MSC.

Nastran Version 68
Basic Dynamic Analysis
User’s Guide

Main Index
 Corporate
MSC.Software Corporation
2 MacArthur Place
Santa Ana, CA 92707 USA
Telephone: (800) 345-2078
Fax: (714) 784-4056

Europe
MSC.Software GmbH
Am Moosfeld 13
81829 Munich, Germany
Telephone: (49) (89) 43 19 87 0
Fax: (49) (89) 43 61 71 6

Asia Pacific
MSC.Software Japan Ltd.
Shinjuku First West 8F
23-7 Nishi Shinjuku
1-Chome, Shinjyku-Ku, Tokyo 160-0023, Japan
Telephone: (81) (03) 6911 1200
Fax: (81) (03) 6911 1201

Worldwide Web
www.mscsoftware.com

Disclaimer
This documentation, as well as the software described in it, is furnished under license and may be used only in accordance
with the terms of such license.
MSC.Software Corporation reserves the right to make changes in specifications and other information contained in this
document without prior notice.
The concepts, methods, and examples presented in this text are for illustrative and educational purposes only, and are not
intended to be exhaustive or to apply to any particular engineering problem or design. MSC.Software 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.
User Documentation: Copyright  2004 MSC.Software Corporation. Printed in U.S.A. All Rights Reserved.
This notice shall be marked on any reproduction of this documentation, in whole or in part. Any reproduction or
distribution of this document, in whole or in part, without the prior written consent of MSC.Software Corporation is
prohibited.
The software described herein may contain certain third-party software that is protected by copyright and licensed from
MSC.Software suppliers.
MSC, MSC/, MSC., MSC.Dytran, MSC.Fatigue, MSC.Marc, MSC.Patran, MSC.Patran Analysis Manager, MSC.Patran
CATXPRES, MSC.Patran FEA, MSC.Patran Laminate Modeler, MSC.Patran Materials, MSC.Patran Thermal, MSC.Patran
Queue Manager and PATRAN are trademarks or registered trademarks of MSC.Software Corporation in the United
States and/or other countries.
NASTRAN is a registered trademark of NASA. PAM-CRASH is a trademark or registered trademark of ESI Group.
SAMCEF is a trademark or registered trademark of Samtech SA. LS-DYNA is a trademark or registered trademark of
Livermore Software Technology Corporation. ANSYS is a registered trademark of SAS IP, Inc., a wholly owned
subsidiary of ANSYS Inc. ABAQUS is a registered trademark of ABAQUS Inc. ACIS is a registered trademark of Spatial
Technology, Inc. CATIA is a registered trademark of Dassault Systemes, SA. EUCLID is a registered trademark of Matra
Datavision Corporation. FLEXlm is a registered trademark of GLOBEtrotter Software, Inc. HPGL is a trademark of
Hewlett Packard. PostScript is a registered trademark of Adobe Systems, Inc. PTC, CADDS and Pro/ENGINEER are
trademarks or registered trademarks of Parametric Technology Corporation or its subsidiaries in the United States and/or
other countries.Unigraphics, Parasolid and I-DEAS are registered trademarks of Electronic Data Systems Corporation or
its subsidiaries in the United States and/or other countries. All other brand names, product names or trademarks belong
to their respective owners.

Main Index
 C O N T E N T S
MSC.Nastran Basic Dynamics User’s Guide

MSC.Nastran Basic Dynamics

User’s Guide

Preface ■ About this Book, xvi

■ Acknowledgements, xvii
■ List of MSC.Nastran Books, xviii
■ Technical Support, xix
■ Internet Resources, xxii
■ Permission to Copy and Distribute MSC Documentation, xxiii

1
sFundamentals of
Dynamic Analysis
■ Overview, 2
■ Equations of Motion, 3
■ Dynamic Analysis Process, 12
■ Dynamic Analysis Types, 14

2
Finite Element ■ Overview, 16
Input Data
■ Mass Input, 17
■ Damping Input, 23
■ Units in Dynamic Analysis, 27
■ Direct Matrix Input, 29
❑ Direct Matrix Input, 29
❑ DMIG Bulk Data User Interface, 30
❑ DMIG Case Control User Interface, 31

3
Real Eigenvalue ■ Overview, 34
Analysis
■ Reasons to Compute Normal Modes, 36
■ Overview of Normal Modes Analysis, 37
■ Methods of Computation, 43
❑ Lanczos Method, 43
❑ Givens and Householder Methods, 44
❑ Modified Givens and Modified Householder Methods, 44
Main Index
 ❑ Automatic Givens and Automatic Householder Methods, 44
❑ Inverse Power Method, 45
❑ Sturm Modified Inverse Power Method, 45
■ Comparison of Methods, 46
■ User Interface for Real Eigenvalue Analysis, 48
❑ User Interface for the Lanczos Method, 48
❑ User Interface for the Other Methods, 50

■ Solution Control for Normal Modes Analysis, 54

❑ Executive Control Section, 54
❑ Case Control Section, 54
❑ Bulk Data Section, 55

■ Examples, 56
❑ Two-DOF Model, 57
❑ Cantilever Beam Model, 60
❑ Bracket Model, 66
❑ Car Frame Model, 71
❑ Test Fixture Model, 76
❑ Quarter Plate Model, 78
❑ DMIG Example, 81

4
Rigid-body Modes ■ Overview, 88
■ SUPORT Entry, 90
❑ Treatment of SUPORT by Eigenvalue Analysis Methods, 90
❑ Theoretical Considerations, 91
❑ Modeling Considerations, 94

■ Examples, 96
❑ Unconstrained Beam Model, 96
❑ Unconstrained Bracket Example, 101

5
Frequency ■ Overview, 104
Response
■ Direct Frequency Response Analysis, 106
Analysis
❑ Damping in Direct Frequency Response, 106

■ Modal Frequency Response Analysis, 108

❑ Damping in Modal Frequency Response, 109
❑ Mode Truncation in Modal Frequency Response Analysis, 112
❑ Dynamic Data Recovery in Modal Frequency Response Analysis, 112

■ Modal Versus Direct Frequency Response, 114

■ Frequency-Dependent Excitation Definition, 115
❑ Frequency-Dependent Loads – RLOAD1 Entry, 115
❑ Frequency-Dependent Loads – RLOAD2 Entry, 116
❑ Spatial Distribution of Loading -- DAREA Entry, 117
❑ Time Delay – DELAY Entry, 117
❑ Phase Lead – DPHASE Entry, 118

Main Index
 ❑ Dynamic Load Tabular Function -- TABLEDi Entries, 118
❑ DAREA Example, 121
❑ Static Load Sets – LSEQ Entry, 122
❑ LSEQ Example, 123
❑ Dynamic Load Set Combination – DLOAD, 124
■ Solution Frequencies, 125
❑ FREQ, 125
❑ FREQ1, 126
❑ FREQ2, 126
❑ FREQ3, 126
❑ FREQ4, 127
❑ FREQ5, 128

■ Frequency Response Considerations, 129

■ Solution Control for Frequency Response Analysis, 130
■ Examples, 133
❑ Two-DOF Model, 133
❑ Cantilever Beam Model, 139
❑ Bracket Model, 145

6
Transient ■ Overview, 150
Response
■ Direct Transient Response Analysis, 151
Analysis
❑ Damping in Direct Transient Response, 152
❑ Initial Conditions in Direct Transient Response, 154

■ Modal Transient Response Analysis, 156

❑ Damping in Modal Transient Response Analysis, 157
❑ Mode Truncation in Modal Transient Response Analysis, 160
❑ Dynamic Data Recovery in Modal Transient Response Analysis, 161

■ Modal Versus Direct Transient Response, 162

■ Transient Excitation Definition, 163
❑ Time-Dependent Loads -- TLOAD1 Entry, 163
❑ Time-Dependent Loads – TLOAD2 Entry, 165
❑ Spatial Distribution of Loading – DAREA Entry, 166
❑ Time Delay -- DELAY Entry, 166
❑ Dynamic Load Tabular Function – TABLEDi Entries, 166
❑ DAREA Example, 169
❑ Static Load Sets -- LSEQ Entry, 170
❑ LSEQ Example, 171
❑ Dynamic Load Set Combination -- DLOAD, 172

■ Integration Time Step, 174

■ Transient Excitation Considerations, 175
■ Solution Control for Transient Response Analysis, 176
■ Examples, 179
❑ Two-DOF Model, 179
❑ Cantilever Beam Model, 182
❑ Bracket Model, 190

Main Index
 7
Enforced Motion ■ Overview, 196
■ The Large Mass Method in Direct Transient and Direct Frequency
Response, 197
■ The Large Mass Method in Modal Transient and Modal Frequency
Response, 199
■ User Interface for the Large Mass Method, 201
❑ Frequency Response, 202
❑ Transient Response, 203

■ Examples, 204
❑ Two-DOF Model, 204
❑ Cantilever Beam Model, 211

8
Restarts In ■ Overview, 218
Dynamic Analysis
■ Automatic Restarts, 219
■ Structure of the Input File, 220
■ User Interface, 221
❑ Cold Start Run, 221
❑ Restart Run, 221

■ Determining the Version for a Restart, 226

■ Examples, 228

9
Plotted Output ■ Overview, 242
■ Structure Plotting, 243
■ X-Y Plotting, 249

10
Guidelines for ■ Overview, 288
Effective Dynamic
■ Overall Analysis Strategy, 289
Analysis
■ Units, 292
■ Mass, 293
■ Damping, 294
■ Boundary Conditions, 297
■ Loads, 298
■ Meshing, 299
■ Eigenvalue Analysis, 301
■ Frequency Response Analysis, 302
Main Index
 ❑ Number of Retained Modes, 302
❑ Size of the Frequency Increment, 302
❑ Relationship of Damping to the Frequency Increment, 302
❑ Verification of the Applied Load, 303
■ Transient Response Analysis, 304
❑ Number of Retained Modes, 304
❑ Size of the Integration Time Step, 304
❑ Duration of the Computed Response, 305
❑ Value of Damping, 305
❑ Verification of the Applied Load, 305

■ Results Interpretation and Verification, 306

■ Computer Resource Requirements, 308

11
Advanced ■ Overview, 312
Dynamic Analysis
■ Dynamic Reduction, 313
Capabilities
■ Complex Eigenvalue Analysis, 314
■ Response Spectrum Analysis, 315
■ Random Vibration Analysis, 317
■ Mode Acceleration Method, 318
■ Fluid Structure Interaction, 319
❑ Hydroelastic Analysis, 319
❑ Virtual Fluid Mass, 319
❑ Coupled Acoustics, 319
❑ Uncoupled Acoustics, 320

■ Nonlinear Transient Response Analysis, 321

❑ Geometric Nonlinearity, 321
❑ Material Nonlinearity, 321
❑ Contact, 322
❑ Nonlinear-Elastic Transient Response Analysis, 322
❑ Nonlinear Normal Modes Analysis, 324

■ Superelement Analysis, 325

❑ Component Mode Synthesis, 325

■ Design Optimization and Sensitivity, 326

■ Control System Analysis, 328
■ Aeroelastic Analysis, 329
❑ Aerodynamic Flutter, 329

■ DMAP, 331

A
Glossary of Terms ■ Glossary of Terms, 334

Main Index
 B
Nomenclature for ■ Nomenclature for Dynamic Analysis, 338
Dynamic Analysis ❑ General, 338
❑ Structural Properties, 339
❑ Multiple Degree-of-Freedom System, 340

C
The Set Notation ■ Overview, 342
System Used in ❑ Displacement Vector Sets, 342
Dynamic Analysis

D
Solution ■ Overview, 348
Sequences for ❑ Structured Solution Sequences for Basic Dynamic Analysis, 348
Dynamic Analysis ❑ Rigid Formats for Basic Dynamic Analysis, 348

E
Case Control ■ Overview, 350
Commands for ❑ Input Specification, 350
Dynamic Analysis ❑ Analysis Specification, 350
❑ Output Specification, 350

E
Case Control ACCELERATION 352
Commands for B2GG 354
Dynamic Analysis BC 354
DISPLACEMENT 355
DLOAD 357
FREQUENCY 358
IC 358
K2GG 360
M2GG 361
METHOD 361
MODES 362
OFREQUENCY 363
OLOAD 365
OTIME 367
SACCELERATION 368
SDAMPING 369
SDISPLACEMENT 370
SUPORT1 371
SVECTOR 371
SVELOCITY 372
TSTEP 373
VELOCITY 374

Main Index
 F
Bulk Data Entries ■ Overview, 378
for Dynamic
Analysis

F
Bulk Data Entries CDAMP1 381
for Dynamic CDAMP2 381
Analysis CDAMP3 382
CDAMP4 383
CMASS1 384
CMASS2 385
CMASS3 386
CMASS4 386
CONM1 387
CONM2 388
CVISC 389
DAREA 390
DELAY 391
DLOAD 392
DMIG 393
DPHASE 396
EIGR 396
EIGRL 400
FREQ 404
FREQ1 404
FREQ2 405
FREQ3 406
FREQ4 408
FREQ5 410
LSEQ 411
PDAMP 412
PMASS 413
PVISC 413
RLOAD1 414
RLOAD2 416
SUPORT 418
SUPORT1 419
TABDMP1 420
TABLED1 422
TABLED2 423
TABLED3 424
TABLED4 426
TIC 426
TLOAD1 427
TLOAD2 429
TSTEP 432

Main Index
 G
Parameters for ■ Overview, 436
Dynamic Analysis

H
File Management ■ Overview, 444
Section ■ Definitions, 445
■ MSC.Nastran Database, 446
■ File Management Commands, 447
❑ INIT, 447
❑ ASSIGN, 448
❑ EXPAND, 450
❑ INCLUDE, 452
❑ Summary, 452

I
Grid Point Weight ■ Overview, 454
Generator ■ Commonly Used Features, 455
■ Example with Direction Dependent Masses, 458

J
Numerical ■ Overview, 470
Accuracy ❑ Linear Equation Solution, 470
Considerations ❑ Eigenvalue Analysis, 470
❑ Matrix Conditioning, 471
❑ Definiteness of Matrices, 472
❑ Numerical Accuracy Issues, 472
❑ Sources of Mechanisms, 473
❑ Sources of Nonpositive Definite Matrices, 474
❑ Detection and Avoidance of Numerical Problems, 474

K
Diagnostic ■ Overview, 480
Messages for
Dynamic Analysis

L
References and ■ Overview, 500
Bibliography ■ General References, 501
■ Bibliography, 502
❑ DYNAMICS – GENERAL, 502

Main Index
 ❑ DYNAMICS – ANALYSIS / TEST CORRELATION, 513
❑ DYNAMICS – COMPONENT MODE SYNTHESIS, 518
❑ DYNAMICS – DAMPING, 522
❑ DYNAMICS – FREQUENCY RESPONSE, 523
❑ DYNAMICS – MODES, FREQUENCIES, AND VIBRATIONS, 524
❑ DYNAMICS – RANDOM RESPONSE, 536
❑ DYNAMICS – REDUCTION METHODS, 537
❑ DYNAMICS – RESPONSE SPECTRUM, 538
❑ DYNAMICS – SEISMIC, 538
❑ DYNAMICS – TRANSIENT ANALYSIS, 540

INDEX MSC.Nastran Basic Dynamics User’s Guide, 543

543

Main Index
 Page xiv

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

Preface

■ About this Book

■ List of MSC.Nastran Books
■ Technical Support
■ Internet Resources
■ Permission to Copy and Distribute MSC Documentation

Main Index
xvi

About this Book

The MSC.Nastran Basic Dynamics User’s Guide is a guide to the proper use of MSC.Nastran for
solving various dynamic analysis problems. This guide serves as both an introduction to
dynamic analysis for the new user and a reference for the experienced user. The major emphasis
focuses on understanding the physical processes in dynamics and properly applying
MSC.Nastran to model dynamic processes while restricting mathematical derivations to a
minimum.
The basic types of dynamic analysis capabilities available in MSC.Nastran are described in this
guide. These common dynamic analysis capabilities include normal modes analysis, transient
response analysis, frequency response analysis, and enforced motion. These capabilities are
described and illustrative examples are presented. Theoretical derivations of the mathematics
used in dynamic analysis are presented only as they pertain to the proper understanding of the
use of each capability.
To effectively use this guide, it is important for you to be familiar with MSC.Nastran’s static
analysis capability and the principles of dynamic analysis. Basic finite element modeling and
analysis techniques are covered only as they pertain to MSC.Nastran dynamic analysis. For
more information on static analysis and modeling, refer to the MSC.Nastran Linear Static Analysis
User’s Guide and to the Getting Started with MSC.Nastran User’s Guide.
This guide is an update to the MSC.Nastran Basic Dynamics User’s Guide for Version 68, which
borrowed much material from the MSC.Nastran Handbook for Dynamic Analysis. However, not
all topics covered in that handbook are covered here. Dynamic reduction, response spectrum
analysis, random response analysis, complex eigenvalue analysis, nonlinear analysis, control
systems, fluid-structure coupling and the Lagrange Multiplyer Method will be covered in the
MSC.Nastran Advanced Dynamic Analysis User’s Guide.

Main Index
 CHAPTER xvii
Preface

Acknowledgements
Much of the basis of this guide was established by Michael Gockel with the MSC.Nastran
Handbook for Dynamic Analysis. This guide has used that information as a starting point. John
Muskivitch began the guide and was the primary contributor. Other major contributors
included David Bella, Franz Brandhuber, Michael Gockel, and John Lee. Other technical
contributors included Dean Bellinger, John Caffrey, Louis Komzsik, Ted Rose, and Candace
Hoecker. Ken Ranger and Sue Rice provided the bracket and test fixture models, respectively,
and John Furno helped to run those models.
This guide benefitted from intense technical scrutiny by Brandon Eby, Douglas Ferg,
John Halcomb, David Herting, Wai Ho, Erwin Johnson, Kevin Kilroy, Mark Miller, and William
Rodden. Gert Lundgren of LAPCAD Engineering provided the car model. Customer feedback
was received from Mohan Barbela of Martin-Marietta Astro-Space, Dr. Robert Norton of Jet
Propulsion Laboratory, Alwar Parthasarathy of SPAR Aerospace Limited, Canada, and Manfred
Wamsler of Mercedes-Benz AG. The efforts of all have made this guide possible, and their
contributions are gratefully acknowledged.
Grant Sitton
June 1997

Main Index
xviii

List of MSC.Nastran Books

Below is a list of some of the MSC.Nastran documents. You may order any of these documents
from the MSC.Software BooksMart site at www.engineering-e.com.

Installation and Release Guides

❏ Installation and Operations Guide
❏ Release Guide

Reference Books
❏ Quick Reference Guide
❏ DMAP Programmer’s Guide
❏ Reference Manual

User’s Guides
❏ Getting Started
❏ Linear Static Analysis
❏ Basic Dynamic Analysis
❏ Advanced Dynamic Analysis
❏ Design Sensitivity and Optimization
❏ Thermal Analysis
❏ Numerical Methods
❏ Aeroelastic Analysis
❏ Superelement
❏ User Modifiable
❏ Toolkit

Main Index
 CHAPTER xix
Preface

Technical Support
For help with installing or using an MSC.Software product, contact your local technical support
services. Our technical support provides the following services:

• Resolution of installation problems

• Advice on specific analysis capabilities
• Advice on modeling techniques
• Resolution of specific analysis problems (e.g., fatal messages)
• Verification of code error.
If you have concerns about an analysis, we suggest that you contact us at an early stage.
You can reach technical support services on the web, by telephone, or e-mail:

Web Go to the MSC.Software website at www.mscsoftware.com, and click on Support. Here, you can
find a wide variety of support resources including application examples, technical application
notes, available training courses, and documentation updates at the MSC.Software Training,
Technical Support, and Documentation web page.

Phone United States Frimley, Camberley

and Telephone: (800) 732-7284 Surrey, United Kingdom
Fax Fax: (714) 784-4343 Telephone: (44) (1276) 67 10 00
Fax: (44) (1276) 69 11 11
Munich, Germany Tokyo, Japan
Telephone: (49) (89) 43 19 87 0 Telephone: (81) (3) 3505 02 66
Fax: (49) (89) 43 61 71 6 Fax: (81) (3) 3505 09 14
Rome, Italy Paris, France
Telephone: (390) (6) 5 91 64 50 Telephone: (33) (1) 69 36 69 36
Fax: (390) (6) 5 91 25 05 Fax: (33) (1) 69 36 45 17

Moscow, Russia Gouda, The Netherlands

Telephone: (7) (095) 236 6177 Telephone: (31) (18) 2543700
Fax: (7) (095) 236 9762 Fax: (31) (18) 2543707

Madrid, Spain
Telephone: (34) (91) 5560919
Fax: (34) (91) 5567280

Main Index
xx

Email Send a detailed description of the problem to the email address below that corresponds to the
product you are using. You should receive an acknowledgement that your message was
received, followed by an email from one of our Technical Support Engineers.
Table 0-1
MSC.Patran Support [email protected]
MSC.Nastran Support [email protected]
MSC.Nastran for Windows Support [email protected]
MSC.visualNastran Desktop 2D Support [email protected]
MSC.visualNastran Desktop 4D Support [email protected]
MSC.Abaqus Support [email protected]
MSC.Dytran Support [email protected]
MSC.Fatigue Support [email protected]
MSC.Interactive Physics Support [email protected]
MSC.Marc Support [email protected]
MSC.Mvision Support [email protected]
MSC.SuperForge Support [email protected]
MSC Institute Course Information [email protected]

Training
The MSC Institute of Technology is the world's largest global supplier of
CAD/CAM/CAE/PDM training products and services for the product design, analysis and
manufacturing market. We offer over 100 courses through a global network of education
centers. The Institute is uniquely positioned to optimize your investment in design and
simulation software tools.
Our industry experienced expert staff is available to customize our course offerings to meet your
unique training requirements. For the most effective training, The Institute also offers many of
our courses at our customer's facilities.

The MSC Institute of Technology is located at:

2 MacArthur Place
Santa Ana, CA 92707
Phone: (800) 732-7211
Fax: (714) 784-4028
The Institute maintains state-of-the-art classroom facilities and individual computer graphics
laboratories at training centers throughout the world. All of our courses emphasize hands-on
computer laboratory work to facility skills development.
We specialize in customized training based on our evaluation of your design and simulation
processes, which yields courses that are geared to your business.
In addition to traditional instructor-led classes, we also offer video and DVD courses, interactive
multimedia training, web-based training, and a specialized instructor's program.

Main Index
 CHAPTER xxi
Preface

Course Information and Registration. For detailed course descriptions, schedule

information, and registration call the Training Specialist at (800) 732-7211 or visit
www.mscsoftware.com.

Main Index
xxii

Internet Resources
MSC.Software (www.mscsoftware.com)
MSC.Software corporate site with information on the latest events, products and services for the
CAD/CAE/CAM marketplace.
Simulation Center (simulate.engineering-e.com)
Simulate Online. The Simulation Center provides all your simulation, FEA, and other
engineering tools over the Internet.
Engineering-e.com (www.engineering-e.com)
Engineering-e.com is the first virtual marketplace where clients can find engineering expertise,
and engineers can find the goods and services they need to do their job
CATIASOURCE (plm.mscsoftware.com)
Your SOURCE for Total Product Lifecycle Management Solutions.

Main Index
 CHAPTER xxiii
Preface

Permission to Copy and Distribute MSC Documentation

If you wish to make copies of this documentation for distribution to co-workers, complete this
form and send it to MSC.Software. MSC will grant written permission if the following conditions
are met:

• All copyright notices must be included on all copies.

• Copies may be made only for fellow employees.
• No copies of this manual, or excerpts thereof, will be given to anyone who is not an
employee of the requesting company.

Please complete and mail to MSC for approval:

MSC.Software
Attention: Legal Department
2 MacArthur Place
Santa Ana, CA 92707

Name:_____________________________________________________________
Title: ______________________________________________________________
Company: _________________________________________________________
Address:___________________________________________________________
__________________________________________________________________
Telephone:_________________Email: __________________________________
Signature:______________________________ Date: ______________________

Please do not write below this line.

APPROVED: MSC.Software Corporation

Name:_____________________________________________________________
Title: ______________________________________________________________
Signature:______________________________ Date: ______________________

Main Index
xxiv

Fold here

_______________________
Place
_______________________
Stamp
_______________________ Here

MSC.Software Corporation
Attention: Legal Department
2 MacArthur Place
Santa Ana, CA 92707

Fold here

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Fundamentals of Dynamic Analysis
1
■ Overview
■ Equations of Motion
■ Dynamic Analysis Process
■ Dynamic Analysis Types

Main Index
2

1.1 Overview
In static structural analysis, it is possible to describe the operation of MSC.Nastran without a
detailed discussion of the fundamental equations. Due to the several types of dynamic analyses
and the different mathematical form of each, some knowledge of both the physics of dynamics
and the manner in which the physics is represented is important to using MSC.Nastran
effectively and efficiently for dynamic analysis.

You should become familiar with the notation and terminology covered in this chapter. This
knowledge will be valuable to understand the meaning of the symbols and the reasons for the
procedures employed in later chapters. “References and Bibliography” on page 499 provides
a list of references for structural dynamic analysis.

Dynamic Analysis Versus Static Analysis. Two basic aspects of dynamic analysis differ
from static analysis. First, dynamic loads are applied as a function of time. Second, this
time-varying load application induces time-varying response (displacements, velocities,
accelerations, forces, and stresses). These time-varying characteristics make dynamic analysis
more complicated and more realistic than static analysis.
This chapter introduces the equations of motion for a single degree-of-freedom dynamic system
(see “Equations of Motion” on page 3), illustrates the dynamic analysis process (see “Dynamic
Analysis Process” on page 12), and characterizes the types of dynamic analyses described in this
guide (see “Dynamic Analysis Types” on page 14). Those who are familiar with these topics
may want to skip to subsequent chapters.

Main Index
 CHAPTER 1 3

1.2 Equations of Motion

The basic types of motion in a dynamic system are displacement u and the first and second
derivatives of displacement with respect to time. These derivatives are velocity and
acceleration, respectively, given below:

· du
u = ------ = v = velocity
dt
2 Eq. 1-1
·· d u dv
u = --------2- = ------ = a = acceleration
dt dt

Velocity and Acceleration. Velocity is the rate of change in the displacement with respect to
time. Velocity can also be described as the slope of the displacement curve. Similarly,
acceleration is the rate of change of the velocity with respect to time, or the slope of the velocity
curve.

Single Degree-of-Freedom System. The most simple representation of a dynamic system is a

single degree-of-freedom (SDOF) system (see Figure 1-1). In an SDOF system, the time-varying
·
displacement of the structure is defined by one component of motion u ( t ) . Velocity u ( t ) and
··
acceleration u ( t ) are derived from the displacement.

p(t)
m = mass (inertia) u(t )
b = damping (energy dissipation
k = stiffness (restoring force) m

p = applied force
k b
u = displacement of mass
·
u = velocity of mass
··
u = acceleration of mass

Figure 1-1 Single Degree-of-Freedom (SDOF) System

Dynamic and Static Degrees-of-Freedom. Mass and damping are associated with the motion
of a dynamic system. Degrees-of-freedom with mass or damping are often called dynamic
degrees-of-freedom; degrees-of-freedom with stiffness are called static degrees-of-freedom. It is
possible (and often desirable) in models of complex systems to have fewer dynamic
degrees-of-freedom than static degrees-of-freedom.
The four basic components of a dynamic system are mass, energy dissipation (damper),
resistance (spring), and applied load. As the structure moves in response to an applied load,
forces are induced that are a function of both the applied load and the motion in the individual
components. The equilibrium equation representing the dynamic motion of the system is
known as the equation of motion.

Main Index
4

Equation of Motion. This equation, which defines the equilibrium condition of the system at
each point in time, is represented as
·· ·
mu ( t ) + bu ( t ) + ku ( t ) = p ( t ) Eq. 1-2

The equation of motion accounts for the forces acting on the structure at each instant in time.
Typically, these forces are separated into internal forces and external forces. Internal forces are
found on the left-hand side of the equation, and external forces are specified on the right-hand
side. The resulting equation is a second-order linear differential equation representing the
motion of the system as a function of displacement and higher-order derivatives of the
displacement.

Inertia Force. An accelerated mass induces a force that is proportional to the mass and the
acceleration. This force is called the inertia force mu·· ( t ) .

Viscous Damping. The energy dissipation mechanism induces a force that is a function of a
dissipation constant and the velocity. This force is known as the viscous damping force bu· ( t ) .
The damping force transforms the kinetic energy into another form of energy, typically heat,
which tends to reduce the vibration.

Elastic Force. The final induced force in the dynamic system is due to the elastic resistance in
the system and is a function of the displacement and stiffness of the system. This force is called
the elastic force or occasionally the spring force ku ( t ) .

Applied Load. The applied load p ( t ) on the right-hand side of Eq. 1-2 is defined as a function
of time. This load is independent of the structure to which it is applied (e.g., an earthquake is
the same earthquake whether it is applied to a house, office building, or bridge), yet its effect on
different structures can be very different.

Solution of the Equation of Motion. The solution of the equation of motion for quantities
such as displacements, velocities, accelerations, and/or stresses—all as a function of time—is the
objective of a dynamic analysis. The primary task for the dynamic analyst is to determine the
type of analysis to be performed. The nature of the dynamic analysis in many cases governs the
choice of the appropriate mathematical approach. The extent of the information required from
a dynamic analysis also dictates the necessary solution approach and steps.
Dynamic analysis can be divided into two basic classifications: free vibrations and forced
vibrations. Free vibration analysis is used to determine the basic dynamic characteristics of the
system with the right-hand side of Eq. 1-2 set to zero (i.e., no applied load). If damping is
neglected, the solution is known as undamped free vibration analysis.

Free Vibration Analysis. In undamped free vibration analysis, the SDOF equation of motion
reduces to
··
mu ( t ) + ku ( t ) = 0 Eq. 1-3

Eq. 1-3 has a solution of the form

u ( t ) = A sin ω n t + B cos ω n t Eq. 1-4

Main Index
 CHAPTER 1 5

The quantity u ( t ) is the solution for the displacement as a function of time t . As shown in Eq. 1-
4, the response is cyclic in nature.

Circular Natural Frequency. One property of the system is termed the circular natural
frequency of the structure ωn . The subscript n indicates the “natural” for the SDOF system. In
systems having more than one mass degree of freedom and more than one natural frequency,
the subscript may indicate a frequency number. For an SDOF system, the circular natural
frequency is given by

k
ωn = ---- Eq. 1-5
m

The circular natural frequency is specified in units of radians per unit time.

Natural Frequency. The natural frequency f n is defined by

ωn
f n = ------ Eq. 1-6
2π

The natural frequency is often specified in terms of cycles per unit time, commonly cycles per
second (cps), which is more commonly known as Hertz (Hz). This characteristic indicates the
number of sine or cosine response waves that occur in a given time period (typically one second).
The reciprocal of the natural frequency is termed the period of response T n given by

1 2π
T n = ---- = ------ Eq. 1-7
fn ωn

The period of the response defines the length of time needed to complete one full cycle of
response.
In the solution of Eq. 1-4, A and B are the integration constants. These constants are determined
by considering the initial conditions in the system. Since the initial displacement of the system
·
u ( t = 0 ) and the initial velocity of the system u ( t = 0 ) are known, A and B are evaluated by
substituting their values into the solution of the equation for displacement and its first derivative
(velocity), resulting in
·
u(t = 0)
B = u(t = 0) and A = ----------------------
ωn
Eq. 1-8

These initial value constants are substituted into the solution, resulting in
·
u(0)
u ( t ) = ----------- sin ω n t + u ( 0 ) cos ω n t Eq. 1-9
ωn

Eq. 1-9 is the solution for the free vibration of an undamped SDOF system as a function of its
initial displacement and velocity. Graphically, the response of an undamped SDOF system is a
sinusoidal wave whose position in time is determined by its initial displacement and velocity as
shown in Figure 1-2.

Main Index
6

Amplitude u ( t )

Time t

Figure 1-2 SDOF System -- Undamped Free Vibrations

If damping is included, the damped free vibration problem is solved. If viscous damping is
assumed, the equation of motion becomes
·· ·
mu ( t ) + bu ( t ) + ku ( t ) = 0 Eq. 1-10

Damping Types. The solution form in this case is more involved because the amount of
damping determines the form of the solution. The three possible cases for positive values of b
are

• Critically damped
• Overdamped
• Underdamped
Critical damping occurs when the value of damping is equal to a term called critical damping
b cr . The critical damping is defined as

b cr = 2 km = 2mω n Eq. 1-11

For the critically damped case, the solution becomes

– bt ⁄ 2m
u ( t ) = ( A + Bt )e Eq. 1-12

Under this condition, the system returns to rest following an exponential decay curve with no
oscillation.

A system is overdamped when b > b cr and no oscillatory motion occurs as the structure returns
to its undisplaced position.

Underdamped System. The most common damping case is the underdamped case where
b < b cr . In this case, the solution has the form

– bt ⁄ 2m
u(t ) = e ( A sin ω d t + B cos ω d t ) Eq. 1-13
Main Index
 CHAPTER 1 7

Again, A and B are the constants of integration based on the initial conditions of the system. The
new term ω d represents the damped circular natural frequency of the system. This term is
related to the undamped circular natural frequency by the following expression:

2
ωd = ωn 1 – ζ Eq. 1-14

The term ζ is called the damping ratio and is defined by

b
ζ = ------- Eq. 1-15
b cr

The damping ratio is commonly used to specify the amount of damping as a percentage of the
critical damping.
In the underdamped case, the amplitude of the vibration reduces from one cycle to the next
following an exponentially decaying envelope. This behavior is shown in Figure 1-3. The
amplitude change from one cycle to the next is a direct function of the damping. Vibration is
more quickly dissipated in systems with more damping.
Amplitude u ( t )

Time t
Figure 1-3 Damped Oscillation, Free Vibration

The damping discussion may indicate that all structures with damping require damped free
vibration analysis. In fact, most structures have critical damping values in the 0 to 10% range,
with values of 1 to 5% as the typical range. If you assume 10% critical damping, Eq. 1-4 indicates
that the damped and undamped natural frequencies are nearly identical. This result is
significant because it avoids the computation of damped natural frequencies, which can involve
a considerable computational effort for most practical problems. Therefore, solutions for
undamped natural frequencies are most commonly used to determine the dynamic
characteristics of the system (see “Real Eigenvalue Analysis” on page 33). However, this does
not imply that damping is neglected in dynamic response analysis. Damping can be included in
other phases of the analysis as presented later for frequency and transient response (see
“Frequency Response Analysis” on page 103 and “Transient Response Analysis” on
page 149).

Main Index
8

Forced Vibration Analysis. Forced vibration analysis considers the effect of an applied load
on the response of the system. Forced vibrations analyses can be damped or undamped. Since
most structures exhibit damping, damped forced vibration problems are the most common
analysis types.
The type of dynamic loading determines the mathematical solution approach. From a numerical
viewpoint, the simplest loading is simple harmonic (sinusoidal) loading. In the undamped
form, the equation of motion becomes
··
mu ( t ) + ku ( t ) = p sin ωt Eq. 1-16

In this equation the circular frequency of the applied loading is denoted by ω . This loading
frequency is entirely independent of the structural natural frequency ω n , although similar
notation is used.
This equation of motion is solved to obtain

p⁄k
u ( t ) = A sin ω n t + B cos ω n t + ----------------------------- sin ωt Eq. 1-17
2 2
1 – ω ⁄ ωn

Initial Condition Steady-State

Solution Solution

where:
A = u· ( t = 0 ) ωp ⁄ k
---------------------- – ------------------------------------------
ωn ( 1 – ω ⁄ ω n )ω n
2 2

B = u(t = 0)

Again, A and B are the constants of integration based on the initial conditions. The third term
in Eq. 1-17 is the steady-state solution. This portion of the solution is a function of the applied
loading and the ratio of the frequency of the applied loading to the natural frequency of the
structure.
The numerator and denominator of the third term demonstrate the importance of the
relationship of the structural characteristics to the response. The numerator p ⁄ k is the static
displacement of the system. In other words, if the amplitude of the sinusoidal loading is applied
as a static load, the resulting static displacement u is p ⁄ k . In addition, to obtain the steady state
solution, the static displacement is scaled by the denominator.
The denominator of the steady-state solution contains the ratio between the applied loading
frequency and the natural frequency of the structure.

Dynamic Amplification Factor for No Damping. The term

1
-----------------------------
2 2
1 – ω ⁄ ωn

Main Index
 CHAPTER 1 9

is called the dynamic amplification (load) factor. This term scales the static response to create an
amplitude for the steady state component of response. The response occurs at the same
frequency as the loading and in phase with the load (i.e., the peak displacement occurs at the
time of peak loading). As the applied loading frequency becomes approximately equal to the
structural natural frequency, the ratio ω ⁄ ωn approaches unity and the denominator goes to zero.
Numerically, this condition results in an infinite (or undefined) dynamic amplification factor.
Physically, as this condition is reached, the dynamic response is strongly amplified relative to
the static response. This condition is known as resonance. The resonant buildup of response is
shown in Figure 1-4.
Amplitude u ( t )

Time t
Figure 1-4 Harmonic Forced Response with No Damping

It is important to remember that resonant response is a function of the natural frequency and the
loading frequency. Resonant response can damage and even destroy structures. The dynamic
analyst is typically assigned the responsibility to ensure that a resonance condition is controlled
or does not occur.
Solving the same basic harmonically loaded system with damping makes the numerical solution
more complicated but limits resonant behavior. With damping, the equation of motion becomes
·· ·
mu ( t ) + bu ( t ) + ku ( t ) = p sin ωt Eq. 1-18

Main Index
10

In this case, the effect of the initial conditions decays rapidly and may be ignored in the solution.
The solution for the steady-state response is

sin ( ωt + θ )
u ( t ) = p ⁄ k ---------------------------------------------------------------------------------- Eq. 1-19
2 2 2 2
(1 – ω ⁄ ωn ) + ( 2ζω ⁄ ω n )

The numerator of the above solution contains a term that represents the phasing of the
displacement response with respect to the applied loading. In the presence of damping, the peak
loading and peak response do not occur at the same time. Instead, the loading and response are
separated by an interval of time measured in terms of a phase angle θ as shown below:

–1 2ζω ⁄ ω n
θ = – tan ----------------------------------- Eq. 1-20
2 2
( 1 – ω ⁄ ωn )

The phase angle θ is called the phase lead, which describes the amount that the response leads
the applied force.

Note: Some texts define θ as the phase lag, or the amount that the response lags the applied
force. To convert from phase lag to phase lead, change the sign of θ in Eq. 1-19 and
Eq. 1-20.

Dynamic Amplification Factor with Damping. The dynamic amplification factor for the
damped case is

1
---------------------------------------------------------------------------------- Eq. 1-21
2 2 2 2
(1 – ω ⁄ ωn ) + ( 2ζω ⁄ ω n )

The interrelationship among the natural frequency, the applied load frequency, and the phase
angle can be used to identify important dynamic characteristics. If ω ⁄ ωn is much less than 1, the
dynamic amplification factor approaches 1 and a static solution is represented with the
displacement response in phase with the loading. If ω ⁄ ω n is much greater than 1, the dynamic
amplification factor approaches zero, yielding very little displacement response. In this case, the
structure does not respond to the loading because the loading is changing too fast for the
structure to respond. In addition, any measurable displacement response will be 180 degrees
out of phase with the loading (i.e., the displacement response will have the opposite sign from
the force). Finally if ω ⁄ ω n = 1 , resonance occurs. In this case, the magnification factor is 1 ⁄ ( 2ζ ) ,
and the phase angle is 270 degrees. The dynamic amplification factor and phase lead are shown
in Figure 1-5 and are plotted as functions of forcing frequency.

Main Index
 CHAPTER 1 11

Amplification Factor

360°
Phase Lead θ (Degrees)

180°

ωn
Forcing Frequency ω

Figure 1-5 Harmonic Forced Response with Damping

In contrast to harmonic loadings, the more general forms of loading (impulses and general
transient loading) require a numerical approach to solving the equations of motion. This
technique, known as numerical integration, is applied to dynamic solutions either with or
without damping. Numerical integration is described in “Transient Response Analysis” on
page 149.

Main Index
12

1.3 Dynamic Analysis Process

Before conducting a dynamic analysis, it is important to define the goal of the analysis prior to
the formulation of the finite element model. Consider the dynamic analysis process to be
represented by the stepsLK in Figure 1-6. The analyst must evaluate the finite element model in
terms of the type of dynamic loading to be applied to the structure. This dynamic load is known
as the dynamic environment. The dynamic environment governs the solution approach (i.e.,
normal modes, transient response, frequency response, etc.). This environment also indicates
the dominant behavior that must be included in the analysis (i.e., contact, large displacements,
etc.). Proper assessment of the dynamic environment leads to the creation of a more refined
finite element model and more meaningful results.

Dynamic
Environment

Finite Element
Model

Modal Yes Results No

Analysis? Satisfactory?
No
Yes

No Results Forced-Response
Satisfactory? Analysis

Yes

End

Figure 1-6 Overview of Dynamic Analysis Process

An overall system design is formulated by considering the dynamic environment. As part of the
evaluation process, a finite element model is created. This model should take into account the
characteristics of the system design and, just as importantly, the nature of the dynamic loading
(type and frequency) and any interacting media (fluids, adjacent structures, etc.). At this point,
the first step in many dynamic analyses is a modal analysis to determine the structure’s natural
frequencies and mode shapes (see “Real Eigenvalue Analysis” on page 33).
In many cases the natural frequencies and mode shapes of a structure provide enough
information to make design decisions. For example, in designing the supporting structure for a
rotating fan, the design requirements may require that the natural frequency of the supporting
structure have a natural frequency either less than 85% or greater than 110% of the operating
speed of the fan. Specific knowledge of quantities such as displacements and stresses are not
required to evaluate the design.
Main Index
 CHAPTER 1 13

Forced response is the next step in the dynamic evaluation process. The solution process reflects
the nature of the applied dynamic loading. A structure can be subjected to a number of different
dynamic loads with each dictating a particular solution approach. The results of a
forced-response analysis are evaluated in terms of the system design. Necessary modifications
are made to the system design. These changes are then applied to the model and analysis
parameters to perform another iteration on the design. The process is repeated until an
acceptable design is determined, which completes the design process.
The primary steps in performing a dynamic analysis are summarized as follows:
1. Define the dynamic environment (loading).
2. Formulate the proper finite element model.
3. Select and apply the appropriate analysis approach(es) to determine the behavior of the
structure.
4. Evaluate the results.

Main Index
14

1.4 Dynamic Analysis Types

This guide describes the following types of dynamic analysis that can be performed with
MSC.Nastran:

• Real eigenvalue analysis (undamped free vibrations).

• Linear frequency response analysis (steady-state response of linear structures to loads
that vary as a function of frequency).
• Linear transient response analysis (response of linear structures to loads that vary as a
function of time).
The above list comprises the basic types of dynamic analysis. Many MSC.Nastran advanced
dynamic analysis capabilities, such as shock/response spectrum analysis, random response
analysis, design sensitivity, design optimization, aeroelasticity, and component mode synthesis,
can be used in conjunction with the above analyses.
In practice, very few engineers use all of the dynamic analysis types in their work. Therefore, it
may not be important for you to become familiar with all of the types. Each type can be
considered independently, although there may be many aspects common to many of the
analyses.
Real eigenvalue analysis is used to determine the basic dynamic characteristics of a structure.
The results of an eigenvalue analysis indicate the frequencies and shapes at which a structure
naturally tends to vibrate. Although the results of an eigenvalue analysis are not based on a
specific loading, they can be used to predict the effects of applying various dynamic loads. Real
eigenvalue analysis is described in “Real Eigenvalue Analysis” on page 33.
Frequency response analysis is an efficient method for finding the steady-state response to
sinusoidal excitation. In frequency response analysis, the loading is a sine wave for which the
frequency, amplitude, and phase are specified. Frequency response analysis is limited to linear
elastic structures. Frequency response analysis is described in “Frequency Response Analysis”
on page 103.
Transient response analysis is the most general method of computing the response to
time-varying loads. The loading in a transient analysis can be of an arbitrary nature, but is
explicitly defined (i.e., known) at every point in time. The time-varying (transient) loading can
also include nonlinear effects that are a function of displacement or velocity. Transient response
analysis is most commonly applied to structures with linear elastic behavior. Transient response
analysis is described in “Transient Response Analysis” on page 149.
Additional types of dynamic analysis are available with MSC.Nastran. These types are
described briefly in “Advanced Dynamic Analysis Capabilities” on page 311 and will be
described fully in the MSC.Nastran Advanced Dynamic Analysis User’s Guide.

