PBAR
101
6
2.000000E+00
2.000000E+00
1.000000E-03
1.000000E+20
NONE
PBAR
101
7
4.629630E-01
4.629630E-01
1.000000E-03
1.000000E+20
NONE
PBAR
102
4
1.500000E+00
1.500000E+00
1.000000E-03
1.000000E+20
NONE
PBAR
102
5
1.736110E-01
1.736110E-01
1.000000E-03
1.000000E+20
NONE
PBAR
102
6
2.000000E+00
2.000000E+00
1.000000E-03
1.000000E+20
NONE
PBAR
102
7
4.629630E-01
4.629630E-01
1.000000E-03
1.000000E+20
NONE
PBAR
103
4
1.500000E+00
1.500000E-01
1.000000E-03
1.000000E+20
WARNING
PBAR
103
5
1.736110E-01
1.736110E-02
1.000000E-03
1.000000E+20
WARNING
PBAR
103
6
2.000000E+00
2.000000E-01
1.000000E-03
1.000000E+20
WARNING
PBAR
103
7
4.629630E-01
4.629630E-02
1.000000E-03
1.000000E+20
WARNING
1. IF FIELD ID IS LESS THAN ZERO, IT IDENTIFIES THE WORD POSITION OF AN ENTRY IN EPT.
2. IF FIELD ID IS GREATER THAN ZERO, IT IDENTIFIES THE FIELD POSITION ON A PROPERTY BULK DATA ENTRY.
3. THE DIFFERENCE FLAG IS USED TO CHARACTERIZE DIFFERENCES BETWEEN ANALYSIS AND DESIGN MODEL PROPERTIES:
IF THE FLAG IS NONE, THEN THERE IS NO SIGNIFICANT DIFFERENCE BETWEEN THE TWO VALUES.
IF THE FLAG IS WARNING, THEN THE USER IS ADVISED THAT DIFFERENCES EXIST.
IF THE FLAG IS FATAL, THEN THE DIFFERENCES ARE GREATER THAN 1.00000E+35 AND THE RUN WILL BE TERMINATED.
31
13-24
13-25
13-26
13-27
13-28
---------------------------------------------------------COLUMN
DRESP2
SUB
FREQ/
NO.
ENTRY ID
CASE
TIME
---------------------------------------------------------86
5
1
87
6
1
88
2401
1
89
5
2
90
6
2
91
2401
2
92
4
6
93
4
6
94
4
6
95
4
6
96
4
6
97
4
6
98
4
6
99
4
6
100
4
6
101
4
6
102
4
6
103
4
6
104
21
7
105
21
7
106
21
7
107
21
7
108
21
7
109
21
7
110
21
7
111
21
7
112
21
7
113
21
7
114
21
7
115
21
7
^^^ DMAP INFORMATION MESSAGE 9029 (DESOPT) DESIGN SENSITIVITY COEFFICIENT MATRIX FOR DIRECT
AND SYNTHETIC RESPONSES - GRADIENTS OF RESPONSES
WITH RESPECT TO INDEPENDENT DESIGN VARIABLES
13-29
MATRIX DSCM2
(GINO NAME 101 ) IS A REAL
115 COLUMN X
3 ROW RECTANG
COLUMN
1
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
6.6667E+02 1.3333E+03 5.0000E+01
COLUMN
2
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
2.3115E-04 5.5879E-06 5.3085E-05
COLUMN
3
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
-5.4311E-03 -4.8257E-03 8.5682E-05
COLUMN
4
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
-5.4313E-03 -4.8257E-03 8.5682E-05
COLUMN
5
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
-1.2751E+06 1.3784E+05 5.2500E+03
COLUMN
6
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
-1.2751E+06 1.3784E+05 5.2500E+03
COLUMN
7
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
-1.2258E+05 -4.9729E+05 2.1250E+03
.
.
.
COLUMN
110
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
5.2690E-03 3.3642E-02 8.6986E-04
COLUMN
111
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
1.4705E+00 1.4853E+00 1.0002E-01
COLUMN
112
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
1.1036E+01 5.7482E+00 3.5277E-01
MATRIX.
13-30
COLUMN
113
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
5.7897E-01 7.2965E-01 2.5005E-02
COLUMN
114
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
9.0316E-03 6.3687E-02 1.4559E-03
COLUMN
115
ROWS
1 THRU
3
-------------------------------------------------ROW
1)
1.2167E-02 8.5035E-02 1.8582E-03
THE NUMBER OF NON-ZERO TERMS IN THE DENSEST COLUMN =
3
THE DENSITY OF THIS MATRIX IS 98.84 PERCENT.
***************************************************************
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
***************************************************************
NUMBER OF FINITE ELEMENT ANALYSES COMPLETED
1
NUMBER OF OPTIMIZATIONS W.R.T. APPROXIMATE MODELS
0
OBJECTIVE AND MAXIMUM CONSTRAINT HISTORY
--------------------------------------------------------------------------------------------------------------OBJECTIVE FROM
OBJECTIVE FROM
FRACTIONAL ERROR
MAXIMUM VALUE
CYCLE
APPROXIMATE
EXACT
OF
OF
NUMBER
OPTIMIZATION
ANALYSIS
APPROXIMATION
CONSTRAINT
--------------------------------------------------------------------------------------------------------------INITIAL
2.005000E+03
1.460522E+01
--------------------------------------------------------------------------------------------------------------DESIGN VARIABLE HISTORY
---------------------------------------------------------------------------------------------------------------------------------INTERNAL |
EXTERNAL
|
|
DV. ID. |
DV. ID.
|
LABEL
|
INITIAL
:
1
:
2
:
3
:
4
:
5
:
---------------------------------------------------------------------------------------------------------------------------------1 |
10
| PBAR101
|
1.0000E+00 :
2 |
20
| PBAR102
|
1.0000E+00 :
3 |
30
| PBAR103
|
1.0000E-01 :
---------------------------------------------------------------------------------------------------------------------------------*** USER INFORMATION MESSAGE 6464 (DOM12E)
RUN TERMINATED DUE TO PARAMETER OPTEXIT =
4.