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Finite Element Input Data
2
■ Overview
■ Mass Input
■ Damping Input
■ Units in Dynamic Analysis

■ Direct Matrix Input

Main Index
16

2.1 Overview
When conducting an MSC.Nastran dynamic analysis, the formulation of the model depends
upon the desired results and the type of dynamic loading. For example, if stresses are of interest,
a finer finite element mesh is required than if only the system level displacement patterns are
needed. Many of the modeling considerations involved in a static analysis are similarly applied
in dynamic analysis. A dynamic analysis, however, requires additional input data which is used
to define the dynamic character of the structure.
In static analysis the stiffness properties are defined by element and material properties. These
same properties are also required for dynamic analysis along with the addition of mass and
damping. “Mass Input” on page 17 describes mass input, and “Damping Input” on page 23
describes damping input.
Correct specification of units is very important for dynamic analysis. Incorrect specification of
units is probably more difficult to diagnose in dynamic analysis than in static analysis. Because
MSC.Nastran has no built-in set of units, you must ensure their consistency (and accuracy).
“Units in Dynamic Analysis” on page 27 describes the common variables and units for dynamic
analysis.
“Direct Matrix Input” on page 29 concludes this chapter with a discussion of direct matrix
input.

Main Index
 CHAPTER 2 17
Finite Element Input Data

2.2 Mass Input

Mass input is one of the major entries in a dynamic analysis. Mass can be represented in a
number of ways in MSC.Nastran. The mass matrix is automatically computed when mass
density or nonstructural mass is specified for any of the standard finite elements (CBAR,
CQUAD4, etc.) in MSC.Nastran, when concentrated mass elements are entered, and/or when
full or partial mass matrices are entered.

Lumped and Coupled Mass. Mass is formulated as either lumped mass or coupled mass.
Lumped mass matrices contain uncoupled, translational components of mass. Coupled mass
matrices contain translational components of mass with coupling between the components. The
CBAR, CBEAM, and CBEND elements contain rotational masses in their coupled formulations,
although torsional inertias are not considered for the CBAR element. Coupled mass can be more
accurate than lumped mass. However, lumped mass is more efficient and is preferred for its
computational speed in dynamic analysis.
The mass matrix formulation is a user-selectable option in MSC.Nastran. The default mass
formulation is lumped mass for most MSC.Nastran finite elements. The coupled mass matrix
formulation is selected using PARAM,COUPMASS,1 in the Bulk Data. Table 2-1 shows the
mass options available for each element type.

Table 2-1 Element Mass Types

Element Type Lumped Mass Coupled Mass*

CBAR X X
CBEAM X X
CBEND X
CONEAX X
CONMi X X
CONROD X X
CRAC2D X X
CRAC3D X X
CHEXA X X
CMASSi X
CPENTA X X
CQUAD4 X X
CQUAD8 X X
CQUADR X X
CROD X X
CSHEAR X
CTETRA X X
Main Index
18

Table 2-1 Element Mass Types (continued)

Element Type Lumped Mass Coupled Mass*

CTRIA3 X X
CTRIA6 X X
CTRIAR X X
CTRIAX6 X X
CTUBE X X
*Couple mass is selected by PARAM,COUPMASS,1.

The MSC.Nastran coupled mass formulation is a modified approach to the classical consistent
mass formulation found in most finite element texts. The MSC.Nastran lumped mass is identical
to the classical lumped mass approach. The various formulations of mass matrices can be
compared using the CROD element. Assume the following properties:

2 (Torsion) 4 (Torsion) Length L

1 (Translation) 3 (Translation) Area A
Torsional Constant J
Young’s Modulus E
L Shear Modulus G
Mass Density ρ
Polar Moment of Inertia I ρ
1-4 are Degrees-of-Freedom

CROD Element Stiffness Matrix. The CROD element’s stiffness matrix [ K ] is given by:

1 2 3 4
AE – AE
------- 0 ----------- 0 1
L L
GJ – GJ
0 ------- 0 ---------- 2
[K] = L L Eq. 2-1
– AE AE
----------- 0 ------- 0 3
L L
– GJ GJ
0 ---------- 0 ------- 4
L 0

The zero entries in the matrix create independent (uncoupled) translational and rotational
behavior for the CROD element, although for most other elements these degrees-of-freedom are
coupled.

CROD Lumped Mass Matrix. The CROD element classical lumped mass matrix is the same as
the MSC.Nastran lumped mass matrix. This lumped mass matrix is

Main Index
 CHAPTER 2 19
Finite Element Input Data

1
--- 0 0 0
2
[ M ] = ρAL 0 0 0 0 Eq. 2-2
1
0 0 --- 0
2
0 0 0 0

The lumped mass matrix is formed by distributing one-half of the total rod mass to each of the
translational degrees-of-freedom. These degrees-of-freedom are uncoupled and there are no
torsional mass terms calculated.
The CROD element classical consistent mass matrix is

1 1
--- 0 --- 0
3 6
Iρ Iρ
0 ------- 0 -------
[ M ] = ρAL 3A 6A Eq. 2-3
1 1
--- 0 --- 0
6 3
Iρ Iρ
0 ------- 0 -------
6A 3A

This classical mass matrix is similar in form to the stiffness matrix because it has both
translational and rotational masses. Translational masses may be coupled to other translational
masses, and rotational masses may be coupled to other rotational masses. However,
translational masses may not be coupled to rotational masses.

CROD Coupled Mass Matrix. The CROD element MSC.Nastran coupled mass matrix is

5 1
------ 0 ------ 0
12 12
[ M ] = ρAL 0 0 0 0 Eq. 2-4
1 5
------ 0 ------ 0
12 12
0 0 0 0

The axial terms in the CROD element coupled mass matrix represent the average of lumped
mass and classical consistent mass. This average is found to yield the best results for the CROD
element as described below. The mass matrix terms in the directions transverse to the element
axes are lumped mass, even when the coupled mass option is selected. Note that the torsional
inertia is not included in the CROD element mass matrix.

Lumped Mass Versus Coupled Mass Example. The difference in the axial mass
formulations can be demonstrated by considering a fixed-free rod modeled with a single CROD
element (Figure 2-1). The exact quarter-wave natural frequency for the first axial mode is

Main Index
20

E⁄ρ
1.5708 ------------
l

Using the lumped mass formulation for the CROD element, the first frequency is predicted to be

E⁄ρ
1.414 ------------
l

which underestimates the frequency by 10%. Using a classical consistent mass approach, the
predicted frequency

E⁄ρ
1.732 ------------
l

is overestimated by 10%. Using the coupled mass formulation in MSC.Nastran, the frequency

E⁄ρ
1.549 ------------
l

is underestimated by 1.4%. The purpose of this example is to demonstrate the possible effects of
the different mass formulations on the results of a simple problem. Remember that not all
dynamics problems have such a dramatic difference. Also, as the model’s m
esh becomes finer, the difference in mass formulations becomes negligible.

u(t )
1

l Single Element Model

π E⁄ρ E⁄ρ
Theoretical Natural Frequency: ω n = --- ---------------- = 1.5708 ----------------
2 l l

E⁄ρ E⁄ρ
MSC.Nastran Lumped Mass: ωn = 2 ---------------- = 1.414 ----------------
l l

E⁄ρ E⁄ρ
Classical Consistent Mass: ωn = 3 ---------------- = 1.732 ----------------
l l

E⁄ρ E⁄ρ
MSC.Nastran Coupled Mass: ωn = 12 ⁄ 5 ---------------- = 1.549 ----------------
l l
Figure 2-1 Comparison of Mass Formulations for a ROD

Main Index
 CHAPTER 2 21
Finite Element Input Data

CBAR, CBEAM Lumped Mass. The CBAR element lumped mass matrix is identical to the
CROD element lumped mass matrix. The CBEAM element lumped mass matrix is identical to
that of the CROD and CBAR mass matrices with the exception that torsional inertia is included.

CBAR, CBEAM Coupled Mass. The CBAR element coupled mass matrix is identical to the
classical consistent mass formulation except for two terms: (1) the mass in the axial direction is
the average of the lumped and classical consistent masses, as explained for the CROD element;
and (2) there is no torsional inertia. The CBEAM element coupled mass matrix is also identical
to the classical consistent mass formulation except for two terms: (1) the mass in the axial
direction is the lumped mass; and (2) the torsional inertia is the lumped inertia.
Another important aspect of defining mass is the units of measure associated with the mass
definition. MSC.Nastran assumes that consistent units are used in all contexts. You must be
careful to specify structural dimensions, loads, material properties, and physical properties in a
consistent set of units.
All mass entries should be entered in mass consistent units. Weight units may be input instead
of mass units, if this is more convenient. However, you must convert the weight to mass by
dividing the weight by the acceleration of gravity defined in consistent units:

ρ m = ( 1 ⁄ g )ρ w Eq. 2-5

where:

ρ m = mass or mass density

g = acceleration of gravity
ρ w = weight or weight density

The parameter
PARAM,WTMASS,factor

performs this conversion. The value of the factor should be entered as 1 ⁄ g . The default value
for the factor is 1.0. Hence, the default value for WTMASS assumes that mass (and mass density)
is entered, instead of weight (and weight density).
When using English units if the weight density of steel is entered as RHO = 0.3 lb/in 3 , using
PARAM,WTMASS,0.002588 converts the weight density to mass density for the acceleration of
gravity g = 386.4 in/sec 2 . The mass density, therefore, becomes 7.765E-4 lb f -sec 2 /in 4 . If the weight
density of steel is entered as RHO = 80000 N/m3 when using metric units, then using
PARAM,WTMASS,0.102 converts the weight density to mass density for the acceleration of
gravity g = 9.8 m/sec 2 . The mass density, therefore, becomes 8160 kg/m3 .
PARAM,WTMASS is used once per run, and it multiplies all weight/mass input (including
CMASSi, CONMi, and nonstructural mass input). Therefore, do not mix input type; use all mass
(and mass density) input or all weight (or weight density) input. PARAM,WTMASS does not
affect direct input matrices M2GG or M2PP (see “Direct Matrix Input” on page 29).
PARAM,CM2 can be used to scale M2GG; there is no parameter scaling for M2PP. PARAM,CM1

Main Index
22

is similar to PARAM,WTMASS since CM1 scales all weight/mass input (except for M2GG and
M2PP), but it is active only when M2GG is also used. In other words, PARAM,CM1 is used in
addition to PARAM,WTMASS if M2GG is used.

MSC.Nastran Mass Input. Mass is input to MSC.Nastran via a number of different entries.
The most common method to enter mass is using the RHO field on the MATi entry. This field is
assumed to be defined in terms of mass density (mass/unit volume). To determine the total
mass of the element, the mass density is multiplied by the element volume (determined from the
geometry and physical properties). For a MAT1 entry, a mass density for steel of
2 4
7.76E-4 lb f -sec /in is entered as follows:

1 2 3 4 5 6 7 8 9 10
$MAT1 MID E G NU RHO A TREF GE

MAT1 2 30.0E6 0.3 7.76E-4

Grid point masses can be entered using the CONM1, CONM2, and CMASSi entries. The
CONM1 entry allows input of a fully coupled 6 x 6 mass matrix. You define half of the terms,
and symmetry is assumed. The CONM2 entry defines mass and mass moments of inertia for a
rigid body. The CMASSi entries define scalar masses.

Nonstructural Mass. An additional way to input mass is to use nonstructural mass, which is
mass not associated with the geometric cross-sectional properties of an element. Examples of
nonstructural mass are insulation, roofing material, and special coating materials.
Nonstructural mass is input as mass/length for line elements and mass/area for elements with
two-dimensional geometry. Nonstructural mass is defined on the element property entry
(PBAR, for example).

Main Index
 CHAPTER 2 23
Finite Element Input Data

2.3 Damping Input

Damping is a mathematical approximation used to represent the energy dissipation observed in
structures. Damping is difficult to model accurately since it is caused by many mechanisms
including

• Viscous effects (dashpot, shock absorber)

• External friction (slippage in structural joints)
• Internal friction (characteristic of the material type)
• Structural nonlinearities (plasticity, gaps)
Because these effects are difficult to quantify, damping values are often computed based on the
results of a dynamic test. Simple approximations are often justified because the damping values
are low.

Viscous and Structural Damping. Two types of damping are generally used for linear-elastic
materials: viscous and structural. The viscous damping force is proportional to velocity, and
the structural damping force is proportional to displacement. Which type to use depends on the
physics of the energy dissipation mechanism(s) and is sometimes dictated by regulatory
standards.
The viscous damping force f v is proportional to velocity and is given by
·
f v = bu Eq. 2-6

where:

b = viscous damping coefficient

·
u = velocity

The structural damping force f s is proportional to displacement and is given by

fs = i ⋅ G ⋅ k ⋅ u Eq. 2-7

where:

G = structural damping coefficient

k = stiffness
u = displacement
i = –1 (phase change of 90 degrees)

For a sinusoidal displacement response of constant amplitude, the structural damping force is
constant, and the viscous damping force is proportional to the forcing frequency. Figure 2-2
depicts this and also shows that for constant amplitude sinusoidal motion the two damping
forces are equal at a single frequency.

Main Index
24

At this frequency,

Gk
G k = b ω∗ or b = ------- Eq. 2-8
ω∗

where ω∗ is the frequency at which the structural and viscous damping forces are equal for a
constant amplitude of sinusoidal motion.

Viscous Damping
·
f v = bu = i b ω u
Structural Damping
fs = i G k u
Damping Force f

ω∗
Forcing Frequency ω
Figure 2-2 Structural Damping and Viscous Damping Forces
for Constant Amplitude Sinusoidal Displacement

If the frequency ω∗ is the circular natural frequency ωn , Eq. 2-8 becomes

G k
b = --------- = G ω n m Eq. 2-9
ωn

Recall the definition of critical damping from Eq. 1-11

b cr = 2 km = 2mω n Eq. 2-10

Some equalities that are true at resonance ( ωn ) for constant amplitude sinusoidal displacement
are

b G
------- = ζ = ---- Eq. 2-11
b cr 2

1 1
and Q = ------ = ---- Eq. 2-12
2ζ G

where Q is the quality or dynamic magnification factor, which is inversely proportional to the
energy dissipated per cycle of vibration.

The Effect of Damping. Damping is the result of many complicated mechanisms. The effect of
damping on computed response depends on the type and loading duration of the dynamic
analysis. Damping can often be ignored for short duration loadings, such as those resulting from
Main Index
 CHAPTER 2 25
Finite Element Input Data

a crash impulse or a shock blast, because the structure reaches its peak response before
significant energy has had time to dissipate. Damping is important for long duration loadings
(such as earthquakes), and is critical for loadings (such as rotating machinery) that continually
add energy to the structure. The proper specification of the damping coefficients can be
obtained from structural tests or from published literature that provides damping values for
structures similar to yours.
As is discussed in detail in “Frequency Response Analysis” on page 103 and “Transient
Response Analysis” on page 149, certain solution methods allow specific forms of damping to
be defined. The type of damping used in the analysis is controlled by both the solution being
performed and the MSC.Nastran data entries. In transient response analysis, for example,
structural damping must be converted to equivalent viscous damping.

Structural Damping Specification. Structural damping is specified on the MATi and

PARAM,G Bulk Data entries. The GE field on the MATi entry is used to specify overall
structural damping for the elements that reference this material entry. This definition is via the
structural damping coefficient GE.
For example, the MAT1 entry:

1 2 3 4 5 6 7 8 9 10
$MAT1 MID E G NU RHO A TREF GE

MAT1 2 30.0E6 0.3 7.764E-4 0.10

specifies a structural damping coefficient of 0.1.

An alternate method for defining structural damping is through PARAM,G,r where r is the
structural damping coefficient. This parameter multiplies the stiffness matrix to obtain the
structural damping matrix. The default value for PARAM,G is 0.0. The default value causes this
source of structural damping to be ignored. Two additional parameters are used in transient
response analysis to convert structural damping to equivalent viscous damping: PARAM,W3
and PARAM,W4.
PARAM,G and GE can both be specified in the same analysis.

Viscous Damping Specification. Viscous damping is defined by the following elements:

CDAMP1 entry Scalar damper between two degrees-of-freedom (DOFs) with reference
to a PDAMP property entry.
CDAMP2 entry Scalar damper between two DOFs without reference to a property entry.
CDAMP3 entry Scalar damper between two scalar points (SPOINTs) with reference to a
PDAMP property entry.
CDAMP4 entry Scalar damper between two scalar points (SPOINTs) without reference
to a property entry.

Main Index
26

CVISC entry Element damper between two grid points with reference to a PVISC
property entry.
CBUSH entry A generalized spring-and-damper structural element that may be
nonlinear or frequency dependent. It references a PBUSH entry.

Viscous damping for modal transient response and modal frequency response is specified with
the TABDMP1 entry.
Note that GE and G by themselves are dimensionless; they are multipliers of the stiffness. The
CDAMPi and CVISC entries, however, have damping units.
Damping is further described in “Frequency Response Analysis” on page 103 and “Transient
Response Analysis” on page 149 as it pertains to frequency and transient response analyses.

Main Index
 CHAPTER 2 27
Finite Element Input Data

2.4 Units in Dynamic Analysis

Because MSC.Nastran does not assume a particular set of units, you must ensure that the units
in your MSC.Nastran model are consistent. Because there is more input in dynamic analysis
than in static analysis, it is easier to make a mistake in units when performing a dynamic
analysis. The most frequent source of error in dynamic analysis is incorrect specification of the
units, especially for mass and damping.

Table 2-2 shows typical dynamic analysis variables, fundamental and derived units, and
common English and metric units. Note that for English units all “lb” designations are lb f . The
use of “lb” for mass (i.e., lb m ) is avoided.

Table 2-2 Engineering Units for Common Variables

Common Common
Variable Dimensions*
English Units Metric Units
Length L in m
2
Mass M lb-sec ⁄ in kg

Time T sec sec

Area 2 2 2
L n m

Volume 3 3 3
L in m

Velocity –1 in / sec m / sec

LT
–2 2 2
Acceleration LT in ⁄ sec m ⁄ sec
Rotation -- rad rad

Rotational Velocity –1 rad / sec rad / sec

T
Rotational Acceleration –2 2 2
T rad ⁄ sec rad ⁄ sec

Circular Frequency –1 rad / sec rad / sec

T
Frequency –1 cps; Hz cps; Hz
T

Eigenvalue –2 2 2 2 2
T rad ⁄ sec rad ⁄ sec
Phase Angle -- deg deg

Force –2 lb N
MLT

Weight –2 lb N
MLT

Moment 2 –2 in-lb N-m

ML T

Main Index
28

Table 2-2 Engineering Units for Common Variables (continued)

Common Common
Variable Dimensions*
English Units Metric Units
Mass Density –3 3 4 3
ML lb-sec ⁄ in kg ⁄ m
–1 –2 2 2
Young’s Modulus ML T lb ⁄ in Pa; N ⁄ m
Poisson’s Ratio -- -- --
–1 –2 2 2
Shear Modulus ML T lb ⁄ in Pa; N ⁄ m

Area Moment of Inertia 4 4 4

L in m

Torsional Constant 4 4 4
L in m

Mass Moment of Inertia 2 2 2

ML in-lb-sec kg-m

Stiffness MT
–2 lb ⁄ in N/m

Viscous Damping Coefficient –1 lb-sec / in N-sec / m

MT

Stress –1 –2 2 2
ML T lb ⁄ in Pa; N ⁄ m
Strain -- -- --
* L Denotes length
M Denotes mass
T Denotes time
-- Denotes dimensionless

Main Index
 CHAPTER 2 29
Finite Element Input Data

2.5 Direct Matrix Input

The finite element approach simulates the structural properties with mathematical equations
written in matrix format. The structural behavior is then obtained by solving these equations.
Usually, all of the structural matrices are generated internally based on the information that you
provide in your MSC.Nastran model. Once you provide the grid point locations, element
connectivities, element properties, and material properties, MSC.Nastran generates the
appropriate structural matrices.

External Matrices. If structural matrices are available externally, you can input the matrices
directly into MSC.Nastran without providing all the modeling information. Normally this is not
a recommended procedure since it requires additional work on your part. However, there are
occasions where the availability of this feature is very useful and in some cases crucial. Some
possible applications are listed below:

• Suppose you are a subcontractor on a classified project. The substructure that you are
analyzing is attached to the main structure built by the main contractor. The stiffness
and mass effects of this main structure are crucial to the response of your component,
but geometry of the main structure is classified. The main contractor, however, can
provide you with the stiffness and mass matrices of the classified structure. By reading
these stiffness and mass matrices and adding them to your MSC.Nastran model, you
can account for the effect of the attached structure without compromising security.
• Perhaps you are investigating a series of design options on a component attached to an
aircraft bulkhead. Your component consists of 500 DOFs and the aircraft model
consists of 100,000 DOFs. The flexibility of the backup structure is somewhat
important. You can certainly analyze your component by including the full aircraft
model (100,500 DOFs). However, as an approximation, you can reduce the matrices for
the entire aircraft down to a manageable size using dynamic reduction (see “Advanced
Dynamic Analysis Capabilities” on page 311). These reduced mass and stiffness
matrices can then be read and added to your various component models. In this case,
you may be analyzing a 2000-DOF system, instead of a 100,500-DOF system.
• The same concept can be extended to a component attached to a test fixture. If the finite
element model of the fixture is available, then the reduced mass and stiffness matrices
of the fixture can be input. Furthermore, there are times whereby the flexibility of the
test fixture at the attachment points can be measured experimentally. The
experimental stiffness matrix is the inverse of the measured flexibility matrix. In this
instance, this experimental stiffness matrix can be input to your model.

One way of reading these external matrices is through the use of the direct matrix input feature
in MSC.Nastran.

Direct Matrix Input

The direct matrix input feature can be used to input stiffness, mass, damping, and load matrices
attached to the grid and/or scalar points in dynamic analysis. These matrices are referenced in
terms of their external grid IDs and are input via DMIG Bulk Data entries. As shown in Table 2-
3, there are seven standard kinds of DMIG matrices available in dynamic analysis.
Main Index
30

Table 2-3 Types of DMIG Matrices in Dynamics

Matrix G Type P Type

Stiffness K2GG K2PP
Mass M2GG M2PP
Damping B2GG B2PP
Load P2G –

The symbols for g-type matrices in mathematical format are [ K 2gg ] , [ M 2gg ] , [ B 2gg ] , and { P 2g }. The
three matrices K2GG, M2GG, and B2GG must be real and symmetric. These matrices are
implemented at the g-set level (see “The Set Notation System Used in Dynamic Analysis” on
page 341 for a description of the set notation for dynamic analysis). In other words, these terms
are added to the corresponding structural matrices at the specified DOFs prior to the application
of constraints (MPCs, SPCs, etc.).
The symbols for p-type matrices in standard mathematical format are [ K 2pp ] , [ M 2pp ] , and [ B 2pp ] .
The p-set is a union of the g-set and extra points. These matrices need not be real or symmetric.
The p-type matrices are used in applications such as control systems. Only the g-type DMIG
input matrices are covered in this guide.

DMIG Bulk Data User Interface

In the Bulk Data Section, the DMIG matrix is defined by a single DMIG header entry followed
by a series of DMIG data entries. Each of these DMIG data entries contains a column of nonzero
terms for the matrix.

Header Entry Format:

1 2 3 4 5 6 7 8 9 10
DMIG NAME “0" IFO TIN TOUT POLAR NCOL

Column Entry Format:

DMIG NAME GJ CJ G1 C1 A1 B1

G2 C2 A2 B2 -etc.-

Example:

DMIG STIF 0 6 1

DMIG STIF 5 3 5 3 250.

5 5 -125. 6 3 -150.

Main Index
 CHAPTER 2 31
Finite Element Input Data

Field Contents
NAME Name of the matrix.
IFO Form of matrix input:
1 = Square
9 or 2 = Rectangular
6 = Symmetric (input only the upper or lower half)
TIN Type of matrix being input:
1 = Real, single precision (one field is used per element)
2 = Real, double precision (one field per element)
3 = Complex, single precision (two fields are used per element)
4 = Complex, double precision (two fields per element)
TOUT Type of matrix to be created:
0 = Set by precision system cell (default)
1 = Real, single precision
2 = Real, double precision
3 = Complex, single precision
4 = Complex, double precision
POLAR Input format of Ai, Bi. (Integer = blank or 0 indicates real, imaginary format;
integer > 0 indicates amplitude, phase format.)
NCOL Number of columns in a rectangular matrix. Used only for IFO = 9.
GJ Grid, scalar, or extra point identification number for the column index or column
number for IFO = 9.
CJ Component number for GJ for a grid point.
Gi Grid, scalar, or extra point identification number for the row index.
Ci Component number for Gi for a grid point.
Ai, Bi Real and imaginary (or amplitude and phase) parts of a matrix element. If the
matrix is real (TIN = 1 or 2), then Bi must be blank.

DMIG Case Control User Interface

In order to include these matrices, the Case Control must contain the appropriate K2GG, M2GG,
or B2GG command. (Once again, only the g-type DMIG input matrices are included in this
guide.)

Examples
1. K2GG = mystiff
The above Case Control command adds terms that are defined by the DMIG Bulk Data
entries with the name “mystiff” to the g-set stiffness matrix.
Main Index
32

2. M2GG = yourmass
The above Case Control command adds terms that are defined by the DMIG Bulk Data
entries with the name “yourmass” to the g-set mass matrix.
3. B2GG = ourdamp
The above Case Control command adds terms that are defined by the DMIG Bulk Data
entries with the name “ourdamp” to the g-set damping matrix.
Use of the DMIG entry for inputting mass and stiffness is illustrated in one of the examples in
“Real Eigenvalue Analysis” on page 33.

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Real Eigenvalue Analysis
3
■ Overview
■ Reasons to Compute Normal Modes
■ Overview of Normal Modes Analysis
■ Methods of Computation

■ Comparison of Methods
■ User Interface for Real Eigenvalue Analysis
■ Solution Control for Normal Modes Analysis
■ Examples

Main Index
34

3.1 Overview
The usual first step in performing a dynamic analysis is determining the natural frequencies and
mode shapes of the structure with damping neglected. These results characterize the basic
dynamic behavior of the structure and are an indication of how the structure will respond to
dynamic loading.

Natural Frequencies. The natural frequencies of a structure are the frequencies at which the
structure naturally tends to vibrate if it is subjected to a disturbance. For example, the strings of
a piano are each tuned to vibrate at a specific frequency. Some alternate terms for the natural
frequency are characteristic frequency, fundamental frequency, resonance frequency, and
normal frequency.

Mode Shapes. The deformed shape of the structure at a specific natural frequency of vibration
is termed its normal mode of vibration. Some other terms used to describe the normal mode are
mode shape, characteristic shape, and fundamental shape. Each mode shape is associated with
a specific natural frequency.
Natural frequencies and mode shapes are functions of the structural properties and boundary
conditions. A cantilever beam has a set of natural frequencies and associated mode shapes
(Figure 3-1). If the structural properties change, the natural frequencies change, but the mode
shapes may not necessarily change. For example, if the elastic modulus of the cantilever beam
is changed, the natural frequencies change but the mode shapes remain the same. If the
boundary conditions change, then the natural frequencies and mode shapes both change. For
example, if the cantilever beam is changed so that it is pinned at both ends, the natural
frequencies and mode shapes change (see Figure 3-2).

z x
1

z x
2

z x
3

z x
4

Figure 3-1 The First Four Mode Shapes of a Cantilever Beam

Main Index
 CHAPTER 3 35
Real Eigenvalue Analysis

z x
1

z x
2

z x
3

z x
4

Figure 3-2 The First Four Mode Shapes of a Simply Supported Beam

Computation of the natural frequencies and mode shapes is performed by solving an eigenvalue
problem as described in “Rigid-Body Mode of a Simple Structure” on page 40. Next, we solve
for the eigenvalues (natural frequencies) and eigenvectors (mode shapes). Because damping is
neglected in the analysis, the eigenvalues are real numbers. (The inclusion of damping makes
the eigenvalues complex numbers; see “Advanced Dynamic Analysis Capabilities” on
page 311.) The solution for undamped natural frequencies and mode shapes is called real
eigenvalue analysis or normal modes analysis.
The remainder of this chapter describes the various eigensolution methods for computing
natural frequencies and mode shapes, and it concludes with several examples.

Main Index
36

3.2 Reasons to Compute Normal Modes

There are many reasons to compute the natural frequencies and mode shapes of a structure. One
reason is to assess the dynamic interaction between a component and its supporting structure.
For example, if a rotating machine, such as an air conditioner fan, is to be installed on the roof of
a building, it is necessary to determine if the operating frequency of the rotating fan is close to
one of the natural frequencies of the building. If the frequencies are close, the operation of the
fan may lead to structural damage or failure.
Decisions regarding subsequent dynamic analyses (i.e., transient response, frequency response,
response spectrum analysis, etc.) can be based on the results of a natural frequency analysis. The
important modes can be evaluated and used to select the appropriate time or frequency step for
integrating the equations of motion. Similarly, the results of the eigenvalue analysis-the natural
frequencies and mode shapes-can be used in modal frequency and modal transient response
analyses (see “Frequency Response Analysis” on page 103 and “Transient Response Analysis”
on page 149).
The results of the dynamic analyses are sometimes compared to the physical test results. A
normal modes analysis can be used to guide the experiment. In the pretest planning stages, a
normal modes analysis can be used to indicate the best location for the accelerometers. After the
test, a normal modes analysis can be used as a means to correlate the test results to the analysis
results.
Design changes can also be evaluated by using natural frequencies and normal modes. Does a
particular design modification cause an increase in dynamic response? Normal modes analysis
can often provide an indication.
In summary, there are many reasons to compute the natural frequencies and mode shapes of a
structure. All of these reasons are based on the fact that real eigenvalue analysis is the basis for
many types of dynamic response analyses. Therefore, an overall understanding of normal
modes analysis as well as knowledge of the natural frequencies and mode shapes for your
particular structure is important for all types of dynamic analysis.

Main Index
 CHAPTER 3 37
Real Eigenvalue Analysis

3.3 Overview of Normal Modes Analysis

The solution of the equation of motion for natural frequencies and normal modes requires a
special reduced form of the equation of motion. If there is no damping and no applied loading,
the equation of motion in matrix form reduces to
··
[M ]{ u} + [K ][ u] = 0 Eq. 3-1
where:

[ M ] = mass matrix
[ K ] = stiffness matrix

This is the equation of motion for undamped free vibration. To solve Eq. 3-1 assume a harmonic
solution of the form

{ u } = { φ } sin ωt Eq. 3-2

where:

{ φ } = the eigenvector or mode shape

ω = is the circular natural frequency

Aside from this harmonic form being the key to the numerical solution of the problem, this form
also has a physical importance. The harmonic form of the solution means that all the
degrees-of-freedom of the vibrating structure move in a synchronous manner. The structural
configuration does not change its basic shape during motion; only its amplitude changes.
If differentiation of the assumed harmonic solution is performed and substituted into the
equation of motion, the following is obtained:

2
– ω [ M ] { φ } sin ωt + [ K ] { φ } sin ωt = 0 Eq. 3-3

which after simplifying becomes

2
([K] – ω [M]){φ} = 0 Eq. 3-4

This equation is called the eigenequation, which is a set of homogeneous algebraic equations for
the components of the eigenvector and forms the basis for the eigenvalue problem. An
eigenvalue problem is a specific equation form that has many applications in linear matrix
algebra. The basic form of an eigenvalue problem is

[ A – λI ]x = 0 Eq. 3-5
where:

A = square matrix
λ = eigenvalues

Main Index
38

I = identity matrix
x = eigenvector

In structural analysis, the representations of stiffness and mass in the eigenequation result in the
physical representations of natural frequencies and mode shapes. Therefore, the eigenequation
is written in terms of K , ω , and M as shown in Eq. 3-4 with ω 2 = λ .
There are two possible solution forms for Eq. 3-4:
1. If det ( [ K ] – ω2 [ M ] ) ≠ 0 , the only possible solution is

{φ} = 0 Eq. 3-6

This is the trivial solution, which does not provide any valuable information from a
physical point of view, since it represents the case of no motion. (“det” denotes the
determinant of a matrix.)
2. If det ( [ K ] – ω 2 [ M ] ) = 0 , then a non-trivial solution ( { φ } ≠ 0 ) is obtained for

2
([K] – ω [M]){φ} = 0 Eq. 3-7
From a structural engineering point of view, the general mathematical eigenvalue
problem reduces to one of solving the equation of the form

2
det ( [ K ] – ω [ M ] ) = 0 Eq. 3-8
or

det ( [ K ] – λ [ M ] ) = 0 Eq. 3-9

2
where λ = ω
The determinant is zero only at a set of discrete eigenvalues λi or ω 2i . There is an eigenvector
{ φ i } which satisfies Eq. 3-7 and corresponds to each eigenvalue. Therefore, Eq. 3-7 can be
rewritten as

2
[ K – ωi M ] { φi } = 0 i = 1, 2, 3… Eq. 3-10

Each eigenvalue and eigenvector define a free vibration mode of the structure. The i-th
eigenvalue λ i is related to the i-th natural frequency as follows:

ωi
f i = ------ Eq. 3-11
2π

where:

f i = i-th natural frequency

ωi = λi

The number of eigenvalues and eigenvectors is equal to the number of degrees-of-freedom that
have mass or the number of dynamic degrees-of-freedom.
Main Index
 CHAPTER 3 39
Real Eigenvalue Analysis

There are a number of characteristics of natural frequencies and mode shapes that make them
useful in various dynamic analyses. First, when a linear elastic structure is vibrating in free or
forced vibration, its deflected shape at any given time is a linear combination of all of its normal
modes

{u} = ∑ ( φ i )ξ i Eq. 3-12

i

where:

{ u } = vector of physical displacements

{ φ i } = i-th mode shape

ξ i = i-th modal displacement

Second, if [ K ] and [ M ] are symmetric and real (as is the case for all the common structural finite
elements), the following mathematical properties hold:

T
{ φ i } [ M ] { φj } = 0 if i ≠ j Eq. 3-13

T
{ φ j } [ M ] { φ j } = m j = j-th generalized mass Eq. 3-14

and

T
{ φj } [ K ] { φj } = 0 if i ≠ j Eq. 3-15

T 2
{ φ j } [ K ] { φ j } = k j = j-th generalized stiffness = ω m j Eq. 3-16

Also, from Eq. 3-14 and Eq. 3-16 Rayleigh’s equation is obtained

T
2 { φj } [ K ] { φj }
ωj = -------------------------------------- Eq. 3-17
T
{ φj } [ M ] { φj }

Eq. 3-13 and Eq. 3-15 are known as the orthogonality property of normal modes, which ensures
that each normal mode is distinct from all others. Physically, orthogonality of modes means that
each mode shape is unique and one mode shape cannot be obtained through a linear
combination of any other mode shapes.
In addition, a natural mode of the structure can be represented by using its generalized mass and
generalized stiffness. This is very useful in formulating equivalent dynamic models and in
component mode synthesis (see “Advanced Dynamic Analysis Capabilities” on page 311).
If a structure is not totally constrained in space, it is possible for the structure to displace (move)
as a rigid body or as a partial or complete mechanism. For each possible component of
rigid-body motion or mechanism, there exists one natural frequency that is equal to zero. The
zero-frequency modes are called rigid-body modes. Rigid-body motion of all or part of a
structure represents the motion of the structure in a stress-free condition. Stress-free, rigid-body
Main Index
40

modes are useful in conducting dynamic analyses of unconstrained structures, such as aircraft
and satellites. Also, rigid-body modes can be indicative of modeling errors or an inadequate
constraint set.
For example, the simple unconstrained structure in Figure 3-3 has a rigid-body mode.

u1 u2

m1 m2
k

 u1  1 
ω1 = 0 { φ1 } =   =  
 2
u 1 

Figure 3-3 Rigid-Body Mode of a Simple Structure

When both masses move the same amount (as a rigid body), there is no force induced in the
connecting spring. A detailed discussion of rigid-body modes is presented in “Rigid-body
Modes” on page 87.
An important characteristic of normal modes is that the scaling or magnitude of the eigenvectors
is arbitrary. Mode shapes are fundamental characteristic shapes of the structure and are
therefore relative quantities. In the solution of the equation of motion, the form of the solution
is represented as a shape with a time-varying amplitude. Therefore, the basic mode shape of the
structure does not change while it is vibrating; only its amplitude changes.
For example, three different ways to represent the two modes of a two-DOF structure are shown
in Figure 3-4. The graphical representation of the eigenvectors in the figure shows the modal
displacements rotated by 90 degrees in order to view the deformation better.

u1 u1 u1

m1

u2
u2 u2
m2

k2

Mode 1 Mode 2
 u1  3   300   – 0.6   u1   – 1.6   160   1.0 
φ1 =   =   =   =  { φ 2 } =   =   =   =  
 u2  1   100   – 0.2   u2   0.8   – 80   – 0.5 

Figure 3-4 Representations of Mode Shapes for a Two-DOF System

Main Index
 CHAPTER 3 41
Real Eigenvalue Analysis

A common misconception about mode shapes is that they define the structural response. Again,
mode shapes are relative quantities. They cannot be used alone to evaluate dynamic behavior.
As described earlier, it is the relation between the structural loading and the natural frequencies
that determines the absolute magnitude of dynamic response. The relation of a specific loading
to a set of natural frequencies provides explicit scale factors that are used to determine the extent
to which each particular mode is excited by the loading. After the individual modal responses
to a particular loading are determined, only then can the various engineering design decisions
be made with the actual (absolute) values of stress and/or displacement. Methods that use the
modal results to determine forced response are termed modal methods or modal superposition
methods. Modal frequency response analysis and modal transient response analysis are
described in “Frequency Response Analysis” on page 103 and “Transient Response
Analysis” on page 149, respectively.

Mode Shape Normalization. Although the scaling of normal modes is arbitrary, for practical
considerations mode shapes should be scaled (i.e., normalized) by a chosen convention. In
MSC.Nastran there are three normalization choices, MASS, MAX, and POINT normalization.
MASS normalization is the default method of eigenvector normalization. This method scales
each eigenvector to result in a unit value of generalized mass

T
{ φ j } [ M ] { φ j } = 1.0 Eq. 3-18

Numerically this method results in a modal mass matrix that is an identity matrix. This
normalization approach is appropriate for modal dynamic response calculations because it
simplifies both computational and data storage requirements. When mass normalization is used
with a model of a heavy, massive structure, the magnitude of each of the terms of the
eigenvectors is very small.
In MAX normalization, each eigenvector is normalized with respect to the largest a-set
component. (“Advanced Dynamic Analysis Capabilities” on page 311 and “The Set Notation
System Used in Dynamic Analysis” on page 341 provide discussions of the a-set.) This
normalization results in the largest a-set displacement value being set to a unit (1.0) value. This
normalization approach can be very useful in the determination of the relative participation of
an individual mode. A small generalized mass obtained using MAX normalization may indicate
such things as local modes or isolated mechanisms.

POINT normalization of eigenvectors allows you to chose a specific displacement component at

which the modal displacement is set to 1 or -1. This method is not recommended because for
complex structures the chosen component in the non-normalized eigenvector may have a very
small value of displacement (especially in higher modes). This small value can cause larger
numbers to be normalized by a small number, resulting in possible numerical roundoff errors in
mode shapes.
Although mode shapes are relative quantities, a number of modal quantities can be helpful in
predicting qualitative responses or in isolating troublesome modal frequencies. Since relative
strains, internal loads, and stresses develop when a structure deforms in a mode shape, you may
recover these quantities during a normal modes analyses. Basically, any quantity that you can
recover for static analysis is also available for normal modes analysis.
Main Index
42

It is important to remember that these output quantities are based on the relative displacements
of a mode shape. The output quantities can be compared for a given mode, but not necessarily
from one mode to another. However, they can still be effectively used in the analysis/design
process.
Modal quantities can be used to identify problem areas by indicating the more highly stressed
elements. Elements that are consistently highly stressed across many or all modes will probably
be highly stressed when dynamic loads are applied.
Modal strain energy is a useful quantity in identifying candidate elements for design changes to
eliminate problem frequencies. Elements with large values of strain energy in a mode indicate
the location of large elastic deformation (energy). These elements are those which most directly
affect the deformation in a mode. Therefore, changing the properties of these elements with
large strain energy should have more effect on the natural frequencies and mode shapes than if
elements with low strain energy were changed.
Structures with two or more identical eigenvalues are said to have repeated roots. Repeated
roots occur for structures that have a plane of symmetry or that have multiple, identical pieces
(such as appendages). The eigenvectors for the repeated roots are not unique because many sets
of eigenvectors can be found that are orthogonal to each other. An eigenvector that is a linear
combination of the repeated eigenvectors is also a valid eigenvector. Consequently, small
changes in the model can make large changes in the eigenvectors for the repeated roots.
Different computers can also find different eigenvectors for the repeated roots. Rigid-body
modes (see “Rigid-body Modes” on page 87) represent a special case of repeated roots.

Main Index
 CHAPTER 3 43
Real Eigenvalue Analysis

3.4 Methods of Computation

Seven methods of real eigenvalue extraction are provided in MSC.Nastran. These methods are
numerical approaches to solving for natural frequencies and modes shapes. The reason for
seven different numerical techniques is because no one method is the best for all problems.
While most of the methods can be applied to all problems, the choice is often based on the
efficiency of the solution process.

The methods of eigenvalue extraction belong to one or both of the following two groups:

• Transformation methods
• Tracking methods
In the transformation method, the eigenvalue equation is first transformed into a special form
from which eigenvalues may easily be extracted. In the tracking method, the eigenvalues are
extracted one at a time using an iterative procedure.
The recommended real eigenvalue extraction method in MSC.Nastran is the Lanczos method.
The Lanczos method combines the best characteristics of both the tracking and transformation
methods. For most models the Lanczos method is the best method to use.
Four of the real eigenvalue extraction methods available in MSC.Nastran are transformation
methods. These methods are

• Givens method
• Householder method
• Modified Givens method
• Modified Householder method
Two of the real eigenvalue extraction methods available in MSC.Nastran are classified as
tracking methods. These methods are

• Inverse power method

• Sturm modified inverse power method
The remainder of this section briefly describes the various methods. The theory and algorithms
behind each method can be found in the MSC.Nastran Numerical Methods User's Guide.

Lanczos Method
The Lanczos method overcomes the limitations and combines the best features of the other
methods. It requires that the mass matrix be positive semidefinite and the stiffness be
symmetric. Like the transformation methods, it does not miss roots, but has the efficiency of the
tracking methods, because it only makes the calculations necessary to find the roots requested
by the user. This method computes accurate eigenvalues and eigenvectors. Unlike the other
methods, its performance has been continually enhanced since its introduction giving it an
advantage. The Lanczos method is the preferred method for most medium- to large-sized
problems, since it has a performance advantage over other methods.

Main Index
44

Givens and Householder Methods

The Givens and Householder modal extraction methods require a positive definite mass matrix
(all degrees-of-freedom must have mass). There is no restriction on the stiffness matrix except
that it must be symmetric. These matrices always result in real (positive) eigenvalues. The
Givens and Householder methods are the most efficient methods for small problems and
problems with dense matrices when a large portion of the eigenvectors are needed. These
methods find all of the eigenvalues and as many eigenvectors as requested. While these
methods do not take advantage of sparse matrices, they are efficient with the dense matrices
sometimes created using dynamic reduction (see “Advanced Dynamic Analysis Capabilities”
on page 311).
The Givens and Householder methods fail if the mass matrix is not positive definite. To
minimize this problem, degrees-of-freedom with null columns are removed by the automatic
application of static condensation (see “Advanced Dynamic Analysis Capabilities” on
page 311) called auto-omit. Applying the auto-omit process is a precaution and may not remove
all possible causes of mass matrix singularity, such as a point mass offset from a grid point, but
it greatly improves the reliability and convenience of the Givens and Householder methods.
Givens and Householder methods use different transformation schemes to obtain the
eigenvalues. For problems in which no spill occurs (i.e., all of the matrices fit in your computer's
main memory), the Householder method costs about half as much as the Givens method for
vector processing computers. In addition, the Householder method can take advantage of
parallel processing computers.

Modified Givens and Modified Householder Methods

The modified Givens and modified Householder methods are similar to their standard methods
with the exception that the mass matrix can be singular. Although the mass matrix is not
required to be nonsingular in the modified methods, a singular mass matrix can produce one or
more infinite eigenvalues. Due to roundoff error, these infinite eigenvalues appear in the output
as very large positive or negative eigenvalues. To reduce the incidence of such meaningless
results, degrees-of-freedom with null masses are eliminated by automatic static condensation as
in the case of the unmodified methods.
The modified methods require more computer time than the standard methods.

Automatic Givens and Automatic Householder Methods

Many times you may not know whether the mass matrix is singular. To assist you in choosing
the appropriate method, two options–automatic Givens and automatic Householder–are
available. Initially the automatic methods use the standard methods. In the first step of the
method, if the mass matrix is not well-conditioned for decomposition, the method shifts to the
corresponding modified method. The modified methods are more expensive and may introduce
numerical noise due to the shift, but they resolve most of the numerical problems of the ill-
conditioned mass matrix. The automatic methods, therefore, use the modified methods when
necessary for numerical stability but use the standard methods when the numerical stability is
accurate.

Main Index
 CHAPTER 3 45
Real Eigenvalue Analysis

Inverse Power Method

The inverse power method is a tracking method since the lowest eigenvalue and eigenvector in
the desired range are found first. Then their effects are "swept" out of the dynamic matrix, the
next higher mode is found, and its effects are "swept" out, and so on. Hence, the term "tracking,"
which means that one root at a time is found. In addition, each root is found via an iterative
procedure. (The classical literature often refers to this method as the inverse iteration method
with sweeping.)
However, the inverse power method can miss modes, making it unreliable. The Sturm modified
inverse power method is a more reliable tracking method.

Sturm Modified Inverse Power Method

This method is similar to the inverse power method except that it uses Sturm sequence logic to
ensure that all modes are found. The Sturm sequence check determines the number of
eigenvalues below a trial eigenvalue, then finds all of the eigenvalues below this trial eigenvalue
until all modes in the designed range are computed. This process helps to ensure that modes are
not missed.
The Sturm modified inverse power method is useful for models in which only the lowest few
modes are needed. This method is also useful as a backup method to verify the accuracy of other
methods.

Main Index
46

3.5 Comparison of Methods

Since MSC.Nastran provides a variety of real eigensolution methods, you must decide which is
best for your application. The best method for a particular model depends on four factors: the
size of the model (the total number of degrees-of-freedom as well as the number of dynamic
degrees-of-freedom), the number of eigenvalues desired, the available real memory of your
computer, and the conditioning of the mass matrix (whether there are massless degrees-of-
freedom). In general, the Lanczos method is the most reliable and efficient, and is the
recommended choice.
For small, dense models whose matrices fit into memory, we recommend using one of the
automatic methods (automatic Householder or automatic Givens). Both automatic Householder
and automatic Givens run modified methods if the mass matrix is singular; however, they run
the unmodified methods, which are faster, if the mass matrix is not singular. Of the two
automatic methods, the automatic Householder method runs faster on computers with vector
processing and also supports parallel processing computers. Note that most real world
problems are not small and dense, unless you use reductive methods such as General Dynamic
Reduction (GDR) or superelements.
The Sturm modified inverse power method can be the best choice when the model is too large
to fit into memory, only a few modes are needed, and a reasonable eigenvalue search range is
specified. This method is also a backup method for the other methods and is used when a check
of the other methods' results is needed.

For medium to large models the Lanczos method is the recommended method. In addition to
its reliability and efficiency, the Lanczos method supports sparse matrix methods that
substantially increase its speed and reduce disk space requirements.

Table 3-1 Comparison of Eigenvalue Methods

Method
Sturm
Modified
Givens, Inverse Modified
Givens, Lanczos
Householder Power Inverse
Householder
Power
Reliability High High Poor (can miss High High
modes)

Relative Cost:

Few Modes Medium Medium Low Low Medium

Many Modes High High High High Medium

Limitations Cannot analyze Expensive for Can miss modes Expensive for Difficulty with
singular [M] many modes many modes massless
mechanisms

Expensive for Expensive for Expensive for

problems that do problems that do many modes
not fit in memory not fit in memory

Main Index
 CHAPTER 3 47
Real Eigenvalue Analysis

Table 3-1 Comparison of Eigenvalue Methods (continued)

Method
Sturm
Modified
Givens, Inverse Modified
Givens, Lanczos
Householder Power Inverse
Householder
Power
Best Small, dense Small, dense To detemine a few To detemine a few Medium to large
Application matrices that fit in matrices that fit in modes modes models
memory memory

Use with Use with Backup method

dynamic dynamic
reduction (see reduction (see
“Advanced “Advanced
Dynamic Dynamic
Analysis Analysis
Capabilities” Capabilities”
in Chapter 11) in Chapter 11)

Main Index
48

3.6 User Interface for Real Eigenvalue Analysis

The EIGR and EIGRL Bulk Data entries define the method and select the parameters that control
the eigenvalue extraction procedure. The EIGRL entry is used for the Lanczos method, and the
EIGR entry is used for all of the other methods.

User Interface for the Lanczos Method

The EIGRL entry has the following format:

Format:

1 2 3 4 5 6 7 8 9 10
EIGRL SID V1 V2 ND MSGLVL MAXSET SHFSCL NORM

option_1=value_1, option_2=value_2, etc.

Example:

EIGRL 1 0.1 3.2

Field Contents
SID Set identification number. (Unique Integer > 0)
V1, V2 The V1 field defines the lower frequency bound; the V2 field defines the upper
frequency field. For vibration analysis: frequency range of interest. For
buckling analysis: eigenvalue range of interest. (Real or blank,
16 16
– 5 × 10 ≤ V1 < V2 ≤ 5. × 10 )

ND Number of eigenvalues and eigenvectors desired. (Integer > 0 or blank)

MSGLVL Diagnostic level. (0 ≤ Integer ≤ 4; Default = 0)
MAXSET Number of vectors in block or set. (1 ≤ Integer ≤ 15; Default = 7)
SHFSCL Estimate of the first flexible mode natural frequency. (Real or blank)
NORM Method for normalizing eigenvectors (Character: "MASS" or "MAX")
MASS Normalize to unit value of the generalized mass. Not available
for buckling analysis. (Default for normal modes analysis.)
MAX Normalize to unit value of the largest displacement in the
analysis set. Displacements not in the analysis set may be larger
than unity. (Default for buckling analysis.)

Examples of the results of using explicit or default values for the V1, V2, and ND fields are
shown in Table 3-2. The defaults on the EIGRL entry are designed to provide the minimum
number of roots in cases where the input is ambiguous. Alternatively, you can write the entry
above using the new free field format. You must specify certain new parameters such as
ALPH,NUMS, and Fi when using the free field format. See the MSC.Nastran Quick Reference
Guide for details.
Main Index
 CHAPTER 3 49
Real Eigenvalue Analysis

1 2 3 4 5 6 7 8 9 10
EIGRL 1

V1 = 0.1, V2 = 3.2

Table 3-2 Number and Type of Roots Found with the EIGRL Entry

Case V1 V2 ND Number and Type of Roots Found

1 V1 V2 ND Lowest ND in range or all in range, whichever is smaller
2 V1 V2 All in range
3 V1 ND Lowest ND in range [V1, ∞]
4 V1 Lowest root in range [V1, ∞]
5 ND Lowest ND roots in [-∞, ∞]
6 Lowest root
7 V2 ND Lowest ND or all in range [-∞, V2], whichever is smaller
8 V2 All below V2

The MSGLVL field of the EIGRL entry is used to control the amount of diagnostic output. The
value of 0 produces no diagnostic output. The values 1, 2, or 3 provide more output with the
higher values providing increasingly more output. In some cases, higher diagnostic levels may
help to resolve difficulties with special modeling problems.
The MAXSET field is used to control the block size. The default value of 7 is recommended for
most applications. There may be special cases where a larger value may result in quicker
convergence of many multiple roots or a lower value may result in more efficiency when the
structure is lightly coupled. However, the default value has been chosen after reviewing the
results from a wide range of problems on several different computer types with the goal of
minimizing the computer time. A common occurrence is for the block size to be reset by
MSC.Nastran during the run because there is insufficient memory for a block size of 7.
Computational efficiency tends to degrade as the block size decreases. Therefore, it is, important
to examine the eigenvalue analysis summary output to determine that MSC.Nastran has
sufficient memory to use an efficient block size. A smaller block size may be more efficient when
only a few roots are requested. The minimum recommended block size is 2.
The SHFSCL field allows a user-designated shift to be used to improve performance especially
when large mass techniques are used in enforced motion analysis (see “Enforced Motion” on
page 195). Large mass techniques can cause a large gap between the rigid body (see
“Rigid-body Modes” on page 87) and flexible frequencies, which can degrade performance of
the Lanczos method or cause System Fatal Message 5299. When SHFSCL is used, its value
should be set close to the expected first nonzero natural frequency.

Main Index
50

The Lanczos method normalizes (i.e., scales) the computed eigenvectors using the MASS or
MAX method. These methods are specified using the NORM field (new for Version 68). The
MASS method normalizes to a unit value of the generalized mass (i.e., m j = 1.0 ). The MAX
method normalizes to a unit value of the largest component in the a-set (see “Advanced
Dynamic Analysis Capabilities” on page 311). The default is MASS.

You can use the continuation entry to specify V1, V2, ND, MSGLVL, MAXSET, SHFSCL and
NORM if you have not specified them on the parent entry. To apply the continuation entry use
the following format: 'option_i=value_i', e.g., ND=6. Using the continuation entry is the only
way to specify the three new options, ALPH, NUMS and Fi.

NUMS The number of segments that a frequency range will be broken into for parallel
processing. You must define a value greater than 1 to take advantage of parallel
processing. You may also specify NUMS using the NUMSEG keyword on the
NASTRAN statement. If you specify both, then NUMS takes precedence.
Fi Directly specifies the upper frequencies of each segment, such that
V1 < F1 < F2 < ... F15 < V2.
ALPH Automatically generates the Fi values based on the following formula:

i NUMS
Fi = ( V2 – V1 ) ⋅ ( 1.0 – ALPH ) ⁄ ( 1.0 – ALPH )

If you specify both ALPH and Fi, then Fi takes precedence over ALPH as long as they are
consistent. If ALPH is multiplied by 100, it may be specified on the FRQSEQ keyword of the
NASTRAN statement.
For a detailed description of the EIGRL input, see the MSC.Nastran Numerical Methods User’s
Guide.

User Interface for the Other Methods

The data entered on the EIGR entry selects the eigenvalue method and the frequency range or
number of required roots. The basic format of the Bulk Data entry is as follows:

1 2 3 4 5 6 7 8 9 10
EIGR SID METHOD F1 F2 NE ND

NORM G C

The METHOD field selects the eigenvalue method from the following list:

AGIV Automatic Givens method

AHOU Automatic Householder method
GIV Givens method
HOU Householder method
INV Inverse power
Main Index
 CHAPTER 3 51
Real Eigenvalue Analysis

MGIV Modified Givens method

MHOU Modified Householder method
SINV Sturm modified inverse power method

The F1 field specifies the lowest frequency of interest in the eigenvalue extraction. The F2 field
specifies the highest frequency of interest in the eigenvalue extraction. The units are cycles per
unit time. The value for F1 must be greater than or equal to 0.0.
The NE field is used by the INV method only. It defines the estimated number of roots in the
range. A good estimate results in a more efficient solution. A high estimate helps to ensure that
all modes are computed within the range.
The ND field is used to specify the desired number of roots (for tracking methods) or
eigenvectors (for transformation methods), beginning with F1.
The NORM field on the continuation entry is used to specify the method of eigenvector
normalization. The choices are

MASS Mass normalization (default–if used, the continuation entry is not required).
MAX Normalization to a unit value of the largest component in the analysis set.
POINT Normalization to a unit value at a user-specified a-set grid pointG and
component C.
There is an interrelationship among the F1, F2, and ND fields on the EIGR entry as defined in
Table 3-3.

Table 3-3 Relationship Between the METHOD Field and Other Fields

METHOD Field
Field GIV, MGIV, AGIV, HOU, MHOU,
INV or SINV
or AHOU
F1, F2 Frequency range of interest. F1 must be Frequency range of interest. If ND is not
input. If METHOD = “SINV” and ND is blank, F1 and F2 are ignored. If ND is
blank, then F2 must be input. blank, eigenvectors are found whose
natural frequencies lie in the range
(Real ≥ 0.0)
between F1 and F2.

(Real ≥ 0.0; F1 < F2)

NE Estimate of number of roots in range Not used
(Required for METHOD = “INV”). Not
used by “SINV” method.
(Integer > 0)

Main Index
52

Table 3-3 Relationship Between the METHOD Field and Other Fields (continued)

METHOD Field
Field GIV, MGIV, AGIV, HOU, MHOU,
INV or SINV
or AHOU
ND Desired number of roots. If this field is Desired number of eigenvectors. If ND
blank and METHOD = “SINV”, then all is zero, the number of eigenvectors is
roots between F1 and F2 are searched, determined from F1 and F2. If all three
and the limit is 600 roots. are blank, then ND is automatically set
to one more than the number of degrees-
(Integer > 0, Default is 3 ∗ NE for
of-freedom listed on SUPORT entries.
METHOD = “INV” only)
(Integer ≥ 0; Default = 0)

The rules for METHOD = GIV, HOU, MGIV, MHOU, AGIV, and AHOU are identical. If any of
these methods are selected, MSC.Nastran finds all of the eigenvalues but only computes the
eigenvectors specified by F1 and F2 or those specified by ND (the desired number). F1 and F2
specify the lower and upper bounds of the frequency range in which eigenvectors are computed,
and ND specifies the number of eigenvectors, beginning with the lowest (or the first rigid-body
mode, if present). If F1, F2, and ND entries are present, ND takes precedence.
If METHOD = SINV, the values of F1, F2, and ND determine the number of eigenvalues and
eigenvectors that are computed. These entries also provide hints to help MSC.Nastran find the
eigenvalues. F1 and F2 specify the frequency range of interest within which MSC.Nastran
searches for modes. MSC.Nastran attempts to find all of the modes in the range between F1 and
F2 or the number specified by ND, whichever is less. If searching stops because ND modes are
found, there is no guarantee that they are the lowest eigenvalues. If ND modes are not found in
the range of interest, SINV usually finds one mode (or possibly more) outside the range F1 and
F2 before stopping the search.
The SINV method is particularly efficient when only a small number of eigenvalues and
eigenvectors are to be computed. Often only the lowest mode is of interest. The following
example illustrates an EIGR entry which extracts only the lowest nonzero eigenvalue.

1 2 3 4 5 6 7 8 9 10
EIGR 13 SINV 0.0 0.01 1

It is assumed in the example above that the frequency of the lowest mode is greater than 0.01
cycles per unit time. MSC.Nastran finds one eigenvalue outside the range F1, F2, and then stops
the search. The eigenvalue found is the lowest nonzero eigenvalue (or a member of the lowest
closely spaced cluster of eigenvalues in cases with close roots) provided that there are no
negative eigenvalues and that the SUPORT entry has been used to specify the correct number of
zero eigenvalues (see “Rigid-body Modes” on page 87).
The following examples demonstrate the use of the EIGR data entry.

Main Index
 CHAPTER 3 53
Real Eigenvalue Analysis

1 2 3 4 5 6 7 8 9 10
EIGR 1 AHOU 10

In this example, the automatic Householder method is selected, and the lowest 10 modes are
requested. Since the default MASS eigenvector normalization is requested, no continuation
entry is needed.

EIGR 2 AHOU 100.

MAX

In this example, the same method is requested, but all the modes below 100 cycles per unit time
are requested with MAX vector normalization.

EIGR 3 SINV 0.1 100.0 6

POINT 32 3

In this example, the Sturm modified inverse power method is requested for the first six modes
found in the range specified (0.1 to 100 Hz). The POINT normalization method is requested with
each eigenvector scaled such that grid point 32 in the T3 direction has a magnitude of 1.0. Note
that this degree-of-freedom must be in the a-set.

Main Index
54

3.7 Solution Control for Normal Modes Analysis

This section describes input required for the selection and control of normal modes analysis.

Executive Control Section

You can run a normal modes analysis, as an independent solution, in SOL 103 of the Structured
Solution Sequences or in SOL 3 of the Rigid Format Solution Sequences. We recommend using
SOL 103 since it contains our most recent enhancements. The Executive Control Section can also
contain diagnostic DIAG16, which prints the iteration information used in the INV or SINV
method. You may also run a normal modes analysis as part of the other solution sequences (such
as modal transient response, modal frequency response, design optimization, and
aeroelasticity).

Case Control Section

The most important eigenanalysis command in the Case Control is the METHOD command.
This command is required. The set identification number specified by the METHOD Case
Control command refers to the set identification number of an EIGR or EIGRL entry in the Bulk
Data.
When a modal analysis is performed, the MSC.Nastran output file contains various diagnostic
messages and an eigenvalue analysis summary. Optional grid and element output are available
using standard Case Control output requests. Eigenvectors are printed only if a
DISPLACEMENT or VECTOR command is included. These requests are summarized in
Table 3-4.

Table 3-4 Eigenvalue Extraction Output Requests

Grid Output
DISPLACEMENT Requests the eigenvector (mode shape) for a set of grid points.
(or VECTOR)
GPFORCE Requests the grid point force balance table to be computed for each
mode for a set of grid points.
GPSTRESS Requests grid point stresses to be computed for a set of grid points. This
request must be accompanied by the ELSTRESS Case Control re quest
and the definition of stress surfaces and/or stress volumes in the
OUTPUT(POST) section of the Case Control. This request also requires
the use of Rigid Format Alter RF3D81 when used in SOL 3.
SPCFORCE Requests forces of single-point constraint to be computed for a set of grid
points for each mode.
Element Output
ELSTRESS (or Requests the computation of modal stresses for a set of elements for each
STRESS) mode.

Main Index
 CHAPTER 3 55
Real Eigenvalue Analysis

Table 3-4 Eigenvalue Extraction Output Requests (continued)

ESE Requests the computation of modal element strain energies for a set of
elements for each mode.
ELFORCE (or Requests the computation of modal element forces for a set of elements
FORCE) for each mode.
STRAIN Requests the computation of modal element strains for a set of ele ments.
Miscellaneous
MODES A special Case Control request that permits selective output requests to
be processed on selective modes.

Bulk Data Section

In addition to Bulk Data entries required to define the structural model, the only other required
Bulk Data entry is the eigenvalue selection entry EIGR or EIGRL. The EIGR entry is used to
select the modal extraction parameters for the inverse power, Sturm modified inverse power,
Givens, Householder, modified Givens, modified Householder, automatic Givens, and
automatic Householder methods. The EIGRL entry is used to select the modal extraction
parameters for the Lanczos method.

Main Index
56

3.8 Examples
This section provides several normal-modes analysis examples showing the input and output.
These examples are as follows:

Number of Element Output Analysis

Model Units
Grid Points Types Requests Method
bd03two 3 CELAS2 DISPLACEMENT AHOU Metric
CONM2 SPCFORCE
ELFORCE
bd03bar1, 11 CBAR DISPLACEMENT SINV Metric
bd03bar2
bd03bkt 236 CQUAD4 DISPLACEMENT Lanczos English
CONM2 STRESS
RBE2 ESE
MODES
bd03car 972 CQUAD4 DISPLACEMENT Lanczos English
CTRIA3 ESE
CELAS2
bd03fix 8157 CHEXA DISPLACEMENT Lanczos English
CPENTA
bd03plt1, 81 CQUAD4 DISPLACEMENT Lanczos Metric
bd03plt2
bd03dmi 7 CBAR None Lanczos Metric
DMIG

These examples are described in the sections that follow.

Main Index
 CHAPTER 3 57
Real Eigenvalue Analysis

Two-DOF Model
This example is a restrained two-DOF model with two springs and two masses as illustrated in
Figure 3-5.

Grid Point 1 m 1 = 0.1 kg

k 1 = 100 N/m

Grid Point 2 m 2 = 10 kg

k 2 = 1.0E4 N/m
z

Figure 3-5 Two-DOF Model

The masses are constrained to deflect in only the y-direction. The example illustrates normal
modes analysis (SOL 103) using automatic selection of the Householder or modified
Householder method (METHOD = AHOU on the EIGR entry). The eigenvectors are normalized
to the unit value of the largest displacement component (NORM = MAX on the EIGR entry). The
input file is shown in Listing 3-1.

Main Index
58

Listing 3-1 Input File for the Two-DOF Model

$ FILE bd03two.dat
$
$ TWO DOF SYSTEM
$ CHAPTER 3, NORMAL MODES
$
TIME 5
SOL 103 $ NORMAL MODES ANALYSIS
CEND
$
TITLE = TWO DOF SYSTEM
SUBTITLE = NORMAL MODES ANALYSIS
$
$ SELECT SPC
SPC = 10
$
$ SELECT EIGR ENTRY
METHOD = 99
$
$ SELECT OUTPUT
SET 1 = 1,2
DISPLACEMENT = 1
SET 2 = 3
SPCFORCE = 2
SET 3 = 11,12
ELFORCE = 3
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$EIGR SID METHOD F1 F2 NE ND +EIG
$+EIG NORM G C
EIGR 99 AHOU 0. 200. +EIG1
+EIG1 MAX
$
GRID 1 0. 2. 0.
GRID 2 0. 1. 0.
GRID 3 0. 0. 0.
GRDSET 13456
CONM2 1 1 0.1
CONM2 2 2 10.0
CELAS2 11 100.0 1 2 2 2
CELAS2 12 1.0E4 2 2 3 2
SPC 10 3 2
$
ENDDATA

The printed output is shown in Listing 3-2. The eigenvalue summary lists the eigenvalue ω 2n ,
circular frequency ω n (radians per second), natural frequency f n (cycles per second), generalized
mass (see Eq. 3-14), and generalized stiffness (see Eq. 3-16) for each mode. The eigenvectors,
SPC forces, and spring forces are shown for each mode.

Main Index
 CHAPTER 3 59
Real Eigenvalue Analysis

Listing 3-2 Output from the Two-DOF System

R E A L E I G E N V A L U E S

MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED

NO. ORDER MASS STIFFNESS
1 1 9.048751E+02 3.008114E+01 4.787562E+00 1.904875E-01 1.723674E+02
2 2 1.105125E+03 3.324342E+01 5.290854E+00 2.105125E-01 2.326426E+02

EIGENVALUE = 9.048751E+02
CYCLES = 4.787562E+00 R E A L E I G E N V E C T O R N O . 1
POINT ID. TYPE T1 T2 T3 R1 R2 R3
1 G 0.0 1.000000E+00 0.0 0.0 0.0 0.0
2 G 0.0 9.512492E-02 0.0 0.0 0.0 0.0

EIGENVALUE = 1.105125E+03
CYCLES = 5.290854E+00 R E A L E I G E N V E C T O R N O . 2

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 1.000000E+00 0.0 0.0 0.0 0.0
2 G 0.0 -1.051249E-01 0.0 0.0 0.0 0.0

EIGENVALUE = 9.048751E+02
F O R C E S O F S I N G L E - P O I N T C O N S T R A I N T

POINT ID. TYPE T1 T2 T3 R1 R2 R3

3 G 0.0 -9.512491E+02 0.0 0.0 0.0 0.0

EIGENVALUE = 1.105125E+03
F O R C E S O F S I N G L E - P O I N T C O N S T R A I N T

POINT ID. TYPE T1 T2 T3 R1 R2 R3

3 G 0.0 1.051249E+03 0.0 0.0 0.0 0.0

EIGENVALUE = 9.048751E+02
F O R C E S I N S C A L A R S P R I N G S ( C E L A S 2 )
ELEMENT FORCE ELEMENT FORCE ELEMENT FORCE ELEMENT FORCE
ID. ID. ID. ID.
11 9.048751E+01 12 9.512491E+02

EIGENVALUE = 1.105125E+03
F O R C E S I N S C A L A R S P R I N G S ( C E L A S 2 )
ELEMENT FORCE ELEMENT FORCE ELEMENT FORCE ELEMENT FORCE
ID. ID. ID. ID.
11 1.105125E+02 12 -1.051249E+03

Main Index
60

Cantilever Beam Model

This example is a fixed-free aluminum cantilever beam with properties as shown in Figure 3-6.

y
y

x z
L
r
4
L = 3.0 m r = 0.014 m J = 6.0E-8 m
2 4 3
A = 6.158E-4 m I1 = I2 = 3.0E-8 m ρ w = 2.65E4 N/m
2
E = 7.1E10 N/m ν = 0.33

Nonstructural = 2.414 N/m

Weight

Figure 3-6 Cantilever Beam Model

The ρw term is the weight density and must be converted to mass density ρ m for consistency of
units. PARAM,WTMASS is used to convert this weight density to mass density.
2 2
WTMASS = 1 ⁄ g = 1 ⁄ 9.81 = 0.102 sec ⁄ m where g is the acceleration of gravity in m/sec .
Therefore, ρ m = ρ w ⋅ WTMASS = 2.65E4 ⋅ ( 0.102 ) = 2703 kg/m3 . The nonstructural weight of 2.414
N/m is added to the beam. This nonstructural weight per length is also scaled by
PARAM,WTMASS.
The example illustrates normal modes analysis (SOL 103) using the Sturm modified inverse
power method (METHOD = SINV on the EIGR entry). Mass normalization (the default) is
chosen for the eigenvectors. All frequencies between 0 and 50 Hz are requested. Two models
are run. In the first model, manufacturing tolerances make the cross section slightly out of
round, making I1 and I2 slightly different. In the second model, the cross section is perfectly
round, making I1 and I2 identical.
Consider the first model. Due to the manufacturing tolerances, I1 = 2.9E-8 m4 and I2 = 3.1E-8 m4 .
The input file is shown in Listing 3-3.

Main Index
 CHAPTER 3 61
Real Eigenvalue Analysis

Listing 3-3 Input File for the First Beam Model

$ FILE bd03bar1.dat
$
$ CANTILEVER BEAM MODEL
$ CHAPTER 3, NORMAL MODES
$
SOL 103 $ NORMAL MODES ANALYSIS
TIME 10
CEND
$
TITLE = CANTILEVER BEAM
SUBTITLE = NORMAL MODES
LABEL = MODEL 1 (I1 NE I2)
$
SPC = 1
$
$ OUTPUT REQUEST
DISPLACEMENT = ALL
$
$ SELECT EIGR ENTRY
METHOD = 10
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$EIGR SID METHOD F1 F2 NE ND +EIG
$+EIG NORM G C
EIGR 10 SINV 0. 50.
$
$ ALUMINUM PROPERTIES:
$ E = 7.1E10 N/m**2, NU = 0.33, RHO = 2.65E4 N/m**3 (W E I G H T DENSITY)
$MAT1 MID E G NU RHO
MAT1 1 7.1+10 0.33 2.65+4
$
$ CONVERT WEIGHT TO MASS: MASS = (1/G)*WEIGHT
$ G = 9.81 m/sec**2 --> WTMASS = 1/G = 0.102
PARAM WTMASS 0.102
$
$ I1 AND I2 SLIGHTLY DIFFERENT DUE TO MANUFACTURING TOLERANCE
$ ADD NONSTRUCTURAL WEIGHT OF 2.414 N/M
PBAR 1 1 6.158-4 2.9-8 3.1-8 6.-8 2.414
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
CBAR 1 1 1 2 0. 1. 0.
CBAR 2 1 2 3 0. 1. 0.
CBAR 3 1 3 4 0. 1. 0.
CBAR 4 1 4 5 0. 1. 0.
CBAR 5 1 5 6 0. 1. 0.
CBAR 6 1 6 7 0. 1. 0.
CBAR 7 1 7 8 0. 1. 0.
CBAR 8 1 8 9 0. 1. 0.
CBAR 9 1 9 10 0. 1. 0.
CBAR 10 1 10 11 0. 1. 0.
GRID 1 0.0 0. 0.
GRID 2 0.3 0. 0.
GRID 3 0.6 0. 0.
GRID 4 0.9 0. 0.
GRID 5 1.2 0. 0.
GRID 6 1.5 0. 0.
GRID 7 1.8 0. 0.
GRID 8 2.1 0. 0.
GRID 9 2.4 0. 0.
GRID 10 2.7 0. 0.
GRID 11 3.0 0. 0.
SPC1 1 123456 1
$
ENDDATA

Main Index
62

The first two resulting y-direction modes are illustrated in Figure 3-7. Displacements in the y-
direction displacements are controlled by the I1 term. Because the structure is also free to
displace in the z-direction, similar modes occur in that direction and are controlled by the I2
term.

Figure 3-7 First Two Mode Shapes in the Y-Direction

Printed output is shown in Listing 3-4. Note that modes 1 and 3 are y-direction (T2) modes and
modes 2 and 4 are z-direction (T3) modes.

Main Index
 CHAPTER 3 63
Real Eigenvalue Analysis

Listing 3-4 Printed Results from the First Model

E I G E N V A L U E A N A L Y S I S S U M M A R Y (STURM INVERSE POWER)

NUMBER OF EIGENVALUES EXTRACTED . . . . . . 6
NUMBER OF TRIANGULAR DECOMPOSITIONS . . . . 9
TOTAL NUMBER OF VECTOR ITERATIONS . . . . . 59
REASON FOR TERMINATION: ALL EIGENVALUES FOUND IN RANGE.

R E A L E I G E N V A L U E S
MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED
NO. ORDER MASS STIFFNESS
1 1 1.629657E+02 1.276580E+01 2.031740E+00 1.000000E+00 1.629657E+02
2 2 1.742047E+02 1.319866E+01 2.100632E+00 1.000000E+00 1.742047E+02
3 3 6.258656E+03 7.911166E+01 1.259101E+01 1.000000E+00 6.258656E+03
4 4 6.690287E+03 8.179417E+01 1.301795E+01 1.000000E+00 6.690287E+03
5 6 4.809111E+04 2.192968E+02 3.490218E+01 1.000000E+00 4.809111E+04
6 5 5.140773E+04 2.267327E+02 3.608563E+01 1.000000E+00 5.140773E+04

EIGENVALUE = 1.629657E+02
CYCLES = 2.031740E+00 R E A L E I G E N V E C T O R N O . 1

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 0.0 1.391872E-02 2.259793E-10 0.0 -1.470464E-09 9.057010E-02
3 G 0.0 5.301210E-02 8.606855E-10 0.0 -2.725000E-09 1.678406E-01
4 G 0.0 1.133067E-01 1.839608E-09 0.0 -3.765826E-09 2.319480E-01
5 G 0.0 1.908986E-01 3.099362E-09 0.0 -4.598456E-09 2.832321E-01
6 G 0.0 2.820258E-01 4.578870E-09 0.0 -5.232973E-09 3.223138E-01
7 G 0.0 3.831632E-01 6.220901E-09 0.0 -5.685052E-09 3.501587E-01
8 G 0.0 4.911331E-01 7.973862E-09 0.0 -5.976744E-09 3.681249E-01
9 G 0.0 6.032288E-01 9.793808E-09 0.0 -6.137029E-09 3.779972E-01
10 G 0.0 7.173455E-01 1.164657E-08 0.0 -6.202180E-09 3.820101E-01
11 G 0.0 8.321185E-01 1.350998E-08 0.0 -6.215970E-09 3.828595E-01

EIGENVALUE = 1.742047E+02
CYCLES = 2.100632E+00 R E A L E I G E N V E C T O R N O . 2

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G -1.363453E-33 2.249974E-10 -1.391872E-02 0.0 9.057010E-02 1.464532E-09
3 G -2.693333E-33 8.574880E-10 -5.301210E-02 0.0 1.678406E-01 2.716750E-09
4 G -3.956893E-33 1.834005E-09 -1.133067E-01 0.0 2.319480E-01 3.758543E-09
5 G -5.123021E-33 3.092054E-09 -1.908986E-01 0.0 2.832321E-01 4.594735E-09
6 G -6.163003E-33 4.571193E-09 -2.820258E-01 0.0 3.223138E-01 5.234397E-09
7 G -7.051232E-33 6.214489E-09 -3.831632E-01 0.0 3.501587E-01 5.692031E-09
8 G -7.765834E-33 7.970303E-09 -4.911331E-01 0.0 3.681249E-01 5.988579E-09
9 G -8.289217E-33 9.794348E-09 -6.032288E-01 0.0 3.779972E-01 6.152242E-09
10 G -8.608492E-33 1.165196E-08 -7.173455E-01 0.0 3.820101E-01 6.219055E-09
11 G -8.715798E-33 1.352052E-08 -8.321185E-01 0.0 3.828595E-01 6.233260E-09

EIGENVALUE = 6.258656E+03
CYCLES = 1.259101E+01 R E A L E I G E N V E C T O R N O . 3

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 0.0 -7.568120E-02 -9.828895E-08 0.0 5.937006E-07 -4.571418E-01
3 G 0.0 -2.464387E-01 -3.200556E-07 0.0 8.257535E-07 -6.358200E-01
4 G 0.0 -4.318525E-01 -5.608564E-07 0.0 7.289605E-07 -5.612913E-01
5 G 0.0 -5.632111E-01 -7.314543E-07 0.0 3.723918E-07 -2.867391E-01
6 G 0.0 -5.916957E-01 -7.684467E-07 0.0 -1.426249E-07 1.098159E-01
7 G 0.0 -4.941767E-01 -6.417950E-07 0.0 -6.984853E-07 5.378201E-01
8 G 0.0 -2.743728E-01 -3.563287E-07 0.0 -1.184615E-06 9.121327E-01
9 G 0.0 4.170797E-02 5.417504E-08 0.0 -1.522753E-06 1.172494E+00
10 G 0.0 4.159041E-01 5.401546E-07 0.0 -1.689206E-06 1.300660E+00
11 G 0.0 8.124724E-01 1.055190E-06 0.0 -1.730573E-06 1.332512E+00

EIGENVALUE = 6.690287E+03
CYCLES = 1.301795E+01 R E A L E I G E N V E C T O R N O . 4

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 3.698388E-22 1.207686E-08 7.568131E-02 0.0 -4.571424E-01 1.427123E-08
3 G 7.305710E-22 -2.529794E-08 2.464390E-01 0.0 -6.358204E-01 -3.100643E-07
4 G 1.073314E-21 -1.866785E-07 4.318528E-01 0.0 -5.612913E-01 -7.615683E-07
5 G 1.389628E-21 -4.615435E-07 5.632114E-01 0.0 -2.867385E-01 -1.006117E-06
6 G 1.671725E-21 -7.409781E-07 5.916957E-01 0.0 1.098168E-01 -7.569575E-07
7 G 1.912658E-21 -8.602044E-07 4.941765E-01 0.0 5.378208E-01 4.733732E-08
8 G 2.106496E-21 -6.796384E-07 2.743725E-01 0.0 9.121327E-01 1.181005E-06
9 G 2.248464E-21 -1.591279E-07 -4.170818E-02 0.0 1.172493E+00 2.239405E-06
10 G 2.335068E-21 6.217935E-07 -4.159040E-01 0.0 1.300659E+00 2.875692E-06
11 G 2.364175E-21 1.520816E-06 -8.124720E-01 0.0 1.332511E+00 3.057266E-0

Main Index
64

Now, consider the second model for which I1 and I2 are identical. Printed output is shown in
Listing 3-5. Note that modes 1 and 2; 3 and 4; 5 and 6; etc., have identical frequencies; this is a
case of repeated roots. Also note that the eigenvectors are not pure y- or pure z-translation (as
they were in the first model); the eigenvectors are linear combinations of the y and z modes since
this model has repeated roots.

Main Index
 CHAPTER 3 65
Real Eigenvalue Analysis

Listing 3-5 Printed Results from the Second Model

E I G E N V A L U E A N A L Y S I S S U M M A R Y (STURM INVERSE POWER)

NUMBER OF EIGENVALUES EXTRACTED . . . . . . 8
NUMBER OF TRIANGULAR DECOMPOSITIONS . . . . 6
TOTAL NUMBER OF VECTOR ITERATIONS . . . . . 68
REASON FOR TERMINATION: ALL EIGENVALUES FOUND IN RANGE.

R E A L E I G E N V A L U E S
MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED
NO. ORDER MASS STIFFNESS
1 1 1.685851E+02 1.298403E+01 2.066473E+00 1.000000E+00 1.685851E+02
2 3 1.685851E+02 1.298403E+01 2.066473E+00 1.000000E+00 1.685851E+02
3 2 6.474471E+03 8.046410E+01 1.280626E+01 1.000000E+00 6.474471E+03
4 5 6.474471E+03 8.046410E+01 1.280626E+01 1.000000E+00 6.474471E+03
5 4 4.974941E+04 2.230458E+02 3.549883E+01 1.000000E+00 4.974941E+04
6 6 4.974941E+04 2.230458E+02 3.549883E+01 1.000000E+00 4.974941E+04
7 7 1.870792E+05 4.325266E+02 6.883875E+01 1.000000E+00 1.870792E+05
8 8 1.870792E+05 4.325266E+02 6.883875E+01 1.000000E+00 1.870792E+05
EIGENVALUE = 1.685851E+02
CYCLES = 2.066473E+00 R E A L E I G E N V E C T O R N O . 1

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 4.198658E-20 9.109908E-03 1.052335E-02 0.0 -6.847622E-02 5.927882E-02
3 G 8.293447E-20 3.469681E-02 4.008020E-02 0.0 -1.268971E-01 1.098529E-01
4 G 1.218317E-19 7.416008E-02 8.566642E-02 0.0 -1.753660E-01 1.518116E-01
5 G 1.577188E-19 1.249444E-01 1.443303E-01 0.0 -2.141397E-01 1.853772E-01
6 G 1.897125E-19 1.845877E-01 2.132276E-01 0.0 -2.436876E-01 2.109563E-01
7 G 2.170278E-19 2.507827E-01 2.896933E-01 0.0 -2.647398E-01 2.291808E-01
8 G 2.389961E-19 3.214496E-01 3.713246E-01 0.0 -2.783232E-01 2.409396E-01
9 G 2.550813E-19 3.948168E-01 4.560753E-01 0.0 -2.857873E-01 2.474010E-01
10 G 2.648916E-19 4.695067E-01 5.423540E-01 0.0 -2.888212E-01 2.500274E-01
11 G 2.681884E-19 5.446261E-01 6.291288E-01 0.0 -2.894634E-01 2.505833E-01
EIGENVALUE = 1.685851E+02
CYCLES = 2.066473E+00 R E A L E I G E N V E C T O R N O . 2

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 1.219015E-33 -1.052334E-02 9.109882E-03 0.0 -5.927866E-02 -6.847615E-02
3 G 2.407892E-33 -4.008016E-02 3.469673E-02 0.0 -1.098526E-01 -1.268970E-01
4 G 3.537263E-33 -8.566635E-02 7.415994E-02 0.0 -1.518114E-01 -1.753659E-01
5 G 4.579275E-33 -1.443302E-01 1.249442E-01 0.0 -1.853771E-01 -2.141396E-01
6 G 5.508284E-33 -2.132276E-01 1.845875E-01 0.0 -2.109563E-01 -2.436876E-01
7 G 6.301480E-33 -2.896932E-01 2.507825E-01 0.0 -2.291809E-01 -2.647399E-01
8 G 6.939436E-33 -3.713246E-01 3.214495E-01 0.0 -2.409399E-01 -2.783234E-01
9 G 7.406566E-33 -4.560753E-01 3.948168E-01 0.0 -2.474014E-01 -2.857874E-01
10 G 7.691474E-33 -5.423541E-01 4.695069E-01 0.0 -2.500278E-01 -2.888214E-01
11 G 7.787220E-33 -6.291289E-01 5.446264E-01 0.0 -2.505838E-01 -2.894636E-01
EIGENVALUE = 6.474471E+03
CYCLES = 1.280626E+01 R E A L E I G E N V E C T O R N O . 3

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G -1.606174E-26 6.893317E-02 3.123881E-02 0.0 -1.886937E-01 4.163812E-01
3 G -3.172614E-26 2.244652E-01 1.017221E-01 0.0 -2.624463E-01 5.791277E-01
4 G -4.660607E-26 3.933468E-01 1.782550E-01 0.0 -2.316833E-01 5.112443E-01
5 G -6.033448E-26 5.129929E-01 2.324757E-01 0.0 -1.183568E-01 2.611722E-01
6 G -7.257351E-26 5.389377E-01 2.442332E-01 0.0 4.532863E-02 -1.000244E-01
7 G -8.302282E-26 4.501138E-01 2.039804E-01 0.0 2.219953E-01 -4.898660E-01
8 G -9.142668E-26 2.499085E-01 1.132523E-01 0.0 3.764997E-01 -8.308035E-01
9 G -9.757999E-26 -3.798927E-02 -1.721591E-02 0.0 4.839685E-01 -1.067950E+00
10 G -1.013329E-25 -3.788206E-01 -1.716723E-01 0.0 5.368715E-01 -1.184688E+00
11 G -1.025940E-25 -7.400293E-01 -3.353632E-01 0.0 5.500190E-01 -1.213700E+00
EIGENVALUE = 6.474471E+03
CYCLES = 1.280626E+01 R E A L E I G E N V E C T O R N O . 4

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G -6.181717E-19 3.123822E-02 -6.893279E-02 0.0 4.163791E-01 1.886905E-01
3 G -1.221121E-18 1.017205E-01 -2.244643E-01 0.0 5.791260E-01 2.624437E-01
4 G -1.794002E-18 1.782531E-01 -3.933455E-01 0.0 5.112445E-01 2.316835E-01
5 G -2.322707E-18 2.324743E-01 -5.129921E-01 0.0 2.611744E-01 1.183602E-01
6 G -2.794219E-18 2.442331E-01 -5.389376E-01 0.0 -1.000213E-01 -4.532376E-02
7 G -3.196926E-18 2.039817E-01 -4.501146E-01 0.0 -4.898637E-01 -2.219917E-01
8 G -3.520915E-18 1.132542E-01 -2.499097E-01 0.0 -8.308033E-01 -3.764996E-01
9 G -3.758207E-18 -1.721460E-02 3.798841E-02 0.0 -1.067952E+00 -4.839722E-01
10 G -3.902959E-18 -1.716725E-01 3.788207E-01 0.0 -1.184692E+00 -5.368777E-01
11 G -3.951610E-18 -3.353654E-01 7.400307E-01 0.0 -1.213704E+00 -5.500259E-01

Main Index
66

This second model was rerun, changing the mass from lumped (the default) to coupled by
adding PARAM,COUPMASS,1 to the Bulk Data. The resulting frequencies are shown below
and are compared to those of the lumped mass model and the theoretical results. Note that the
frequency difference is greater at higher frequencies. For most production-type models (i.e.,
complex three-dimensional structures), this difference is negligible.

Frequencies (Hz)
Lumped Mass Coupled Mass
Theory
Model Model
2.076 2.066 2.076
13.010 12.806 13.010
36.428 35.499 36.437
71.384 68.838 71.451

Bracket Model
This example is a steel bracket as shown in Figure 3-8.

Figure 3-8 Bracket Model

A concentrated mass is suspended from the center of the hole in bracket. This mass has the
following properties:

2
m = 0.0906 lb-sec /in
2
I11 = 0.35 in-lb-sec
2
I22 = 0.56 in-lb-sec
2
I33 = 0.07 in-lb-sec

The concentrated mass (grid point 999) is connected to the bracket by an RBE2 element
connecting 24 grid points, as shown in Figure 3-9.

Main Index
 CHAPTER 3 67
Real Eigenvalue Analysis

Figure 3-9 Bracket Model Showing RBE2 Element (Dashed Lines)

The bracket is clamped by constraining six degrees-of-freedom for each of 12 grid points near
the base.
This example illustrates a normal modes analysis (SOL 103) using the Lanczos method (EIGRL
entry). All frequencies below 100 Hz are requested. The MODES Case Control command is used
to specify the number of times a subcase is repeated and therefore enables different output
requests for each mode. The output requests for this problem are eigenvectors for all modes
(DISPLACEMENT = ALL above the subcase level), corner stresses for the first two modes
(STRESS(CORNER) = ALL and MODES = 2 in Subcase 1), and element strain energies for the
third mode (ESE = ALL). An abridged version of the input file is shown in Listing 3-6.