PAGE
159
SUBCASE 7
13-31
********************************************************************
*
*
*
*
*
D E S I G N
O P T I M I Z A T I O N
*
*
*
*
*
********************************************************************
*******************************************
*
*
*
D E S I G N
C Y C L E
1
*
*
*
*******************************************
*****
DESIGN VARIABLES
-----
--------------------------------------------------------------------------------------------------------------------------INTERNAL
DESVAR
LOWER
INPUT
OUTPUT
UPPER
ID
ID
LABEL
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------1
10
PBAR101
1.0000E-03
1.0000E+00
2.0000E+00
1.0000E+02
2
20
PBAR102
1.0000E-03
1.0000E+00
1.6390E+00
1.0000E+02
3
30
PBAR103
1.0000E-03
1.0000E-01
2.0000E-01
1.0000E+02
-----
DESIGNED PROPERTIES
-----
13-32
--------------------------------------------------------------------------------------------------------------------------PROPERTY
PROPERTY
FIELD
TYPE OF
LOWER
INPUT
OUTPUT
UPPER
TYPE
ID
ID
PROPERTY
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------PBAR
101
4
DVPREL1
1.0000E-03
1.5000E+00
3.0000E+00
1.0000E+20
PBAR
101
5
DVPREL1
1.0000E-03
1.7361E-01
3.4722E-01
1.0000E+20
PBAR
101
6
DVPREL1
1.0000E-03
2.0000E+00
4.0000E+00
1.0000E+20
PBAR
101
7
DVPREL1
1.0000E-03
4.6296E-01
9.2593E-01
1.0000E+20
PBAR
102
4
DVPREL1
1.0000E-03
1.5000E+00
2.4586E+00
1.0000E+20
PBAR
102
5
DVPREL1
1.0000E-03
1.7361E-01
2.8455E-01
1.0000E+20
PBAR
102
6
DVPREL1
1.0000E-03
2.0000E+00
3.2781E+00
1.0000E+20
PBAR
102
7
DVPREL1
1.0000E-03
4.6296E-01
7.5881E-01
1.0000E+20
PBAR
103
4
DVPREL1
1.0000E-03
1.5000E-01
3.0000E-01
1.0000E+20
PBAR
103
5
DVPREL1
1.0000E-03
1.7361E-02
3.4722E-02
1.0000E+20
PBAR
103
6
DVPREL1
1.0000E-03
2.0000E-01
4.0000E-01
1.0000E+20
PBAR
103
7
DVPREL1
1.0000E-03
4.6296E-02
9.2593E-02
1.0000E+20
-----
13-33
--------------------------------------------------------------------------------------------------------------------------INTERNAL
INTERNAL
INTERNAL
DCONSTR
RESPONSE
RESPONSE
L/U
REGION
SUBCASE
INPUT
OUTPUT
ID
ID
ID
TYPE
FLAG
ID
ID
VALUE
VALUE
--------------------------------------------------------------------------------------------------------------------------1
50
2
EQUA
UPPER
6
1
-3.9543E-01
-6.6547E-01
2
50
1
EQUA
UPPER
5
1
-3.9543E-01
-6.6440E-01
3
50
3
EQUA
LOWER
2401
1
1.7848E-02
-3.4503E-02
4
60
4
EQUA
LOWER
2401
2
1.1241E-02
-3.9465E-02
5
6
5
EQUA
UPPER
4
6
-4.3507E-01
-3.6965E-01
6
6
6
EQUA
UPPER
4
6
1.4605E+01**
1.3839E+00**
7
7
7
EQUA
UPPER
21
7
8.1231E+00
-8.8540E+00
8
7
8
EQUA
UPPER
21
7
-4.7149E-01
-4.1885E-01
-----
-----
--------------------------------------------------------------------------------------------------------------------------INTERNAL
PROPERTY
FIELD
L/U
CYCLE
INPUT
OUTPUT
ID
ID
FLAG
FLAG
LIMIT
VALUE
VALUE
--------------------------------------------------------------------------------------------------------------------------9
101
4
LOWER
1.2000E+00
-2.5000E-01
-1.5000E+00
10
101
5
LOWER
1.3889E-01
-2.5000E-01
-1.5000E+00
11
101
6
LOWER
1.6000E+00
-2.5000E-01
-1.5000E+00
12
101
7
LOWER
3.7037E-01
-2.5000E-01
-1.5000E+00
13
102
4
LOWER
1.2000E+00
-2.5000E-01
-1.0488E+00
-----
13-34
--------------------------------------------------------------------------------------------------------------------------INTERNAL
PROPERTY
FIELD
L/U
CYCLE
INPUT
OUTPUT
ID
ID
FLAG
FLAG
LIMIT
VALUE
VALUE
--------------------------------------------------------------------------------------------------------------------------14
102
5
LOWER
1.3889E-01
-2.5000E-01
-1.0488E+00
15
102
6
LOWER
1.6000E+00
-2.5000E-01
-1.0488E+00
16
102
7
LOWER
3.7037E-01
-2.5000E-01
-1.0488E+00
17
103
4
LOWER
1.2000E-01
-2.5000E-01
-1.5000E+00
18
103
5
LOWER
7.3611E-03
-1.3585E+00
-3.7170E+00
19
103
6
LOWER
1.6000E-01
-2.5000E-01
-1.5000E+00
20
103
7
LOWER
3.6296E-02
-2.7551E-01
-1.5510E+00
21
101
4
UPPER
1.8000E+00
-1.6667E-01
6.6667E-01
22
101
5
UPPER
2.0833E-01
-1.6667E-01
6.6667E-01
23
101
6
UPPER
2.4000E+00
-1.6667E-01
6.6667E-01
24
101
7
UPPER
5.5556E-01
-1.6667E-01
6.6667E-01
25
102
4
UPPER
1.8000E+00
-1.6667E-01
3.6586E-01
26
102
5
UPPER
2.0833E-01
-1.6667E-01
3.6586E-01
27
102
6
UPPER
2.4000E+00
-1.6667E-01
3.6586E-01
28
102
7
UPPER
5.5556E-01
-1.6667E-01
3.6586E-01
29
103
4
UPPER
1.8000E-01
-1.6667E-01
6.6667E-01
30
103
5
UPPER
2.7361E-02
-3.6548E-01
2.6904E-01
31
103
6
UPPER
2.4000E-01
-1.6667E-01
6.6667E-01
32
103
7
UPPER
5.6296E-02
-1.7763E-01
6.4474E-01
WEIGHT RESPONSE
-----
--------------------------------------------------------------------------------------------------------------------------INTERNAL
DRESP1
RESPONSE
LOWER
INPUT
OUTPUT
UPPER
ID
ID
LABEL
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------1
10
WEIGHT
N/A
2.0050E+03
3.5277E+03
N/A
13-35
D E S I G N
C Y C L E =
-----
S U B C A S E =
DISPLACEMENT RESPONSES
-----
--------------------------------------------------------------------------------------------------------------------------INTERNAL
DRESP1
RESPONSE
GRID
COMPONENT
LOWER
INPUT
OUTPUT
UPPER
ID
ID
LABEL
ID
NO.
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------2
100
RTROT
100
5
N/A
1.1291E-02
1.1403E-02
N/A
3
101
RTIPROT
120
5
N/A
2.1841E-02
1.7259E-02
N/A
4
201
LTIPROT
220
5
N/A
2.1841E-02
1.7240E-02
N/A
-----
-----
--------------------------------------------------------------------------------------------------------------------------RESTRAINED /
AESTAT /
INTERNAL
DRESP1
RESPONSE UNRESTRAINED COMPONENT AESURF
LOWER
INPUT
OUTPUT
UPPER
ID
ID
LABEL
FLAG
NO.