Main Index
68

Listing 3-6 Abridged Input File for the Bracket Model

$ FILE bd03bkt.dat
$
$ BRACKET MODEL
$ CHAPTER 3, NORMAL MODES
$
TIME 10
SOL 103 $ NORMAL MODES ANALYSIS
CEND
$
TITLE = BRACKET MODEL
SUBTITLE = NORMAL MODES ANALYSIS
$
SPC = 1
$
$ SELECT EIGRL
METHOD = 777
$
$ OUTPUT REQUESTS
DISPLACEMENT = ALL
SUBCASE 1
MODES = 2 $ USE FOR FIRST TWO MODES
STRESS(CORNER) = ALL
SUBCASE 3
ESE = ALL
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$EIGRL SID V1 V2 ND MSGLVL MAXSET SHFSCL NORM
EIGRL 777 100.
$
$ CONCENTRATED MASS, SUPPORTED WITH AN RBE2 ELEMENT
$
GRID 999 3.0 3.0 3.8
$
$CONM2 EID G CID M X1 X2 X3 +CON1
$+CON1 I11 I21 I22 I31 I32 I33
CONM2 999 999 0.0906 +CM1
+CM1 0.35 0.56 0.07
$
RBE2 999 999 123456 126 127 91 81 80 +RB1
+RB1 90 95 129 128 96 86 85 228 +RB2
+RB2 229 199 189 190 200 195 227 226 +RB3
+RB3 194 184 185
$
$ STEEL, M A S S DENSITY FOR RHO
MAT1 1 3.+7 1.153+7 7.76-4
$
... basic model ...
$
ENDDATA

Listing 3-7 shows an abridged version of the resulting MSC.Nastran output. The circular total
element strain energy (ESE) for each mode of the entire model is ESE = ω 2i ⁄ 2 for the i-th mode
when φ Ti Mφi = 1 . The frequency of the third mode is 234.49 radians per second squared;
therefore, the total strain energy of the model is ESE total = ( 234.49 )2 ⁄ 2 = 27493 for the third mode.

Main Index
 CHAPTER 3 69
Real Eigenvalue Analysis

The printed ESE total = 27492 . The eigenvectors are printed for each mode, the element corner
stresses are printed for the first and second modes, and the element strain energies are printed
for the third mode. (Only the headers are shown in the figure in order to save space.)

Listing 3-7 Abridged Output from the Bracket Model

E I G E N V A L U E A N A L Y S I S S U M M A R Y (LANCZOS ITERATION)
BLOCK SIZE USED ...................... 6
NUMBER OF DECOMPOSITIONS ............. 2
NUMBER OF ROOTS FOUND ................ 3
NUMBER OF SOLVES REQUIRED ............ 3
TERMINATION MESSAGE : REQUIRED NUMBER OF EIGENVALUES FOUND.

R E A L E I G E N V A L U E S
MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED
NO. ORDER MASS STIFFNESS
1 1 3.930304E+03 6.269214E+01 9.977763E+00 1.000000E+00 3.930304E+03
2 2 2.878402E+04 1.696586E+02 2.700200E+01 1.000000E+00 2.878402E+04
3 3 5.498442E+04 2.344876E+02 3.731985E+01 1.000000E+00 5.498442E+04

EIGENVALUE = 3.930304E+03
CYCLES = 9.977763E+00 R E A L E I G E N V E C T O R N O . 1

EIGENVALUE = 2.878402E+04
CYCLES = 2.700200E+01 R E A L E I G E N V E C T O R N O . 2

EIGENVALUE = 5.498442E+04
CYCLES = 3.731985E+01 R E A L E I G E N V E C T O R N O . 3

EIGENVALUE = 3.930304E+03
S T R E S S E S I N Q U A D R I L A T E R A L E L E M E N T S ( Q U A D 4 ) OPTION = CUBIC

EIGENVALUE = 2.878402E+04
S T R E S S E S I N Q U A D R I L A T E R A L E L E M E N T S ( Q U A D 4 ) OPTION = CUBIC

E L E M E N T S T R A I N E N E R G I E S

ELEMENT-TYPE = QUAD4 * TOTAL ENERGY OF ALL ELEMENTS IN PROBLEM = 2.749221E+04

MODE 3 TOTAL ENERGY OF ALL ELEMENTS IN SET -1 = 2.749221E+04
*
TYPE = QUAD4 SUBTOTAL 2.749221E+04 100.0000

The deformed shape resulting from the first mode is illustrated in Figure 3-10 and is overlaid on
the undeformed shape. Figure 3-11 illustrates the stress contours plotted on the deformed shape
of the second mode. The element strain energy contour plot for the third mode is shown in
Figure 3-12.

Main Index
70

Figure 3-10 Deformed Shape of the First Mode

Figure 3-11 Second Mode Deformation with Element Stress Contours

Main Index
 CHAPTER 3 71
Real Eigenvalue Analysis

Figure 3-12 Element Strain Energy Contours for the Third Mode

Car Frame Model

Figure 3-13 shows a model of an aluminum car frame. The frame model is comprised of plate
elements (CQUAD4 and CTRIA3), with springs (CELAS2) representing the suspension. Spring
stiffnesses are input in the three translational directions; a stiffness of 500 lb/in is used in the
vertical direction (T2), and stiffnesses of 1000 lb/in are used in the other translational directions
(T1 and T3). When using CELASi elements to connect two grid points, it is recommended that
the coordinates of the two grid points be identical in order to represent coaxial springs
(noncoincident coordinates can lead to errors). The goal of the analysis is to compute resonant
frequencies up to 50 Hz using the Lanczos method. Element strain energies are computed for
the springs in order to help characterize the resulting modes.
Listing 3-8 shows the input file. Modal displacements are written to the plot file and are not
printed. The rigid body mass matrix (see “Parameters for Dynamic Analysis” on page 435) is
computed via the PARAM,GRDPNT,0 entry. "Include" files are used to partition the input file
into several smaller files. The INCLUDE statement inserts an external file into the input file. The
basic file is bd03car.dat. The springs are contained in file bd03cars.dat (Listing 3-9), and the rest
of the input file is contained in file bd03carb.dat (not shown).

Main Index
72

Figure 3-13 Car Frame Model

Listing 3-8 Basic Input File for the Car Model

$ FILE bd03car.dat
$
$ CAR FRAME MODEL
$ CHAPTER 3, NORMAL MODES
$
$ MODEL COURTESY LAPCAD ENGINEERING
$ CHULA VISTA, CALIFORNIA
$
SOL 103 $ NORMAL MODES ANALYSIS
TIME 30
CEND
$
TITLE = CAR MODEL WITH SUSPENSION SPRINGS
SUBTITLE = MODAL ANALYSIS CASE CONTROL
ECHO = UNSORT
$
METHOD = 1
$
DISPLACEMENT(PLOT) = ALL
SET 99 = 1001,1002,1003,1011,1012,1013,
1021,1022,1023,1031,1032,1033
ESE = 99
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
INCLUDE 'bd03cars.dat' $ Car springs
$
$EIGRL SID V1 V2 ND MSGLVL MAXSET SHFSCL NORM
EIGRL 1 -1. 50.
$
$MAT1 MID E G NU RHO
MAT1 1 1.0E7 0.33 2.59E-4
$
$ PRINT RIGID-BODY MASS
PARAM,GRDPNT,0
$
ECHOOFF
INCLUDE 'bd03carb.dat' $ Rest of Bulk Data
$
ENDDATA

Main Index
 CHAPTER 3 73
Real Eigenvalue Analysis

Listing 3-9 Input File for the Springs

$ FILE bd03cars.dat
$
$ SPRINGS CONNECTED TO GROUND
$ GRIDS 1059,1562,1428,1895 HAVE THE SAME COORDS. AS 59,562,428,895
$
GRID 1059 152.012 32.7964 -2.90000 123456
GRID 1562 152.012 32.7964 -67.1000 123456
GRID 1428 35.6119 30.8257 -0.66667 123456
GRID 1895 35.6119 30.8257 -69.3333 123456
$
CELAS2 1001 1000. 59 1 1059 1
CELAS2 1002 500. 59 2 1059 2
CELAS2 1003 1000. 59 3 1059 3
CELAS2 1011 1000. 562 1 1562 1
CELAS2 1012 500. 562 2 1562 2
CELAS2 1013 1000. 562 3 1562 3
CELAS2 1021 1000. 428 1 1428 1
CELAS2 1022 500. 428 2 1428 2
CELAS2 1023 1000. 428 3 1428 3
CELAS2 1031 1000. 895 1 1895 1
CELAS2 1032 500. 895 2 1895 2
CELAS2 1033 1000. 895 3 1895 3

Listing 3-10 shows the grid point weight generator output. The grid point weight generator
indicates that the translational mass is 0.8027 lb-sec 2 ⁄ in. . This type of information is useful in
verifying your model. Listing 3-11 shows some of the rest of the output. The eigenvalue
analysis summary indicates that there are 12 modes below 50 Hz. The large element strain
energies in the first six modes indicate that these are primarily suspension modes, comprised of
the car frame acting as a rigid body supported by the flexible springs. Approximately 95% of the
total strain energy is represented by spring deformation as opposed to frame deformation.
Modes 7 and above show insignificant spring strain energy, indicating that these are primarily
frame modes.

Listing 3-10 Output from the Grid Point Weight Generator

O U T P U T F R O M G R I D P O I N T W E I G H T G E N E R A T O R
REFERENCE POINT = 0
M O
* 8.027376E-01 0.000000E+00 0.000000E+00 0.000000E+00 -2.808118E+01 -2.179473E+01 *
* 0.000000E+00 8.027376E-01 0.000000E+00 2.808118E+01 0.000000E+00 7.610537E+01 *
* 0.000000E+00 0.000000E+00 8.027376E-01 2.179473E+01 -7.610537E+01 0.000000E+00 *
* 0.000000E+00 2.808118E+01 2.179473E+01 2.325008E+03 -2.153940E+03 2.662697E+03 *
* -2.808118E+01 0.000000E+00 -7.610537E+01 -2.153940E+03 1.020870E+04 7.626398E+02 *
* -2.179473E+01 7.610537E+01 0.000000E+00 2.662697E+03 7.626398E+02 9.437676E+03 *
S
* 1.000000E+00 0.000000E+00 0.000000E+00 *
* 0.000000E+00 1.000000E+00 0.000000E+00 *
* 0.000000E+00 0.000000E+00 1.000000E+00 *
DIRECTION
MASS AXIS SYSTEM (S) MASS X-C.G. Y-C.G. Z-C.G.
X 8.027376E-01 0.000000E+00 2.715050E+01 -3.498177E+01
Y 8.027376E-01 9.480728E+01 0.000000E+00 -3.498177E+01
Z 8.027376E-01 9.480728E+01 2.715050E+01 0.000000E+00
I(S)
* 7.509408E+02 8.764180E+01 -3.962963E-01 *
* 8.764180E+01 2.011031E+03 -2.217236E-01 *
* -3.962963E-01 -2.217236E-01 1.630595E+03 *
I(Q)
* 2.017097E+03 *
* 7.448741E+02 *
* 1.630595E+03 *
Q
* 6.905332E-02 9.976128E-01 5.012719E-04 *
* -9.976128E-01 6.905355E-02 -4.555999E-04 *
* -4.891269E-04 -4.686146E-04 9.999998E-01 *

Main Index
74

Listing 3-11 Abridged Output from the Car Model

E I G E N V A L U E A N A L Y S I S S U M M A R Y (LANCZOS ITERATION)
BLOCK SIZE USED ...................... 7
NUMBER OF DECOMPOSITIONS ............. 3
NUMBER OF ROOTS FOUND ................ 12
NUMBER OF SOLVES REQUIRED ............ 11
TERMINATION MESSAGE : REQUIRED NUMBER OF EIGENVALUES FOUND.

R E A L E I G E N V A L U E S
MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED
NO. ORDER MASS STIFFNESS
1 1 2.346479E+03 4.844047E+01 7.709540E+00 1.000000E+00 2.346479E+03
2 2 2.654886E+03 5.152559E+01 8.200551E+00 1.000000E+00 2.654886E+03
3 3 3.769821E+03 6.139887E+01 9.771934E+00 1.000000E+00 3.769821E+03
4 4 4.633242E+03 6.806792E+01 1.083335E+01 1.000000E+00 4.633242E+03
5 5 5.078395E+03 7.126286E+01 1.134184E+01 1.000000E+00 5.078395E+03
6 6 8.485758E+03 9.211817E+01 1.466106E+01 1.000000E+00 8.485758E+03
7 7 2.805541E+04 1.674975E+02 2.665805E+01 1.000000E+00 2.805541E+04
8 8 5.350976E+04 2.313218E+02 3.681600E+01 1.000000E+00 5.350976E+04
9 9 5.940912E+04 2.437399E+02 3.879240E+01 1.000000E+00 5.940912E+04
10 10 8.476198E+04 2.911391E+02 4.633622E+01 1.000000E+00 8.476198E+04
11 11 9.134271E+04 3.022296E+02 4.810133E+01 1.000000E+00 9.134271E+04
12 12 9.726959E+04 3.118807E+02 4.963736E+01 1.000000E+00 9.726959E+04

E L E M E N T S T R A I N E N E R G I E S

ELEMENT-TYPE = ELAS2 * TOTAL ENERGY OF ALL ELEMENTS IN PROBLEM = 1.173240E+03

MODE 1 TOTAL ENERGY OF ALL ELEMENTS IN SET 99 = 1.104569E+03
*
ELEMENT-ID STRAIN-ENERGY PERCENT OF TOTAL STRAIN-ENERGY-DENSITY
1002 2.735009E+02 23.3116
1003 4.059090E-02 .0035
1012 2.685884E+02 22.8929
1013 3.240471E-02 .0028
1021 8.017746E-02 .0068
1022 2.833448E+02 24.1506
1023 3.963123E-01 .0338
1031 7.903841E-02 .0067
1032 2.781467E+02 23.7076
1033 3.573737E-01 .0305

TYPE = ELAS2 SUBTOTAL 1.104569E+03 94.1469

E L E M E N T S T R A I N E N E R G I E S

ELEMENT-TYPE = ELAS2 * TOTAL ENERGY OF ALL ELEMENTS IN PROBLEM = 1.402770E+04

MODE 7 TOTAL ENERGY OF ALL ELEMENTS IN SET 99 = 3.172818E+03
*
ELEMENT-ID STRAIN-ENERGY PERCENT OF TOTAL STRAIN-ENERGY-DENSITY
1001 1.278503E+02 .9114
1002 5.884620E+02 4.1950
1003 8.020268E+01 .5717
1011 1.278420E+02 .9114
1012 5.883779E+02 4.1944
1013 8.022697E+01 .5719
1021 7.525866E+01 .5365
1022 4.885996E+02 3.4831
1023 2.255234E+02 1.6077
1031 7.528390E+01 .5367
1032 4.896509E+02 3.4906
1033 2.255398E+02 1.6078

TYPE = ELAS2 SUBTOTAL 3.172818E+03 22.6182

Mode shapes for modes 7, 8, 9, and 10 are shown in Figure 3-14. Mode 7 is an overall twisting
mode; mode 8 is a "roof collapse" mode; mode 9 is a local (front) roof mode; and mode 10 is a
local rear mode. Plots such as these, in conjunction with element strain energies, help to
illustrate each of the mode shapes.

Main Index
 CHAPTER 3 75
Real Eigenvalue Analysis

Figure 3-14 Mode Shapes for Modes 7, 8, 9, and 10

Main Index
76

Test Fixture Model

This example is an aluminum test fixture, which is shown in Figure 3-15. The model is
comprised of 8157 grid points, 5070 CHEXA elements, and 122 CPENTA elements. The primary
plates are 1 inch thick, and the gusset plates are 0.5 inch thick. The base of the fixture is
constrained to have no vertical (y) motion, and the bolt holes at the base are constrained to also
have no horizontal (x and z) motion.

Figure 3-15 Test Fixture Model

A portion of the input file is shown in Listing 3-12. The Lanczos method is used to compute the
modes. The first six modes are requested (ND is 6, with V1 and V2 blank). The GRDSET Bulk
Data entry removes the rotational DOFs (456) from the analysis since the solid elements have no
rotational stiffness. The Bulk Data is in free format.

Main Index
 CHAPTER 3 77
Real Eigenvalue Analysis

Listing 3-12 Abridged Input File for Test Fixture Model

$ FILE bd03fix.dat
$
$ TEST FIXTURE
$ CHAPTER 3, NORMAL MODES
$
TIME 240
SOL 103
CEND
ECHO = NONE
DISPLACEMENT(PLOT) = ALL
SPC = 1
$
METHOD = 1
$
BEGIN BULK
$
$GRDSET, ,CP, , , , CD, PS
GRDSET , , , , , , , 456
$
$EIGRL, SID, V1, V2, ND, MSGLVL, MAXSET, SHFSCL, NORM
EIGRL , 1, , , 6
$
$MAT1, MID, E, G, NU, RHO
MAT1 , 3, 1.00E7, , 0.334, 2.5383-4
$
... basic model ...
$
ENDDATA

Figure 3-16 shows the first four mode shapes. The first mode is a bending mode, the second and
third modes are twist modes, and the fourth mode is a bending mode.

Main Index
78

Figure 3-16 Test Fixture Mode Shapes

Quarter Plate Model

This example is a quarter model of a simply supported flat plate, shown in Figure 3-17. This
example illustrates the use of multiple boundary conditions (new in Version 68) for modeling
symmetric structures. In this case the plate is doubly symmetric.

Main Index
 CHAPTER 3 79
Real Eigenvalue Analysis

Lines of Symmetry

ss ss

ss ss
ss ss

ss ss
Full Model
1
z
y ss
2
x
ss
Quarter Model
Figure 3-17 Derivation of Quarter Plate Model

A portion of the input file is shown in Listing 3-13. Four subcases are used—one for each of the
following sets of boundary conditions for the quarter model:

• Symmetric-antisymmetric
• Antisymmetric-symmetric
• Symmetric-symmetric
• Antisymmetric-antisymmetric
The BC Case Control command identifies multiple boundary conditions. The SPCADD Bulk
Data entry defines a union of SPC sets.

Main Index
80

Listing 3-13 Input File (Abridged) for the Quarter Plate Model

$ FILE bd03plt1.dat
$
$ QUARTER PLATE MODEL
$ CHAPTER 3, NORMAL MODES
$
SOL 103 $ NORMAL MODES ANALYSIS
TIME 10
CEND
TITLE = SIMPLY SUPPORTED PLATE USING SYMMETRY
SUBTITLE = NORMAL MODES CASE CONTROL
LABEL = QUARTER PLATE MODEL
$
DISPLACEMENT = ALL
$
SUBCASE 1
LABEL = SYM-ASYM
BC = 1
METHOD = 1
SPC = 101
SUBCASE 2
LABEL = ASYM-SYM
BC = 2
SPC = 102
METHOD = 1
SUBCASE 3
LABEL = SYM-SYM
BC = 3
SPC = 103
METHOD = 1
SUBCASE 4
LABEL = ASYM-ASYM
BC = 4
SPC = 104
METHOD = 1
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10......
$
$ SYM-ASYM
SPCADD 101 11 1 4
$ ASYM-SYM
SPCADD 102 11 2 3
$ SYM-SYM
SPCADD 103 11 1 3
$ ASYM-ASYM
SPCADD 104 11 2 4
$
$EIGRL SID V1 V2
EIGRL 1 -0.1 100.
$
SPC 1 1 246 0.00
SPC 1 2 246 0.00
... etc. ...
SPC 11 80 12356 0.00
SPC 11 81 123456 0.00
$
... basic model ...
$
ENDDATA

Figure 3-18 shows the quarter plate mode shapes and the corresponding mode shapes for a full
model of the same structure. Note that the quarter plate modes match the full plate modes.

Main Index
 CHAPTER 3 81
Real Eigenvalue Analysis

Figure 3-18 Mode Shapes for the Quarter Model (Left) and Full Model (Right)

DMIG Example
This example illustrates the use of a DMIG entry to input external mass and stiffness. The
cantilever beam model shown in Figure 3-19 is used for this purpose. The model consists of
CBAR elements 1 through 4. Element 5 is a model from another subcontractor that is input via
DMIG entries. The model contains two DOFs (R2 and T3) per grid point.

Main Index
82

X
1 2 3 4 5 6
1 2 3 4 5
= Grid Point ID
L = Element ID

Figure 3-19 Planar Cantilever Beam

The stiffness and mass matrices from the contractor for element 5 are as follows:

5, 3 5, 5 6, 3 6, 5
5, 3  500038.8 
 
K5 = 5, 5  – 250019.4 166679.6 
6, 3  – 500038.8 250019.4 500038.8 
 
6, 5  – 250019.4 83339.8 250019.4 166679.6 

5, 3 5, 5 6, 3 6, 5
5, 3  3.5829 0. 0. 0. 
 
M5 = 5, 5  0. 0. 0. 0. 
6, 3  0. 0. 3.5829 0. 
 
6, 5  0. 0. 0. 0. 

Since the matrices are symmetric, only the lower or upper triangular portion of the matrices need
to be provided via the DMIG entries.
The corresponding input file is shown in Listing 3-14.

Main Index
 CHAPTER 3 83
Real Eigenvalue Analysis

Listing 3-14 Input File for the DMIG Example

$ FILE bd03dmi.dat
$
$ DMIG EXAMPLE
$ CHAPTER 3, NORMAL MODES
$
SOL 103 $ NORMAL MODES ANALYSIS
TIME 10
CEND
TITLE = DMIG TO READ STIFFNESS AND MASS FOR ELEM 5
SUBTITLE = PLANAR PROBLEM
$
SPC = 10
$
$ SPECIFY K2GG AND M2GG
K2GG = EXSTIF
M2GG = EXMASS
$
METHOD = 10
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$EIGRL SID V1 V2 ND
EIGRL 10 2
$
CBAR 1 1 1 2 10
CBAR 2 1 2 3 10
CBAR 3 1 3 4 10
CBAR 4 1 4 5 10
$
$ HEADER ENTRY FOR STIFFNESS
DMIG EXSTIF 0 6 1
$
DMIG EXSTIF 5 3 5 3 500039. +000001
++0000015 5 -250019. 6 3 -500039. +000002
++0000026 5 -250019. +000003
$
DMIG EXSTIF 5 5 5 5 166680. +000004
++0000046 3 250019. 6 5 83340. +000005
$
DMIG EXSTIF 6 3 6 3 500039. +000006
++0000066 5 250019. +000007
$
DMIG EXSTIF 6 5 6 5 166680.
$
$ HEADER ENTRY FOR MASS
DMIG EXMASS 0 6 1
$
$ DATA ENTRIES FOR MASS
$
DMIG EXMASS 5 3 5 3 3.5829
DMIG EXMASS 6 3 6 3 3.5829
$
GRID 1 0. 0. 0. 1246
GRID 2 1. 0. 0. 1246
GRID 3 2. 0. 0. 1246
GRID 4 3. 0. 0. 1246
GRID 5 4. 0. 0. 1246
GRID 6 5. 0. 0. 1246
GRID 10 0. 0. 10. 123456
MAT1 1 7.1+10 .33 2700.
PBAR 1 1 2.654-3 5.869-7
SPC1 10 123456 1
$
ENDDATA

Main Index
84

In this example, EXSTIF is chosen as the name of the input stiffness matrix. Therefore, in order
to bring in this stiffness matrix and add it to the global stiffness matrix, the Case Control
command K2GG = EXSTIF is required. Note that these stiffness terms are additions to any
existing terms in the global stiffness matrix at that location and not replacements of the stiffness
terms at that location. In the Bulk Data Section, five DMIG entries are required—one for the
header entry and four for the data column entries since there are four non-null columns in the
above matrix.
For the header entry, the same name EXSTIF must be used to match the name called out in the
Case Control Section. The third field is "0", which must be the value used for the header entry.
The fourth field (IFO) is set to "6" to denote a symmetric matrix input. The fifth field (TIN) is set
to "1" to denote that the matrix is real, single precision.
The terms in the matrix are referenced in terms of their external grid IDs when using the DMIG
entries. Physically, each term in a particular column of the stiffness matrix ( K ij ) represents the
induced reactive load in the i-th degree-of-freedom due to a unit displacement in the j-th
direction with all other displacement degrees-of-freedom held to zero. Since the matrix is
symmetric, only the lower triangular portion of the matrix is input.
The first DMIG data column entry defines the first column of the above matrix. Field 2 of this
DMIG entry must have the same name EXSTIF as referenced by the Case Control K2GG =
EXSTIF command. Fields 3 and 4 of this entry identify this column in terms of its external grid
ID and corresponding degree-of-freedom, respectively. In this case, it is grid point5, degree-of-
freedom 3 (z-translation at grid point 5). Once this column is defined, follow the format
description as described in the section for column data entry format, and then you can input the
four terms in this column row-by-row. These four terms are defined by sets of three fields. They
are the external grid ID number, corresponding degree-of-freedom, and the actual matrix term,
respectively. The first term of column one is defined by external grid ID 5, degree-of-freedom 3
(z-translation at grid point 5) with a stiffness value of 500039. The second term of column one is
defined by external grid ID 5, degree-of-freedom 5 (y-rotation at grid point 5) with a stiffness
value of -250019. The third term of column one is defined by external grid ID 6, degree-of-
freedom 3 (z-translation at grid point 6) with a stiffness value of -500039. The fourth term of
column one is defined by external grid ID 6, degree-of-freedom 5 (y-rotation at grid point 6) with
a stiffness value of -250019.
The next DMIG entry defines the second column of the above matrix. Fields 3 and 4 of this entry
identify this column in terms of its external grid ID and corresponding degree-of-freedom,
respectively. In this case, it is grid point 5, degree-of-freedom 5 (y-rotation at grid point 5). The
rest of the procedure is similar to that of column one with the exception that only three terms
need to be input due to symmetry.
The next two DMIG entries defines columns three and four of the stiffness matrix, respectively.
Note that due to symmetry, one less row needs to be defined for each additional column.
The mass matrix is input in a similar manner as the stiffness matrix with the following
exceptions:

Main Index
 CHAPTER 3 85
Real Eigenvalue Analysis

• The command M2GG = EXMASS instead of K2GG = EXSTIF is used in the Case Control
Section. In this case, EXMASS is the name of the mass matrix referenced in field two of
the DMIG Bulk Data entries.
• The matrix defined in the DMIG entries is expressed in the mass matrix terms rather
than in stiffness matrix terms.
• Since there are only two non-null columns for the mass matrix, only two DMIG data
entries are required instead of the four entries needed for the stiffness matrix.
Mass matrices input using DMIG are not scaled by PARAM,WTMASS.
In this example the small-field input format is used, and the maximum number of characters that
can be input are eight (including sign and decimal point). Greater input precision can be
achieved by using the large-field format and by changing the TIN field to 2 for the DMIG entries.
The first two computed natural frequencies for this example are 1.676 Hz and 10.066 Hz.

Main Index
86

Main Index
 MSC.Nastran Basic Dynamics User’s GuideMSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Rigid-body Modes
4
■ Overview
■ SUPORT Entry
■ Examples

Main Index
88

4.1 Overview
A structure or a portion of a structure can displace without developing internal loads or stresses
if it is not sufficiently tied to ground (constrained). These stress-free displacements are
categorized as rigid-body modes or mechanism modes.

Rigid-body Modes. Rigid-body modes occur in unconstrained structures, such as satellites and
aircraft in flight. For a general, unconstrained 3-D structure without mechanisms, there are six
rigid-body modes often described as T1, T2, T3, R1, R2, and R3, or combinations thereof.
Rigid-body modes can also be approximated for certain kinds of dynamic or modal tests in
which the test specimen is supported by very flexible supports, such as bungee cords or
inflatable bags. In this case the test specimen itself does not distort for the lowest mode(s) but
instead displaces as a rigid body with all of the deformation occurring in the flexible support.
Rigid-body modes can improperly occur if a structure that should be constrained is not fully
constrained (for example, in a building model for which the boundary conditions (SPCs) were
forgotten).

Mechanism Modes. A mechanism mode occurs when a portion of the structure can displace as
a rigid body, which can occur when there is an internal hinge in the structure. An example of a
mechanism is a ball and socket joint or a rudder in an airplane. A mechanism mode can also
occur when two parts of a structure are improperly joined. A common modeling error resulting
in a mechanism is when a bar is cantilevered from a solid element; the bar has rotational stiffness
and the solid has no rotational stiffness, resulting in a pinned connection when the two are
joined.
The presence of rigid-body and/or mechanism modes is indicated by zero frequency
eigenvalues. Due to computer roundoff, the zero frequency eigenvalues are numerical zeroes
on the order of 1.0E-4 Hz or less for typical structures. The same unconstrained model may give
different values of the rigid-body frequencies when run on different computer types.
Rigid-body modes generated by MSC.Nastran are orthogonal with respect to each other and
with respect to the flexible (or elastic) modes. Each rigid-body mode has the following property:

T
{ φ rig } [ M ] { φ rig } > 0
Eq. 4-1
T
{ φ rig } [ K ] { φ rig } = 0

where { φrig } denotes the rigid-body mode shapes, [ M ] denotes the mass matrix, and [ K ] denotes
the stiffness matrix. The rigid-body modes contain no strain energy.
Figure 4-1 depicts the rigid-body modes for a two-dimensional model of a straight beam.
Because the two-dimensional model has three DOFs per grid point (x-translation, y-translation,
and z-rotation), there are three rigid-body modes. The figure shows classical rigid-body modes,
whereby one mode is purely x-translation, another is purely y-translation, and another is purely
z-rotation about the center of the beam. Because rigid-body modes are a special case of repeated
roots, any linear combination of these displacement shapes also comprises a valid set of
rigid-body modes.

Main Index
 CHAPTER 4 89
Rigid-body Modes

Figure 4-1 Rigid-Body Modes of a 2-D Beam

Main Index
90

4.2 SUPORT Entry

Rigid-body modes are computed in MSC.Nastran without requiring special user intervention,
although the use of a SUPORT Bulk Data entry makes the mode shapes look cleaner. The
SUPORT entry also makes the rigid-body mode shapes repeatable when the mass or stiffness of
the model changes and the model is reanalyzed. The SUPORT (note spelling!) entry does not
constrain the model; it simply provides a frame of reference for the rigid-body shape calculations
by defining the r-set components of motion. The SUPORT entry is not required for any of the
dynamic analysis methods except for response spectrum analysis (see “Advanced Dynamic
Analysis Capabilities” on page 311).
If the SUPORT is used, one DOF should be specified for each rigid-body mode or mechanism.
The form of the SUPORT entry is shown below:

1 2 3 4 5 6 7 8 9 10
SUPORT ID1 C1 ID2 C2 ID3 C3 ID4 C4

Field Contents
IDi Grid or scalar point identification number.
Ci Component numbers. (0 or blank for scalar points and any unique
combination of the integers 1 through 6 for grid points).

Treatment of SUPORT by Eigenvalue Analysis Methods

The eigenvalue extraction methods treat the SUPORT entry differently as described below.

EIGR Methods
Each of the eigenvalue extraction methods selected on the EIGR Bulk Data entry (AGIV, AHOU,
GIV, HOU, INV, MGIV, MHOU, and SINV) treats the SUPORT in the same manner.
Eigenvalues are first computed using the information on the EIGR Bulk Data entry. The
frequencies of the first N r modes (where N r is the number of SUPORT DOFs) are replaced with
a value of 0.0 Hz. The first N r eigenvectors are replaced by modes that are calculated by moving
each SUPORT DOF a unit distance and then mass orthogonalizing them with respect to the other
modes. The fact that the eigenvectors are calculated via kinematics is the reason that the
SUPORT entry produces cleaner rigid-body modes (as opposed to the rigid-body modes
computed without the use of the SUPORT entry).
Note that MSC.Nastran has no built-in checks to warn if any of the N r eigenvalues are not
rigid-body modes. This replacement can hide potential modeling problems. The results from
UIM 3035 should be carefully checked (see “Examples” on page 96). The strain energy term for
each SUPORT DOF should be zero.

Main Index
 CHAPTER 4 91
Rigid-body Modes

A poor choice of DOFs on the SUPORT entry can adversely affect the computation of the flexible
modes for the INV and SINV methods. Flexible modes computed with the other methods
(AGIV, AHOU, GIV, HOU, MGIV, and MHOU) are not adversely affected by a poor choice of
SUPORT DOFs. Again, the results of UIM 3035 should be carefully checked to ensure that a
proper choice of SUPORT DOFs is made.

Lanczos Method
In the Lanczos method, the SUPORT entry attempts to give 0.0 Hz modes. The Lanczos routine
compares the values of the calculated modes (without the SUPORT entry) and determines if the
calculated frequencies for the N r modes are near 0.0 Hz. If the computed frequencies are near
0.0 Hz, the computed values are replaced with 0.0 Hz. If a computed frequency is not near
0.0 Hz, then it is retained. (Note that this may cause problems in response spectrum analysis
where 0.0 Hz is required for the SUPORT frequencies.) The Lanczos-computed, rigid-body
eigenvectors are used, i.e., the rigid-body eigenvectors computed by the SUPORT entry are not
used.
In the above paragraph, “near” means that the eigenvalues are less than 0.01 times the shift scale
(the SHFSCL field on the EIGRL entry). All computed eigenvalues less than this threshold are
candidate rigid-body modes. There is some logic to determine the relationship of these
candidate rigid-body modes to the number of DOFs on the SUPORT entry. Suppose that there
are three eigenvalues less than the threshold of 0.01 times the shift scale. If your SUPORT entry
defines two DOFs, then the first two frequencies become 0.0 Hz, and the third becomes
something that is nonzero but small (on the order of 1.0E-6, for example). On the other hand, if
your SUPORT entry defines four DOFs, then only the first three are treated as rigid-body modes,
and the fourth is not replaced. Furthermore, the use of a SUPORT entry forces an extra
decomposition, which increases the computer run time.

Because the SUPORT entry is not used to compute the rigid-body eigenvectors, there is probably
little to be gained by using the SUPORT entry with the Lanczos method unless response
spectrum analysis is being performed.

Theoretical Considerations
Degrees-of-freedom defined on the SUPORT entry are placed in the r-set. When an r-set is
present, static rigid-body vectors are calculated in MSC.Nastran by first partitioning the a-set
into the r- and l-sets

 ul 
{ ua } =   Eq. 4-2
 ur 

Introducing this partitioning in the stiffness matrix results in

K ll K lr  u l  0 
  = P  Eq. 4-3
K rl K rr  u r   r

for the rigid-body modes defined in the r-set.

Main Index
92

There is no load on l-set DOFs. The load { P r } on the r-set is not needed in subsequent equations.
Then solve for u l in terms of u r

{ ul } = [ D ] { ur } Eq. 4-4

–1
where: [ D ] = – [ K ll ] [ K lr ]

The matrix [ D ] is used to construct a set of rigid-body vectors [ ψrig ]

D
[ ψ rig ] = Eq. 4-5
Ir

where [ ψrig ] represents the motion of the a-set for a unit motion of each SUPORT DOF with all
other SUPORT DOFs constrained and [ I r ] is an r × r identity matrix.
The rigid-body vectors can be used to create a rigid-body mass matrix [ M r ]

T D T D
[ M r ] = [ ψ rig ] [ M aa ] [ ψ rig ] = [ M aa ] Eq. 4-6
Ir Ir

To improve the quality of the rigid-body mode shapes, orthogonalization is applied to create a
diagonal mass matrix [ M o ] by

T
[ M o ] = [ φ ro ] [ M r ] = [ φ ro ] Eq. 4-7

where: [ φ ro ] is a transformation matrix.

This transformation matrix is used to construct the final set of rigid-body mode shape vectors by

Dφ ro
[ φ rig ] = [ ψ rig ] [ φ ro ] = Eq. 4-8
φ ro

such that

T
[ φ rig ] [ K aa ] [ φ rig ] ≈ 0 Eq. 4-9

T
[ φ rig ] [ M aa ] [ φ rig ] = Mo Eq. 4-10

where: M o is a diagonal matrix.

Care must be taken when selecting SUPORT DOFs. Each SUPORT DOF must be able to displace
independently without developing internal stresses. In other words, the SUPORT DOFs must
be statically determinate. The SUPORT is used only to facilitate the calculation of rigid-body
vectors. If you do not specify the r-set DOFs, the rigid-body modes are calculated directly by the
method selected for the flexible frequency modes. If an insufficient number of r-set DOFs are
specified, the calculation of rigid-body modes is unreliable.
Main Index
 CHAPTER 4 93
Rigid-body Modes

As a modeling aid, MSC.Nastran calculates equivalent internal strain energy (work) for each
rigid-body vector as follows:

K ll K lr D
[X] = D
T
Ir Eq. 4-11
K rl K rr Ir

which can be simplified as

T
[ X ] = [ D ] [ K ll ] [ D ] + [ K rr ] Eq. 4-12

When r-set DOFs exist, the printed strain energies are the diagonal elements of [ X ] divided by 2
and should be approximately zero.
Note that [ X ] is the transformation of the stiffness matrix [ K aa ] to r-set coordinates, which by
definition of rigid-body (i.e., zero frequency) vector properties should be null. If this is not the
case, the equilibrium may be violated by the r-set choice or other modeling errors may exist. The
matrix [ X ] is also called the rigid-body check matrix.
MSC.Nastran also calculates a rigid-body error ratio

T
[ K rr ] + [ K lr ] [ D ]
ε = ------------------------------------------------------- Eq. 4-13
K rr

2
where: = Euclidean norm of the matrix = ∑ ∑ x ij
i j

One value of ε is calculated using Eq. 4-13 based on all SUPORT DOFs. Therefore, in UIM 3035
the same ε is printed for every supported DOF.
The rigid-body error ratio and the strain energy should be zero if a set of statically determinate
SUPORT DOFs is chosen. Roundoff error may lead to computational zero values for these
quantities. ("Computational zero" is a small number ( 10 –5 , for example) that normally is 0.0
except for numerical roundoff.) The rigid-body error ratio and strain energy may be
significantly nonzero for any of the following reasons:

• Roundoff error accumulation.

• The ur set is overdetermined, leading to redundant supports. The condition gives high
strain energy.
• The ur set is underspecified, leading to a singular reduced-stiffness matrix and a
MAXRATIO error. This condition gives a high rigid-body error ratio.
• The multipoint constraints are statically indeterminate. This condition gives high
strain energy and a high rigid-body error ratio.
• There are too many single-point constraints. This condition gives high strain energy
and a high rigid-body error ratio.

Main Index
94

• [ K rr ] is null. This condition gives a unit value for the rigid-body error but low strain
energy. This is an acceptable condition and may occur when generalized dynamic
reduction is used (see “Advanced Dynamic Analysis Capabilities” on page 311).

Modeling Considerations
When using a SUPORT you must select a set of DOFs that is capable of constraining all the rigid-
body modes. Another way to state this requirement is that the r-set must be able to constrain the
structure in a statically determinate manner. There are usually many choices of DOFs that
satisfy this requirement. Two choices that work for simple three-dimensional structures are:

• Six DOFs on one grid point when all its degrees-of-freedom have stiffness.
• Three translation DOFs normal to one plane, two translation DOFs normal to an
orthogonal plane, and then one translational DOF normal to the last orthogonal plane.
Such a system can be used, for instance, on a model composed entirely of solid
elements that have no inherent stiffness for grid point rotation. See Figure 4-2.

Plane 1

2
4

Plane 3
Plane 2 5 6
1, 2, 3 Normal to Plane 1
4, 5 Normal to Plane 2
6 Normal to Plane 3

Figure 4-2 Statically Determinate r-set

There are special cases where a model need not have six rigid-body modes. A planar model has
only three rigid-body modes, while an airplane with a free rudder has seven, for example. If you
use the SUPORT, it is your responsibility to determine all the modes of rigid-body motion, then
provide r-set DOFs that define these rigid-body modes. Another special case is the application
of enforced motion by the large mass technique (see “Enforced Motion” on page 195). If the
input points describe redundant load paths, diagnostics are produced that indicate
overconstraint. For this case, these diagnostics may be safely ignored.
Poorly-constrained rigid-body modes result from either constraining DOFs with relatively little
stiffness or from constraining a set of DOFs that are almost linearly dependent on one another.
An example of the former is a model of a very thin cylindrical shell. The degrees-of-freedom
normal to the shell and their associated bending degrees-of-freedom may all be too soft to avoid
numerical conditioning problems. A modeling cure for this condition is to connect many grid

Main Index
 CHAPTER 4 95
Rigid-body Modes

points to a new reference grid point with an RBE3 element and then to place the reference grid
point in the r-set. The RBE3 element does not affect the flexible modes when applied in this
manner.
An example of a structure whose r-set shows poor linear independence is a slightly curved bar
modeled using coordinate systems that follow the curve of the bar such that the x-axis is always
tangent to the bar. The x DOFs at each end of the bar describe linearly independent DOFs in a
mathematical sense. However, numerical truncation produces poor conditioning if the angle
between the ends is less than a few degrees. This condition is detected by the automatic
diagnostics discussed earlier. This problem can be corrected (or better yet, avoided) by making
a careful sketch of all r-set DOFs, including their locations in space and the orientation of their
global coordinates. Then apply the three-plane test described earlier.
Using a physical analogy, a good r-set can be chosen by finding one grid point that sustains all
possible loadings well if it is tied to ground in an actual hardware test. If there is no such grid
point, the ties to ground should be spread over enough grid points to sustain the loads without
damaging the structure. An RBE3 element used for this purpose can provide good rigid-body
modes without affecting the flexible modes.

Main Index
96

4.3 Examples
This section provides several rigid-body modes examples showing input and output. These
examples are as follows:

Table 4-1 Unconstrained Beam Model Summary

Analysis SUPORT/ Redundancy of

Model
Method No SUPORT SUPORT
bd04bar1 Lanczos No SUPORT –
bd04bar2 Lanczos SUPORT Statically determinate
bd04bar3 Lanczos SUPORT Underdetermined
bd04bar4 Lanczos SUPORT Overdetermined
bd04bar5 SINV No SUPORT –
bd04bar6 SINV SUPORT Statically determinate
bd04bar7 SINV SUPORT Underdetermined
bd04bar8 SINV SUPORT Overdetermined
bd04bkt Lanczos No SUPORT –

These examples are described in the sections that follow.

Unconstrained Beam Model

The constraints (SPCs) on the example cantilever beam model from “Cantilever Beam Model”
on page 139 are removed to create an unconstrained structure as shown in Figure 4-3. A
GRDSET entry is added with the z-translation, x-rotation, and y-rotation directions constrained
to make the problem two-dimensional. Therefore, there are three DOFs per grid point (x-
translation, y-translation, and z-rotation) and three rigid-body modes.

Figure 4-3 Unconstrained Beam Model

Modes are computed using two methods (Lanczos and SINV), with and without a SUPORT
entry. The SUPORT entry is used in three ways:

• Statically determinate (grid point 1, components 1, 2, and 6)

• Underdetermined (grid point 1, components 1 and 2)
• Overdetermined (grid point 1, components 1, 2, and 6, plus grid point 11, component1)
Listing 4-1 shows a portion of the input file for the statically determinate SUPORT and the
Lanczos method.
Main Index
 CHAPTER 4 97
Rigid-body Modes

Listing 4-1 Input File for Cantilever Beam Model

$ FILE bd04bar2.dat
$
$ CANTILEVER BEAM MODEL
$ CHAPTER 4, RIGID-BODY MODES
$
SOL 103
TIME 10
CEND
TITLE = CANTILEVER BEAM
SUBTITLE = NORMAL MODES
LABEL = USE SUPORT, STATICALLY DETERMINATE
$
$ OUTPUT REQUEST
DISPLACEMENT = ALL
$
$ SELECT EIGRL ENTRY
METHOD = 10
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$ STATICALLY DETERMINATE SUPORT
$SUPORT G C
SUPORT 1 126
$
$ MAKE 2D MODEL
GRDSET 345
$
$EIGRL SID V1 V2
EIGRL 10 -0.1 50.
$
... basic model ...
$
ENDDATA

Table 4-2 lists the computed frequencies. The overdetermined run for the Lanczos method
works well; the same run for the SINV method gives an extra zero frequency mode that does not
really exist. The overdetermined runs have redundant SUPORTs in the x-direction; therefore,
two rigid-body modes are computed in this direction when using the SINV method. In all cases
the flexible (greater than 0) frequencies are correct.