ID
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------5
1401 CLDELTA
0
4
517
N/A
2.6340E-01
2.6922E-01
N/A
6
1402 CLP
0
4
513
N/A
-4.4697E-01
-4.3374E-01
N/A
C Y C L E =
S U B C A S E =
-----
--------------------------------------------------------------------------------------------------------------------------RESTRAINED /
AESTAT /
INTERNAL
DRESP1
RESPONSE UNRESTRAINED COMPONENT AESURF
LOWER
INPUT
OUTPUT
UPPER
ID
ID
LABEL
FLAG
NO.
ID
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------7
1401 CLDELTA
0
4
517
N/A
2.1655E-01
2.2531E-01
N/A
8
1402 CLP
0
4
513
N/A
-5.0933E-01
-5.0408E-01
N/A
D E S I G N
C Y C L E =
13-36
-----
S U B C A S E =
FLUTTER RESPONSES
-----
-----
-----------------------------------------------------------------------------------------------------------------------INTERNAL
DRESP1
RESPONSE
MODE
MACH
LOWER
INPUT
OUTPUT
UPPER
ID
ID
LABEL
NO.
VELOCITY
NO.
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------9
1 FLUTTER
6
1.0000E+03
9.0000E-01
N/A
-1.3052E-02
-1.1089E-02
N/A
10
1 FLUTTER
6
1.5000E+03
9.0000E-01
N/A
4.3816E-01
4.1516E-02
N/A
D E S I G N
C Y C L E =
-----
S U B C A S E =
FLUTTER RESPONSES
-----
--------------------------------------------------------------------------------------------------------------------------INTERNAL
DRESP1
RESPONSE
MODE
MACH
LOWER
INPUT
OUTPUT
UPPER
ID
ID
LABEL
NO.
VELOCITY
NO.
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------11
11 FLUTTER
5
1.5000E+03
1.2000E+00
N/A
2.4369E-01
-2.6562E-01
N/A
12
11 FLUTTER
6
1.0000E+03
1.2000E+00
N/A
-1.4145E-02
-1.2566E-02
N/A
13-37
--------------------------------------------------------------------------------------------------------------------------INTERNAL
DRESP2
RESPONSE
EQUATION
LOWER
INPUT
OUTPUT
UPPER
ID
ID
LABEL
ID
BOUND
VALUE
VALUE
BOUND
--------------------------------------------------------------------------------------------------------------------------1
5
RHSTWIST
5
N/A
1.0550E-02
5.8562E-03
1.7450E-02
2
6
LHSTWIST
5
N/A
1.0550E-02
5.8376E-03
1.7450E-02
3
2401
ROLLEFF
103
6.0000E-01
5.8929E-01
6.2070E-01
N/A
4
2401
ROLLEFF
103
4.3000E-01
4.2517E-01
4.4697E-01
N/A
5
4
GDAMP
4
N/A
-4.3052E-01
-4.1089E-01
-3.0000E-01
6
4
GDAMP
4
N/A
4.0816E+00
1.1516E-01
-3.0000E-01
7
21
GDAMP
4
N/A
2.1369E+00
-2.9562E+00
-3.0000E-01
8
21
GDAMP
4
N/A
-4.4145E-01
-4.2566E-01
-3.0000E-01
****************************************************************************************
INSPECTION OF CONVERGENCE DATA FOR THE OPTIMAL DESIGN WITH RESPECT TO APPROXIMATE MODELS
(SOFT CONVERGENCE DECISION LOGIC)
****************************************************************************************
RELATIVE CHANGE IN OBJECTIVE
7.5947E-01 MUST BE LESS THAN
1.0000E-03
ABSOLUTE CHANGE IN OBJECTIVE
1.5227E+03 MUST BE LESS THAN
1.0000E-02
MAX OF RELATIVE PROP.CHANGES
1.0000E+00 MUST BE LESS THAN
1.0000E-03
--- AND --MAXIMUM CONSTRAINT VALUE
1.3839E+00 MUST BE LESS THAN
5.0000E-03
OR
MAX OF RELATIVE D.V. CHANGES
1.0000E+00 MUST BE LESS THAN
1.0000E-03
****************************************************************************************
OR
AND
13-38
13-39
13-40
APPENDIX A
CASE CONTROL COMMANDS AND BULK
DATA ENTRIES
DESIGN MODELING INPUT DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
CASE CONTROL COMMANDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-3
AUXCASE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-4
AUXMODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-5
DESGLB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-6
DESOBJ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-7
DESSUB . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-8
MODTRAK. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-9
BULK DATA ENTRIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-10
BNDGRID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-11
DCONADD. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-12
DCONSTR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-13
DEQATN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-15
DESVAR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-19
DLINK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-20
DOPTPRM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-21
DRESP1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-25
DRESP2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-31
DSAPRT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-34
APPENDIX A
CASE CONTROL COMMANDS AND BULK DATA
ENTRIES (Cont.)
DSCREEN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-37
DTABLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-39
DVBSHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-40
DVGRID . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-41
DVPREL1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-43
DVPREL2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-46
DVSHAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-48
MODTRAK. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-49
DLINK
Linear relations
DVPREL2
Nonlinear relations
DRESP2
A-1
Constraint functions
DCONADD
DESSUB Case
Control command
DESGLB Case
Control command
DOPTPRM
DTABLE
Constants
User-definition equation
A-2
A-3
AUXCASE
Format:
AUXCASE
Examples:
AUXCAS
AUXC
Remarks:
1.
AUXCASE indicates the beginning of Case Control commands for an auxiliary model.
AUXCASE must follow the primary model Case Control commands.
2.
All Case Control commands following this entry are applicable until the next AUXCASE or
BEGIN BULK command. Commands from preceding Case Control Sections are ignored.
3.
Each auxiliary model Case Control must be delimited with the AUXCASE command.
4.
The AUXMODEL command is used to associate the auxiliary model Case Control with a
particular auxiliary model.
5.
6.
A-4
AUXMODEL
Format:
AUXMODEL = n
Examples:
AUXMODEL = 4
AUXM = 4
Describer
Meaning
Remarks:
1.
AUXMODEL references a particular auxiliary model for analysis and may only be specified
in the auxiliary model Case Control Section.
2.
See the BEGIN BULK command for the Bulk Data definition of an auxiliary model.
A-5
DESGLB
Format:
DESGLB = n
Examples:
DESGLB = 10
DESG = 25
Describer
Meaning
Remarks:
1.
A DESGLB command is optional and invokes constraints that are to be applied independent of a particular subcase. These constraints could be based on responses that are
independent of subcases (e.g., WEIGHT or VOLUME).
2.
The DESGLB command can be used to invoke constraints that are not a function of
DRESP1 entries; e.g., DRESP2 responses that are not functions of DRESP1 responses
are subcase independent.
A-6
DESOBJ
Design Objective
DESOBJ the DRESP1 or DRESP2 entry to be used as the design objective.
Selects
Design Objective
Format:
DESOBJ (
MAX
) = N
MIN
Examples:
DESOBJ = 10
DESO = 25
Describer
Meaning
MIN
MAX
Remarks:
1.
A DESOBJ command is required for a design optimization task and is optional for a sensitivity task.
2.
If the DESOBJ command is specified within a SUBCASE, the identified DRESPi Bulk Data
entry use a response only from that subcase. If DESOBJ appears above all SUBCASE
commands and there are multiple subcases, it uses a global response.
3.
A-7
DESSUB
Format:
DESSUB = n
Examples:
DESSUB = 10
DESS = 25
Describer
Meaning
Remark:
1.
A DESSUB command is required for every subcase for which constraints are to be
applied.
A-8
MODTRAK
Format:
MODTRAK = n
Example:
MODTRAK = 100
Describer
Meaning
Remark:
1.
Selection of a MODTRAK Bulk Data entry with the MODTRAK Case Control command
activates mode tracking for the current subcase. This request is limited to normal modes
subcases (ANALYSIS = MODES) in design optimization (SOL 200).
A-9
A-10
BNDGRID
Format:
1
BNDGRID
GP1
GP2
GP3
GP4
GP5
GP6
GP7
GP8
-etc.-
123
41
42
43
44
45
46
47
10
Example:
BNDGRID
49
GP1
"THRU"
GP2
BNDGRID
123
41
THRU
49
Field
Contents
GPi
Shape boundary grid point identification number. (0 < Integer < 1000000; For
THRU option, GP1< GP2)
Remarks:
1.
C specifies the components for the listed grid points for which boundary motion is
prescribed.
2.
Multiple BNDGRID entries may be used to specify the shape boundary grid point identification numbers.
3.
Both fixed and free shape boundary grid point identification numbers are listed on this
entry.
4.
The degrees of freedom specified on BNDGRID entries must be sufficient to statically constrain the model.
5.
Degrees of freedom specified on this entry form members of the mutually exclusive s-set.
They may not be specified on other entries that define mutually exclusive sets. See the
MSC/NASTRAN Quick Reference Guide, Appendix B for a list of these entries.
A-11
DCONADD
Format:
1
DCONADD
DCID
DC1
DC2
DC3
DC4
DC5
DC6
DC7
DC8
-etc.-
10
10
Example:
DCONADD
12
Field
Contents
DCID
DCi
Remarks:
1.
2.
A-12
DCONSTR
Design Constraints
DCONSTR
Define design
Design Constraints
constraints.
Format:
1
DCONSTR
DCID
RID
10
10
LALLOW UALLOW
Example:
DCONSTR
1.25
Field
Contents
DCID
RID
LALLOW
UALLOW
Remarks:
1.
The DCONSTR entry may be selected in the Case Control Section by the DESSUB or
DESGLB command.
2.
3.
For a given DCID, the associated RID can be referenced only once.
4.
The units of LALLOW and UALLOW must be consistent with the referenced response defined on the DRESPi entry. If RID refers to an EIGN response, then the imposed bounds
must be expressed in units of eigenvalue, (radian/time)2. If RID refers to a FREQ response, then the imposed bounds must be expressed in cycles/time.
5.
LALLOW and UALLOW are unrelated to the stress limits specified on the MATi entry.
6.
A-13
DCONSTR
Design Constraints
> GSCAL
As Remark 6 indicates, small values of UALLOW and LALLOW require special processing
and should be avoided. Bounds of exactly zero are particularly troublesome. This can be
avoided by using a DRESP2 entry that offsets the constrained response from zero.
A-14
DEQATN
DEQATN one or more equations for use in design sensitivity or p-element analysis.
Design Equation Definition
Defines
Format:
1
DEQATN
EQID
10
EQUATION
EQUATION (Cont.)
Example:
DEQATN
14
Field
Contents
EQID
EQUATION
Remarks:
1.
A-15
DEQATN
3
A
F 1 = A + B C ( D + 10 ) + sin ( PI ( 1 ) R ) + -------------BC
F = A + B F1 D
where SIN and PI are intrinsic functions. See Remark 4.
2.
EQUATION may contain embedded blanks. EQUATION must contain less than 12,500
nonblank characters. This is equivalent to approximately 195 continuation entries.
3.
The syntax of the expressions follows FORTRAN language standards. The allowable
arithmetic operations are shown in Table 1 in the order of execution precedence. Parenthesis are used to change the order of precedence. Operations within parentheses are
performed first with the usual order of precedence being maintained within the
parentheses.
Operator
Operation
Sample Expressions
Interpreted As
, +
XY
X(Y)
Exponentiation
XY
(XY)
, +
Negative or Positive
XY
(X)Y
, /
Multiplication or Division
XY+Z
(XY)+Z
+,
Addition or Subtraction
X+Y
X+Y
4.
The expressions may contain intrinsic functions. Table 2 contains the format and descriptions of functions that may appear in the expressions. The use of functions that may be
discontinuous must be used with caution because they can cause discontinuous derivatives. These are ABS, DIM, MAX, MIN, and MOD. For examples and further details see
the MSC/NASTRAN DMAP Module Dictionary.
(Continued)
A-16
DEQATN
Format
Description
Mathematical Expression
ABS(x)
absolute value
|x|
ACOS(x)
arccosine
cos-1 x
ACOSH(x)
hyperbolic arccosine
cosh-1 x
ASIN(x)
arcsine
sin-1 x
ASINH(x)
hyperbolic arcsine
sinh-1 x
ATAN(x)
arctangent
tan-1 x
ATAN2(x,y)
arctangent of quotient
tan-1 (x/y)
ATANH(x)
hyperbolic arctangent
tanh-1 x
ATANH2(x,y)
tanh-1 (x/y)
COS(x)
cosine
cos x
COSH(x)
hyperbolic cosine
cosh x
DIM(x,y)
positive difference
x MIN(x,y)
EXP(x)
exponential
ex
LOG(x)
natural logarithm
loge x
LOG10(x)
common logarithm
log10 x
LOGX(x,y)
base x logarithm
logx y
maximum
minimum
MOD(x,y)
remainder (modulo)
x y (INT(x/y))
PI(x)
multiples of pi (p)
SIN(x)
sine
sin x
SINH(x)
hyperbolic sine
sinh x
SQRT(x)
square root
TAN(x)
tangent
tan x
TANH(x)
hyperbolic tangent
tanh x
(Continued)
A-17
DEQATN
5.
b.
DRESP2 entry, then xi represents the DVIDj, LABLk, NRm, and Gp fields in that
order.
c.
d.
e.
6.
7.
The DMAP logical operators NOT, AND, OR < XOR, and XQV cannot be used as Xi
names.
8.