Table 4-2 Frequencies for the Unconstrained Beam Models

Lanczos Method Frequencies (Hz) SINV Method Frequencies (Hz)

Mode Stat. Over- Stat. Over

No Under No Under
Det. Det. Det. Det.
SUPORT Det. SUPORT Det.
SUPORT SUPORT SUPORT SUPORT
1 2.15E-7 0.0 0.0 0.0 6.88E-7 0.0 0.0 0.0
2 4.02E-7 0.0 0.0 0.0 6.79E-6 0.0 0.0 0.0
3 8.07E-6 3.07E-5 0.0 0.0 7.12E-6 1.45E-6 0.0 0.0

Main Index
98

Table 4-2 Frequencies for the Unconstrained Beam Models

Lanczos Method Frequencies (Hz) SINV Method Frequencies (Hz)

Mode Stat. Over- Stat. Over

No Under No Under
Det. Det. Det. Det.
SUPORT Det. SUPORT Det.
SUPORT SUPORT SUPORT SUPORT
4 12.82 12.82 12.82 12.82 12.82 12.82 12.82 0.0
5 34.62 34.62 34.62 34.62 34.62 34.62 34.62 12.82
6 66.60 66.60 66.60 34.62
7 66.60

Figure 4-4 shows the output for the rigid-body modes computed for the SINV method when
using no SUPORT. The rigid-body frequencies are denoted by computational zeroes on the
order of 10 –5 Hz or less. Note that the magnitude may be different when the same problem is
run on a different computer type.
Figure shows the output for the rigid-body modes computed for the SINV method when using
the statically determinate SUPORT. The accuracy of the statically determinate SUPORT DOFs
is verified by the computational zeroes for epsilon and the strain energy printed in UIM 3035.
Note that the three rigid-body modes have frequencies of 0.0 Hz. The SUPORT entry provides
cleaner mode shapes than those shown in Figure 4-4 as illustrated by the purely x-translation, y-
translation, and z-rotation eigenvectors shown in Figure 4-5.

Main Index
 CHAPTER 4 99
Rigid-body Modes

Listing 4-2

R E A L E I G E N V A L U E S
MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED
NO. ORDER MASS STIFFNESS
1 1 1.866121E-11 4.319862E-06 6.875275E-07 1.000000E+00 1.866121E-11
2 2 1.819082E-09 4.265070E-05 6.788069E-06 1.000000E+00 1.819082E-09
3 3 2.000299E-09 4.472470E-05 7.118156E-06 1.000000E+00 2.000299E-09
4 4 6.483918E+03 8.052278E+01 1.281560E+01 1.000000E+00 6.483918E+03
5 5 4.732272E+04 2.175379E+02 3.462222E+01 1.000000E+00 4.732272E+04
6 6 1.751285E+05 4.184836E+02 6.660372E+01 1.000000E+00 1.751285E+05

EIGENVALUE = 1.866121E-11
CYCLES = 6.875275E-07 R E A L E I G E N V E C T O R N O . 1

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 4.273806E-34 -2.470432E-01 0.0 0.0 0.0 -1.067134E-01
2 G 4.273806E-34 -2.790573E-01 0.0 0.0 0.0 -1.067134E-01
3 G 4.273806E-34 -3.110713E-01 0.0 0.0 0.0 -1.067134E-01
4 G 4.273806E-34 -3.430853E-01 0.0 0.0 0.0 -1.067134E-01
5 G 4.273806E-34 -3.750993E-01 0.0 0.0 0.0 -1.067134E-01
6 G 4.273806E-34 -4.071134E-01 0.0 0.0 0.0 -1.067134E-01
7 G 4.273806E-34 -4.391274E-01 0.0 0.0 0.0 -1.067134E-01
8 G 4.273806E-34 -4.711414E-01 0.0 0.0 0.0 -1.067134E-01
9 G 4.273806E-34 -5.031555E-01 0.0 0.0 0.0 -1.067134E-01
10 G 4.273806E-34 -5.351695E-01 0.0 0.0 0.0 -1.067134E-01
11 G 4.273806E-34 -5.671835E-01 0.0 0.0 0.0 -1.067134E-01

EIGENVALUE = 1.819082E-09
CYCLES = 6.788069E-06 R E A L E I G E N V E C T O R N O . 2

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G -2.953214E-01 -5.597336E-01 0.0 0.0 0.0 3.291537E-01
2 G -2.953214E-01 -4.609875E-01 0.0 0.0 0.0 3.291537E-01
3 G -2.953214E-01 -3.622414E-01 0.0 0.0 0.0 3.291537E-01
4 G -2.953214E-01 -2.634953E-01 0.0 0.0 0.0 3.291537E-01
5 G -2.953214E-01 -1.647492E-01 0.0 0.0 0.0 3.291537E-01
6 G -2.953214E-01 -6.600305E-02 0.0 0.0 0.0 3.291537E-01
7 G -2.953214E-01 3.274305E-02 0.0 0.0 0.0 3.291537E-01
8 G -2.953214E-01 1.314892E-01 0.0 0.0 0.0 3.291538E-01
9 G -2.953214E-01 2.302354E-01 0.0 0.0 0.0 3.291538E-01
10 G -2.953214E-01 3.289815E-01 0.0 0.0 0.0 3.291538E-01
11 G -2.953214E-01 4.277276E-01 0.0 0.0 0.0 3.291538E-01

EIGENVALUE = 2.000299E-09
CYCLES = 7.118156E-06 R E A L E I G E N V E C T O R N O . 3

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G -2.953611E-01 5.596586E-01 0.0 0.0 0.0 -3.291096E-01
2 G -2.953611E-01 4.609257E-01 0.0 0.0 0.0 -3.291096E-01
3 G -2.953611E-01 3.621928E-01 0.0 0.0 0.0 -3.291096E-01
4 G -2.953611E-01 2.634599E-01 0.0 0.0 0.0 -3.291096E-01
5 G -2.953611E-01 1.647270E-01 0.0 0.0 0.0 -3.291096E-01
6 G -2.953611E-01 6.599414E-02 0.0 0.0 0.0 -3.291095E-01
7 G -2.953611E-01 -3.273870E-02 0.0 0.0 0.0 -3.291095E-01
8 G -2.953611E-01 -1.314715E-01 0.0 0.0 0.0 -3.291095E-01
9 G -2.953611E-01 -2.302044E-01 0.0 0.0 0.0 -3.291095E-01
10 G -2.953611E-01 -3.289373E-01 0.0 0.0 0.0 -3.291095E-01
11 G -2.953611E-01 -4.276701E-01 0.0 0.0 0.0 -3.291095E-01

Figure 4-4 Unconstrained Beam Modes Without SUPORT (SINV Method)

Main Index
100

Listing 4-3

*** USER INFORMATION MESSAGE 3035 FOR DATA BLOCK KLR

SUPPORT PT.NO. EPSILON STRAIN ENERGY EPSILONS LARGER THAN 0.001 ARE FLAGGED WITH ASTERISKS
1 7.7496606E-17 5.5879354E-09
2 7.7496606E-17 0.0000000E+00
3 7.7496606E-17 7.1622708E-11

R E A L E I G E N V A L U E S
MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED
NO. ORDER MASS STIFFNESS
1 1 0.0 0.0 0.0 1.000000E+00 0.0
2 2 0.0 0.0 0.0 1.000000E+00 0.0
3 3 0.0 0.0 0.0 1.000000E+00 0.0
4 4 6.483918E+03 8.052278E+01 1.281560E+01 1.000000E+00 6.483918E+03
5 5 4.732272E+04 2.175379E+02 3.462222E+01 1.000000E+00 4.732272E+04
6 6 1.751285E+05 4.184836E+02 6.660372E+01 1.000000E+00 1.751285E+05
7 7 4.616299E+05 6.794335E+02 1.081352E+02 1.000000E+00 4.616299E+05

EIGENVALUE = 0.000000E+00
CYCLES = 0.000000E+00 R E A L E I G E N V E C T O R N O . 1

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
2 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
3 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
4 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
5 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
6 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
7 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
8 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
9 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
10 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0
11 G 4.176756E-01 0.0 0.0 0.0 0.0 0.0

EIGENVALUE = 0.000000E+00
CYCLES = 0.000000E+00 R E A L E I G E N V E C T O R N O . 2

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 4.176756E-01 0.0 0.0 0.0 0.0
2 G 0.0 4.176756E-01 0.0 0.0 0.0 -3.593777E-15
3 G 0.0 4.176756E-01 0.0 0.0 0.0 -6.723840E-15
4 G 0.0 4.176756E-01 0.0 0.0 0.0 -9.575675E-15
5 G 0.0 4.176756E-01 0.0 0.0 0.0 -1.220145E-14
6 G 0.0 4.176756E-01 0.0 0.0 0.0 -1.439249E-14
7 G 0.0 4.176756E-01 0.0 0.0 0.0 -1.596912E-14
8 G 0.0 4.176756E-01 0.0 0.0 0.0 -1.698929E-14
9 G 0.0 4.176756E-01 0.0 0.0 0.0 -1.745155E-14
10 G 0.0 4.176756E-01 0.0 0.0 0.0 -1.770949E-14
11 G 0.0 4.176756E-01 0.0 0.0 0.0 -1.758923E-14

EIGENVALUE = 0.000000E+00
CYCLES = 0.000000E+00 R E A L E I G E N V E C T O R N O . 3

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 -7.163078E-01 0.0 0.0 0.0 4.775385E-01
2 G 0.0 -5.730462E-01 0.0 0.0 0.0 4.775385E-01
3 G 0.0 -4.297847E-01 0.0 0.0 0.0 4.775385E-01
4 G 0.0 -2.865231E-01 0.0 0.0 0.0 4.775385E-01
5 G 0.0 -1.432615E-01 0.0 0.0 0.0 4.775385E-01
6 G 0.0 1.508516E-14 0.0 0.0 0.0 4.775385E-01
7 G 0.0 1.432615E-01 0.0 0.0 0.0 4.775385E-01
8 G 0.0 2.865231E-01 0.0 0.0 0.0 4.775385E-01
9 G 0.0 4.297847E-01 0.0 0.0 0.0 4.775385E-01
10 G 0.0 5.730463E-01 0.0 0.0 0.0 4.775385E-01
11 G 0.0 7.163078E-01 0.0 0.0 0.0 4.775385E-01

Figure 4-5 Unconstrained Beam Modes With Statically Determinate SUPORT

(SINV Method)

Figure 4-6 shows the epsilon and strain energy printed in UIM 3035 for the three SUPORT cases
(statically determinate, overdetermined, and underdetermined).

Main Index
 CHAPTER 4 101
Rigid-body Modes

It can be seen from this example that aside from clean rigid-body vectors there is no advantage
to using a SUPORT entry to compute rigid-body modes.

Statically determinate SUPORT:

*** USER INFORMATION MESSAGE 3035 FOR DATA BLOCK KLR

SUPPORT PT.NO. EPSILON STRAIN ENERGY EPSILONS LARGER THAN 0.001 ARE FLAGGED WITH ASTERISKS
1 7.7496606E-17 5.5879354E-09
2 7.7496606E-17 0.0000000E+00
3 7.7496606E-17 7.1622708E-11

Overdetermined SUPORT:

*** USER INFORMATION MESSAGE 3035 FOR DATA BLOCK KLR

SUPPORT PT.NO. EPSILON STRAIN ENERGY EPSILONS LARGER THAN 0.001 ARE FLAGGED WITH ASTERISKS
1 1.9913979E-01 7.2869660E+06 ****
2 1.9913979E-01 0.0000000E+00 ****
3 1.9913979E-01 7.1850081E-11 ****
4 1.9913979E-01 7.2869660E+06 ****

Underdetermined SUPORT:

*** USER INFORMATION MESSAGE 4158---STATISTICS FOR SYMMETRIC DECOMPOSITION OF DATA BLOCK KLL FOLLOW

MAXIMUM RATIO OF MATRIX DIAGONAL TO FACTOR DIAGONAL = 9.1E+13 AT ROW NUMBER 31

*** USER WARNING MESSAGE 4698. STATISTICS FOR DECOMPOSITION OF MATRIX KLL .
THE FOLLOWING DEGREES OF FREEDOM HAVE FACTOR DIAGONAL RATIOS GREATER THAN
1.00000E+05 OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.

GRID POINT ID DEGREE OF FREEDOM MATRIX/FACTOR DIAGONAL RATIO MATRIX DIAGONAL

11 R3 9.49483E+13 2.84000E+04

*** USER INFORMATION MESSAGE 3035 FOR DATA BLOCK KLR

SUPPORT PT.NO. EPSILON STRAIN ENERGY EPSILONS LARGER THAN 0.001 ARE FLAGGED WITH ASTERISKS
1 7.6288287E-17 5.5879354E-09
2 7.6288287E-17 -7.2759576E-12

Figure 4-6 UIM 3035 Results

Unconstrained Bracket Example

The constraints (SPCs) on the example bracket model from “Real Eigenvalue Analysis” on
page 33 are removed to create an unconstrained model (see “Examples” on page 56 for a
description of the model). Figure 4-7 shows the bracket model. The model is a
three-dimensional model and therefore produces six rigid-body modes. The MSC.Nastran
results are shown in Listing 4-4 for the first eight modes. The Lanczos method is used.

Figure 4-7 Unconstrained Bracket Model

Main Index
102

Listing 4-4 Unconstrained Bracket Frequencies

R E A L E I G E N V A L U E S
MODE EXTRACTION EIGENVALUE RADIANS CYCLES GENERALIZED GENERALIZED
NO. ORDER MASS STIFFNESS
1 1 -1.690642E-07 4.111741E-04 6.544039E-05 1.000000E+00 -1.690642E-07
2 2 -9.807991E-09 9.903530E-05 1.576196E-05 1.000000E+00 -9.807991E-09
3 3 -5.515176E-09 7.426423E-05 1.181952E-05 1.000000E+00 -5.515176E-09
4 4 -3.390596E-09 5.822883E-05 9.267405E-06 1.000000E+00 -3.390596E-09
5 5 1.266017E-08 1.125174E-04 1.790770E-05 1.000000E+00 1.266017E-08
6 6 1.726585E-08 1.313996E-04 2.091289E-05 1.000000E+00 1.726585E-08
7 7 2.649932E+06 1.627861E+03 2.590821E+02 1.000000E+00 2.649932E+06
8 8 4.279463E+06 2.068686E+03 3.292416E+02 1.000000E+00 4.279463E+06

The six rigid-body modes have computational zero frequencies on the order of 10 –5 Hz. Note
that the magnitudes of the rigid-body modes may be different when the same problem is run on
a different computer type. Also note that the output is sorted by the value of the eigenvalue in
ascending order.

Main Index
 MSC.Nastran Basic Dynamics User’s GuideMSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Frequency Response Analysis
5
■ Overview
■ Direct Frequency Response Analysis
■ Modal Frequency Response Analysis
■ Modal Versus Direct Frequency Response

■ Frequency-Dependent Excitation Definition

■ Solution Frequencies
■ Frequency Response Considerations
■ Solution Control for Frequency Response Analysis
■ Examples

Main Index
104

5.1 Overview
Frequency response analysis is a method used to compute structural response to steady-state
oscillatory excitation. Examples of oscillatory excitation include rotating machinery,
unbalanced tires, and helicopter blades. In frequency response analysis the excitation is
explicitly defined in the frequency domain. All of the applied forces are known at each forcing
frequency. Forces can be in the form of applied forces and/or enforced motions (displacements,
velocities, or accelerations).

Phase Shift. Oscillatory loading is sinusoidal in nature. In its simplest case, this loading is
defined as having an amplitude at a specific frequency. The steady-state oscillatory response
occurs at the same frequency as the loading. The response may be shifted in time due to
damping in the system. The shift in response is called a phase shift because the peak loading
and peak response no longer occur at the same time. An example of phase shift is shown in
Figure 5-1.

Phase
Shift
Loading Response

Time

Figure 5-1 Phase Shift

Complex Numbers. The important results obtained from a frequency response analysis
usually include the displacements, velocities, and accelerations of grid points as well as the
forces and stresses of elements. The computed responses are complex numbers defined as
magnitude and phase (with respect to the applied force) or as real and imaginary components,
which are vector components of the response in the real/imaginary plane. These quantities are
graphically presented in Figure 5-2.

Main Index
 CHAPTER 5 105
Frequency Response Analysis

i (Imaginary)

u
ui

θ
ur r (Real)

Figure 5-2 Complex Plane

where:

2 2
u = magnitude = u r + u i
–1
θ = phase angle = tan ( u i ⁄ u r )
u r = real component = u cos θ

u i = imaginary component = u sin θ

Two different numerical methods can be used in frequency response analysis. The direct
method solves the coupled equations of motion in terms of forcing frequency. The modal
method utilizes the mode shapes of the structure to reduce and uncouple the equations of
motion (when modal or no damping is used); the solution for a particular forcing frequency is
obtained through the summation of the individual modal responses. The choice of the method
depends on the problem. The two methods are described in “Direct Frequency Response
Analysis” on page 106 and “Modal Frequency Response Analysis” on page 108.

Main Index
106

5.2 Direct Frequency Response Analysis

In direct frequency response analysis, structural response is computed at discrete excitation
frequencies by solving a set of coupled matrix equations using complex algebra. Begin with the
damped forced vibration equation of motion with harmonic excitation

·· · iωt
[ M ] { x ( t ) } + [ B ] { x ( t ) } + [ K ] { x ( t ) } = { P ( ω ) }e Eq. 5-1

The load in Eq. 5-1 is introduced as a complex vector, which is convenient for the mathematical
solution of the problem. From a physical point of view, the load can be real or imaginary, or
both. The same interpretation is used for response quantities.
For harmonic motion (which is the basis of a frequency response analysis), assume a harmonic
solution of the form:

iωt
{ x } = { u ( ω ) }e Eq. 5-2

where { u ( ω ) } is a complex displacement vector. Taking the first and second derivatives of Eq. 5-
2, the following is obtained:

· iωt
{ x } = iω { u ( ω ) }e

Eq. 5-3

·· 2 iωt
{ x } = – ω { u ( ω ) }e

When the above expressions are substituted into Eq. 5-1, the following is obtained:

2 iωt iωt iωt iωt

– ω [ M ] { u ( ω ) }e + iω [ B ] { u ( ω ) }e + [ K ] { u ( ω ) }e = { P ( ω ) }e Eq. 5-4

which after dividing by e iωt simplifies to

2
[ – ω M + iωB + K ] { u ( ω ) } = { P ( ω ) } Eq. 5-5

The equation of motion is solved by inserting the forcing frequency ω into the equation of
motion. This expression represents a system of equations with complex coefficients if damping
is included or the applied loads have phase angles. The equations of motion at each input
frequency are then solved in a manner similar to a statics problem using complex arithmetic.

Damping in Direct Frequency Response

Damping simulates the energy dissipation characteristics of a structure. Damping in direct
frequency response is represented by the damping matrix [ B ] and additions to the stiffness
matrix [ K ] .
The damping matrix is comprised of several matrices

1 2
[B] = [B ] + [B ] Eq. 5-6

Main Index
 CHAPTER 5 107
Frequency Response Analysis

where:

1
[ B ] = damping elements (CVISC, CDAMPi) and B2GG
2
[ B ] = B2PP direct input matrix and transfer functions

In frequency response, PARAM,G and GE on the MATi entry do not form a damping matrix.
Instead, they form the following complex stiffness matrix:

[ K ] = ( 1 + iG ) [ K ] + i ∑ G E [ K E ] Eq. 5-7

where:

[ K ] = global stiffness matrix

G = overall structural damping coefficient (PARAM,G)
[ K E ] = element stiffness matrix

G E = element structural damping coefficient (GE on the MATi entry)

When the above parameters and/or coefficients are specified, they are automatically
incorporated into the stiffness matrix and therefore into the equation of motion for the solution.
All of the forms of damping can be used in the same analysis, and their effects are added
together.
In frequency response analysis, it is not necessary to assume an equivalent viscous form for
structural damping since the solution is complex. Therefore, a complex stiffness matrix is
allowed.

Main Index
108

5.3 Modal Frequency Response Analysis

Modal frequency response analysis is an alternate approach to computing the frequency
response of a structure. This method uses the mode shapes of the structure to reduce the size,
uncouple the equations of motion (when modal or no damping is used), and make the numerical
solution more efficient. Since the mode shapes are typically computed as part of the
characterization of the structure, modal frequency response is a natural extension of a normal
modes analysis.
As a first step in the formulation, transform the variables from physical coordinates { u ( ω ) } to
modal coordinates { ξ ( ω ) } by assuming

iωt
{ x } = [ φ ] { ξ ( ω ) }e Eq. 5-8

The mode shapes [ φ ] are used to transform the problem in terms of the behavior of the modes
as opposed to the behavior of the grid points. Eq. 5-8 represents an equality if all modes are
used; however, because all modes are rarely used, the equation usually represents an
approximation.
To proceed, temporarily ignore all damping, which results in the undamped equation for
harmonic motion

2
–ω [ M ] { x } + [ K ] { x } = { P ( ω ) } Eq. 5-9

at forcing frequency ω .
Substituting the modal coordinates in Eq. 5-8 for the physical coordinates in Eq. 5-9 and
dividing by e iωt , the following is obtained:

2
–ω [ M ] [ φ ] { ξ ( ω ) } + [ K ] [ φ ] { ξ ( ω ) } = { P ( ω ) } Eq. 5-10

Now this is the equation of motion in terms of the modal coordinates. At this point, however,
the equations remain coupled.
To uncouple the equations, premultiply by [ φ T ] to obtain

2 T T T
– ω [ φ] [M ][φ ]{ ξ(ω )} + [φ ] [K][φ ]{ ξ(ω )} = [φ] { P(ω )} Eq. 5-11
where:

T
[ φ ] [ M ] [ φ ] = modal (generalized) mass matrix
T
[ φ ] [ K ] [ φ ] = modal (generalized) stiffness matrix
T
[ φ ] { P } = modal force vector

Main Index
 CHAPTER 5 109
Frequency Response Analysis

The final step uses the orthogonality property of the mode shapes to formulate the equation of
motion in terms of the generalized mass and stiffness matrices, which are diagonal matrices.
These diagonal matrices do not have the off-diagonal terms that couple the equations of motion.
Therefore, in this form the modal equations of motion are uncoupled. In this uncoupled form,
the equations of motion are written as a set of uncoupled single degree-of-freedom systems as

2
–ω mi ξi ( ω ) + ki ξ ( ω ) = pi ( ω ) Eq. 5-12

where:

m i = i-th modal mass

k i = i-th modal stiffness

p i = i-th modal force

The modal form of the frequency response equation of motion is much faster to solve than the
direct method because it is a series of uncoupled single degree-of-freedom systems.
Once the individual modal responses ξ i ( ω ) are computed, physical responses are recovered as
the summation of the modal responses using

iωt
{ x } = [φ ]{ ξ(ω )} e Eq. 5-13

These responses are in complex form (magnitude/phase or real/imaginary) and are used to
recover additional output quantities requested in the Case Control Section.

Damping in Modal Frequency Response

If a damping matrix [ B ] exists, the orthogonality property (see “Overview of Normal Modes
Analysis” on page 37) of the modes does not, in general, diagonalize the generalized damping
matrix

T
[ φ ] [ B ] [ φ ] ≠ diagonal Eq. 5-14

If structural damping is used, the orthogonality property does not, in general, diagonalize the
generalized stiffness matrix

T
[ φ ] [ K ] [ φ ] ≠ diagonal Eq. 5-15

where K = ( 1 + iG ) [ K ] + i ∑ G E [ K E ]
In the presence of a [ B ] matrix or a complex stiffness matrix, the modal frequency approach
solves the coupled problem in terms of modal coordinates using the direct frequency approach
described in “Direct Frequency Response Analysis” on page 106

2 T T T T
[ – ω [ φ ] [ M ] [ φ ] + iω [ φ ] [ B ] [ φ ] + [ φ ] [ K ] [ φ ] ] { ξ ( ω ) } = [ φ ] { P ( ω ) } Eq. 5-16

Main Index
110

Eq. 5-16 is similar to Eq. 5-5 for the direct frequency response analysis method except that Eq. 5-
16 is expressed in terms of modal coordinates ξ . Since the number of modes used in a solution
is typically much less than the number of physical variables, using the coupled solution of the
modal equations is less costly than using physical variables.
If damping is applied to each mode separately, the uncoupled equations of motion can be
maintained. When modal damping is used, each mode has damping b i where b i = 2m i ωi ζi . The
equations of motion remain uncoupled and have the form

2
– ω m i ξ i ( ω ) + iωb i ξ i ( ω ) + k i ξ i ( ω ) = p i ( ω ) Eq. 5-17

for each mode.

Each of the modal responses is computed using

pi( ω )
ξ i ( ω ) = ------------------------------------------------ Eq. 5-18
2
– m i ω + ib i ω + k i

The TABDMP1 Bulk Data entry defines the modal damping ratios. A table is created by the
frequency/damping pairs specified on the TABDMP1 entry. The solution refers to this table for
the damping value to be applied at a particular frequency. The TABDMP1 Bulk Data entry has
a Table ID. A particular TABDMP1 table is activated by selecting the Table ID with the
SDAMPING Case Control command.

1 2 3 4 5 6 7 8 9 10
TABDMP1 TID TYPE

f1 g1 f2 g2 f3 g3 -etc.- ENDT

Field Contents
TID Table identification number.
TYPE Type of damping units:
G (default)
CRIT
Q
fi Frequency value (cycles per unit time).
gi Damping value in the units specified.

Main Index
 CHAPTER 5 111
Frequency Response Analysis

At resonance, the three types of damping are related by the following equations:

bi Gi
ζ i = ------- = ------
b cr 2

b cr = 2m i ω i Eq. 5-19

1 1
Q i = ------------- = ------
( 2ζ i ) Gi

Note that the i subscript is for the i-th mode, and not the i-th excitation frequency.

The values of fi and gi define pairs of frequencies and dampings. Note that gi can be entered as
one of the following: structural damping (default), critical damping, or quality factor. The
entered damping is converted to structural damping internally using Eq. 5-19. Straight-line
interpolation is used for modal frequencies between consecutive fi values. Linear extrapolation
is used at the ends of the table. ENDT ends the table input.
For example, if modal damping is entered using Table 5-1 and modes exist at 1.0, 2.5, 3.6, and
5.5 Hz, MSC.Nastran interpolates and extrapolates as shown in Figure 5-3 and in the table. Note
that there is no table entry at 1.0 Hz; MSC.Nastran uses the first two table entries at f = 2.0 and
f = 3.0 to extrapolate the value for f = 1.0 .

0.20
+
ζ +
+ +
0.10
+ = entered value
= computed value

0.0
2.0 4.0 6.0
f (Hz)

Figure 5-3 Example TABDMP1

Table 5-1 Example TABDMP1 Interpolation/Extrapolation

Entered Computed
f z f z
2.0 0.16 1.0 0.14
3.0 0.18 2.5 0.17
4.0 0.13 3.6 0.15
6.0 0.13 5.5 0.13

Main Index
112

1 2 3 4 5 6 7 8 9 10
TABDMP1 10 CRIT +TAB1

+TAB1 2.0 0.16 3.0 0.18 4.0 0.13 6.0 0.13 +TAB2

+TAB2 ENDT

Modal damping is processed as a complex stiffness when PARAM,KDAMP is entered as -1. The
uncoupled equation of motion becomes

2
– ω m i ξ i ( ω ) + ( 1 + iG ( ω ) ) k i ξ i ( ω ) = p i ( ω ) Eq. 5-20

The default for PARAM,KDAMP is 1, which processes modal damping as a damping matrix as
shown in Eq. 5-17.
The decoupled solution procedure used in modal frequency response can be used only if either
no damping is present or modal damping alone (via TABDMP1) is used. Otherwise, the modal
method uses the coupled solution method on the smaller modal coordinate matrices if
nonmodal damping (i.e., CVISC, CDAMPi, GE on the MATi entry, or PARAM,G) is present.

Mode Truncation in Modal Frequency Response Analysis

It is possible that not all of the computed modes are required in the frequency response solution.
You need to retain, at a minimum, all the modes whose resonant frequencies lie within the range
of forcing frequencies. For example, if the frequency response analysis must be between 200 and
2000 Hz, all modes whose resonant frequencies are in this range should be retained. This
guideline is only a minimum requirement, however. For better accuracy, all modes up to at least
two to three times the highest forcing frequency should be retained. In the example where a
structure is excited to between 200 and 2000 Hz, all modes from 0 to at least 4000 Hz should be
retained.
The frequency range selected on the eigenvalue entry (EIGRL or EIGR) is one means to control
the modes used in the modal frequency response solution. Also, three parameters are available
to limit the number of modes included in the solution. PARAM,LFREQ gives the lower limit on
the frequency range of retained modes, and PARAM,HFREQ gives the upper limit on the
frequency range of retained modes. PARAM,LMODES gives the number of lowest modes to be
retained. These parameters can be used to include the proper set of modes. Note that the default
is for all computed modes to be retained.

Dynamic Data Recovery in Modal Frequency Response Analysis

In modal frequency response analysis, two options are available for recovering displacements
and stresses: the mode displacement method and the matrix method. Both methods give the
same answers, although with differences in cost.
The mode displacement method computes the total physical displacements for each excitation
frequency from the modal displacements, and then computes element stresses from the total
physical displacements. The number of operations is proportional to the number of excitation
frequencies.
Main Index
 CHAPTER 5 113
Frequency Response Analysis

The matrix method computes displacements per mode and element stresses per mode, and then
computes physical displacements and element stresses as the summation of modal
displacements and element stresses. Costly operations are proportional to the number of modes.
Since the number of modes is usually much less that the number of excitation frequencies, the
matrix method is usually more efficient and is the default. The mode displacement method can
be selected by using PARAM,DDRMM,-1 in the Bulk Data. The mode displacement method is
required when “frequency-frozen” structural plots are requested (see “Plotted Output” on
page 241).
The mode acceleration method (“Advanced Dynamic Analysis Capabilities” on page 311) is
another data recovery method for modal frequency response analysis. This method can provide
better accuracy since detailed local stresses and forces are subject to mode truncation and may
not be as accurate as the results computed with the direct method.

Main Index
114

5.4 Modal Versus Direct Frequency Response

Some general guidelines can be used when selecting modal frequency response analysis versus
direct frequency response analysis. These guidelines are summarized in Table 5-2.

Table 5-2 Modal Versus Direct Frequency Response

Modal Direct
Small Model X
Large Model X
Few Excitation Frequencies X
Many Excitation Frequencies X
High Frequency Excitation X
Nonmodal Damping X
Higher Accuracy X

In general, larger models may be solved more efficiently in modal frequency response because
the numerical solution is a solution of a smaller system of uncoupled equations. The modal
method is particularly advantageous if the natural frequencies and mode shapes were computed
during a previous stage of the analysis. In that case, you simply perform a restart (see “Restarts
In Dynamic Analysis” on page 217). Using the modal approach to solve the uncoupled
equations is very efficient, even for very large numbers of excitation frequencies. On the other
hand, the major portion of the effort in a modal frequency response analysis is the calculation of
the modes. For large systems with a large number of modes, this operation can be as costly as a
direct solution. This result is especially true for high-frequency excitation. To capture high
frequency response in a modal solution, less accurate, high-frequency modes must be computed.
For small models with a few excitation frequencies, the direct method may be the most efficient
because it solves the equations without first computing the modes. The direct method is more
accurate than the modal method because the direct method is not concerned with mode
truncation.
Table 5-2 provides an overview of which method to use. Many additional factors may be
involved in the choice of a method, such as contractual obligations or local standards of practice.

Main Index
 CHAPTER 5 115
Frequency Response Analysis

5.5 Frequency-Dependent Excitation Definition

An important aspect of a frequency response analysis is the definition of the loading function.
In a frequency response analysis, the force must be defined as a function of frequency. Forces
are defined in the same manner regardless of whether the direct or modal method is used.
The following Bulk Data entries are used for the frequency-dependent load definition:

RLOAD1 Tabular input-real and imaginary

RLOAD2 Tabular input-magnitude and phase
DAREA Spatial distribution of dynamic load
LSEQ Generates the spatial distribution of dynamic loads from static load entries
DLOAD Combines dynamic load sets
TABLEDi Tabular values versus frequency
DELAY Time delay
DPHASE Phase lead
The particular entry chosen for defining the dynamic loading is largely a function of user
convenience for concentrated loads. Pressure and distributed loads, however, require a more
complicated format.
There are two important aspects of dynamic load definition. First, the location of the loading on
the structure must be defined. Since this characteristic locates the loading in space, it is called
the spatial distribution of the dynamic loading. Secondly, the frequency variation in the loading
is the characteristic that differentiates a dynamic load from a static load. This frequency
variation is called the temporal distribution of the load. A complete dynamic loading is a
product of spatial and temporal distributions.
Using Table IDs and Set IDs in MSC.Nastran makes it possible to apply many complicated and
temporally similar loadings with a minimum of input. Combining simple loadings to create
complicated loading distributions that vary in position as well as frequency is also a
straightforward task.
The remainder of this section describes the Bulk Data entries for frequency-dependent
excitation. The description is given in terms of the coefficients that define the dynamic load. See
“Bulk Data Entries for Dynamic Analysis” on page 377 for more complete Bulk Data
descriptions.

Frequency-Dependent Loads – RLOAD1 Entry

The RLOAD1 entry is a general form in which to define a frequency-dependent load. It defines
a dynamic loading of the form

i { θ – 2πfτ }
{ P ( f ) } = { A [ C ( f ) + iD ( f ) ]e } Eq. 5-21

Main Index
116

The values of the coefficients are defined in tabular format on a TABLEDi entry. You need not
explicitly define a force at every excitation frequency. Only those values that describe the
character of the loading are required. MSC.Nastran will interpolate for intermediate values.

1 2 3 4 5 6 7 8 9 10
RLOAD1 SID DAREA DELAY DPHASE TC TD

Field Contents
SID Set ID defined by a DLOAD Case Control command or a DLOAD Bulk Data
entry.
DAREA Identification number of the DAREA entry set that defines A.
(Integer > 0)
DELAY Identification number of the DELAY entry set that defines τ .
DPHASE Identification number of the DPHASE entry set that defines θ .
TC TABLEDi entry that defines C(f).
TD TABLEDi entry that defines D(f).

Note that f is the frequency in cycles per unit time and that e iwt = ( cot wt + i sin wt ) .

Frequency-Dependent Loads – RLOAD2 Entry

The RLOAD2 entry is a variation of the RLOAD1 entry used for defining a frequency-dependent
load. Whereas the RLOAD1 entry defines the real and imaginary parts of the complex load, the
RLOAD2 entry defines the magnitude and phase.
The RLOAD2 entry defines dynamic excitation in the form

i { φ ( f ) + θ – 2πfτ }
{ P ( f ) } = { AB ( f )e } Eq. 5-22

The RLOAD2 definition may be related to the RLOAD1 definition by

iφ ( f )
C ( f ) + iD ( f ) = B ( f )e Eq. 5-23

RLOAD1 RLOAD2
Definition Definition

1 2 3 4 5 6 7 8 9 10
RLOAD2 SID DAREA DELAY DPHASE TB TP

Field Contents
SID Set ID defined by a DLOAD Case Control command.
DAREA Identification number of the DAREA entry set that defines A.
(Integer > 0)
Main Index
 CHAPTER 5 117
Frequency Response Analysis

Field Contents
DELAY Identification number of the DELAY entry set that defines τ .
(Integer > 0)
DPHASE Identification number of the DPHASE entry set that defines θ in degrees.
(Integer > 0)
TB TABLEDi entry defining amplitude versus frequency pairs for B ( f ) .
(Integer > 0)
TP TABLEDi entry defining phase angle versus frequency pairs for φ ( f ) in
degrees. (Integer > 0)

Note that f is the frequency in cycles per unit time.

Spatial Distribution of Loading -- DAREA Entry

The DAREA entry defines the degrees-of-freedom where the dynamic load is to be applied and
the scale factor to be applied to the loading. The DAREA entry provides the basic spatial
distribution of the dynamic loading.

1 2 3 4 5 6 7 8 9 10
DAREA SID P1 C1 A1 P2 C2 A2

Field Contents
SID Set ID specified by RLOADi entires.
Pi Grid, extra, or scalar point ID.
Ci Component number.
Ai Scale factor.

A DAREA entry is selected by RLOAD1 or RLOAD2 entries. Any number of DAREA entries
may be used; all those with the same SID are combined.

Time Delay – DELAY Entry

The DELAY entry defines the time delay τ in an applied load.

1 2 3 4 5 6 7 8 9 10
DELAY SID P1 C1 P2 C2 τ2

Field Contents
SID Set ID specified by an RLOADi entry.
Pi Grid, extra, or scalar point ID.

Main Index
118

Field Contents
Ci Component number.
τi Time delay for Pi, Ci. (Default = 0.0)

A DAREA entry must be defined for the same point and component. Any number of DELAY
entries may be used; all those with the same SID are combined.

Phase Lead – DPHASE Entry

The DPHASE entry defines the phase lead θ .

1 2 3 4 5 6 7 8 9 10
DPHASE SID P1 C1 θ1 P2 C2 θ2

Field Contents
SID Set ID specified by an RLOADi entry.
Pi Grid, extra, or scalar point ID.
Ci Component number.
θi Phase lead (in degrees) for Pi, Ci. (Default = 0.0)

A DAREA entry must be defined for the same point and component. Any number of DPHASE
entires may be used; all those with the same SID are combined.

Dynamic Load Tabular Function -- TABLEDi Entries

The TABLEDi entries (i = 1 through 4) each define a tabular function for use in generating
frequency-dependent dynamic loads. The form of each TABLEDi entry varies slightly,
depending on the value of i, as does the algorithm for y(x). The x values need not be evenly
spaced.
The TABLED1, TABLED2, and TABLED3 entries linearly interpolate between the end points
and linearly extrapolate outside of the endpoints, as shown in Figure 5-4. The TABLED1 entry
also performs logrithmic interpolation between points. The TABLED4 entry assigns the
endpoint values to any value beyond the endpoints.

Main Index
 CHAPTER 5 119
Frequency Response Analysis

Linear Interpolation Between Endpoints

y

Linear Extrapolation of Segment x 6 – x 7

Linear Extrapolation of Segment x 2 – x 1

x
x1 x2 x3 x4 x5 x 6 x7

Figure 5-4 Interpolation and Extrapolation for TABLED1,

TABLED2, and TABLED3 Entries

The TABLED1 entry has the following format:

1 2 3 4 5 6 7 8 9 10
TABLED1 TID XAXIS YAXIS

x1 y1 x2 y2 x3 y3 -etc.- ENDT

Field Contents
TID Table identification number.
XAXIS Specifies a linear or logarithmic interpolation for the x-axis. (Character:
“LINEAR” or “LOG”; default = “LINEAR”)
YAXIS Specifies a linear or logarithmic interpolation for the y-axis. (Character:
“LINEAR” or “LOG”; default = “LINEAR”)
xi, yi Tabular values. Values of x are frequency in cycles per unit time.
ENDT Ends the table input.

The TABLED1 entry uses the algorithm

y = yT ( x ) Eq. 5-24

The algorithms used for interpolation and extrapolation are as follows:

XAXIS YAXIS y
T
(x)

xj – x x – xi
LINEAR LINEAR ----------------- yi + ----------------- yj
xj – xi xj – xi

ln ( xj ⁄ x ) ln ( x ⁄ x i )
LOG LINEAR --------------------------- yi + --------------------------- yj
ln ( xj ⁄ xi ) ln ( xj ⁄ xi )

Main Index
120

XAXIS YAXIS y
T
(x)

xj – x x – xi
LINEAR LOG exp ----------------- ln yi + ----------------- ln yj
xj – xi xj – xi

ln ( xj ⁄ x ) ln ( x ⁄ x i )
LOG LOG exp --------------------------- ln y i + --------------------------- yj
ln ( xj ⁄ xi ) ln ( xj ⁄ xi )

The TABLED2 entry has the following format:

1 2 3 4 5 6 7 8 9 10
TABLED2 TID X1

x1 y1 x2 y2 x3 y3 -etc.- ENDT

Field Contents
TID Table identification number.
X1 Table parameter.
xi, yi Tabular values. Values of x are frequency in cycles per unit time.

The TABLED2 entry uses the algorithm

y = y T ( x – X1 ) Eq. 5-25

ENDT ends the table input.

The TABLED3 entry has the following format:

1 2 3 4 5 6 7 8 9 10
TABLED3 TID X1 X2

x1 y1 x2 y2 x3 y3 -etc.- ENDT

Field Contents
TID Table identification number.
X1, X2 Table parameters (X2 ≠ 0.0).
xi, yi Tabular values. Values of x are frequency in cycles per unit time.

The TABLED3 entry uses the algorithm

y = y T  -----------------
x – X1
Eq. 5-26
 X2 

ENDT ends the table input.

Main Index
 CHAPTER 5 121
Frequency Response Analysis

The TABLED4 entry has the following format:

1 2 3 4 5 6 7 8 9 10
TABLED4 TID X1 X2 X3 X4

A0 A1 A2 A3 A4 A5 -etc.- ENDT

Field Contents
TID Table identification number.
Xi Table parameters (X2 ≠ 0.0; X3 < X4).
Ai Coefficients.

The TABLED4 entry uses the algorithm

N
i
Ai  -----------------
x – X1
y = ∑  X2 
Eq. 5-27
i = 0

N is the degree of the power series. When x < X3, X3 is used for x; when x > X4, X4 is used for
x. This condition has the effect of placing bounds on the table; there is no extrapolation outside
of the table boundaries.
ENDT ends the table input.

DAREA Example
Suppose the following command is in the Case Control Section:
DLOAD = 35
in addition to the following entries in the Bulk Data Section:

1 2 3 4 5 6 7 8 9 10
$RLOAD1 SID DAREA DELAY DPHASE TC TD

RLOAD1 35 29 31 40

$DAREA SID POINT COMPONENT SCALE

DAREA 29 30 1 5.2

$DELAY SID POINT COMPONENT LAG

DELAY 31 30 1 0.2

$TABLED1 ID XAXIS YAXIS

$ x1 y1 x2 y2 x3 y3 x4 y4

TABLED1 40 LINEAR LINEAR

0.0 4.0 2.0 8.0 6.0 8.0 ENDT

Main Index
122

The DLOAD Set ID 35 in Case Control selects the RLOAD1 entry in the Bulk Data having a Set ID
35. On the RLOAD1 entry is a reference to DAREA Set ID 29, DELAY Set ID 31, and TABLED1
Set ID 40. The DAREA entry with Set ID 29 positions the loading on grid point 30 in the 1
direction with a scale factor of 5.2 applied to the load. The DELAY entry with Set ID 31 delays
the loading on grid point 30 in the 1 direction by 0.2 units of time. The TABLED1 entry with Set
ID 40 defines the load in tabular form. This table is shown graphically in Figure 5-5. The result
of these entries is a dynamic load applied to grid point 30, component T1, scaled by 5.2 and
delayed by 0.2 units of time.

10

Amplitude
on 5
TABLED1

0 2 4 6
Frequency (Hz)
Figure 5-5 TABLED1 - Amplitude Versus Frequency

Static Load Sets – LSEQ Entry

MSC.Nastran does not have specific data entries for many types of dynamic loads. Only
concentrated forces and moments can be specified directly using DAREA entries. To
accommodate more complicated loadings conveniently, the LSEQ entry is used to define static
load entries that define the spatial distribution of dynamic loads.

1 2 3 4 5 6 7 8 9 10
LSEQ SID DAREA ID LOAD ID TEMP ID

The LSEQ Bulk Data entry contains a reference to a DAREA Set ID and a static Load Set ID. The
static loads are combined with any DAREA entries in the referenced set. The DAREA Set ID
does not need to be defined with a DAREA Bulk Data entry. The DAREA Set ID is referenced
by an RLOADi entry. This reference defines the temporal distribution of the dynamic loading.
The Load Set ID may refer to one or more static load entries (FORCE, PLOADi, GRAV, etc.). All
static loads with the Set ID referenced on the LSEQ entry define the spatial distribution of the
dynamic loading. MSC.Nastran converts this information to equivalent dynamic loading.

Figure 5-6 demonstrates the relationships of these entries. To activate a load set defined in this
manner, the DLOAD Case Control command refers to the Set ID of the selected DLOAD or
RLOADi entry, and the LOADSET Case Control command refers to the Set ID of the selected
LSEQ entries. The LSEQ entries point to the static loading entries that are used to define
dynamic loadings with DAREA Set ID references. Together this relationship defines a complete
dynamic loading. To apply dynamic loadings in this manner, the DLOAD and LOADSET Case
Control commands and the RLOADi and LSEQ Bulk Data entries must be defined. A DAREA

Main Index
 CHAPTER 5 123
Frequency Response Analysis

Bulk Data entry does not need to be defined since the RLOADi and LSEQ entries reference a
common DAREA ID. The LSEQ entry can also be interpreted as an internal DAREA entry
generator for static load entries.