Input errors on the DEQATN entry often result in poor messages. Substituting a [ for a
parenthesis or violating the restriction against large field format are examples. Known
messages are UFM 215, SFM 233 and UFM 5199. If any of these messages are encountered then review the DEQATN entry input.
A-18
DESVAR
Design Variable
DESVAR
Defines
Design Variable
Format:
1
DESVAR
ID
LABEL
XINIT
XLB
XUB
DELXV
BARA1
35.0
10.
100.
0.2
10
Example:
DESVAR
Field
Contents
ID
LABEL
XINIT
XLB
XUB
DELXV
Remarks:
1.
DELXV can be used to control the change in the design variable during one optimization
cycle.
2.
If DELXV is blank, the default is taken from the specification of the DELX parameter on
the DOPTPRM entry. If DELX is not specified, then the default is 1.0.
A-19
DLINK
DLINK
Relates
one design variable to one or more other design variables.
Format:
1
DLINK
ID
DDVID
C0
CMULT
IDV1
C1
IDV2
C2
IDV3
C3
-etc.-
10
0.1
0.33
2.0
1.0
7.0
10
Example:
DLINK
Field
Contents
ID
DDVID
C0
CMULT
IDVi
Ci
Remarks:
1.
i Ci IDVi
2.
This capability provides a means of linking physical design variables such as element
thicknesses to nonphysical design variables such as the coefficients of interpolating
functions.
3.
CMULT provides a simple means of scaling the Ci. For example if Ci = 1/7, 2/7, 4/7, etc.
is desired, then CMULT = 1/7 and Ci = 1, 2, 4, etc., may be input.
4.
An independent IDVi must not occur on the same DLINK entry more than once.
5.
A-20
DOPTPRM
Format:
1
DOPTPRM PARAM1
VAL1
PARAM2
VAL2
PARAM3
VAL3
PARAM4
VAL4
PARAM5
VAL5
-etc.-
DESMAX
10
Example:
DOPTPRM IPRINT
10
Field
Contents
PARAMi
VALi
Remarks:
1.
Name
APRCOD
CONV1
CONV2
CONVDV
Relative convergence criterion on design variables. (Real > 0.0; Default = 0.001)
CONVPR
(Continued)
A-21
DOPTPRM
Name
CT
CTMIN
DABOBJ
DELB
DELOBJ
DELP
Fractional change allowed in each property during any optimization design cycle.
This provides constraints on property moves. (Real > 0.0; Default = 0.2)
DELX
Fractional change allowed in each design variable during any optimization cycle.
(Real > 0.0; Default = 1.0)
DESMAX
DOBJ1
DOBJ2
DPMIN
DX1
DX2
DXMIN
Minimum design variable move limit (Real > 0.0; Default = 0.05).
GMAX
Maximum constraint violation allowed at the converged optimum. (Real > 0.0;
Default = 0.005)
GSCAL
(Continued)
A-22
DOPTPRM
Name
IGMAX
IPRINT
IPRNT1
IPRNT2
ISCAL
Design variables are rescaled every ISCAL iterations. Set ISCAL= 1 to turn off
scaling. (Integer; Default = NDV (number of design variables))
ITMAX
Maximum number of iterations allowed at optimizer level during each design cycle.
(Integer; Default = 40)
ITRMOP
ITRMST
JTMAX
Maximum number of iterations allowed at the optimizer level for the Sequential
Linear Programming Method. This is the number of linearized subproblems
solved. (Integer 0; Default = 20)
JPRINT
JWRITE
If JWRITE > 0, file number on which iteration history will be written. (Integer > 0;
Default = 0)
(Continued)
A-23
DOPTPRM
Name
METHOD
P1
Print control items specified for P2. (Integer 0; Default = 0) Initial results are
always printed prior to the first approximate optimization. If an optimization task
is performed, final results are always printed for the final analysis unless
PARAM,SOFTEXIT,YES is specified. These two sets of print are not controllable.
n: Print at every n-th design cycle.
P2
PTOL
STPSCL
Scaling factor for shape finite difference step sizes, to be applied to all shape
design variables. (Real > 0.0; Default = 1.0)
A-24
DRESP1
DRESP1
Sensitivity
Responseor
Quantities
Defines a set of structural responses that is used in the designDesign
either
as constraints
as an
objective.
Format:
1
DRESP1
ID
LABEL
RTYPE
PTYPE
REGION
ATTA
ATTB
ATT1
ATT2
-etc.-
DX1
STRESS
PROD
10
Example:
DRESP1
102
103
Field
Contents
ID
LABEL
RTYPE
PTYPE
Element flag (PTYPE = ELEM) or property entry name. Used with element type
responses (stress, strain, force, etc.) to identify the property type, since property
entry IDs are not unique across property types. (Character: ELEM, PBAR,
PSHELL, etc.)
REGION
ATTA, ATTB,
ATTi
(Continued)
A-25
DRESP1
Response
Type
(RTYPE)
Response Attributes
ATTA (Integer > 0)
WEIGHT
Blank
Blank
SEIDi or ALL
VOLUME
Blank
Blank
SEIDi or ALL
EIGN
Approximation Code.
See Remark 19.
Blank
FREQ
Approximation Code.
See Remark 19.
Blank
LAMA
Approximation Code.
See Remark 19.
Blank
DISP
Displacement Component
Blank
Grid ID
STRAIN
Blank
STRESS
Blank
FORCE
Blank
CSTRAIN
LAMINA Number
(Integer; Default = 1)
CSTRESS
LAMINA Number
(Integer; Default = 1)
CFAILURE
LAMINA Number
(Integer; Default = 1)
FRDISP
Displacement Component
Grid ID
FRVELO
Velocity Component
Grid ID
FRACCL
Acceleration Component
Grid ID
FRSPCF
Grid ID
FRSTRE
(Continued)
A-26
DRESP1
Response
Type
(RTYPE)
Response Attributes
ATTA (Integer > 0)
FRFORC
TDISP
Displacement Component
Grid ID
TVELO
Velocity Component
Grid ID
TACCL
Acceleration Component
Grid ID
TSPCF
Grid ID
TSTRE
TFORC
TRIM
AESTAT or AESURF
Entry ID
Blank
Blank
STABDER
AESTAT or AESURF
Entry ID
Component
FLUTTER
Blank
Remarks:
1.
Stress, strain, and force item codes can be found in the MSC/NASTRAN Quick Reference
Guide, Appendix A. For stress or strain item codes that have dual meanings, such as von
Mises or maximum shear, the option specified in the Case Control Section will be used;
i.e., STRESS(VONM) or STRESS(MAXS).
2.
RTYPE=CSTRESS, CSTRAIN, and CFAILURE are used only with the PCOMP entry.
CSTRESS and CSTRAIN item codes are described under Table 1. (Element
Stress/Strain Item Codes) in the MSC/NASTRAN Quick Reference Guide, Appendix A.
CFAILURE item codes are described under Table Table 2. (Element Force Item Codes)
in the MSC/NASTRAN Quick Reference Guide, Appendix A. Only force item codes that
refer to failure indices of direct stress and interlaminar shear stress are valid.
(Continued)
A-27
DRESP1
The CFAILURE response type requires the following specifications on the applicable
entries:
a.
b.
c.
Stress limits in the ST, SC, and SS fields on all MATi entries.
3.
ATTB is used only for responses of composite laminae, dynamics, and stability derivatives. For other responses, this field must be blank.
4.
All grids associated with a DRESP1 entry are considered to be in the same region for
screening purposes. Only up to NSTR displacement constraints (see DSCREEN entry)
per group per load case will be retained in the design optimization phase.
5.
6.
7.
If RTYPE = DISP, TDISP, TVELO, TACCL or TSPCF, multiple component numbers (any unique combination of the digits 1 through 6 with no embedded blanks) may be
specified on a single entry. Multiple response components may not be used on any other
response types.
8.
9.
Real/imaginary representation is the default for complex response types. Magnitude/phase representation must be requested by the corresponding Case Control
command; e.g., DlSP(PHASE) = ALL.
10.
REGION is used for constraint screening. The NSTR field on DSCREEN entries gives the
maximum number of constraints retained for each region per load case.
IF RTYPE = WEIGHT, VOLUME, LAMA, EIGN or FREQ, no REGION identification
number should be specified. For all other responses, if the REGION field is left blank, the
default specified in Table 2 is used. Usually, the default value is appropriate.
If the REGION field is not blank, all the responses on this entry as well as all responses on
other DRESP1 entries that have the same RTYPE and REGION identification number will
be grouped into the same region.
(Continued)
A-28
DRESP1
Response Type
Default Region
WEIGHT
No region
VOLUME
No region
LAMA
No region
EIGN
No region
FREQ
No region
DISP
FRDISP
FRVELO
FRACCL
FRSPCF
TDISP
TVELO
TACCL
TSPCF
FLUTTER
OTHER
11.
REGION is valid only among the same type of responses. Responses of different types
will never be grouped into the same region, even if they are assigned the same REGION
identification number by the user.
12.
If RTYPE = WEIGHT or VOLUME, field ATTi = ALL implies total weight/volume of all
superelements except external superelements.
13.
RTYPE = STABDER identifies a stability derivative response. ATTB is the restraint flag
for the stability derivative. ATTB = 0 means unrestrained, and ATTB = 1 means restrained. For example, ATTA = 4000, ATTB = 0, and ATT1 = 3 reference the unrestrained
Cz derivative for the AESTAT (or AESURF) entry ID = 4000.
(Continued)
A-29
DRESP1
14.
RTYPE = FLUTTER identifies a set of damping responses. The set is specified by ATTi:
ATT1 = Identification number of a SET1 entry that specifies a set of modes.
ATT2 = Identification number of an FLFACT entry that specifies a list of densities.
ATT3 = Identification number of an FLFACT entry that specifies a list of Mach numbers.
ATT4 = Identification number of an FLFACT entry that specifies a list of velocities.
15.
16.
For RTYPE = TDISP, TVELO, TACCL, TSPCF, TFORC, and TSTRE, ATTB
specifies a time value. If ATTB is specified, then the responses are evaluated at the closest time selected by the OTIME command. The default for ATTB is all time steps selected
by the OTIME command.
17.
Intermediate station responses on CBAR elements due to PLOAD1 and/or CBARAO entries may not be defined on the DRESP1 entry.
18.
RTYPE = EIGN refers to normal modes response in terms of eigenvalue (radian/time)2 while RTYPE = FREQ refers to normal modes response in terms of natural
frequency or units of cycles per unit time.
19.
For RTYPE = LAMA, EIGN or FREQ, the response approximation used for optimization
can be individually selected. (Approximation Code = 1 = direct linearization, = 2 = Inverse
Linearization).
A-30
DRESP2
Format:
1
DRESP2
ID
LABEL
EQID
REGION
DESVAR
DVID1
DVID2
DVID3
DVID4
DVID5
DVID6
DVID7
DVID8
-etc.-
LABL1
LABL2
LABL3
LABL4
LABL5
LABL6
LABL7
LABL8
-etc.-
NR1
NR2
NR3
NR4
NR5
NR6
NR7
NR8
-etc.-
G1
C1
G2
C2
G3
C3
G4
C4
-etc.-
DPIP1
DPIP2
DPIP3
DPIP4
DPIP5
DPIP6
DPIP7
DPIP8
DPIP9
-etc.-
LBUCK
DESVAR
101
205
209
DTABLE
DRESP1
DNODE
DVPREL1
Example:
DRESP2
201
DTABLE
PI
YM
DRESP1
14
22
33
DNODE
14
22
43
101
102
DVPREL1
Field
Contents
ID
LABEL
A-31
10
DRESP2
EQID
REGION
DESVAR
DVIDi
DTABLE
Flag indicating that the labels for the constants in a DTABLE entry follow.
(Character)
LABLj
DRESP1
NRk
DNODE
Flag signifying that the following fields are designed grid points. See Remark 7.
(Character)
Gm
Cm
DVPREL1
DPIPi
Remarks:
1.
DRESP2 entries may only reference DESVAR, DTABLE, DRESP1, and DVPREL1 entries. They may not reference other DRESP2 entries.
2.
3.
DRESP2 entries must have unique identification numbers with respect to DRESP1
entries.
4.
The DESVAR, DTABLE, DRESP1, DNODE, and DVPREL1 flags in field 2 must
appear in the order given above. Any of these words, along with the identification numbers associated with them, may be omitted if they are not involved in this DRESP2
relationship. However, at least one of these four types of arguments must exist.
5.
The REGION field follows the same rules as for the DRESP1 entries. DRESP1 and
DRESP2 responses will never be contained in the same region, even if they are assigned
the same REGION identification number. The default is to put all responses referenced
by one DRESP2 entry in the same region.
(Continued)
A-32
DRESP2
The variables identified by DVIDi, LABLj, NRk, the Gm, Cm pairs and DPIPi are assigned
(in that order) to the variable names (x1, x2, x3, etc.) specified in the left-hand side of the
first equation on the DEQATN entry referenced by EQID. The variable names x1 through
xN (N = m + n + p + q) are assigned in the order DVID1, ..., DVIDm, LABL1, ..., LABLn,
NR1, ..., NRp, G1, ..., Gq, DPIP1, ..., DPIPr. In the example below,
DESVARs 101 and 3 are assigned to arguments A and B.
DTABLEs PI and YM are assigned to arguments C and D.
Grid 14, Component 1 is assigned to argument R.
DRESP2
LBUCK
DESVAR
101
DTABLE
PI
YM
DNODE
14
DEQATN
7.
(Gm, Cm) refer to a designed grid component. Depending on the scheme used in generating basis vectors, a designed grid component can be one of the following:
a.
b.