DLOAD LOADSET

Case Control
Bulk Data
RLOADi LSEQ

Dynamic Load DAREA Static Load Entries

(Temporal (ReferenceLink) (Spatial
Distribution) Distribution)
Figure 5-6 Relationship of Dynamic and Static Load Entries

LSEQ Example
Suppose the following commands are in the Case Control Section:
LOADSET = 27
DLOAD = 25
and the following entries are in the Bulk Data Section:

1 2 3 4 5 6 7 8 9 10
$LSEQ SID DAREA LID

LSEQ 27 28 26

$RLOAD1 SID DAREA DELAY DPHASE TC TD

RLOAD1 25 28 29

DAREA
$STATIC LOAD SID Reference
Link
PLOAD1 26 etc.

FORCE 26 etc.

$TABLED1 TID

TABLED1 29 etc.

In the above, the LOADSET request in Case Control selects the LSEQ Set ID 27 entry. The
DLOAD request in Case Control selects the RLOAD1 Set ID 25 entry. This RLOAD1 entry refers
to a TABLED1 ID 29, which is used to define the frequency-dependent variation in the loading.
DAREA Set ID 28 links the LSEQ and RLOAD1 entries. In addition, the LSEQ entry refers to
static Load Set ID 26, which is defined by FORCE and PLOAD1 entries. The FORCE and

Main Index
124

PLOAD1 entries define the spatial distribution of the dynamic loading and through the DAREA
link refer to the RLOAD1/TABLED1 combination for the frequency-varying characteristics of
the load. Note that there is no DAREA entry.

Dynamic Load Set Combination – DLOAD

One of the requirements of frequency-dependent loads is that RLOAD1s and RLOAD2s must
have unique SIDs. If they are to be applied in the same analysis, they must be combined using
the DLOAD Bulk Data entry. The total applied load is constructed from a linear combination of
component load sets as follows:

{ P } = S ∑ Si { Pi } Eq. 5-28
i

where:

S = overall scale factor

S i = scale factor for the i-th load set

P i = i-th load set

P = total applied load

The DLOAD Bulk Data entry has the following format:

1 2 3 4 5 6 7 8 9 10
DLOAD SID S S1 L1 S2 L2

Field Contents
SID Load Set ID.
S Overall scale factor.
Si Individual scale factors.
Li Load Set ID numbers for RLOAD1 and RLOAD2 entries.

As an example, in the following DLOAD entry:

\

$DLOAD SID S S1 L1 S2 L2 -etc.-

DLOAD 33 3.25 0.5 14 2.0 27

a dynamic Load Set ID of 33 is created by taking 0.5 times the loads in the Load Set ID of 14,
adding to it 2.0 times the loads in the Load Set ID of 27, and multiplying that sum by an overall
scale factor of 3.25.

As with other frequency-dependent loads, a dynamic load combination defined by the DLOAD
Bulk Data entry is selected by the DLOAD Case Control command.

Main Index
 CHAPTER 5 125
Frequency Response Analysis

5.6 Solution Frequencies

A major consideration when you conduct a frequency response analysis is selecting the
frequency at which the solution is to be performed. There are six Bulk Data entries that you can
use to select the solution frequencies. It is important to remember that each specified frequency
results in an independent solution at the specified excitation frequency.
To select the loading frequencies, use the FREQ, FREQ1, FREQ2, FREQ3, FREQ4 and FREQ5
Bulk Data entries.

FREQ Defines descrete excitation frequencies.

FREQ1 Defines a starting frequency F start , a frequency incitement ∆f , and the
number of frequency increments to solve NDF.
FREQ2 Defines a starting frequency F start , and ending frequency F end , and the
number of logarithmic intervals, NF, to be used in the frequency range.
FREQ3* Defines the number of excitation frequencies used between modal pairs in a
given range.
FREQ4* Defines excitation frequencies using a spread about each normal mode
within a range.
FREQ5* Defines excitation frequencies as all frequencies in a given range as a defined
fraction of the normal modes.
*
Used for modal solution only.

The FREQUENCY Case Control command selects FREQi Bulk Data entries. All FREQi entries
with the same selected Set ID are applied in the analysis; therefore, you can use any combination
of FREQ, FREQ1, FREQ2, FREQ3, FREQ4 and FREQ5 entries.
The examples that follow show the formats of the FREQi entries. Notice that the six sets of
excitation frequencies shown in the examples will be combined in a single analysis if the Set IDs
are identical.

FREQ
This FREQ entry specifies ten specific (unequally spaced loading frequencies to be analyzed.

1 2 3 4 5 6 7 8 9 10
$FREQ SID F F F F F F F

$ F F F F F F F F

FREQ 3 2.98 3.05 17.9 21.3 25.6 28.8 31.2

29.2 22.4 19.3

Main Index
126

Field Contents
SID Set ID specified by a FREQUENCY Case Control command.
F Frequency value (cycles per unit time).

FREQ1
The FREQ1 example specifies 14 frequencies between 2.9 Hz and 9.4 Hz in increments of 0.5 Hz.

$FREQ1 SID F start ∆f NDF

FREQ1 6 2.9 0.5 13

Field Contents
SID Set ID specified by a FREQUENCY Case Control command.
F start Starting frequency in set (cycles per unit time).

∆f Frequency increment (cycles per unit time).

NDF Number of frequency increments.

FREQ2
The FREQ2 example specifies six logarithmic frequency intervals between 1.0 and 8.0 Hz,
resulting in frequencies at 1.0, 1.4142, 2.0, 2.8284, 4.0, 5.6569, and 8.0 Hz being used for the
analysis.

1 2 3 4 5 6 7 8 9 10
$FREQ2 SID F start F end NF

FREQ2 9 1.0 8.0 6

Field Contents
SID Set ID specified by a FREQUENCY Case Control command.
F start Starting frequency (cycles per unit time).

F end Ending frequency (cycles per unit time).

NF Number of logarithmic intervals.

FREQ3
The FREQ3 example requests 10 frequencies between each set of modes within the range 20 and
2000, plus ten frequencies between 20 and the lowest mode in the range, plus 10 frequencies
between the highest mode in the range and 2000.

Main Index
 CHAPTER 5 127
Frequency Response Analysis

1 2 3 4 5 6 7 8 9 10
$FREQ3 SID F1 F2 TYPE NEF CLUSTER

FREQ3 6 20.0 2000.0 LINEAR 10 2.0

Field Contents
SID Set ID specified by a FREQUENCY Case Control command.
F1 Lower bound of modal frequency range in cycles per unit time.
(Real > 0.0)
F2 Upper bound of modal frequency range in cycles per unit time.
(Real > 0.0, F2 ≥ F1, Default = F1)
TYPE LINEAR or LOG. Specifies linear or logarithmic interpolation between
frequencies. (Character; Default = “LINEAR”)
NEF Number of excitation frequencies within each subrange including the
end points. (Integer > 1, Default = 10)
CLUSTER A CLUSTER value greater than 1 provides closer spacing of excitation
frequencies near the modal frequencies, where greater resolution is
needed. (Real > 0.0; Default = 1.0)

FREQ4
The example FREQ4 chooses 21 equally spaced frequencies across a frequency band of 0.7 ⋅ fN to
1.3 ⋅ f N for each natural frequency between 20 and 2000.

1 2 3 4 5 6 7 8 9 10
$FREQ4 SID F1 F2 FSPD NFM

FREQ4 6 20.0 2000.0 0.30 21

Field Contents
SID Set ID specified by a FREQUENCY Case Control command.
F1 Lower bound of modal frequency range in cycles per unit time.
(Real > 0.0)
F2 Upper bound of modal frequency range in cycles per unit time.
(Real > 0.0, F2 ≥ F1, Default = F1)
FSPD Frequency spread, +/– the fractional amount specified for each mode
which occurs in the frequency range F1 to F2. (1.0 > Real > 0.0,
Default = 0.10)
NFM Number of evenly spaced frequencies per “spread” mode. (Integer > 0;
Default = 3; If NFM is even, NFM + 1 will be used.)

Main Index
128

FREQ5
The example FREQ5 will compute excitation frequencies which are 0.6, 0.8, 0.9, 0.95, 1.0, 1.05, 1.1,
and 1.2 times the natural frequencies for all natural frequencies, but use only the computed
frequencies that fall within the range 20 and 2000.

1 2 3 4 5 6 7 8 9 10
$FREQ5 SID F1 F2 FR1 FR2 FR3 FR4 FR5

FR6 FR7 -etc.-

FREQ5 6 20.0 2000.0 1.0 0.6 0.8 0.9 0.95

1.05 1.1 1.2

Field Contents
SID Set ID specified by a FREQUENCY Case Control command.
F1 Lower bound of modal frequency range in cycles per unit time.
(Real > 0.0)
F2 Upper bound of modal frequency range in cycles per unit time.
(Real > 0.0, F2 ≥ F1, Default = F1)
FRi Fractions of the natural frequencies in the range F1 to F2. (Real > 0.0)

Main Index
 CHAPTER 5 129
Frequency Response Analysis

5.7 Frequency Response Considerations

Exciting an undamped (or modal or viscous damped) system at 0.0 Hz using direct frequency
response analysis gives the same results as a static analysis and also gives almost the same
results when using modal frequency response (depending on the number of retained modes).
Therefore, if the maximum excitation frequency is much less than the lowest resonant frequency
of the system, a static analysis is probably sufficient.

Undamped or very lightly damped structures exhibit large dynamic responses for excitation
frequencies near resonant frequencies. A small change in the model (or running it on another
computer) may result in large changes in such responses.
Use a fine enough frequency step size ( ∆f ) to adequately predict peak response. Use at least five
points across the half-power bandwidth (which is approximately 2ζf n for an SDOF system) as
shown in Figure 5-7.

Peak Response

Peak
------------ = Half-Power Point
Response 2

Frequency

Half-Power Bandwidth
= Excitation Frequency

Figure 5-7 Half-Power Bandwidth

For maximum efficiency, an uneven frequency step size should be used. Smaller frequency
spacing should be used in regions near resonant frequencies, and larger frequency step sizes
should be used in regions away from resonant frequencies.

Main Index
130

5.8 Solution Control for Frequency Response Analysis

The following tables summarize the data entries that can be used to control a frequency response
analysis. Certain data entries are required, some data entries are optional, and others are user
selectable.
In the Executive Control Section of the MSC.Nastran input file, a solution must be selected using
the SOL i statement where i is an integer value chosen from Table 5-3.

Table 5-3 Frequency Response Solutions in MSC.Nastran

Structured
Rigid
Solution
Formats
Sequences
Direct 8 108
Modal 11 111

Solutions 108 and 111 are the preferred SOLs; these are the ones used in the examples that follow.
In the Case Control Section of the MSC.Nastran input file, you must select the solution
parameters associated with the current analysis (i.e., frequencies, loads, and boundary
conditions), and also the output quantities required from the analysis. The Case Control
commands directly related to frequency response analysis are listed in Table 5-4. They can be
combined in the standard fashion with the more generic entries, such as SPC, MPC, etc.

Table 5-4 Case Control Commands for Frequency Response Solution Control

Case
Direct or
Control Description Required/Optional
Modal
Command
DLOAD Both Select the dynamic load set from Required
Bulk Data
FREQUENCY Both Select FREQi entries from Bulk Data Required
METHOD Modal Select the eigenvalue extraction Required
parameters
LOADSET Both Select the LSEQ set from Bulk Data Optional
SDAMPING Modal Select the modal damping table Optional
from Bulk Data
OFREQUENCY Both Select the frequencies for output Optional
(default = all)

The types of results available from a frequency response analysis are similar to those for a typical
static analysis except that the results are a complex function of the applied loading frequency.
Additional quantities (characteristic of dynamic problems) are also available. The output
quantities are summarized in Table 5-5 and Table 5-6.
Main Index
 CHAPTER 5 131
Frequency Response Analysis

Table 5-5 Grid Output from a Frequency Response Analysis

Case Control
Description
Command
ACCELERATION Grid point acceleration results for a set of grid points.
DISPLACEMENT Grid point displacement results for a set of grid points.
(or VECTOR)
OLOAD Requests the applied load table to be output for a set of grid
points.
SACCELERATION Requests the solution set acceleration output: d-set in direct
solutions and modal variables in modal solutions.
SDISPLACEMENT Requests the solution set displacement output: d-set in direct
solutions and modal variables in modal solutions.
SVECTOR Requests the real eigenvector output for the a-set in modal
solutions.
SVELOCITY Requests the solution set velocity output: d-set in direct solutions
and modal variables in modal solutions.
SPCFORCES Requests the forces of a single-point constraint for a set of grid
points.
VELOCITY Grid point velocity results for a set of grid points.

Frequency response output is in real/imaginary format (the default) or magnitude/phase

format (the phase angle is in degrees). Frequency response output is also in SORT1 or SORT2
format. In SORT1 format, the results are listed by frequency; in SORT2 format, the results are
listed by grid point or element number. SORT1 is the default for direct frequency response
analysis (SOL 108), and SORT2 is the default for modal frequency response analysis (SOL 111).
PARAM,CURVPLOT,1 and PARAM,DDRMM,-1 are necessary to obtain SORT1 output in
SOL 111. These output formats are specified with the Case Control commands. The command
DISPLACEMENT(PHASE, SORT2) = ALL
prints displacements in magnitude/phase and SORT2 formats. The output formats are
illustrated in the first example in “Examples” on page 133.

Table 5-6 Element Output from a Frequency Response Analysis

Case Control Command Description

ELSTRESS (or STRESS) Element stress results for a set of elements.
ELFORCE (or FORCE) Element force results for a set of elements.
STRAIN Element strain results for a set of elements.

Main Index
132

A number of Bulk Data entries are unique to frequency response analysis. These entries can be
combined with other generic entries in the Bulk Data. Bulk Data entries directly related to
frequency response analysis are summarized in Table 5-7.

Table 5-7 Bulk Data Entries for Frequency Response Analysis

Bulk Data Direct Required/

Description
Entry or Modal Optional
FREQ, FREQi Both Excitation frequencies Required
RLOADi Both Dynamic loading Required
EIGR or EIGRL Modal Eigenvalue analysis parameters Required
LSEQ Both Dynamic loading from static loads Optional
TABLEDi Both Frequency-dependent tables Both*
DAREA Both Load component and scale factor Required†
DELAY Both Time delay on dynamic load Optional
DPHASE Both Phase angle on dynamic load Optional
DLOAD Both Dynamic load combination, Optional
required if RLOAD1 and RLOAD2
are used
TABDMP1 Modal Modal damping table Optional
*
Required for RLOAD1; optional for RLOAD2.
†
The DAREA ID is required; the DAREA Bulk Data entry is not required if an
LSEQ entry is used.

Main Index
 CHAPTER 5 133
Frequency Response Analysis

5.9 Examples
This section provides several examples showing the input and output. These examples are:

Frequency Response Bulk

Model Output
Data Entries
bd05two EIGRL, FREQ1, TABDMP1, X-Y plots (linear), printed
RLOAD1, DAREA, TABLED1 results (SORT1, SORT2)
bd05bar EIGRL, FREQ1, TABDMP1, X-Y plots (log)
DLOAD, RLOAD2, DAREA,
DPHASE, TABLED1
bd05bkt EIGRL, FREQ1, TABDMP1, X-Y plot (log)
RLOAD1, LSEQ, TABLED1,
PLOAD4

These examples are described in the sections that follow.

Two-DOF Model
Consider the two-DOF system shown in Figure 5-8. Modal frequency response (SOL 111) is run
with a 20 N load applied to the primary mass (grid point 2) across a frequency range of 2 to 10
Hz with an excitation frequency increment of 0.05 Hz. Uniform modal damping of 5% critical
damping is used. Listing 5-1 shows part of the input file.

Auxiliary Structure,
Grid Point 1
P = 20 N sin ωt

Primary Structure,
Grid Point 2

x
Figure 5-8 Two-DOF Model

Main Index
134

Listing 5-1 Input File (Abridged) for the Two-DOF Example

$ FILE bd05two.dat
$
$ TWO-DOF SYSTEM
$ CHAPTER 5, FREQUENCY RESPONSE
$
TIME 5
SOL 111 $ MODAL FREQUENCY RESPONSE
CEND
TITLE = TWO-DOF SYSTEM
SUBTITLE = MODAL FREQUENCY RESPONSE
LABEL = 20 N FORCE APPLIED TO PRIMARY MASS
$
$ SPECIFY SPC
SPC = 996
$
$ SPECIFY MODAL EXTRACTION
METHOD = 10
$
$ SPECIFY DYNAMIC INPUT
DLOAD = 999
FREQ = 888
SDAMPING = 777
$
$ SELECT OUTPUT
SET 11 = 1,2
DISPLACEMENT(PHASE,PLOT) = 11
$
$ XYPLOTS
$
... X-Y plot commands ...
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$ ENTRIES FOR FREQUENCY RESPONSE
$
$ LOAD DEFINITION
$RLOAD1 SID DAREA TC
RLOAD1 999 997 901
$DAREA SID P1 C1 A1
DAREA 997 2 2 20.0
$TABLED1 TID +TABL1
$+TABL1 X1 Y1 X2 Y3 ETC.
TABLED1 901 +TAB901
+TAB901 0.0 1.0 10.0 1.0 ENDT
$
$ ALTERNATE LOAD DEFINITION USING DLOAD
$DLOAD SID S S1 RLOAD1
$DLOAD 999 1.0 1.0 998
$RLOAD1 SID DAREA TC
$RLOAD1 998 997 901
$
$ FREQUENCY RANGE 2-10 HZ
$FREQ1 SID F1 DF NDF
FREQ1 888 2. 0.05 160
$
$ MODAL DAMPING OF 5% CRITICAL
$TABDMP1 TID TYPE +TABD1
$+TABD1 F1 G1 F2 G2 ETC.
TABDMP1 777 CRIT +TABD7
+TABD7 0. 0.05 100. 0.05 ENDT
$
$ MODAL EXTRACTION
$EIGRL SID V1 V2 ND MSGLVL
EIGRL 10 -0.1 20. 0
$
... basic model ...
$
ENDDATA

Main Index
 CHAPTER 5 135
Frequency Response Analysis

Table 5-8 shows the relationship between the Case Control commands and the Bulk Data
entries. Note that the RLOAD1 entry references the DAREA and TABLED1 entries. The input
file also shows an alternate way to specify the dynamic load, by using a DLOAD Bulk Data entry.
Because there is only a single RLOAD1 entry, the DLOAD Bulk Data entry is not required.

Table 5-8 Relationship Between the Case Control Commands and Bulk Data Entries for
the Two-DOF Model

Case Control Bulk Data

METHOD EIGRL
FREQUENCY FREQ1
SDAMPING TABDMP1
DLOAD RLOAD1

DAREA
TABLED1

The RLOAD1 entry describes a sinusoidal load in the form

i { θ – 2πfτ }
{ P ( f ) } = { A [ C ( f ) + iD ( f ) ]e } Eq. 5-29

where:

A = 20.0 (entered on the DAREA entry)

C = 1.0 for all frequencies entered on the TABLED1 entry
D = 0.0 (field 7 of the RLOAD1 entry is blank)
θ = 0.0 (field 5 of the RLOAD1 entry is blank)
τ = 0.0 (field 4 of the RLOAD1 entry is blank)

Output can be printed in either real/imaginary or magnitude/phase format and in either SORT1
or SORT2 format. These formats are illustrated in Figure 5-9, Figure 5-10, and Figure 5-11
showing a portion of their printed output.

Main Index
136

POINT-ID = 1
C O M P L E X D I S P L A C E M E N T V E C T O R
(REAL/IMAGINARY)

FREQUENCY TYPE T1 T2 T3 R1 R2 R3
2.000000E+00 G 0.0 2.813052E-03 0.0 0.0 0.0 0.0
0.0 -2.107985E-04 0.0 0.0 0.0 0.0
2.050000E+00 G 0.0 2.866642E-03 0.0 0.0 0.0 0.0
0.0 -2.229164E-04 0.0 0.0 0.0 0.0
2.100000E+00 G 0.0 2.923141E-03 0.0 0.0 0.0 0.0
0.0 -2.358382E-04 0.0 0.0 0.0 0.0
2.150000E+00 G 0.0 2.982732E-03 0.0 0.0 0.0 0.0
0.0 -2.496362E-04 0.0 0.0 0.0 0.0
2.200000E+00 G 0.0 3.045609E-03 0.0 0.0 0.0 0.0
0.0 -2.643908E-04 0.0 0.0 0.0 0.0

POINT-ID = 2
C O M P L E X D I S P L A C E M E N T V E C T O R
(REAL/IMAGINARY)

FREQUENCY TYPE T1 T2 T3 R1 R2 R3
2.000000E+00 G 0.0 2.374954E-03 0.0 0.0 0.0 0.0
0.0 -1.129933E-04 0.0 0.0 0.0 0.0
2.050000E+00 G 0.0 2.397706E-03 0.0 0.0 0.0 0.0
0.0 -1.180853E-04 0.0 0.0 0.0 0.0
2.100000E+00 G 0.0 2.421475E-03 0.0 0.0 0.0 0.0
0.0 -1.234173E-04 0.0 0.0 0.0 0.0
2.150000E+00 G 0.0 2.446311E-03 0.0 0.0 0.0 0.0
0.0 -1.290072E-04 0.0 0.0 0.0 0.0
2.200000E+00 G 0.0 2.472262E-03 0.0 0.0 0.0 0.0
0.0 -1.348744E-04 0.0 0.0 0.0 0.0

Figure 5-9 Real/Imaginary Output in SORT2 Format

Main Index
 CHAPTER 5 137
Frequency Response Analysis

POINT-ID = 1
C O M P L E X D I S P L A C E M E N T V E C T O R
(MAGNITUDE/PHASE)

FREQUENCY TYPE T1 T2 T3 R1 R2 R3
2.000000E+00 G 0.0 2.820939E-03 0.0 0.0 0.0 0.0
0.0 355.7145 0.0 0.0 0.0 0.0
2.050000E+00 G 0.0 2.875296E-03 0.0 0.0 0.0 0.0
0.0 355.5535 0.0 0.0 0.0 0.0
2.100000E+00 G 0.0 2.932640E-03 0.0 0.0 0.0 0.0
0.0 355.3874 0.0 0.0 0.0 0.0
2.150000E+00 G 0.0 2.993161E-03 0.0 0.0 0.0 0.0
0.0 355.2159 0.0 0.0 0.0 0.0
2.200000E+00 G 0.0 3.057064E-03 0.0 0.0 0.0 0.0
0.0 355.0386 0.0 0.0 0.0 0.0

POINT-ID = 2
C O M P L E X D I S P L A C E M E N T V E C T O R
(MAGNITUDE/PHASE)

FREQUENCY TYPE T1 T2 T3 R1 R2 R3
2.000000E+00 G 0.0 2.377640E-03 0.0 0.0 0.0 0.0
0.0 357.2761 0.0 0.0 0.0 0.0
2.050000E+00 G 0.0 2.400612E-03 0.0 0.0 0.0 0.0
0.0 357.1805 0.0 0.0 0.0 0.0
2.100000E+00 G 0.0 2.424619E-03 0.0 0.0 0.0 0.0
0.0 357.0823 0.0 0.0 0.0 0.0
2.150000E+00 G 0.0 2.449710E-03 0.0 0.0 0.0 0.0
0.0 356.9813 0.0 0.0 0.0 0.0
2.200000E+00 G 0.0 2.475939E-03 0.0 0.0 0.0 0.0
0.0 356.8773 0.0 0.0 0.0 0.0

Figure 5-10 Magnitude/Phase Output in SORT2 Format

Main Index
138

FREQUENCY = 2.000000E+00
C O M P L E X D I S P L A C E M E N T V E C T O R
(REAL/IMAGINARY)

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 2.813051E-03 0.0 0.0 0.0 0.0
0.0 -2.107985E-04 0.0 0.0 0.0 0.0
2 G 0.0 2.374954E-03 0.0 0.0 0.0 0.0
0.0 -1.129933E-04 0.0 0.0 0.0 0.00

FREQUENCY = 2.050000E+00
C O M P L E X D I S P L A C E M E N T V E C T O R
(REAL/IMAGINARY)

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 2.866640E-03 0.0 0.0 0.0 0.0
0.0 -2.229163E-04 0.0 0.0 0.0 0.0
2 G 0.0 2.397706E-03 0.0 0.0 0.0 0.0
0.0 -1.180853E-04 0.0 0.0 0.0 0.00

FREQUENCY = 2.100000E+00
C O M P L E X D I S P L A C E M E N T V E C T O R
(REAL/IMAGINARY)

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 2.923141E-03 0.0 0.0 0.0 0.0
0.0 -2.358381E-04 0.0 0.0 0.0 0.0
2 G 0.0 2.421475E-03 0.0 0.0 0.0 0.0
0.0 -1.234173E-04 0.0 0.0 0.0 0.00

FREQUENCY = 2.150000E+00
C O M P L E X D I S P L A C E M E N T V E C T O R
(REAL/IMAGINARY)

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 2.982731E-03 0.0 0.0 0.0 0.0
0.0 -2.496362E-04 0.0 0.0 0.0 0.0
2 G 0.0 2.446311E-03 0.0 0.0 0.0 0.0
0.0 -1.290072E-04 0.0 0.0 0.0 0.00

FREQUENCY = 2.200000E+00
C O M P L E X D I S P L A C E M E N T V E C T O R
(REAL/IMAGINARY)

POINT ID. TYPE T1 T2 T3 R1 R2 R3

1 G 0.0 3.045608E-03 0.0 0.0 0.0 0.0
0.0 -2.643907E-04 0.0 0.0 0.0 0.0
2 G 0.0 2.472263E-03 0.0 0.0 0.0 0.0
0.0 -1.348744E-04 0.0 0.0 0.0 0.0

Figure 5-11 Real/Imaginary Output in SORT1 Format

Figure 5-12 shows the plots of the the resulting displacement magnitudes for grid points 1 and
2. Note that the response for grid point 1 is nearly an order of magnitude larger than that of grid
point 2. This large difference in response magnitudes is characteristic of dynamic absorbers (also
called tuned mass dampers), in which an auxiliary structure (i.e., the small mass and stiffness)
is attached to the primary structure in order to decrease the dynamic response of the primary
structure. If this same model is rerun without the auxiliary structure, the response of the
primary structure (grid point 2) at 5.03 Hz is twice what it was with the auxiliary structure
attached, as shown in Figure 5-13.

Main Index
 CHAPTER 5 139
Frequency Response Analysis

0.12

Displacement 1 (m)

0.014
Displacement 2 (m)

0
2 Frequency (Hz) 10
Figure 5-12 Displacement Response Magnitudes With the Auxiliary Structure

0.02
Displacement 2 (m)

0
2 Frequency (Hz) 10
Figure 5-13 Displacement Response Magnitude Without the Auxiliary Structure

Cantilever Beam Model

Consider the cantilever beam shown in Figure 5-14. This model is a planar model of the
cantilever beam introduced in “Real Eigenvalue Analysis” on page 33 with unrestrained DOFs
in the T2 and R3 directions. Two loads are applied: one at grid point 6 and the other at grid point
11. The loads have the frequency variation shown in Figure 5-15. The loads in the figure are
indicated with a heavy line in order to emphasize their values. The load at grid point 6 has a
45-degree phase lead, and the load at grid point 11 is scaled to be twice that of the load at grid

Main Index
140

point 6. Modal frequency response is run across a frequency range of 0 to 20 Hz. Modal
damping is used with 2% critical damping between 0 and 10 Hz and 5% critical damping above
10 Hz. Modes to 500 Hz are computed using the Lanczos method.

6 11
x

p(f) 2p(f)
Figure 5-14 Cantilever Beam Model with Applied Loads

Main Index
 CHAPTER 5 141
Frequency Response Analysis

4.0

Load Magnitude 6 (N)

3.0

0
100
Load Phase 6 (deg)

45

0
8.0
Load Magnitude 11 (N)

6.0

0
100
Load Phase 11 (deg)

0
0 Frequency (Hz) 20
Figure 5-15 Applied Loads

Listing 5-2 shows the abridged input file. The output quantities, as defined in the Case Control
Section, are the applied loads (OLOAD) for grid points 6 and 11, physical displacements
(DISPLACEMENT) for grid points 6 and 11, solution set displacements (SDISPLACEMENT) for
modes 1 and 2, and element forces (ELFORCE) for element 6. These output quantities are plotted
rather than printed.

Main Index
142

Listing 5-2 Input File (Abridged) for the Beam Example

$ FILE bd05bar.dat
$
$ CANTILEVER BEAM MODEL
$ CHAPTER 5, FREQUENCY RESPONSE
$
SOL 111 $ MODAL FREQUENCY RESPONSE
TIME 10
CEND
TITLE = CANTILEVER BEAM
SUBTITLE = MODAL FREQUENCY RESPONSE
$
SPC = 21
$
DLOAD = 22
FREQ = 27
SDAMPING = 20
$
METHOD = 10
$
SET 15 = 6,11
OLOAD(PHASE,PLOT) = 15
$
$ PHYSICAL OUTPUT REQUEST
SET 11 = 6,11
DISPLACEMENT(PHASE,PLOT) = 11
$
$ MODAL SOLUTION SET OUTPUT
SET 12 = 1,2
SDISP(PHASE,PLOT) = 12
$
$ ELEMENT FORCE OUTPUT
SET 13 = 6
ELFORCE(PHASE,PLOT) = 13
$
$ XYPLOTS
$
... X-Y plot commands ...
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
EIGRL 10 -0.1 500. 0
$
FREQ1 27 0.0 0.05 400
$
TABDMP1 20 CRIT +TABD1
+TABD1 0.0 0.02 10.0 0.02 10.01 0.05 25.0 0.05 +TABD2
+TABD2 ENDT
$
$ DYNAMIC LOADING
$DLOAD SID S S1 L1 S2 L2
DLOAD 22 1.0 1.0 231 1.0 232
$RLOAD2 SID DAREA DELAY DPHASE TB TP
RLOAD2 231 241 261 25
RLOAD2 232 242 25
$DAREA SID P1 C1 A1
DAREA 241 6 2 1.0
DAREA 242 11 2 2.0
$DPHASE SID P1 C1 TH1
DPHASE 261 6 2 45.
$TABLED1 TID +TABL1
$+TABL1 X1 Y1 X2 Y2 ETC.
TABLED1 25 +TABL1
+TABL1 0. 1. 5.0 3. 15.0 3.0 20.0 1. +TABL2
+TABL2 25.0 1. ENDT
$
... basic model ...
$
ENDDATA

Main Index
 CHAPTER 5 143
Frequency Response Analysis

Table 5-9 shows the relationship between the Case Control commands and the Bulk Data
entries. Note that the DLOAD Bulk Data entry references two RLOAD2 entries, each of which
references a separate DAREA entry and a common TABLED1 entry. The RLOAD2 entry for grid
point 6 also references a DPHASE entry that defines the 45-degree phase lead.

Table 5-9 Relationship Between Case Control Commands and Bulk Data Entries for the
Beam Model

Case Control Bulk Data

METHOD EIGRL
FREQUENCY FREQ1
SDAMPING TABDMP1
DLOAD DLOAD

RLOAD2 231

DAREA
DPHASE
TABLED1

RLOAD2 232
DAREA

The RLOAD2 entry describes a sinusoidal load in the form

i { φ ( f ) + θ – 2πfτ }
{ P ( f ) } = { AB ( f ) e } Eq. 5-30

where:

A = 1.0 for grid point 6 and 2.0 for grid point 11 (entered on the DAREA entry)
B = function defined on the TABLED1 entry
φ = 0.0 (field 7 of the RLOAD2 entry is blank)
θ = phase lead of 45 degrees for grid point 6 (entered on the DPHASE entry)
τ = 0.0 (field 4 of the RLOAD2 entry is blank)

Logarithmic plots of the output are shown in the following figures. Figure 5-16 shows the
magnitude of the displacements for grid points 6 and 11. Figure 5-17 shows the magnitude of
the modal displacements for modes 1 and 2. Figure 5-18 shows the magnitude of the bending
moment at end A in plane 1 for element 6. Logarithmic plots are especially useful for displaying
frequency response results since there can be several orders of magnitude between the
maximum and minimum response values.

Main Index
144

1.0

Displacement 6 (m)
1.0E-5

1.0
Displacement 11 (m)

1.0E-4
0 Frequency (Hz) 20

Figure 5-16 Displacement Magnitude (Log)

1.0
SDisplacement 1

1.0E-4
1.0E-2
SDisplacement 2

1.0E-4

0 Frequency (Hz) 20
Figure 5-17 Modal Displacement Magnitude (Log)

Main Index
 CHAPTER 5 145
Frequency Response Analysis

1000

Moment A1 (N-m)

0.1
0 Frequency (Hz) 20

Figure 5-18 Bending Moment Magnitude at End A, Plane 1 (Log)

Bracket Model
Consider the bracket model shown in Figure 5-19. An oscillating pressure load of 3 psi is
applied to the elements on the top face in the z-direction. The model is constrained at its base.
Modal frequency response is run from 0 to 100 Hz with a frequency step size of 0.2 Hz.
Eigenvalues to 1000 Hz are computed using the Lanczos method. Modal damping is applied as
2% critical damping for all modes.

Figure 5-19 Bracket Model

Listing 5-3 shows the abridged input file. The LSEQ entry is used to apply the pressure loads
(PLOAD4 entries). Note that the LSEQ and RLOAD1 entries reference a common DAREA
ID (999) and that there is no explicit DAREA entry. Table 5-10 shows the relationship between
the Case Control commands and the Bulk Data entries.

Main Index
146

Listing 5-3 Input File (Abridged) for the Bracket Model

$ FILE bd05bkt.dat
$
$ BRACKET MODEL
$ CHAPTER 5, FREQUENCY RESPONSE
$
SOL 111 $ MODAL FREQUENCY RESPONSE
TIME 100
CEND
TITLE = BRACKET MODEL
SUBTITLE = MODAL FREQUENCY RESPONSE ANALYSIS
$
SPC = 1
$
METHOD = 777
DLOAD = 2
LOADSET = 3
SDAMPING = 4
FREQUENCY = 5
$
$ OUTPUT REQUEST
SET 123 = 999
DISPLACEMENT(PHASE,PLOT)=123
$
$ XYPLOTS
$
... X-Y plot commands ...
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10......
$
$ NORMAL MODES TO 1000 HZ
$EIGRL SID V1 V2
EIGRL 777 -0.1 1000.

$
$ EXCITATION FREQUENCY DEFINITION 0 TO 100 HZ
$FREQ1 SID F1 DF NDF
FREQ1 5 0.0 0.2 500
$
$ MODAL DAMPING OF 2% CRITICAL FOR ALL MODES
$TABDMP1 TID TYPE +TABD1
$+TABD1 F1 G1 F2 G2 ETC.
TABDMP1 4 CRIT +TABD1
+TABD1 0.0 0.02 1000.0 0.02 ENDT
$
$ LOAD DEFINITION
$
$RLOAD1 SID DAREA DELAY DPHASE TC TD
RLOAD1 2 999 22
$
$LSEQ SID DAREA LID TID
LSEQ 3 999 1
$
$TABLED1 TID +TABL1
$+TABL1 X1 Y1 X2 Y2 ETC.
TABLED1 22 +TABL1
+TABL1 0.0 1.0 1000.0 1.0 ENDT
$
$ PRESURE LOAD OF 3 PSI PER ELEMENT
$PLOAD4 SID EID P1
PLOAD4 1 171 -3.
PLOAD4 1 172 -3.
PLOAD4 1 160 -3.
etc.
$
... basic model ...
$
ENDDATA

Main Index
 CHAPTER 5 147
Frequency Response Analysis

Table 5-10 Relationship Between Case Control Commands and

Bulk Data Entries for the Bracket Model

Case Control Bulk Data

METHOD EIGRL
FREQUENCY FREQ1
SDAMPING TABDMP1
LOADSET PLOAD4
LSEQ

(DAREA ID)
DLOAD
RLOAD1
TABLED1

Figure 5-20 shows a logarithmic plot of the z-displacement magnitude of grid point 999, which
is the concentrated mass at the center of the cutout.

10

1.000

0.100
Displacement (m)

0.010

0.001

0.0001
0 Frequency (Hz) 100
Figure 5-20 Displacement Magnitude (Log)

Main Index
148

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Transient Response Analysis
6
■ Overview
■ Direct Transient Response Analysis
■ Modal Transient Response Analysis
■ Modal Versus Direct Transient Response

■ Transient Excitation Definition

■ Integration Time Step
■ Transient Excitation Considerations
■ Solution Control for Transient Response Analysis
■ Examples

Main Index
 6.1 Overview
Transient response analysis is the most general method for computing forced dynamic response.
The purpose of a transient response analysis is to compute the behavior of a structure subjected
to time-varying excitation. The transient excitation is explicitly defined in the time domain. All
of the forces applied to the structure are known at each instant in time. Forces can be in the form
of applied forces and/or enforced motions (see “Enforced Motion” on page 195).

The important results obtained from a transient analysis are typically displacements, velocities,
and accelerations of grid points, and forces and stresses in elements.
Depending upon the structure and the nature of the loading, two different numerical methods
can be used for a transient response analysis: direct and modal. The direct method performs a
numerical integration on the complete coupled equations of motion. The modal method utilizes
the mode shapes of the structure to reduce and uncouple the equations of motion (when modal
or no damping is used); the solution is then obtained through the summation of the individual
modal responses. The choice of the approach is problem dependent. The two methods are
described in “Direct Transient Response Analysis” on page 151 and “Modal Transient
Response Analysis” on page 156.

Main Index
 CHAPTER 6 151
Transient Response Analysis

6.2 Direct Transient Response Analysis

In direct transient response, structural response is computed by solving a set of coupled
equations using direct numerical integration. Begin with the dynamic equation of motion in
matrix form
·· ·
[M]{u(t) } + [B]{u(t )} + [K]{u(t )} = {P(t )} Eq. 6-1

The fundamental structural response (displacement) is solved at discrete times, typically with a
fixed integration time step ∆t .
By using a central finite difference representation for the velocity { u· ( t ) } and the acceleration
··
{ u ( t ) } at discrete times,

· 1
{ u n } = --------- { u n + 1 – u n – 1 }
2∆t

·· 1
{ u n } = -------- { u n + 1 – 2u n + u }
2 n–1
∆t Eq. 6-2

and averaging the applied force over three adjacent time points, the equation of motion can be
rewritten as:

M B
-------- ( u n + 1 – 2u n + u n – 1) + --------- ( u n + 1 – u n – 1 ) Eq. 6-3
2 2∆t
∆t
K 1
+ ---- ( u n + 1 + u n + u n – 1) = --- ( P n + 1 + P n + P n – 1 )
3 3

Collecting terms, the equation of motion can be rewritten as:

[ A1 ] { un + 1 } = [ A2 ] + [ A3 ] { un } + [ A4 ] { un – 1 } Eq. 6-4

where:

M B K
A1 = -------- + --------- + ----
2 2∆t 3
∆t

[ A 2 ] = 1--- { P n + 1 + P n + P n – 1 }
3
2M K
[ A3 ] = -------2- – ----
3
∆t

M B K
[ A4 ] = – -------2- + --------- – ----
2∆t 3
∆t

Matrix [ A 1 ] is termed the dynamic matrix, and [ A 2 ] is the applied force (averaged over three
adjacent time points). This approach is similar to the classical Newmark-Beta direct integration
method except that { P ( t ) } is averaged over three time points and [ K ] is modified such that the
dynamic equation of motion reduces to a static solution [ K ] { u n } = { P n } if no [ M ] or [ B ] exists.

Main Index
152

The transient solution is obtained by decomposing [ A 1 ] and applying it to the right-hand side of
the above equation. In this form, the solution behaves like a succession of static solutions with
each time step performing a forward-backward substitution (FBS) on a new load vector. Note
that the transient nature of the solution is carried through by modifying the applied force matrix
[ A 2 ] with the [ A 3 ] and [ A 4 ] terms.

In its simplest form, the [ M ] , [ B ] , and [ K ] matrices are assumed to be constant throughout the
analysis and do not change with time. Special solution methods are available in MSC.Nastran
for variations in these matrices (see the MSC.Nastran Advanced Dynamic Analysis User’s
Guide).
A significant benefit presents itself if ∆t remains constant during the analysis. With a constant
∆t , the [ A 1 ] matrix needs to be decomposed only once. Each progressive step in the analysis is
only an FBS of a new load vector. If ∆t is changed, [ A 1 ] must be redecomposed, which can be a
costly operation in large problems.
Another efficiency in the direct transient solution is that the output time interval may be greater
than the solution time interval. In many cases it is not necessary to sample output response at
each solution time. For example, if the solution is performed every 0.001 second the results can
be output every fifth time step or every 0.005 second. This efficiency reduces the amount of
output.

Damping in Direct Transient Response

The damping matrix [ B ] is used to represent the energy dissipation characteristics of a structure.
In the general case, the damping matrix is comprised of several matrices

1 2 G 1
[ B ] = [ B ] + [ B ] + ------- [ K ] + ------- ∑ G E [ K E ] Eq. 6-5
W3 W4

where:

1 = damping elements (CVISC, CDAMPi) + B2GG

[B ]
2 = B2PP direct input matrix + transfer functions
[B ]
G = overall structural damping coefficient (PARAM,G)
W 3 = frequency of interest in radians per unit time (PARAM,W3) for the conversion of
overall structural damping into equivalent viscous damping
[ K ] = global stiffness matrix
G E = element structural damping coefficient (GE on the MATi entry)

W 4 = frequency of interest in radians per unit time (PARAM,W4) for conversion of element
structural damping into equivalent viscous damping
[ K E ] = element stiffness matrix

Main Index
 CHAPTER 6 153
Transient Response Analysis

Transient response analysis does not permit the use of complex coefficients. Therefore,
structural damping is included by means of equivalent viscous damping. To appreciate the
impact of this on the solution, a relation between structural damping and equivalent viscous
damping must be defined.
The viscous damping force is a damping force that is a function of a damping coefficient b and
the velocity. It is an induced force that is represented in the equation of motion using the [B]
matrix and velocity vector.
·· ·
[M]{u(t) } + [B]{u(t )} + [K]{u(t )} = {P(t )} Eq. 6-6

The structural damping force is a displacement-dependent damping. The structural damping

force is a function of a damping coefficient G and a complex component of the structural
stiffness matrix.
··
[ M ] { u ( t ) } + ( 1 + iG ) [ K ] { u ( t ) } = { P ( t ) } Eq. 6-7

Assuming constant amplitude oscillatory response for an SDOF system, the two damping forces
are identical if

Gk = bω Eq. 6-8

or

Gk
b = ------- Eq. 6-9
ω

Therefore, if structural damping G is to be modeled using equivalent viscous damping b , then

the equality Eq. 6-9 holds at only one frequency (see Figure 6-1).
Two parameters are used to convert structural damping to equivalent viscous damping. An
overall structural damping coefficient can be applied to the entire system stiffness matrix using
PARAM,W3,r where r is the circular frequency at which damping is to be made equivalent. This
parameter is used in conjunction with PARAM,G. The default value for W3 is 0.0, which causes
the damping related to this source to be ignored in transient analysis.

PARAM,W4 is an alternate parameter used to convert element structural damping to equivalent

viscous damping. PARAM,W4,r is used where r is the circular frequency at which damping is
to be made equivalent. PARAM,W4 is used in conjunction with the GE field on the MATi entry.
The default value for W4 is 0.0 which causes the related damping terms to be ignored in transient
analysis.

Units for PARAM,W3 and PARAM,W4 are radians per unit time. The choice of W3 or W4 is
typically the dominant frequency at which the damping is active. Often, the first natural
frequency is chosen, but isolated individual element damping can occur at different frequencies
and can be handled by the appropriate data entries.