A grid component that is free to move in a full auxiliary model when the external displacement fields are used.
c.
When the auxiliary model boundary shape method is used, a designed grid component can be either a loaded grid component in the auxiliary model or a grid
component with motion that is obtained from interpolation of boundary shapes.
d.
When the interface using geometric boundary shapes is used, a designed grid component is either a grid component defined on a DVGRID entry or one with motion that
is obtained from interpolation of boundary shapes.
A-33
DSAPRT
DSAPRT
Specifies
design sensitivity output parameters.
Format:
ALL
FORMATTED
NOEXPORT
DSAPRT ( UNFORMATTED ,
, [ START = i ], [ BY = j ], [ END = k ] ) = n
EXPORT
NOPRINT
NONE
Examples:
DSAPRT(FORMATTED,EXPORT)
DSAPRT(FORMATTED,START = FIRST,BY = 3,END = LAST) = 101
DSAPRT(UNFORMATTED,START = FIRST)
DSAPRT(UNFORMATTED,EXPORT)
DSAPRT(FORMATTED,END = 4) = ALL
DSAPRT(UNFORMATTED,END = SENS) = ALL
DSAPRT(NOPRINT,EXPORT)
Describers
Meaning
FORMATTED
UNFORMATTED Output will be printed as a matrix print (see description of the MATPRN module
in the MSC/NASTRAN DMAP Module Dictionary).
NOPRINT
EXPORT
NOEXPORT
START = i
Specifies the first design cycle for output. (Integer > 0 or Character: FIRST or
LAST; Default = 1 or FIRST)
BY = j
Specifies the design cycle interval for output. (Integer 1, or > 0; Default = 0)
See Remark 2.
END = k
Specifies the last design cycle for output. (Integer > 0 or Character: FIRST,
LAST, or SENS; Default = LAST)
ALL
All design responses (defined in DRESP1 and DRESP2 entries) will be output.
A-34
DSAPRT
Remarks:
1.
Only one DSAPRT may appear in the Case Control Section and should appear above all
SUBCASE commands.
2.
Sensitivity data will be output at design cycles i, i + j, i + 2j, ..., k. Note that the BY = 0
default implies no sensitivity analysis at the intermediate design cycles.
3.
END = SENS requests design sensitivity analysis, and no optimization will be performed.
4.
DSAPRT(NOPRINT,EXPORT,END = SENS)
DSAPRT(UNFORMATTED,START = LAST)
A-35
A-36
DSCREEN
Format:
1
DSCREEN
RTYPE
TRS
NSTR
0.7
10
Example:
DSCREEN STRESS
Field
Contents
RTYPE
Response type for which the screening criteria apply. See Table 1. (Character)
TRS
NSTR
Maximum number of constraints to be retained per region per load case. See
Remark 3. (Integer > 0; Default = 20)
Remarks:
1.
Displacement and displacement derivative constraints associated with one particular load
case are grouped by the specification of DRESP1 entries. From each group, a maximum
of NSTR constraints are retained per load case.
Default Region
RTYPE Specification
DISP
STRESS
STRAIN
FORCE
EQUA
CSTRESS
CSTRAIN
CFAILURE
FLUTTER
FRDISP
Default Region
RTYPE Specification
DRESP1
Property ID
Property ID
Property ID
DRESP2
Property ID
Property ID
Property ID
DRESP1
DRESP1
FRVELO
FRACCL
FRSTRE
FRFORC
FRSPC
TACCL
TDISP
TVELO
TSTRE
TFORC
TSPC
(Continued)
A-37
DRESP1
DRESP1
Property ID
Property ID
DRESP1
DRESP1
DRESP1
DRESP1
Property ID
Property ID
DRESP1
DSCREEN
Stress-strain constraints are grouped by the property; i.e., all elements belonging to the
set of PIDs specified under ATTi on a DRESPi entry are regarded as belonging to the
same region. In superelement sensitivity analysis, if the property (PID) is defined in more
than one superelement, then separate regions are defined. A particular stress constraint
specification may be applied to many elements in a region generating many stress constraints, but only up to NSTR constraints per load case will be retained.
3.
4.
If a certain type of constraint exists but no corresponding DSCREEN entry is specified, all
the screening criteria used for this type of constraint will be furnished by the default
values.
5.
Constraints can be retained only if they are greater than TRS. See the Remarks under
the DCONSTR entry for a definition of constraint value.
6.
A-38
DTABLE
Table Constants
DTABLE a table of real constants that are used in equations (see DEQATN entry).
Defines
Table Constants
Format:
1
DTABLE
LABL1
VALU1
LABL2
VALU2
LABL3
VALU3
LABL4
VALU4
LABL5
VALU5
LABL6
VALU6
-etc.-
PI
3.142
10.1
1.0E6
0.1
5.5E5
100.
Example:
DTABLE
Field
Contents
LABLi
VALUi
Remarks:
1.
Only one DTABLE entry may be specified in the Bulk Data Section.
2.
A-39
10
DVBSHAP
DVBSHAP
Design Variable
Boundary Shapes
Associates a design variable identification number to a linear combination
of to
boundary
shape
vectors from a particular auxiliary model.
Format:
1
DVBSHAP
DVID
AUXMOD
COL1
SF1
COL2
SF2
COL3
SF3
1.6
10
Example:
DVBSHAP
Field
Contents
DVID
AUXMOD
COLi
SFi
Scaling factor for load sequence identification number. (Real; Default = 1.0)
Remarks:
1.
2.
Multiple references to the same DVID and/or COLi will result in the vector addition of the
referenced boundary shape vectors.
3.
A-40
DVGRID
DVGRID the relationship between design variables and grid point locations.
Design Variable to Grid Point Relation
Defines
Format:
1
DVGRID
DVID
GID
CID
COEFF
N1
N2
N3
108
0.2
0.5
0.3
1.0
10
Example:
DVGRID
Field
Contents
DVID
GID
Grid point (GRID) or geometric point (POINT) identification number. (Integer > 0)
CID
COEFF
Ni
Remarks:
1.
A CID of zero or blank (the default) references the basic coordinate system.
2.
Multiple references to the same grid ID and design variable result in vectorial addition of
the participation vectors defined by CID, COEFF, and Ni. There is no restriction on the
number of DVGRID entries that may reference a given grid (GID) or design variable
(DVID).
3.
{ g }i { g }i =
The vector { N } = [ N x N y N z ] is determined from CID and Ni. Note that it is a change
0
in a design variable from its initial value X , and not the absolute value of the design
0
A-41
The DVGRID entry defines the participation coefficients (basis vectors) of each design
variable for each of the coordinates affected by the design process in the relationship
{ g }i =
5.
DVGRID
j { T }ij X j
DVGRID entries that reference grid points on MPCs or RSSCON entries produce incorrect
sensitivities. Often the sensitivities are 0.0 which may result in a warning message indicating zero gradients which may be followed by UFM 6499. Other rigid elements produce
correct results.
A-42
DVPREL1
DVPREL1 the relation between an analysis model property and design variables.
Design Variable to Property Relation
Defines
Format:
1
DVPREL1
ID
TYPE
PID
FID
PMIN
PMAX
C0
DVID1
COEF1
DVID2
COEF2
DVID3
-etc.-
12
PBAR
612
0.2
3.0
0.25
20
20.0
0.3
10
Example:
DVPREL1
Field
Contents
ID
TYPE
PID
FlD
Field position of the property entry, or word position in the element property table
of the analysis model. (Integer 0)
PMIN
Minimum value allowed for this property. If FID references a stress recovery
location, then the default value for PMIN is 1.0+35. PMIN must be explicitly set
to a negative number for properties that may be less than zero (for example, field
ZO on the PCOMP entry). (Real; Default = 1.0E20)
PMAX
C0
DVIDi
COEFi
Remarks:
1.
The relationship between the analysis model property and design variables is given by:
P i = C0 +
2.
i COEFi DVIDi
A-43
DVPREL1
PTYPE = PBEND is not supported, either directly through FIDs or indirectly via word positions in the element property table.
4.
FID may be either a positive or a negative number. If FID > 0, it identifies the field position
on a property entry. If FID < 0, it identifies the word position of an entry in the element
property table. For example, to specify the area of a PBAR, either FID = +4 or FID = 3
can be used. However, if PTYPE = PBEAM, FID must be negative. See the following
element property table for the word positions for PBEAM.
Word
Type
Item
1
2
3
4
5
6
7
8 through 13
14 through 21
22
23
24 through 29
30 through 37
38
39
40 through 45
46 through 53
54
55
56 through 61
62 through 69
70
71
72 through 77
78 through 85
86
87
88 through 93
94 through 101
I
I
I
I
I
I
R
R
R
I
R
R
R
I
R
R
R
I
R
R
R
I
R
R
R
I
R
R
R
Property ID
Material ID
Number of segments
Constant cross section flag (1 = yes, 0 = no)
Unused
Stress output request flag, SO (1 = yes, 0 = no)
X/XB ratio; at end A, X/XB = 0.0
A, I1, I2, I12, J, NSM
C1, C2, D1, D2, E1, E2, F1, F2
Repeat of words 6 through 21 for the 1st intermediate
station
(Continued)
A-44
DVPREL1
Word
Type
102
103
104 through 109
110 through 117
I
R
R
R
118
119
120 through 125
126 through 133
134
135
136 through 141
142 through 149
150
151
152 through 157
158 through 165
166
167
168 through 173
174 through 181
182 through 189
190 through 197
I
R
R
R
I
R
R
R
I
R
R
R
I
R
R
R
R
R
Item
End B
K1, K2, S1, S2, NSI(A), NSI(B), CW(A), CW(B)
M1(A), M2(A), M1(B), M2(B), N1(A), N2(A), N1(B),
N2(B)
A-45
DVPREL2
DVPREL2
Variable
Relation
Defines the relation between an analysis model property and design Design
variables
withtoaProperty
user-supplied
equation.
Format:
1
DVPREL2
ID
TYPE
PID
FID
PMIN
PMAX
EQID
DESVAR
DVID1
DVID2
DVID3
-etc.-
DTABLE
LABL1
LABL2
LABL3
-etc.-
13
PBAR
712
0.2
DESVAR
11
13
DTABLE
PI
YM
10
Example:
DVPREL2
Field
Contents
ID
TYPE
PID
FID
PMIN
Minimum value allowed for this property. If FID references a stress recovery
location field, then the default value for PMIN is 1.0+35. PMIN must be
explicitly set to a negative number for properties that may be less than zero (for
example, field ZO on the PCOMP entry). (Real; Default = 1.E20)
PMAX
EQID
DESVAR
DESVAR flag. Indicates that the IDs of DESVAR entries follow. (Character)
DVIDi
DTABLE
DTABLE flag. Indicates that the IDs for the constants in a DTABLE entry follow.
This field may be omitted if there are no constants involved in this relation.
(Character)
LABLi
A-46
DVPREL2
Remarks:
1.
The variables identified by DVIDi and LABLi correspond to variable names (x1, x2, etc.)
listed in the left-hand side of the first equation on the DEQATN entry identified by EQID.
The variable names x1 through xN (where N = m + n) are assigned in the order DVID1,
DVID2, ..., DVIDn, LABL1, LABL2, ..., LABLm.
2.
If both DESVAR and DTABLE are specified in field 2, DESVAR must appear first.
3.
FID may be either a positive or a negative number. If FID > 0, it identifies the field position
on a property entry. If FID < 0, it identifies the word position of an entry in EPT. For
example, to specify the area of a PBAR, either FID = +4 or FID = 3 may be used. However, if PTYPE = PBEAM, FID must be negative. See Remark 4 on the DVPREL1 entry
description for specification of the PBEAM element property tables.
4.
PTYPE = PBEND is not supported, either directly through FIDs or indirectly via word positions in the element property table.
A-47
DVSHAP
DVSHAP
Variable
to Basis
Vector(s)
Defines a shape basis vector by relating a design variable identificationDesign
number
(DVID)
to columns
of a displacement matrix.
Format:
1
DVSHAP
DVID
COL1
SF1
COL2
SF2
COL3
SF3
2.0
1.0
10
Example:
DVSHAP
Field
Contents
DVID
COLi
SFi
Scaling factor applied to the COLi-th column of the displacement matrix. (Real;
Default = 1.0)
Remarks:
1.
2.
3.
Multiple references to the same DVID and/or COLi will result in a linear combination of
displacement vectors. In the example above, the shape basis vector is a linear combination of the fourth column and twice the second column.
4.
The displacement matrix must have been created by MSC/NASTRAN and be available on
a database, which is attached via the DBLOCATE FMS statement shown below:
ASSIGN
DBLOCATE
A-48
MODTRAK
Format:
1
MODTRAK
SID
10
Example:
MODTRAK
100
26
0.80
Field
Contents
SID
Sets identification number that is selected in the Case Control Section with the
MODTRAK command. (Integer; No Default) See Remark 1.
LOWRNG
HIGHRNG
MTFILTER
(Integer 0,
Remarks:
1.
Only the designed modes for the subcase will be tracked. A designed mode is one that is
used in the design model (in connection with either objective or constraints) and, therefore, identified on a DRESP1 entry.
2.
The range of modes LOWRNG through HIGHRNG, inclusive, will be used to track the designed modes. If LOWRNG and HIGHRNG are both blank, then all computed modes will
be used to search for the designed modes. Since large numbers of computed modes will
result in higher computational costs, limiting the search range with LOWRNG and HIGHRNG is recommended.
3.
Modes are considered to correlate if their mass normalized cross orthogonalities are
greater than MTFILTER.
A-49