Main Index
154

Damping Structural Damping fs = iGku

Force
Equivalent b = Gk ⁄ ω 3 (or ω 4 )
Viscous
Damping

ω 3 (or ω 4 ) ω
fv = bu· = ib ω u

Figure 6-1 Structural Damping Versus Viscous Damping

(Constant Oscillatory Displacement)

Initial Conditions in Direct Transient Response

You may impose initial displacements and/or velocities in direct transient response. The TIC
Bulk Data entry is used to define initial conditions on the components of grid points. The IC
Case Control command is used to select TIC entries from the Bulk Data.
If initial conditions are used, initial conditions should be specified for all DOFs having nonzero
values. Initial conditions for any unspecified DOFs are set to zero.
·
Initial conditions { u 0 } and { u 0 } are used to determine the values of { u – 1 } , { P 0 } , and { P –1 } used
in Eq. 6-4 to calculate { u 1 } .
·
{ u – 1 } = { u 0 } – { u 0 }∆t Eq. 6-10

·
{ P–1 } = [ K ] { u –1 } + [ B ] { u0 } Eq. 6-11

In the presence of initial conditions, the applied load specified at t = 0 is replaced by

·
{ P0 } = [ K ] { u0 } + [ B ] { u0 } Eq. 6-12

Regardless of the initial conditions specified, the initial acceleration for all points in the structure
is assumed to be zero (constant initial velocity).
The format for the TIC entry is

1 2 3 4 5 6 7 8 9 10
TIC SID G C U0 V0

Field Contents
SID Set ID specified by the IC Case Control command.
G Grid, scalar, or extra point.
C Component number.
Main Index
 CHAPTER 6 155
Transient Response Analysis

Field Contents
U0 Initial displacement.
V0 Initial velocity.

Initial conditions may be specified only in direct transient response. In modal transient response
all initial conditions are set to zero. Initial conditions may be specified only in the a-set (see
“Advanced Dynamic Analysis Capabilities” on page 311).

Main Index
156

6.3 Modal Transient Response Analysis

Modal transient response is an alternate approach to computing the transient response of a
structure. This method uses the mode shapes of the structure to reduce the size, uncouple the
equations of motion (when modal or no damping is used), and make the numerical integration
more efficient. Since the mode shapes are typically computed as part of the characterization of
the structure, modal transient response is a natural extension of a normal modes analysis.

As a first step in the formulation, transform the variables from physical coordinates { u } to modal
coordinates { ξ } by

{u(t )} = [φ ]{ξ(t )} Eq. 6-13

The mode shapes [ φ ] are used to transform the problem in terms of the behavior of the modes
as opposed to the behavior of the grid points. Eq. 6-13 represents an equality if all modes are
used; however, because all modes are rarely used, the equation usually represents an
approximation.
To proceed, temporarily ignore the damping, resulting in the equation of motion
··
[M]{u(t) } + [K]{u(t )} = {P(t) } Eq. 6-14

If the physical coordinates in terms of the modal coordinates (Eq. 6-13 is substituted into Eq. 6-
14), the following equation is obtained:

··
[ M][ φ]{ ξ( t) } + [ K ]{ φ}{ ξ( t )} = { P( t) } Eq. 6-15

This is now the equation of motion in terms of the modal coordinates. At this point, however,
the equations remain coupled.
To uncouple the equations, premultiply by [ φ ] T to obtain

T ·· T T
[ φ ] [ M ][φ ]{ ξ} + [ φ ] [ K ][ φ] {ξ } = [ φ ] { P( t) } Eq. 6-16

where:

T
[ φ ] [ M ] [ φ ] = modal (generalized) mass matrix
T
[ φ ] [ K ] [ φ ] = modal (generalized) stiffness matrix
T
[ φ ] { P } = modal force vector

The final step uses the orthogonality property of the mode shapes to formulate the equation of
motion in terms of the generalized mass and stiffness matrices that are diagonal matrices. These
matrices do not have off-diagonal terms that couple the equations of motion. Therefore, in this
form, the modal equations of motion are uncoupled. In this uncoupled form, the equations of
motion are written as a set on uncoupled SDOF systems as

··
mi ξi ( t ) + ki ξi ( t ) = pi ( t ) Eq. 6-17

Main Index
 CHAPTER 6 157
Transient Response Analysis

where:

m i = i-th modal mass

k i = i-th modal stiffness

p i = i-th modal force

Note that there is no damping in the resulting equation. The next subsection describes how to
include damping in modal transient response.
Once the individual modal responses ξ i ( t ) are computed, physical responses are recovered as
the summation of the modal responses

{u(t )} = [φ ]{ξ(t )} Eq. 6-18

Since numerical integration is applied to the relatively small number of uncoupled equations,
there is not as large a computational penalty for changing ∆t as there is in direct transient
response analysis. However, a constant ∆t is still recommended.
Another efficiency option in the modal transient solution is that the output time interval may be
greater than the solution time interval. In many cases, it is not necessary to sample output
response at each solution time. For example, if the solution is performed every 0.001 second, the
results can be output every fifth time step or every 0.005 second. This efficiency reduces the
amount of output.

Damping in Modal Transient Response Analysis

If the damping matrix [ B ] exists, the orthogonality property (see “Overview of Normal Modes
Analysis” on page 37) of the modes does not, in general, diagonalize the generalized damping
matrix

T
[ φ ] [ B ] [ φ ] ≠ diagonal Eq. 6-19

In the presence of a [ B ] matrix, the modal transient approach solves the coupled problem in
terms of modal coordinates using the direct transient numerical integration approach described
in Section 4.2 as follows:

[ A1 ] { ξn + 1 } = [ A2 ] + [ A3 ] { ξn } + [ A4 ] { ξn – 1 } Eq. 6-20

where:

T M B K
[ A 1 ] = [ φ ] -------2- + 2∆t
--------- + ---- [ φ ]
3
∆t

[ A 2 ] = 1--- [ φ ]T { P n + 1 + P n + P n – 1 }
3
T 2M K
[ A 3 ] = [ φ ] -------2- – ---3- [ φ ]
∆t

T M B K
[ A 4 ] = [ φ ] – -------2- + 2∆t
--------- – ---- [ φ ]
3
∆t
Main Index
158

These equations are similar to the direct transient method except that they are in terms of modal
coordinates. Since the number of modes used in a solution is typically much less than the
number of physical variables, the direct integration of the modal equations is not as costly as
with physical variables.
If damping is applied to each mode separately, the decoupled equations of motion can be
maintained. When modal damping is used, each mode has damping b i . The equations of
motion remain uncoupled and have the following form for each mode:

·· ·
mi ξi ( t ) + b i ξi ( t ) + k i ξi ( t ) = p i ( t ) Eq. 6-21

or

·· · 2 1
ξ i ( t ) + 2ζ i ω i ξ i ( t ) + ω i ξ i ( t ) = ------ p i ( t ) Eq. 6-22
mi

where:

ζ i = b i ⁄ ( 2m i ω i ) ≡ modal damping ratio

ω i = k i ⁄ m i ≡ modal frequency (eigenvalue)

2

The TABDMP1 Bulk Data entry defines the modal damping ratios. A table is created by the
frequency-damping pairs specified on a TABDMP1 entry. The solution refers to this table for the
damping value to be applied at a particular frequency. The TABDMP1 Bulk Data entry has a Set
ID. A particular TABDMP1 table is activated by selecting the Set ID with SDAMPING = Set ID
Case Control command.

1 2 3 4 5 6 7 8 9 10
TABDMP1 ID TYPE

f1 g1 f2 g2 f3 g3 -etc.- ENDT

Field Contents
TID Table identification number.
TYPE Type of damping units:
G (default)
CRIT
Q
fi Frequency value (cycles per unit time).
gi Damping value in the units specified.

At resonance, the three types of damping are related by the following equations:

Main Index
 CHAPTER 6 159
Transient Response Analysis

bi Gi
ζ i = ------- = ------
b cr 2

b cr = 2m i ω i Eq. 6-23

1 1
Q i = -------- = ------
2ζ i Gi

The values of fi (units = cycles per unit time) and gi define pairs of frequencies and dampings.
Note that gi can be entered as structural damping (default), critical damping, or quality factor.
The entered damping is internally converted to structural damping using Eq. 6-23. Straight-line
interpolation is used for modal frequencies between consecutive fi values. Linear extrapolation
is used at the ends of the table. ENDT ends the table input.
For example, if modal damping is entered using Table 6-1 and if modes exist at 1.0, 2.5, 3.6, and
5.5 Hz, MSC.Nastran interpolates and extrapolates as shown in Figure 6-2 and the table. Note
that there is no table entry at 1.0 Hz; MSC.Nastran uses the first two table entries at f = 2.0 and
f = 3.0 to extrapolate the value for f = 1.0 .

0.20
+
+
+ +
0.10
+= entered value
= computed value

0.0
2.0 4.0 6.0
f(Hz)

Figure 6-2 Example TABDMP1

Main Index
160

Table 6-1 Example TABDMP1 Interpolation/Extrapolation

Entered Computed
f z f z
2.0 0.16 1.0 0.14
3.0 0.18 2.5 0.17
4.0 0.13 3.6 0.15
6.0 0.13 5.5 0.13

1 2 3 4 5 6 7 8 9 10
TABDMP1 10 CRIT +TAB1

+TAB1 2.0 0.16 3.0 0.18 4.0 0.13 6.0 0.13 +TAB2

+TAB2 ENDT

With the modal equations in the form of Eq. 6-22, an efficient uncoupled analytical integration
algorithm is used to solve for modal response as decoupled SDOF systems. Each of the modal
responses is computed using

0, No initial conditions for

modal transient response.

·
– bt ⁄ 2m  ξ o + b ⁄ ( 2m ) ξ o 
ξ(t) = e  ξ o cos ω d t + ----------------------------------------- sin ω d t Eq. 6-24
 ω d 
t
– bt ⁄ 2m 1 bτ ⁄ 2m
+e ----------- ∫ e p ( τ ) sin ω d ( t – τ )dτ
mω d
0

In a modal transient analysis, you may add nonmodal damping (CVISC, CDAMPi, GE on the
MATi entry, or PARAM,G). With nonmodal damping, there is a computational penalty due to
the coupled [ B ] matrix, causing the coupled solution algorithm to be used. In modal transient
response analysis, it is recommended that you use only modal damping (TABDMP1). If discrete
damping is desired, direct transient response analysis is recommended.

Note that there are no nonzero initial conditions for modal transient response analysis.

Mode Truncation in Modal Transient Response Analysis

It is possible that not all of the computed modes are required in the transient response solution.
Often, only the lowest few suffice for dynamic response calculation. It is quite common to
evaluate the frequency content of transient loads and determine a frequency above which no
modes are noticeably excited. This frequency is called the cutoff frequency. The act of
specifically not using all of the modes of a system in the solution is termed mode truncation.

Main Index
 CHAPTER 6 161
Transient Response Analysis

Mode truncation assumes that an accurate solution can be obtained using a reduced set of
modes. The number of modes used in a solution is controlled in a modal transient response
analysis through a number of methods.
The frequency range selected on the eigenvalue entry (EIGRL or EIGR) is one means to control
the frequency range used in the transient response solution. Also, three parameters are available
to limit the number of modes included in the solution. PARAM,LFREQ gives the lower limit on
the frequency range of retained modes, and PARAM,HFREQ gives the upper limit on the
frequency range of retained modes. PARAM,LMODES gives the number of the lowest modes
to be retained. These parameters can be used to include the desired set of modes. Note that the
default is for all computed modes to be retained.
It is very important to remember that truncating modes in a particular frequency range may
truncate a significant portion of the behavior in that frequency range. Typically, high-frequency
modes are truncated because they are more costly to compute. So, truncating high-frequency
modes truncates high frequency response. In most cases, high-frequency mode truncation is not
of concern. You should evaluate the truncation in terms of the loading frequency and the
important characteristic frequencies of the structure.

Dynamic Data Recovery in Modal Transient Response Analysis

In modal transient response analysis, two options are available for recovering displacements
and stresses: mode displacement method and matrix method. Both methods give the same
answers, although with cost differences.
The mode displacement method computes the total physical displacements for each time step
from the modal displacements and then computes element stresses from the total physical
displacements. The number of operations is proportional to the number of time steps.
The matrix method computes displacements per mode and element stresses per mode, and then
computes physical displacements and element stresses as the summation of modal
displacements and element stresses. Costly operations are proportional to the number of modes.
Since the number of modes is usually much less that the number of time steps, the matrix
method is usually more efficient and is the default. The mode displacement method can be
selected by using PARAM,DDRMM,-1 in the Bulk Data. The mode displacement method is
required when “time frozen” deformed structure plots are requested (see “Plotted Output” on
page 241).

The mode acceleration method (“Advanced Dynamic Analysis Capabilities” on page 311) is
another form of data recovery for modal transient response analysis. This method can provide
better accuracy since detailed local stresses and forces are subject to mode truncation and may
not be as accurate as the results computed with the direct method.

Main Index
162

6.4 Modal Versus Direct Transient Response

Some general guidelines can be used in selecting modal transient response analysis versus direct
transient response analysis. These guidelines are summarized in Table 6-2.

Table 6-2 Modal Versus Direct Transient Response

Modal Direct
Small Model X
Large Model X
Few Time Steps X
Many Time Steps X
High Frequency Excitation X
Normal Damping X
Higher Accuracy X
Initial Conditions X

In general, larger models may be solved more efficiently in modal transient response because the
numerical solution is a solution of a smaller system of uncoupled equations. This result is
certainly true if the natural frequencies and mode shape were computed during a previous stage
of the analysis. Using Duhamel’s integral to solve the uncoupled equations is very efficient even
for very long duration transients. On the other hand, the major portion of the effort in a modal
transient response analysis is the calculation of the modes. For large systems with a large
number of modes, this operation can be as costly as direct integration. This is especially true for
high-frequency excitation. To capture high frequency response in a modal solution, less accurate
high-frequency modes must be computed. For small models with a few time steps, the direct
method may be the most efficient because it solves the equations without first computing the
modes. The direct method is more accurate than the modal method because the direct method
is not concerned with mode truncation. For systems with initial conditions, direct transient
response is the only choice.
Table 6-2 provides a starting place for evaluating which method to use. Many additional factors
may be involved in the choice of a method, such as contractual obligations or local standards of
practice.

Main Index
 CHAPTER 6 163
Transient Response Analysis

6.5 Transient Excitation Definition

An important aspect of a transient response analysis is the definition of the loading function. In
a transient response analysis, the force must be defined as a function of time. Forces are defined
in the same manner whether the direct or modal method is used.
The following Bulk Data entries are used for the transient load definition:

TLOAD1 Tabular input

TLOAD2 Analytical function
DAREA Spatial distribution of dynamic load
TABLEDi Tabular values versus time
LSEQ Generates the spatial distribution of dynamic loads from static load
entries
DLOAD Combines dynamic load sets
DELAY Time delay
The particular entry chosen for defining the dynamic loading is largely a function of user
convenience for concentrated loads. Pressure and distributed loads, however, require a more
complicated format.

There are two important aspects of dynamic load definition. First, the location of the loading on
the structure must be defined. Since this characteristic locates the loading in space, it is called
the spatial distribution of the dynamic loading. Secondly, the time variation in the loading is the
characteristic that differentiates a dynamic load from a static load. This time variation is called
the temporal distribution of the load. A complete dynamic loading is a product of spatial and
temporal distributions.
Using Table IDs and Set IDs in MSC.Nastran makes it possible to apply many complicated and
temporally similar loadings with a minimum of input. Combining simple loadings to create
complicated loading distributions that vary in position as well as time is also a straightforward
task.
The remainder of this section describes the Bulk Data entries for transient excitation. The
description is given in terms of the coefficients that define the dynamic load. See Appendix F
for more complete Bulk Data descriptions.

Time-Dependent Loads -- TLOAD1 Entry

The TLOAD1 entry is the most general form in which to define a time-dependent load. It defines
a dynamic loading of the form

{P(t)} = {A ⋅ F(t – τ)} Eq. 6-25

Main Index
164

The coefficients of the force are defined in tabular format. You need not explicitly define a force
at every instant in time for which the transient solution is evaluated. Only those values which
describe the character of the loading are required. MSC.Nastran interpolates linearly for
intermediate values.

1 2 3 4 5 6 7 8 9 10
TLOAD1 SID DAREA DELAY TYPE TID

Field Contents
SID Set ID defined by a DLOAD Case Control command or a DLOAD Bulk Data
entry.
DAREA Identification number of DAREA entry that defines A. (Integer > 0)
DELAY Identification number of DAREA entry that defines τ . (Integer > 0)
TYPE Excitation function as defined below. Additional information is in “Enforced
Motion” on page 195.

Main Index
 CHAPTER 6 165
Transient Response Analysis

Integer Excitation Function

0 or blank Force or Moment
1 Enforced Displacement
2 Enforced Velocity
3 Enforced Acceleration

Values 1, 2, and 3 apply only to the large mass method for enforced
motion.
TID TABLEDi entry that defines F(t).

Time-Dependent Loads – TLOAD2 Entry

The TLOAD2 entry is a general analytical form with which to define a time-dependent load. The
value of the force at a particular instant in time is determined by evaluating the analytic function
at the specific time. You enter the appropriate constants in the function.
The TLOAD2 entry defines dynamic excitation in the form:


 0 , t < ( T1 + τ ) or t > ( T2 + τ )
{ P( t) } =  Eq. 6-26
 A t̃ B e C t̃ cos ( 2πF t̃ + P ) , ( T1 + τ ) ≤ t ≤ ( T2 + τ )


where: t̃ = t – ( T1 – τ ) .

1 2 3 4 5 6 7 8 9 10
TLOAD2 SID DAREA DELAY TYPE T1 T2 F P

C B

Field Contents
SID Set ID defined by a DLOAD Case Control command.
DAREA Identification number of DAREA entry that defines A. (Integer > 0)
DELAY Identification number of DAREA entry that defines τ . (Integer > 0)
TYPE Defined as on the TLOAD1 entry.
T1, T2 Time constants.
F Frequency (cycles per unit time).
P Phase angle (degrees).
C Exponential coefficient.
B Growth coefficient.

Main Index
166

Spatial Distribution of Loading – DAREA Entry

The DAREA entry defines the degrees-of-freedom where the dynamic load is to be applied and
a scale factor to be applied to the loading. The DAREA entry provides the basic spatial
distribution of the dynamic loading.

1 2 3 4 5 6 7 8 9 10
DAREA SID P1 C1 A1 P2 C2 A2

Field Contents
SID Set ID specified by TLOADi entries.
Pi Grid, extra, or scalar point ID.
Ci Component number.
Ai Scale factor.

A DAREA entry is selected by the TLOAD1 or TLOAD2 entry. Any number of DAREA entries
may be used; all those with the same SID are combined.

Time Delay -- DELAY Entry

The DELAY entry defines the time delay τ in an applied load.

1 2 3 4 5 6 7 8 9 10
DELAY SID P1 C1 τ1 P2 C2 τ2

Field Contents
SID Set ID specified by TLOADi entry.
Pi Grid, extra, or scalar point ID.
Ci Component number.
τi Time delay for Pi, Ci.

A DAREA entry must be defined for the same point and component.
Any number of DELAY entries may be used; all those with the same SID are combined.

Dynamic Load Tabular Function – TABLEDi Entries

The TABLEDi entries (i = 1 through 4) each define a tabular function for use in generating
frequency-dependent dynamic loads. The form of each TABLEDi entry varies slightly,
depending on the value of i, as does the algorithm for y(x). The x values need not be evenly
spaced.

Main Index
 CHAPTER 6 167
Transient Response Analysis

The TABLED1, TABLED2, and TABLED3 entries linearly interpolate between the end points
and linearly extrapolate outside of the endpoints as shown in Figure 6-3. In Version 69 the
TABLED1 added the option to perform logrithmic interpolation between points, also. The
TABLED4 entry uses the endpoint values for values beyond the endpoints.

Linear Interpolation Between Endpoints

y

Linear Extrapolation of Segment x6 – x7

Linear Extrapolation of Segment x2 – x1

x
x1 x2 x3 x4 x5 x6 x7

Figure 6-3 Interpolation and Extrapolation for TABLED1, TABLED2,

and TABLED3 Entries

The TABLED1 entry has the following format:

1 2 3 4 5 6 7 8 9 10
TABLED1 TID XAXIS YAXIS

x1 y1 x2 y2 x3 y3 -etc.- ENDT

Field Contents
TID Table identification number.
XAXIS Specifies a linear or logarithmic interpolation for the x-axis.
(Character: “LINEAR” or “LOG”; Default = “LINEAR”)
YAXIS Specifies a linear or logarithmic interpolation for the y-axis.
(Character: “LINEAR” or “LOG”; Sefault = “LINEAR”)
xi, yi Tabular values. Values of x are frequency in cycles per unit time.
ENDT Ends the table input.

The TABLED1 entry uses the algorithm

y = yT ( x ) Eq. 6-27

The algorithms used for interpolation and extrapolation are as follows:

XAXIS YAXIS yT ( x )

xj – x x – xi
LINEAR LINEAR ----------------- yi + ----------------- yj
xj – xi xj – xi

ln ( xj ⁄ x ) ln ( x ⁄ xi )
LOG LINEAR --------------------------- yi + --------------------------- yj
ln ( xj ⁄ xi ) ln ( xj ⁄ xi )
Main Index
168

XAXIS YAXIS yT ( x )

xj – x x – xi
LINEAR LOG exp ----------------- ln yi + ----------------- ln yj
xj – xi xj – xi

ln ( xj ⁄ x ) ln ( x ⁄ xi )
LOG LOG exp --------------------------- ln yi + --------------------------- ln yj
ln ( xj ⁄ xi ) ln ( xj ⁄ xi )

The TABLED2 entry has the following format:

1 2 3 4 5 6 7 8 9 10
TABLED2 TID X1

x1 y1 x2 y2 x3 y3 -etc.- ENDT

Field Contents
TID Table identification number.
X1 Table parameter.
xi, yi Tabular values.
ENDT Ends the table input.

The TABLED2 entry uses the algorithm

y = y T ( x – X1 ) Eq. 6-28

The TABLED3 entry has the following format:

1 2 3 4 5 6 7 8 9 10
TABLED3 TID X1 X2

x1 y1 x2 y2 x3 y3 -etc.- ENDT

Field Contents
TID Table identification number.
X1, X2 Table parameters.
xi, yi Tabular values.
ENDT Ends the table input.

The TABLED3 entry uses the algorithm

y = y T  -----------------
x – X1
Eq. 6-29
X2

The TABLED4 entry has the following format:

Main Index
 CHAPTER 6 169
Transient Response Analysis

1 2 3 4 5 6 7 8 9 10
TABLED4 TID X1 X2 X3 X4

A0 A1 A2 A3 A4 A5 -etc.- ENDT

Field Contents
TID Table identification number.
Xi Table parameters. ( X2 ≠ 0.0; X3 < X4 ) .
Ai Coefficients.

The TABLED4 entry uses the algorithm

N
x – X1 i
y = ∑ Ai  ----------------- Eq. 6-30
 X2 
i = 0

N is the degree of the power series. When x < X3 , X3 is used for x ; when x > X4 , X4 is used for x .
This condition has the effect of placing bounds on the table; there is no extrapolation outside of
the table boundaries.
ENDT ends the table input.

DAREA Example
Suppose the following command is in the Case Control Section:
DLOAD = 35
in addition to the following entries in the Bulk Data Section:

1 2 3 4 5 6 7 8 9 10
$TLOAD1 SID DAREA DELAY TYPE TID

TLOAD1 35 29 31 40

$DAREA SID POINT COMPONENT SCALE

DAREA 29 30 1 4.0

$DELAY SID POINT COMPONENT LAG

DELAY 31 30 1 0.2

$TABLED1 ID

$ X1 y1 x2 y2 x3 y3 x4 y4

TABLED1 40 LINEAR LINEAR

0.0 0.0 0.3 1.0 2.0 1.0 ENDT

Main Index
170

The DLOAD Set ID 35 in the Case Control selects the TLOAD1 entry in the Bulk Data having a
Set ID 35. On the TLOAD1 entry is a reference to DAREA Set ID 29, DELAY Set ID 31, and
TABLED1 Set ID 40. The DAREA entry with Set ID 29 positions the loading on grid point 30 in
the 1 direction with a scale factor of 4.0 applied to the load. The DELAY entry with Set ID 31
delays the loading on grid point 30 in the 1 direction by 0.2 units of time. The TABLED1 entry
with Set ID 40 defines the load time history in tabular form. The result of these entries is a
dynamic load applied to grid point 30, component T1, scaled by 4.0 and delayed by 0.2 units of
time.
Figure 6-4 shows the TABLED1 time history and the applied load (scaled by the DAREA entry
and time shifted by the DELAY entry).

1.0

Amplitude
on
TABLED1

0
0.5 1.0 1.5 2.0 2.5
Time (sec)

4.0
Force
Amplitude

0
0.5 1.0 1.5 2.0 2.5
Time (sec)

Figure 6-4 Time History from the TABLED1 Entry (Top) and Applied Load (Bottom)

Static Load Sets -- LSEQ Entry

MSC.Nastran does not have specific data entries for many types of dynamic loads. Only
concentrated forces and moments can be specified directly using DAREA entries. To
accommodate more complicated loadings conveniently, the LSEQ entry is used to define static
load entries that define the spatial distribution of dynamic loads.

1 2 3 4 5 6 7 8 9 10
LSEQ SID DAREA ID LOAD ID TEMP ID

Main Index
 CHAPTER 6 171
Transient Response Analysis

The LSEQ Bulk Data entry contains a reference to a DAREA Set ID and a static Load Set ID. The
static loads are combined with any DAREA entry in the referenced set. The DAREA Set ID does
not need to be defined with a DAREA Bulk Data entry. The DAREA Set ID is referenced by a
TLOADi entry. This reference defines the temporal distribution of the dynamic loading. The
Load Set ID may refer to one or more static load entries (FORCE, PLOADi, GRAV, etc.). All static
loads with the Set ID referenced on the LSEQ entry define the spatial distribution of the dynamic
loading. MSC.Nastran converts this information to equivalent dynamic loading.
Figure 6-5 demonstrates the relationships of these entries. To activate a load set defined in this
manner, the DLOAD Case Control command refers to the Set ID of the selected DLOAD or
TLOADi entry, and the LOADSET Case Control command refers to the Set ID of the selected
LSEQ entries. The LSEQ entries point to the static loading entries that are used to define
dynamic loadings with DAREA Set ID references. Together this relationship defines a complete
dynamic loading. To apply dynamic loadings in this manner, the DLOAD and LOADSET Case
Control commands and the TLOADi and LSEQ Bulk Data entries must be defined. A DAREA
Bulk Data entry does not need to be defined since the TLOADi and LSEQ entries reference a
common DAREA ID. The LSEQ entry can also be interpreted as an internal DAREA entry
generator for static load entries.

DLOAD LOADSET

Case Control
Bulk Data
TLOADi LSEQ

Dynamic Load DAREA Static Load Entries

(Temporal (Reference (Spatial
Distribution) Link) Distribution)
Figure 6-5 Relationship of Dynamic and Static Load Entries

LSEQ Example
Suppose the following commands are in the Case Control Section:

LOADSET = 27
DLOAD = 25
in addition to the following entries in the Bulk Data Section:

1 2 3 4 5 6 7 8 9 10
$LSEQ SID DAREA LID

LSEQ 27 28 26

$TLOAD1 SID DAREA DELAY TYPE TID

TLOAD1 25 28 29
DAREA
Main Index Reference
172

$STATIC SID
LOAD

PLOAD1 26 -etc.-

FORCE 26 -etc.-

$TABLED1 TID

TABLED1 29 -etc.-

In the above, the LOADSET request in Case Control selects the LSEQ Set ID 27 entry. The
DLOAD request in Case Control selects the TLOAD1 Set ID 25 entry. This TLOAD1 entry refers
to a TABLED1 ID 29, which is used to define the temporal variation in the loading. DAREA
Set ID 28 links the LSEQ and TLOAD1 entries. In addition, the LSEQ entry refers to static Load
Set ID 26, which is defined by FORCE and PLOAD1 entries. The FORCE and PLOAD1 entries
define the spatial distribution of the dynamic loading and through the DAREA link refer to the
TLOAD1/TABLED1 combination for the time-varying characteristics of the load.

Dynamic Load Set Combination -- DLOAD

One of the requirements of transient loads is that all TLOAD1s and TLOAD2s must have unique
SIDs. If they are to be applied in the same analysis, they must be combined using the DLOAD
Bulk Data entry. The total applied load is constructed from a combination of component load
sets as follows:

{ P } = S ∑ Si { Pi } Eq. 6-31
i

where:

S = overall scale factor

S i = scale factor for i-th load set

P i = i-th set of loads

P = total applied load

The DLOAD Bulk Data entry has the following format:

1 2 3 4 5 6 7 8 9 10
DLOAD SID S S1 L1 S2 L2 -etc.-

Field Contents
SID Load set ID.
S Overall scale factor.

Main Index
 CHAPTER 6 173
Transient Response Analysis

Field Contents
Si Individual scale factors.
Li Load set ID number for TLOAD1 and TLOAD2 entries.

As an example, in the following DLOAD entry:

1 2 3 4 5 6 7 8 9 10
$DLOAD SID S S1 L1 S2 L2 -etc.-

DLOAD 33 3.25 0.5 14 2.0 27

A dynamic Load Set ID of 33 is created by taking 0.5 times the loads in Load Set ID of 14, adding
to it 2.0 times the loads in Load Set ID of 27, and multiplying that sum by an overall scale factor
of 3.25.

As with other transient loads, a dynamic load combination defined by the DLOAD Bulk Data
entry is selected by the DLOAD Case Control command.

Main Index
174

6.6 Integration Time Step

The TSTEP Bulk Data entry is used to select the integration time step for direct and modal
transient response analysis. This entry also controls the duration of the solution and which time
steps are to be output. The TSTEP Bulk Data entry is selected by the Set ID referenced on the
TSTEP Case Control command.
The integration time step must be small enough to represent accurately the variation in the
loading. The integration time step must also be small enough to represent the maximum
frequency of interest. The maximum frequency of interest is often called the cut-off frequency.
It is recommended to use at least ten solution time steps per period of response for the cutoff
frequency. For a given integration time step, integration errors increase with increasing natural
frequency because there is an upper limit to the frequency that can be represented by a given
time step. Also, integration errors accumulate with total time.
In both direct and modal transient analysis, the cost of integration is directly proportional to the
number of time steps. For example, doubling the load duration doubles the integration effort.
In specifying the duration of the analysis on the TSTEP entry, it is important to use an adequate
length of time to properly capture long period (low frequency) response. In many cases, the
peak dynamic response does not occur at the peak value of load nor necessarily during the
duration of the loading function. A good rule is: always solve for at least one cycle of response
for the lowest frequency mode after the peak excitation.

You may change ∆t during a run but doing so causes the dynamic matrix to be redecomposed,
which can be costly in direct transient response analysis.
The TSTEP Bulk Data entry has the following format:

1 2 3 4 5 6 7 8 9 10
TSTEP SID N1 ∆t 1 NO1

N2 ∆t 2 NO2

-etc.-

Field Contents
SID Set ID specified by a TSTEP Case Control command.
Ni Number of time steps of value ∆t i .

∆t i Integration time step.

NOi Output every NOi-th time step.

Main Index
 CHAPTER 6 175
Transient Response Analysis

6.7 Transient Excitation Considerations

A number of important considerations must be remembered when applying transient loads.
The averaging of applied loads (Eq. 6-3) in the integration smooths the force and decreases the
apparent frequency content. Very sharp spikes in a loading function induce a high-frequency
transient response. If the high-frequency transient response is of primary importance in an
analysis, a very small integration time step must be used.

It is also important to avoid defining discontinuous forcing functions when describing applied
loads. The numerical integration of discontinuous forcing functions may cause different results
for the same analysis run on different computers because of slight numerical differences on
different computer types. If the analysis calls for loadings with sharp impulses, it is best to
smooth the impulse over at least one integration time increment.
The loading function must accurately describe the spatial and temporal distribution of the
dynamic load. Simplifying assumptions must not change the character of the load in magnitude,
location, or frequency content.

Main Index
176

6.8 Solution Control for Transient Response Analysis

The following tables summarize the data entries that can be used to control a transient response
analysis. Certain data entries are required, some data entries are optional, while others are user
selectable.
In the Executive Control Section of the MSC.Nastran input file, a solution must be selected using
the SOL i statement where i is an integer value chosen from Table 6-3.

Table 6-3 Transient Response Solutions in MSC.Nastran

Structured
Rigid
Solution
Formats
Sequences
Direct 9 109
Modal 12 112

We have applied these solutions in the examples that follow.

In the Case Control Section of the MSC.Nastran input file, you must select the solution
parameters associated with the current analysis (i.e., time steps, loads, and boundary conditions)
and also the output quantities required from the analysis. The Case Control commands directly
related to transient response analysis are listed in Table 6-4. They can be combined in the
standard fashion with the more generic commands, such as SPC, MPC, etc.

Table 6-4 Transient Response Case Control Commands

Case Direct
Required/
Control or Description
Optional
Commands Modal
DLOAD Both Select the dynamic load set from the Bulk Data Required*
TSTEP Both Select the TSTEP entry from the Bulk Data Required
METHOD Modal Select the eigenvalue extraction parameters Required
LOADSET Both Select the LSEQ set from the Bulk Data Optional
SDAMPING Modal Select the modal damping table from the Bulk Data Optional
IC Direct Select TIC entries for initial conditions from the Optional
Bulk Data
OTIME Both Select the times for output (default = all) Optional
*
Not required when using initial conditions.

The types of results available from a transient response analysis are similar to those for a static
analysis except that the results vary with time. Additional quantities are also available, which
is characteristic of dynamics problems. The output quantities are summarized in Table 6-5 and
Table 6-6.
Main Index
 CHAPTER 6 177
Transient Response Analysis

Table 6-5 Grid Point Output from a Transient Response Analysis

Case Control
Description
Command
ACCELERATION Grid point acceleration time history for a set of grid points
DISPLACEMENT Grid point displacement time history for a set of grid points
(or VECTOR)
GPSTRESS Grid point stress time history (requires SURFACE/VOLUME
definition in the OUTPUT(POST) section of the Case Control)
OLOAD Requests applied load table to be output for a set of grid points
SACCELERATION Requests solution set acceleration output: d-set in direct solutions
and modal variables in modal solutions
SDISPLACEMENT Requests solution set displacement output: d-set in direct
solutions and modal variables in modal solutions
SVECTOR Requests real eigenvector output for the a-set in modal solutions
SVELOCITY Requests solution set velocity output: d-set in direct solutions and
modal variables in modal solutions
SPCFORCES Requests forces of single-point constraint for a set of grid points
VELOCITY Grid point velocity time history for a set of grid points

Table 6-6 Element Output from a Transient Analysis

Case Control
Description
Command
ELSTRESS (or STRESS) Element stress time history for a set of elements
ELFORCE (or FORCE) Element force time history for a set of elements
STRAIN Element strain time history for a set of elements

Because the results may be output for many time steps, the volume of output can be very large.
Prudent selection of the output quantities is recommended.
A number of Bulk Data entries are unique to transient response analysis. They can be combined
with other generic entries in the Bulk Data. Bulk Data entries directly related to transient
analysis are summarized in Table 6-7.

Table 6-7 Bulk Data Entries for Transient Response Analysis

Bulk Data Direct or Required/

Description
Entry Modal Optional
TSTEP Both Integration time step and solution control Required
TLOADi Both Dynamic loading Required*
Main Index
178

Table 6-7 Bulk Data Entries for Transient Response Analysis

Bulk Data Direct or Required/

Description
Entry Modal Optional
EIGR or EIGRL Modal Eigenvalue analysis parameters Required
LSEQ Both Dynamic loading from static loads Optional
TABLEDi Both Time-dependent tables for TLOADi Optional*
TIC Direct Initial conditions on grid, scalar, and Optional
extra points
DAREA Both Load component and scale factor Required†
DELAY Both Time delay on dynamic load Optional
DLOAD Both Dynamic load combination, required if Optional
TLOAD1 and TLOAD2 are used
TABDMP1 Modal Modal damping table Optional
*
Not required for initial conditions.
†
The DAREA ID is required; the DAREA Bulk Data entry is not required if an LSEQ
entry is used.

Main Index
 CHAPTER 6 179
Transient Response Analysis

6.9 Examples
This section provides several examples showing the input and output. These examples are

Transient Response
Model Output
Bulk Data Entries
bd06two TSTEP, TIC X-Y plots
bd06bar EIGRL, TSTEP, X-Y plots
TABDMP1, DLOAD,
TLOAD2, DAREA,
DELAY
bd06bkt EIGRL, TSTEP, X-Y plot
TABDMP1, TLOAD1,
LSEQ, TABLED1,
PLOAD4

These examples are described in the sections that follow.

Two-DOF Model
Consider the two-DOF system shown in Figure 6-6. Direct transient response (SOL 109) is run
with an initial displacement of 0.1 meter at grid point 2. The analysis is run for a duration of
10 seconds with a ∆t of 0.01 second. Damping is neglected in the analysis. Listing 6-1 shows
part of the input file.

Grid Point 1 m1

k1

Grid Point 2 m2

k2

Figure 6-6 Two-DOF Model

Main Index
180

Listing 6-1 Input File (Abridged) for the Two-DOF Example

$ FILE bd06two.dat
$
$ TWO-DOF SYSTEM
$ CHAPTER 6, TRANSIENT RESPONSE
$
TIME 5
SOL 109 $ DIRECT TRANSIENT RESPONSE
CEND
TITLE = TWO-DOF SYSTEM
SUBTITLE = DIRECT FREQUENCY RESPONSE
LABEL = INITIAL DISPL. AT GRID 2
$
$ SPECIFY SPC
SPC = 996
$
$ SPECIFY DYNAMIC INPUT
TSTEP = 888
IC = 777
$
$ SELECT OUTPUT
SET 11 = 1,2
DISPLACEMENT(PLOT) = 11
$
$ XYPLOTS
$
... X-Y plot commands ...
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$ ENTRIES FOR TRANSIENT RESPONSE
$
$ INITIAL CONDITION
$TIC SID G C U0 V0
TIC 777 2 2 0.1
$
$ TIME STEP
$TSTEP SID N1 DT1 NO1
TSTEP 888 1000 0.01 1
$
... basic model ...
$
ENDDATA

Table 6-8 shows the relationship between the Case Control commands and the Bulk Data
entries. This example represents the simplest form of dynamic response input. The only
required entries are those that define the time step and the initial conditions. Note that the
unspecified initial conditions are assumed to be zero. Note, too, that the initial conditions are
available only for direct transient response analysis.

Table 6-8 Relationship Between Case Control Commands

and Bulk Data Entries for the Two-DOF Model

Case Control Bulk Data

TSTEP TSTEP
IC TIC
Main Index
 CHAPTER 6 181
Transient Response Analysis

Figure 6-7 shows the plots of the resulting displacements for grid points 1 and 2. Note that there
are two frequencies of response: a higher frequency of about 5 Hz, and a lower frequency of
about 0.25 Hz. The energy (and hence response) appears to be transferred repetitively between
grid points 1 and 2 as represented by the lower frequency response. This energy transfer is
called beating. Beating occurs when there are closely-spaced modes (in this case, 4.79 Hz and
5.29 Hz) in which energy transfer can readily occur. The response is comprised of two
frequencies as given below:

f1 + f2
f higher = ------------------
2
Eq. 6-32
f2 – f1
f lower = -----------------
2

where:
f1 = lower of the closely-spaced mode frequencies
f2 = higher of the closely-spaced mode frequencies

In this example, f higher is 5.04 Hz and f lower is 0.25 Hz. The lower frequency is called the beat
frequency and is the frequency at which energy transfer occurs.

Main Index
182

Figure 6-7 Displacements of Grid Points 1 and 2

Cantilever Beam Model

Consider the cantilever beam shown in Figure 6-8. This beam model is the same as in
“Examples” in Chapter 5. Modal transient response (SOL 112) is run with loads applied to grid
points 6 and 11 as shown in Figure 6-9. The analysis is run for a duration of 2 seconds with a ∆t
of 0.001 second. Modal damping of 5% critical damping is used for all modes. Modes up to 3000
Hz are computed using the Lanczos method. Listing 6-2 shows part of the input file.

Main Index
 CHAPTER 6 183
Transient Response Analysis

6 11
x

p6(t) p11(t)

Figure 6-8 Cantilever Beam Model with Applied Loads

Force 6 (N)
Time
(sec)
-3
0.1 0.5

Force 11 (N)
Time
(sec)
-6
0.5
Figure 6-9 Applied Loads for the Beam Model

Main Index
184

Listing 6-2 Input File (Abridged) for the Beam Example

$ FILE bd06bar.dat
$
$ CANTILEVER BEAM MODEL
$ CHAPTER 6, TRANSIENT RESPONSE
$
SOL 112 $ MODAL TRANSIENT RESPONSE
TIME 10
CEND
TITLE = CANTILEVER BEAM
SUBTITLE = MODAL TRANSIENT RESPONSE
$
SPC = 21
DLOAD = 22
TSTEP = 27
SDAMPING = 25
$
METHOD = 10
$
$ PHYSICAL OUTPUT REQUEST
SET 11 = 6,11
DISPLACEMENT(PLOT) = 11
ACCELERATION(PLOT) = 11
$
$ MODAL SOLUTION SET OUTPUT
SET 12 = 1,2
SDISP(PLOT) = 12
$
$ ELEMENT FORCE OUTPUT
SET 13 = 6
ELFORCE(PLOT) = 13
$
$ APPLIED LOAD OUTPUT
SET 15 = 6,11
OLOAD(PLOT) = 15
$
$ XYPLOTS
$
... X-Y plot commands ...
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$EIGRL SID V1 V2 MSGLVL
EIGRL 10 -0.1 3000. 0
$
$TSTEP SID N1 DT1 NO1
TSTEP 27 2000 0.001 1
$
$ MODAL DAMPING OF 5% IN ALL MODES
$TABDMP1 TID TYPE +TABD
$+TABD F1 G1 F2 G2 ETC.
TABDMP1 25 CRIT +TABD
+TABD 0. 0.05 1000. 0.05 ENDT
$
$ DYNAMIC LOADING
$DLOAD SID S S1 L1 S2 L2
DLOAD 22 1.0 1.0 231 1.0 232
$TLOAD2 SID DAREA DELAY TYPE T1 T2 F P +TL1
$+TL1 C B
TLOAD2 231 241 0 0.0 0.5 2.0 90.
TLOAD2 232 242 262 0 0.0 0.5 4.0 90.
$DAREA SID P1 C1 A1
DAREA 241 11 2 6.0
DAREA 242 6 2 3.0
$DELAY SID P1 C1 T1
DELAY 262 6 2 0.1
$
... basic model ...
$
ENDDATA

Main Index
 CHAPTER 6 185
Transient Response Analysis

Table 6-9 shows the relationship between the Case Control commands and the Bulk Data
entries. The DLOAD Bulk Data entry references two TLOAD2 entries, each of which references
separate DAREA entries. A TLOAD2 entry also references a DELAY entry to apply the time
delay to the load at grid point 6.

Table 6-9 Relationship Between Case Control Commands

and Bulk Data Entriesfor the Bar Model

Case Control Bulk Data

METHOD EIGRL
TSTEP TSTEP
SDAMPING TABDMP1
DLOAD DLOAD
TLOAD2 231

DAREA
TLOAD2 232

DAREA
DELAY

Plotted output is shown in the following figures. Figure 6-10 shows the applied loads at grid
points 6 and 11. Figure 6-11 shows the plots of the displacements for grid points 6 and 11.
Figure 6-12 shows the accelerations for grid points 6 and 11. Figure 6-13 shows the bending
moment at end A in plane 1 for element 6. Figure 6-14 shows the modal displacements for
modes 1 and 2.

Main Index
186

3.0

Force 6 (N)
0.0

-3.0

6.0
Force 11 (N)

0.0

-6.0
0.0 2.0
Time (sec)
Figure 6-10 Applied Loads at Grid Points 6 and 11

Main Index
 CHAPTER 6 187
Transient Response Analysis

0.03

Displacement 6 (m)
0.0

-0.02

0.08
Displacement 11 (m)

0.0

-0.06
0.0 2.0
Time (sec)
Figure 6-11 Displacements at Grid Points 6 and 11

Main Index
188

4.0

Acceleration 6 ( m 2 ⁄ sec )
0.0

-6.0

10.0
Acceleration 11 ( m 2 ⁄ sec )

0.0

-15.0
0.0 2.0
Time (sec)
Figure 6-12 Accelerations at Grid Points 6 and 11

Main Index
 CHAPTER 6 189
Transient Response Analysis

25.0

Bending Moment A1 (Nm) 0.0

-20.0
0.0 2.0
Time (sec)
Figure 6-13 Bending Moment A1 for Element 6

Main Index
190

0.08

SDisplacement 1
0.0

-0.10

0.0015
SDisplacement 2

0.0

-0.0010
0.0 2.0
Time (sec)
Figure 6-14 Modal Displacements for Modes 1 and 2

Bracket Model
Consider the bracket model shown in Figure 6-15. A pressure load of 3 psi is applied to the
elements in the top face in the z-direction with the time history shown in Figure 6-16. The modal
transient analysis is run for 4 seconds with a time step size of 0.005 second. Modal damping of
2% critical damping is used for all modes. Modes up to 3000 Hz are computed with the Lanczos
method. The model is constrained near the base.

Main Index
 CHAPTER 6 191
Transient Response Analysis

Figure 6-15 Bracket Model

Pressure

3 psi

0 0.10 0.15 5
Time (sec)
Figure 6-16 Time Variation for Applied Load

Listing 6-3 shows the abridged input file. The LSEQ entry is used to apply the pressure loads
(PLOAD4 entries). Note that the LSEQ and TLOAD1 entries reference a common DAREA ID
(999) and that there is no explicit DAREA entry. Table 6-10 shows the relationship between the
Case Control commands and the Bulk Data entries.

Main Index
192

Listing 6-3 Input File (Abridged) for the Bracket Model

$ FILE bd06bkt.dat
$
$ BRACKET MODEL
$ CHAPTER 6, TRANSIENT RESPONSE
$
SOL 112 $ MODAL TRANSIENT RESPONSE
TIME 100
CEND
TITLE = BRACKET MODEL
SUBTITLE = MODAL TRANSIENT RESPONSE ANALYSIS
$
SPC = 1
$
METHOD = 777
$
DLOAD = 2
LOADSET = 3
SDAMPING = 4
TSTEP = 5
$
$ OUTPUT REQUEST
SET 123 = 999
DISPLACEMENT(PLOT)=123
$
$ XYPLOTS
$
... X-Y plot commands ...
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10......
$
$ NORMAL MODES TO 3000 HZ
$EIGRL SID V1 V2
EIGRL 777 -0.1 3000.
$
$ 4 SECONDS OF RESPONSE
$TSTEP SID N1 DT1 NO1
TSTEP 5 800 0.005 1
$
$ MODAL DAMPING OF 2% CRITICAL
$TABDMP1 TID TYPE +TABD1
$+TABD1 F1 G1 F2 G2 ETC.
TABDMP1 4 CRIT +TABD1
+TABD1 0.0 0.02 3000.0 0.02 ENDT
$
$ LOAD DEFINITION
$
$TLOAD1 SID DAREA DELAY TYPE TID
TLOAD1 2 999 22
$
$LSEQ SID DAREA LID TID
LSEQ 3 999 1
$
$ TIME HISTORY
$TABLED1 TID +TABL1
$+TABL1 X1 Y1 X2 Y2 ETC.
TABLED1 22 +TABL1
+TABL1 0.0 0.0 0.1 0.0 0.15 1.0 5.0 1.0 +TABL2
+TABL2 ENDT
$
$ PRESSURE LOAD OF 3 PSI PER ELEMENT
$PLOAD4 SID EID P1
PLOAD4 1 171 -3.
PLOAD4 1 172 -3.
PLOAD4 1 160 -3.
etc.
$
... basic model ...
$
ENDDATA

Main Index
 CHAPTER 6 193
Transient Response Analysis

Table 6-10 Relationship Between Case Control Commands and

Bulk Data Entries for the Bracket Model

Case Control Bulk Data

METHOD EIGRL
FREQUENCY FREQ1
SDAMPING TABDMP1
LOADSET PLOAD4
LSEQ

(DAREA ID)
DLOAD
RLOAD1
TABLED1

Figure 6-17 shows a plot of the z-displacement of grid point 999, which is the concentrated mass
at the center of the cutout.

0
Displacement (m)

-0.04

-0.08
0 Time (sec) 4
Figure 6-17 Displacement Time History for Grid Point 999

Main Index
194

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Enforced Motion
7
■ Overview
■ The Large Mass Method in Direct Transient and Direct Frequency Response
■ The Large Mass Method in Modal Transient and Modal Frequency Response
■ User Interface for the Large Mass Method

■ Examples

Main Index
196

7.1 Overview
Enforced motion specifies the displacement, velocity, and/or acceleration at a set of grid points
for frequency and transient response. Enforced motion is used when base motion is specified
instead of or in conjunction with applied loads. A common application is an earthquake
excitation applied to a building. In this case there are no applied loads, instead the base of the
building undergoes an enforced displacement or acceleration time history.

MSC.Nastran does not include a completely automatic method for prescribing enforced motion
in dynamics. Instead, the procedures described in “Frequency Response Analysis” on page 103
and “Transient Response Analysis” on page 149 for specifying applied forces are used in
conjunction with techniques that convert applied forces into enforced motion. One such method
that is applicable to both transient response and frequency response is described in this
chapter—the large mass method. Another method for enforced motion, described in the
MSC.Nastran Advanced Dynamic Analysis User’s Guide, is the Lagrange multiplier technique.

Main Index
 CHAPTER 7 197
Enforced Motion

7.2 The Large Mass Method in Direct Transient and Direct

Frequency Response
If a very large mass m o , which is several orders of magnitude larger than the mass of the entire
structure, is connected to a degree-of-freedom and a dynamic load p is applied to the same
degree-of-freedom, then the acceleration of the degree-of-freedom, to a close approximation, is
as follows:

·· 1
u = ------- p Eq. 7-1
mo
··
In other words, the load that produces a desired acceleration u is approximately
··
p = mo u Eq. 7-2

The accuracy of this approximation increases as m o is made larger in comparison to the mass of
the structure. The only limit for the size of m o is numeric overflow in the computer. MSC
recommends that the value of m o be approximately 10 6 times the mass of the entire structure for
an enforced translational degree-of-freedom and 10 6 times the mass moment of inertia of the
entire structure for a rotational DOF. The factor 10 6 is a safe limit that should produce
approximately six digits of numerical accuracy.
The large mass method is implemented in direct transient and frequency response analysis by
placing large masses m o on all enforced degrees-of-freedom and supplying applied dynamic
loads specified by Eq. 7-2; that is, the function u·· is input on entries normally used for the input
of loads, and the scale factor m o can be input on DAREA or DLOAD Bulk Data entries,
whichever is more convenient. CMASSi or CONMi entries should be used to input the large
masses.

Eq. 7-2 is not directly helpful if enforced displacement or enforced velocity is specified rather
than enforced acceleration. However, Eq. 7-2 can be made serviceable in frequency response
analysis by noting that

·· · 2
u = iωu = – ω u Eq. 7-3

so that

·· · 2
p = m o u = iωm o u = – ω m o u Eq. 7-4

The added factor ( iω or – ω 2 ) can be carried by the function tabulated on the TABLEDi entry used
to specify the frequency dependence of the dynamic load.
In the case of transient analysis, provision is made on the TLOAD1 and TLOAD2 entries for you
to indicate whether an enforced displacement, velocity, or acceleration is supplied (TYPE = 1, 2,
or 3). MSC.Nastran then automatically differentiates a specified velocity once or a specified
displacement twice to obtain an acceleration. The remaining required user actions are the same
as for enforced acceleration.

Main Index
198

In summary, the user actions for direct frequency and direct transient response are
• Remove any constraints from the enforced degrees-of-freedom.
• Apply large masses mo with CMASSi or CONMi Bulk Data entries to the DOFs where
the motion is enforced. The magnitude of m o should be approximately 10 6 times the
entire mass of the structure (or approximately 10 6 times the entire mass moment of
inertia of the structure if the component of enforced motion is a rotation).
• In the case of direct frequency response, apply a dynamic load computed according to
Eq. 7-4 to each enforced degree-of-freedom.
• In the case of direct transient response,
• Indicate in field 5 of the TLOAD1 and TLOAD2 entries whether the enforced
motion is a displacement, velocity, or acceleration.
• Apply a dynamic load to each enforced degree-of-freedom equal to mo u , mo u· , or
··
m o u , depending on whether the enforced motion is a displacement, velocity, or
acceleration.
Be careful when using PARAM,WTMASS. The WTMASS parameter multiplies the large mass
value, which changes the effective enforced acceleration to

·· p
u = -------------------------------------- Eq. 7-5
m o ⋅ WTMASS

Enforced velocity and enforced displacement are changed likewise.

You may well ask whether a stiff spring may be used instead of a large mass. In that case the
applied load is

p = ko u Eq. 7-6

where k o is the stiffness of the stiff spring and u is the enforced displacement. The large stiffness
method certainly works, but the large mass method is preferred because it is easier to estimate a
good value for the large mass than to estimate a good value for the stiff spring. In addition and
more importantly, the large mass method is far superior when modal methods are used. If very
stiff springs are used for modal analysis rather than very large masses, the vibration modes
corresponding to the very stiff springs have very high frequencies and in all likelihood, are not
included among the modes used in the response analysis. This is the main reason that large
masses should be used instead of stiff springs.
The stiff spring method is advantageous in the case of enforced displacement because it avoids
the roundoff error that occurs while differentiating the displacement to obtain acceleration in the
large mass method. The stiff spring method also avoids the problem of rigid-body drift when
applying enforced motion on statically determinate support points. (Rigid-body drift means
that the displacement increases continuously with time, which is often caused by the
accumulation of small numerical errors when integrating the equations of motion.)

Main Index
 CHAPTER 7 199
Enforced Motion

7.3 The Large Mass Method in Modal Transient and Modal

Frequency Response
The steps described in the previous section must also be followed when a modal method of
response analysis is used. Also, if the enforced degrees of freedom are not sufficient to suppress
all rigid-body motions, which may be the case for an airplane in flight, additional DOFs that
describe the remaining rigid-body motions can also be entered on the SUPORT entry. Use of the
SUPORT entry is discussed in “Rigid-body Modes” on page 87. Note that the use of the
SUPORT entry is optional.
The rigid-body mode(s) can be removed from consideration either by not computing them or by
using PARAM,LFREQ,r where r is a small positive number (0.001 Hz, for example). If this is
done, the displacements, velocities, and accelerations obtained are relative to the overall motion
of the structure and are not absolute response quantities. Stresses and element forces are the
same as when the rigid-body modes are included because the rigid-body modes do not
contribute to them. Rigid-body modes can be discarded to remove rigid-body drift.
Rigid-body modes occur when the structure is unconstrained and large masses are applied at the
DOFs which, if constrained, result in a statically determinate structure. Redundant constrained
DOFs, which result in a statically indeterminate structure, present a different situation when the
constraints are removed and large masses are applied at those redundant DOFs. In that case,
very low-frequency modes occur but they are not all rigid-body modes; some are modes that
represent the motion of one large mass relative to the others. These very low-frequency,
nonrigid-body modes do contribute to element forces and stresses, and they must be retained in
the solution. In some cases, their frequencies are not necessarily small; they may be only an
order or two in magnitude less than the frequency of the first flexible mode. If PARAM,LFREQ,r
is used to remove the true rigid-body modes, then r must be set below the frequency of the first
relative motion mode.
Consider the two-dimensional clamped bar in Figure 7-1. Assume that each end of the bar is
subjected to the same enforced acceleration time history in the y-direction. One way to model
the bar is to use two large masses (one at each end), which are unconstrained in the y-direction.
This model provides two very low-frequency modes: one that is a rigid-body mode and one that
is not. The second mode contributes to element forces and stresses, and removing its
contribution leads to an error because with two such large masses, one mass can drift over time
relative to the other mass. A better way to model the case of identical inputs at multiple locations
is to use one large mass connected to the end points by an RBE2 element. This model provides
only one rigid-body mode, which can be safely discarded if only the answers relative to the
structure are desired.

Main Index
200

··
u(t) ··
u(t)

x
Basic Model

p(t ) p(t)

Two Large Masses at Each End

RB E 2
E2 RB

p(t)

One Large Mass and One RBE2 Element

Figure 7-1 Clamped-Clamped Bar Undergoing Enforced Acceleration

Main Index
 CHAPTER 7 201
Enforced Motion

7.4 User Interface for the Large Mass Method

There is no special user interface for the large mass method other than to specify a large mass at
excitation DOFs and to specify the large force. For transient response the type of enforced
motion (displacement, velocity, or acceleration) must be specified on the TLOADi entries. The
remainder of the input is identical to that of frequency response (“Frequency Response
Analysis” on page 103) or transient response (“Transient Response Analysis” on page 149)
analysis.
The force applied at a point is the product of terms from the DLOAD, DAREA, and TABLEDi
entries. The scaling of the large force is arbitrary; it can be on any one or more of these entries.
These entries follow:

DLOAD Format:

1 2 3 4 5 6 7 8 9 10
DLOAD SID S S1 L1 S2 L2 -etc.-

DAREA Format:

DAREA SID P1 C1 A1 P2 C2 A2

TABLED1 Format:

TABLED1 TID

x1 y1 x2 y2 x3 y3 -etc.- ENDT

TABLED2 Format:

TABLED2 TID X1

x1 y1 x2 y2 x3 y3 -etc.-

TABLED3 Format:

TABLED3 TID X1 X2

x1 y1 x2 y2 x3 y3 -etc.-

TABLED4 Format:

TABLED4 TID X1 X2 X3 X4

A0 A1 A2 A3 A4 A5 -etc.-

The TABLED4 entry defines a power series and is convenient in frequency response for enforced
constant velocity or displacement.

Main Index
202

Frequency Response
If a DLOAD entry is used to scale RLOAD1 input, the applied force magnitude in terms of
MSC.Nastran input is

Pi ( f ) = S ⋅ ∑ ( Si ⋅ Ai [ Ci ( f ) + i ⋅ Di ( f ) ] ) Eq. 7-7
i

where S and Si are input on the DLOAD Bulk Data entry, A i is input on the DAREA entry, and
C i ( f ) and D i ( f ) are input on the TABLEDi entries. Note that the “i” non-subscript term in the
expression i ⋅ D i ( f ) is – 1 .
If a DLOAD entry is used to scale RLOAD2 input, the applied force magnitude in terms of
MSC.Nastran input is

Pi( f ) = S ⋅ ∑ ( Si ⋅ Ai ⋅ Bi ( f ) ) Eq. 7-8

i

where S and S i are input on the DLOAD Bulk Data entry, A i is input on the DAREA entry, and
B i ( f ) is input on the TABLEDi entry.

Specification of the large force value depends upon whether acceleration, velocity, or
displacement is enforced.

Enforced Acceleration
Enforced acceleration is the easiest to apply since the required force is directly proportional to
the desired acceleration times the large mass:
··
p ( ω ) = mo u ( ω ) Eq. 7-9

Enforced Velocity
Enforced velocity requires a conversion factor
· ·
p ( ω ) = iωm o u = i ( 2πf )m o u ( ω ) Eq. 7-10
·
For constant velocity u ( ω ) , it may be easiest to use the RLOAD1 and TABLED4 entries because
the imaginary term i ⋅ D ( f ) of Eq. 7-7 and the frequency-dependent term 2πf can be specified
directly.

Enforced Displacement
Enforced displacement also requires a conversion factor

2 2
p ( ω ) = – ω m o u = – ( 2πf ) m o u ( ω ) Eq. 7-11

For constant displacement u ( ω ) , it may be easiest to use the TABLED4 entry because the
frequency-dependent term – ( 2πf ) 2 can be specified directly.
Main Index
 CHAPTER 7 203
Enforced Motion

Transient Response
For transient response, the type of enforced motion (displacement, velocity, or acceleration) is
specified with the TYPE field (field 5) on the TLOAD1 and TLOAD2 Bulk Data entries.

TLOAD1 Format

1 2 3 4 5 6 7 8 9 10
TLOAD1 SID DAREA DELAY TYPE TID

TLOAD2 Format

TLOAD2 SID DAREA DELAY TYPE T1 T2 F P

C B

where:

TYPE 0 (or blank) = Applied force (default)

TYPE 1 = Enforced displacement
TYPE 2 = Enforced velocity
TYPE 3 = Enforced acceleration

MSC.Nastran converts enforced displacements and velocities into accelerations by

differentiating once for velocity and twice for displacement. Note that for enforced
acceleration, you can specify either force (TYPE = 0 or blank) or acceleration (TYPE = 3); they
are the same for the large mass method.
You still need to use the large mass when specifying any type of enforced motion in transient
response analysis.

Main Index
204

7.5 Examples
This section provides several examples showing the input and output. These examples are:

Model Analysis Type Enforced Motion

bd07two Frequency Response Constant Acceleration
bd07bar1 Transient Response Ramp, Acceleration
bd07bar2 Transient Response Ramp, Displacement
bd07bar3 Transient Response Ramp, Displacement,
Discard Rigid-Body Mode

These examples are described in the sections that follow.

Two-DOF Model
Consider the two-DOF model first introduced in “Real Eigenvalue Analysis” on page 33 and
shown below in Figure 7-2. For this example, apply a constant magnitude base acceleration of
2
1.0 m/sec over the frequency range of 2 to 10 Hz and run modal frequency response with 5%
critical damping in all modes. The acceleration input is applied to the large mass (grid point 3).
The input file for this model is shown in Listing 7-1.

Grid Point 1 m 1 = 0.1 kg

Grid Point 2 m2 = 10.0 kg

Grid Point 3 Large 7

m 3 = 10 kg
Mass
x

Figure 7-2 Two-DOF Model with Large Mass

Main Index
 CHAPTER 7 205
Enforced Motion

Listing 7-1 Input File for Enforced Constant Acceleration

$ FILE bd07two.dat
$
$ TWO-DOF SYSTEM
$ CHAPTER 7, ENFORCED MOTION
$
TIME 5
SOL 111
CEND
TITLE = TWO-DOF SYSTEM
SUBTITLE = MODAL FREQUENCY RESPONSE
LABEL = ENFORCED CONSTANT ACCELERATION MAGNITUDE
$
$ SPECIFY MODAL EXTRACTION
METHOD = 10
$
$ SPECIFY DYNAMIC INPUT
DLOAD = 999
FREQ = 888
SDAMPING = 777
$
$ SELECT OUTPUT
DISPLACEMENT(PHASE,PLOT) = ALL
ACCELERATION(PHASE,PLOT) = ALL
$
$ XYPLOTS
$
... X-Y plot commands ...
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$ LARGE MASS AT BASE GRID POINT
CONM2 999 3 1.0E7
$
$ LOAD DEFINITION (INCLUDES SCALE FACTORS FOR ENFORCED ACCELERATION)
$DLOAD SID S S1 RLOAD1
DLOAD 999 1.0E7 1.0 998
$RLOAD1 SID DAREA TC
RLOAD1 998 997 901
$DAREA SID P1 C1 A1
DAREA 997 3 2 1.0
$TABLED4 TID X1 X2 X3 X4 +TAB4
TABLED4 901 0. 1. 0. 100. +TAB901
$+TAB4 A0 A1 A2 A3 A4 A5
+TAB901 1.0 ENDT
$
$ MODAL EXTRACTION
$EIGRL SID V1 V2 ND MSGLVL
EIGRL 10 -1. 30. 0
$
$ FREQUENCY RANGE 2-10 HZ
$FREQ1 SID F1 DF NDF
FREQ1 888 2. 0.05 160
$
$ MODAL DAMPING OF 5% CRITICAL
$TABDMP1 TID TYPE
$+TAB1 F1 G1 F2 G2 ETC
TABDMP1 777 CRIT +TABD7
+TABD7 0. 0.05 100. 0.05 ENDT
$
... basic model ...
$
ENDDATA

The large mass value is chosen as 1.0E7 kilograms and is input via the CONM2 entry. The scale
factor for the load (1.0E7) is input on the DLOAD Bulk Data entry. The factor of 1.0E7 is
approximately six orders of magnitude greater than the overall structural mass (10.1 kg). The
Main Index
206

TABLED4 entry defines the constant acceleration input. (One of the other TABLEDi entries can
also be used, but the TABLED4 entry is chosen to show how to use it for enforced constant
velocity and displacement later in this example.)
Figure 7-3 shows the X-Y plots resulting from the input point (grid point 3) and an output point
(grid point 1). The plots show acceleration and displacement magnitudes. Note that the
acceleration input is not precisely 1.0 m/sec 2 ; there is a very slight variation between 0.9999 and
1.0000 due to the large mass approximation.

Figure 7-3 Displacements and Accelerations for the Two-DOF Model

Main Index
 CHAPTER 7 207
Enforced Motion

Figure 7-3 Displacements and Accelerations for the Two-DOF Model (continued)
This model was analyzed with several values of large mass. Table 7-1 shows the results. Note
that the model with the 10 6 mass ratio is the model discussed earlier. Peak frequency response
results are compared for each model, and the natural frequencies are compared to those of the
constrained model in “Real Eigenvalue Analysis” on page 33. The table shows that a mass ratio
of 10 6 is a good value to use for this model.

Table 7-1 Models with Different Large Mass Ratios

Ratio of Natural Response Peaks (m/sec2)

Large Mass Frequencies* ·· ·· ·· ··
to Structure (Hz) X1 X2 X3 X3
max max max min

102
10 -9 52.0552 6.5531 1.0335 0.9524

4.8011
5.3025

104 10-10 52.2823 6.7641 1.0003 0.9995

4.7877
5.2910

Main Index
208

Table 7-1 Models with Different Large Mass Ratios (continued)

Ratio of Natural Response Peaks (m/sec2)

Large Mass Frequencies* ·· ·· ·· ··
to Structure (Hz) X1 X2 X3 X3
max max max min

106 0.0 52.2836 6.7661 1.0000 0.9999

4.7876
5.2909

108 0.0 52.2836 6.7662 1.0000 1.0000

4.7876
5.2909
* Resonant frequencies for the constrained model are 4.7876 and 5.2909 Hz.

This model can also be changed to apply constant velocity or constant displacement at its base.
Listing 7-2 is an abridged input file for the model, showing the Bulk Data entries required for
enforced constant acceleration, enforced constant velocity, and enforced constant displacement.
Note that only one of these is usually applied to any model, but all three are shown here for
comparison purposes.

Main Index
 CHAPTER 7 209
Enforced Motion

Listing 7-2 Bulk Data Entries for Enforced Constant Motion

$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$ ENTRIES FOR ENFORCED MOTION
$
$ LARGE MASS AT BASE GRID POINT
CONM2 999 3 1.0E7
$
$ LOAD DEFINITION
$DLOAD SID S S1 RLOAD1
DLOAD 999 1.0E7 1.0 998
$DAREA SID P1 C1 A1
DAREA 997 3 2 1.0
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$ UNIQUE ENTRIES FOR ENFORCED CONSTANT ACCELERATION MAGNITUDE
$
$RLOAD1 SID DAREA TC
RLOAD1 998 997 901
$TABLED4 TID X1 X2 X3 X4 +TAB4
TABLED4 901 0. 1. 0. 100. +TAB901
$+TAB4 A0 A1 A2 A3 A4 A5
+TAB901 1.0 ENDT
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$ UNIQUE ENTRIES FOR ENFORCED CONSTANT VELOCITY MAGNITUDE
$
$RLOAD1 SID DAREA TD
RLOAD1 998 997 902
$TABLED4 TID X1 X2 X3 X4 +TAB4
TABLED4 902 0. 1. 0. 100. +TAB902
$+TAB4 A0 A1 A2 A3 A4 A5
+TAB902 0.0 6.283185 ENDT
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$ $
$
$ UNIQUE ENTRIES FOR ENFORCED CONSTANT DISPLACEMENT MAGNITUDE
$
$RLOAD1 SID DAREA TC
RLOAD1 998 997 903
$TABLED4 TID X1 X2 X3 X4 +TAB4
TABLED4 903 0. 1. 0. 100. +TAB903
$+TAB4 A0 A1 A2 A3 A4 A5
+TAB903 0.0 0.0 -39.4784 ENDT

Each input utilizes the TABLED4 entry. The TABLED4 entry uses the algorithm

N
i
A i  -----------------
x – X1
Y = ∑  X2 
Eq. 7-12
i = 0

where x is input to the table, Y is returned, and N is the degree of the power series. When x < X3 ,
X3 is used for x ; when x > X4 , X4 is used for x . This condition has the effect of placing bounds
on the TABLED4 entry; note that there is no extrapolation outside of the table boundaries. There
are N+1 entries to this table.
Constant acceleration is the easiest to apply since the force is proportional to the mass for all
Main Index
frequencies. The power series for this case becomes
210

A0 + A1  -----------------
x – X1
Eq. 7-13
 X2 

where:

A0 = 1.0
X1 = 0.0
X2 = 1.0

Therefore, these terms define a constant (1.0 in this case).

Constant velocity involves a scale factor that is directly proportional to circular frequency ( 2πf ).
The power series for this case becomes

A0 + A1  -----------------
x – X1
Eq. 7-14
 X2 

where:

A0 = 0.0
A1 = 2π = 6.283185
X1 = 0.0
X2 = 1.0

Note that a phase change of 90 degrees is also required; this change is input using the TD field
(field 7) of the RLOAD1 entry.
Constant displacement involves a scale factor that is proportional to the circular frequency
squared ( 2πf ) 2 with a sign change. The power series for this case becomes

x – X1 2
A0 + A1  ----------------- + A2  -----------------
x – X1
Eq. 7-15
 X2   X2 

where:

A0 = 0.0
A1 = 0.0

A2 = – ( 2π ) 2 = – 39.4784

X1 = 0.0
X2 = 1.0

Table 7-2 summarizes the coefficients for the power series.

Main Index
 CHAPTER 7 211
Enforced Motion

Table 7-2 Coefficients for the Power Series

Type of
A0 A1 A2
Excitation
2
Enforced u 0.0 0.0 – ( 2π )
· 0.0 2π –
Enforced u
·· 1.0 – –
Enforced u

Cantilever Beam Model

Consider the cantilever beam first introduced in “Frequency Response Analysis” on page 103
and shown in Figure 7-4. In this case the planar model is analyzed for bending; therefore, only
three DOFs per grid point are considered: T1 (x-translation), T2 (y-translation), and R3 (z-
rotation). An acceleration ramp function in the y-direction is enforced at the base (grid point 1)
by applying a large mass and a force. T1 and R3 are constrained at grid point 1 since the enforced
motion is in only the T2-direction. Modal transient response analysis (SOL 112) is run with 5%
critical damping used for all modes. Modes up to 3000 Hz are computed with the Lanczos
method. Figure 7-5 shows the idealized ramp function and the MSC.Nastran implementation.
The excitation is not cut off abruptly; instead it is cut off over two time steps. A time step of 0.001
second is used, and the analysis is run for 1.0 second. Listing 7-3 shows the abridged input file.

Large Mass x
1.0E9 kg

Grid Point 1
Figure 7-4 Beam Model with Large Mass

Main Index
212

0.15

Idealized Input
·· 2
Y (m/sec )

0 0.05 Time (sec)

0.15

MSC.Nastran Input
·· 2
Y (m/sec )

Time (sec)
0.05
0.052

Figure 7-5 Idealized Ramp Function Versus MSC.Nastran Ramp Function

Main Index
 CHAPTER 7 213
Enforced Motion

Listing 7-3 Abridged Input File for Enforced Acceleration

$ FILE bd07bar1.dat
$
$ CANTILEVER BEAM MODEL
$ CHAPTER 7, ENFORCED MOTION
$
SOL 112
TIME 10
CEND
TITLE = CANTILEVER BEAM
SUBTITLE = MODAL TRANSIENT RESPONSE
LABEL = ENFORCED ACCELERATION
$
SPC = 21
DLOAD = 22
TSTEP = 27
SDAMPING = 25
$
METHOD = 10
$
$ PHYSICAL OUTPUT REQUEST
SET 11 = 1,11
DISPLACEMENT(PLOT) = 11
ACCELERATION(PLOT) = 11
$
$ XYPLOTS
$
... X-Y plot commands ...
$
BEGIN BULK
$
$.......2.......3.......4.......5.......6.......7.......8.......9.......10.....$
$
$ LARGE MASS OF 1.0E9
$CONM2 EID G CID M
CONM2 15 1 1.0E9
$
$ CONSTRAIN MASS IN 1,6 DIRECTIONS
SPC 21 1 16
$
$ DYNAMIC LOADING
$DLOAD SID S S1 L1
DLOAD 22 1.0E9 0.102 23
$TLOAD1 SID DAREA DELAY TYPE TID
TLOAD1 23 24 0 25
$DAREA SID P1 C1 A1
DAREA 24 1 2 0.15
$TABLED1 TID +TABL1
$+TABL1 X1 Y1 X2 Y2 ETC.
TABLED1 25 +TABL1
+TABL1 0.0 0.0 0.05 1.0 0.052 0.0 0.1 0.0 +TABL2
+TABL2 ENDT
$
$ CONVERT WEIGHT TO MASS: MASS = (1/G)*WEIGHT
$ G = 9.81 m/sec**2 --> WTMASS = 1/G = 0.102
PARAM WTMASS 0.102
$
$EIGRL SID V1 V2 MSGLVL
EIGRL 10 -1. 3000. 0
$
$TSTEP SID N1 DT1 NO1
TSTEP 27 1000 0.001 1
$
$ MODAL DAMPING OF 5% IN ALL MODES
$TABDMP1 TID TYPE +TABD
$+TABD F1 G1 F2 G2 ETC.
TABDMP1 25 CRIT +TABD
+TABD 0. 0.05 1000. 0.05 ENDT
$
... basic model ...
$
ENDDATA

A large mass of 1.0E9 kg is placed at grid point 1. This grid point is constrained in the T1- and
R3-directions but is free in the T2-direction. The load is scaled to give a peak input acceleration
of 0.15 m/sec 2 . This scaling is performed by applying a scale factor of 1.0E9 in the S field (field 3)
Main Index
214

of the DLOAD entry, a scale factor of 0.102 in the S1 field (field 4) of the DLOAD entry, and a
factor of 0.15 in the A1 field (field 5) of the DAREA entry. The applied load is scaled by 0.102
because the large mass is also scaled by 0.102 due to the PARAM,WTMASS entry (see Eq. 7-5).
The time variation is specified with the TABLED1 entry. The TLOAD1 entry specifies the type
of loading (field 5) as 0 (applied force); this gives the same answers if the type is specified as 3
(enforced acceleration).
Figure 7-6 shows the displacement and acceleration response at grid points 1 (base) and 11 (tip).
Note that at the end of the acceleration pulse the base has a constant velocity, and therefore, a
linearly increasing displacement.

Figure 7-6 Response for Enforced Acceleration

Next, consider the same model with a 0.015 meter displacement imposed instead of an
acceleration. The same ramp time history function is used (with a peak enforced displacement
of 0.015 meter) so that the only change to the input file is to change the excitation type from 0
(applied force) to 1 (enforced displacement on field 5 of the TLOAD1 entry) and the amplitude
in the DAREA entry from 0.15 to 0.015. Figure 7-7 shows the idealized input displacement time
history. Figure 7-8 shows the displacement and acceleration response at grid points 1 and 11.

Main Index
 CHAPTER 7 215
Enforced Motion

0.15

Idealized Input
Y(m)

0 0.05 Time (sec)

Figure 7-7 Response for Enforced Displacement (With the Rigid-Body Mode)

Figure 7-8 Response for Enforced Displacement (Without the Rigid-Body Mode)

Now, consider a change to the enforced displacement run. In this case, remove the rigid-body
mode's contribution either by not computing the rigid-body mode (by setting V1 to a small
positive value, such as 0.01 Hz) or by neglecting the rigid-body mode in the transient response
(by setting PARAM,LFREQ to a small positive number, such as 0.01 Hz). Figure 7-9 shows the
resulting displacement and acceleration responses at grid points 1 and 11. Note that the
responses are relative to the structure and are not absolute. The relative displacement of grid
point 1 should be zero, and it is very close to zero (i.e., 10 –10 ) as a result of the sufficiently large
mass.

Main Index
216

Figure 7-9 Response for Enforced Displacement (Without the Rigid-Body Mode)

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Restarts In Dynamic Analysis
8
■ Overview
■ Automatic Restarts
■ Structure of the Input File
■ User Interface

■ Determining the Version for a Restart

■ Examples

Main Index
218

8.1 Overview
A restart is a logical way of continuing from a previous run without having to start from the
beginning. The savings can be in terms of both time and money. Some forms of restarts are used
practically every day perhaps without you realizing it. An example of a restart can be as simple
as reading a book. Normally, you probably do not finish reading a book in one continuous
stretch. You may read a hundred pages today and another fifty pages tomorrow, and so on.
Each time that you continue from where you left off previously is a restart. It is much more time
consuming and impractical to start from page one every time that you pick up the book.
This analogy can be applied to MSC.Nastran. In the case of a static analysis, the most expensive
and time-consuming part of the run is the decomposition of the stiffness matrix. This fact is
especially true for large models. Now suppose after completing the original run, you want to
obtain additional output (e.g., stresses, displacements, etc.) or add more load cases. You can
always start from the beginning and redo the whole problem, or you can perform a restart at a
fraction of the time and cost. In the case of additional output requests and additional load
conditions, the decomposition of the stiffness matrix, which was performed in the previous run,
is not redone if restart is used.
In dynamic analysis, the calculation of normal modes is, in general, the most expensive
operation. Therefore, a common application of restart is the performance of a transient or
frequency response analysis by restarting from the normal modes calculation, which was saved
in the database from a previous run. This restart process avoids the recalculation of the normal
modes.
The cost of restarting is measured in the disk space required to store data blocks for subsequent
use. Judging whether to not save data blocks and simply rerun the analysis or to save data
blocks for restarting is determined by several factors, such as the amount of available disk space
and your computer’s solution speed. Note that the database can be copied to tape which then
provides you with more free disk space. When you are ready to perform a restart, this database
can then be copied from the tape back to your disk.

Main Index
 CHAPTER 8 219
Restarts In Dynamic Analysis

8.2 Automatic Restarts

In Version 66, we introduced a new Executive System that more effectively uses modern
database technology. Among its many features, the new Executive System implements the
NASTRAN Data Definition Language (NDDL). With NDDL, MSC.Nastran is able to use
automatic restart logic with the superelement solution sequences. For more information on the
NDDL, refer to the MSC.Nastran 2005 DMAP Programmer’s Guide.

In order to maintain upward compatibility, we retained the old solution sequences and added a
new series of superelement solution sequences in Version 66. These solution sequences are also
known as the Structured Solution Sequences (solution sequences numbers greater than 100).
The automatic restart logic is available only for these solution sequences. Restart is no longer
available for the rigid format solution sequences. Use these new Structured Solution Sequences
rather than the old ones since all the new features are automatically implemented in these new
solution sequences. Improvements made to the automatic restart logic for Version 67 and later,
have made it even more efficient and robust. In this chapter, we will address the restart logic for
the Structured Solution Sequences. Note that you are not required to use superelements in order
to use the superelement solution sequences. If you are familiar with the rigid format solutions,
converting to the new Structured Solution Sequences only requires that you replace the solution
command. For example, a normal modes run only requires replacing the SOL 3 command with
the SOL 103 command in the Executive Control Section.

Main Index
220

8.3 Structure of the Input File

Before presenting details on how restarts work, the following simple flow diagram of the
MSC.Nastran input file structure may be beneficial:

NASTRAN Statements Optional

File Management Optional

Statements

Executive Control
Statements

CEND Required Delimiter

Case Control
Commands

BEGIN BULK Required Delimiter

Bulk Data Entries

ENDDATA Required Delimiter

This order must be followed as shown for all MSC.Nastran input files. For details regarding the
statements, commands, and entries, see the MSC.Nastran Quick Reference Guide.

Main Index
 CHAPTER 8 221
Restarts In Dynamic Analysis

8.4 User Interface

There are two types of runs--cold start and restart runs--which are described below.

Cold Start Run

The first run, which is called the cold start run, is identical to what you usually do for submitting
an MSC.Nastran job with the exception that the database must be saved. In addition to your
normal output files (e.g., F06 file), four database files are created as a result of this run, assuming
that the default database initialization values are used. The convention for the filenames is
machine dependent; therefore, you should refer to the machine-dependent documentation for
your specific computer for the exact syntax. For a typical UNIX machine with an MSC.Nastran
input file called “run1.dat”, the following sample submittal command can be used:
nastran run1
where “nastran” is the name of the shell script for executing MSC.Nastran. Note that the above
submittal command assumes that you have not modified the “scr=no” option as your default
nastran submittal option. The “scr=no” option saves the database at the end of the run, which
is required if you intend to perform restarts. The following four database files are created as a
result of the above command.
run1.DBALL
run1.MASTER
run1.USROBJ
run1.USRSOU
The last two files are not needed and can be deleted if you do not want to store your own DMAP,
which is usually the case.

Restart Run
NASTRAN Statement Section
This section is normally the same as your cold start run. The BUFFSIZE must not be changed in
a restart run.

File Management Section (FMS)

This section tells MSC.Nastran that you are performing a restart run. The RESTART statement
is required in any restart run. The general format for the RESTART statement is as follows:
RESTART VERSION=a,b
where “a” is the version from which you restart (default value for “a” is LAST) and “b” indicates
whether version “a” is to be kept (KEEP) or deleted (NOKEEP) at the end of the run. The default
value for “b” is NOKEEP. Due to the default values, the following two restart statements are
identical:
RESTART VERSION=LAST,NOKEEP
RESTART

Main Index
222

Whenever a restart is performed, a new version number is automatically created in the database.
For each restart, the current version number is incremented by one regardless of whether the job
ran successfully or not. There are two exceptions to this rule and they are discussed later on.
You also need to tell MSC.Nastran which database you want to attach to your current run. There
are several ways to accomplish this; one way is to use the ASSIGN statement. For example, if
you are restarting from the database created by run1.dat, the following FMS statements can be
used in your current run:
RESTART
ASSIGN MASTER=’run1.MASTER’
For UNIX machines, the filenames are case sensitive. They should be entered exactly as they
were created in the cold start run and enclosed with single quotes as shown above.
An alternate way is to use the DBS statement on the submittal line instead of the ASSIGN
statement. Assuming the current run is called run2.dat, then the equivalent submittal statement
is as follows:
nastran run2 dbs=run1
The ASSIGN statement is not needed in this case as long as the run1 database files are contained
in the submittal directory. Again the syntax is for a typical UNIX machine. You should refer to
the machine-dependent documentation for your particular computer for specific details.

Executive Control Section

This section is the same as your normal run with the exception of perhaps the “SOL x”
command. For example, if you are performing a normal modes analysis in run1.dat, then the
“SOL x” command in run1.dat should reference “SOL 103". In run2.dat, if you are performing a
modal transient restart from run1.dat, then the “SOL x” statement in this case should then
reference “SOL 112". DIAGs can be turned on or off.

Case Control Section

The automatic restart logic compares the modifications made to the Case Control and Bulk Data
Sections in determining which operations need to be processed upon restart. Therefore, you
must be very careful with the changes that you make in your restart run. Adhering to the
following rules will avoid unnecessary reprocessing of previously completed operations.

• You must include all “solution-type” related Case Control commands, which are
unchanged as compared to the cold start run, in your restart run. In other words, do
not make unnecessary LOAD, SPC, MPC, or METHOD command changes or remove
them from the Case Control Section unless these are actual changes. This process is
clarified later with the example problems.
• Output requests can be modified. A typical example can be a request of the eigenvector
printout which was not requested in the cold start run.

Main Index
 CHAPTER 8 223
Restarts In Dynamic Analysis

Bulk Data Section

As mentioned in the previous section, the automatic restart logic compares the changes made in
the Bulk Data Section and determines the path that it follows. A copy of the Bulk Data is stored
for each version. The restart run must not contain any Bulk Data entry that was included in the
previous runs and saved in the database. The Bulk Data Section in the current restart run should
contain only new entries, changed entries, and/or the deletion of old entries from the database.
This philosophy is slightly different than the one used in the Case Control Section.
Table 8-1 shows an example where a normal modes run is performed in the cold start run. The
eigenvalue table and the grid point weight generator output are requested in this run. For the
restart run, the eigenvector output is desired; furthermore, reprinting the grid point weight
generator output is not desired. One way to accomplish this is to delete the existing
“PARAM,GRDPNT,0" entry from the database.

Main Index
224

Table 8-1 Listing of the Cold Start and Restart Input Files

Cold Start Run Restart Run

$ $
$ FILENAME - run1.dat $ FILENAME - run2.dat
$ $
$ TYPICAL UNIX SUBMITTAL $ TYPICAL UNIX SUBMITTAL
$ COMMAND $ COMMAND
$ $
$ nastran run1 $ nastran run2
$ $
SOL 103 RESTART
TIME 5 ASSIGN MASTER=’run1.MASTER’
CEND SOL 103
TITLE = COLDSTART RUN TIME 5
METHOD = 10 CEND
SPC = 1 TITLE = RESTART RUN
BEGIN BULK METHOD = 10
$ SPC = 1
GRID,1,,0.,0.,0. $
GRID,2,,10.,0.,0. $ REQUEST FOR EIGENVECTOR
CROD,1,10,1,2 $ PRINTOUT
PROD,10,1,1.0 $
MAT1,1,1.+7,,.32 DISP = ALL
EIGRL,10,,,1 BEGIN BULK
CONM2,100,2,,10. $
PARAM,WTMASS,.00259 $ SKIP GRID POINT WEIGHT
$ $ GENERATOR OUTPUT
$ REQUEST FOR GRID POINT $
$ WEIGHT GENERATOR /,7
$ OUTPUT $
$ ENDDATA
PARAM,GRDPNT,0
$
SPC1,1,123456,1
SPC1,1,23456,2
$
ENDDATA

The format for the delete entry is “/,K1,K2" where K1 and K2 are the sorted Bulk Data sequence
number of the first and last entries in the sequence to be removed, respectively. In the event that
K2 = K1, the following two entries are identical:
/,K1,K2
/,K1
The values of K1 and K2 can be obtained in the F06 file of your cold start run as long as the Bulk
Data is echoed with the “ECHO = SORT” Case Control command, which is the default option.
From Figure 8-1, the sorted Bulk Data count for “PARAM,GRDPNT,0" is “7"; therefore, the
delete entry in this case can be either “/,7,7" or “/,7". No additional Bulk Data entry is required

Main Index
 CHAPTER 8 225
Restarts In Dynamic Analysis

for this restart run since there are no other changes involved for this model. Note that the same
SPC and METHOD commands are used in both the cold start and restart runs since neither the
actual boundary condition nor the desired eigenvalue calculation has changed.

Listing 8-1

S O R T E D B U L K D A T A E C H O

CARD
COUNT . 1 .. 2 .. 3 .. 4 .. 5 .. 6 .. 7 .. 8 .. 9 .. 10 .
1- CONM2 100 2 10.
2- CROD 1 10 1 2
3- EIGRL 10 1
4- GRID 1 0. 0. 0.
5- GRID 2 10. 0. 0.
6- MAT1 1 1.+7 .32 .097
7- PARAM GRDPNT 0
8- PARAM WTMASS .00259
9- PROD 10 1 1.0
10- SPC1 1 23456 2
11- SPC1 1 123456 1
ENDDATA
TOTAL COUNT= 12

Figure 8-1 Echo of the Sorted Bulk Data Input for the Cold Start Run

You can include as many of these delete entries as necessary. However, if the case requires many
changes, it is probably more convenient to delete the entire old Bulk Data Section from the
database and replace it with the fully revised Bulk Data Section, including the new and modified
entries. This operation can be accomplished by including the full new Bulk Data plus the
following entry in the Bulk Data Section of the restart run:
/1,i
where i is any positive integer that is greater than or equal to the number of Bulk Data entries
from your cold start run.
For conventional dynamic analysis (i.e., non-superelement), restarts involving model changes
(e.g., changing the thickness of a plate) are not very efficient. Therefore, the savings is probably
minimal, if any. However, in the case of additional output requests or a restart from a modes
run to a response run, the savings can be substantial. This type of restart is covered extensively
in “Examples” on page 228. For superelement analysis, even restarts involving model changes
can be beneficial as long as these changes are localized.

Main Index
226

8.5 Determining the Version for a Restart

Not all versions in the database are restartable. For each run, a message is printed near the top
of the F04 file indicating the version number of the current job. In general, if a job completes
without any fatal messages, then that particular version is restartable. It is a good idea to keep
a brief log of all the restartable versions since this is probably the most reliable source of
information regarding whether a version is restartable. If a restart job failed, e.g., due to Bulk
Data error, then this newly created version is not restartable, which is indicated by the following
error message at the bottom of the F06 file.

Listing 8-2

*** USER WARNING MESSAGE 784 (XCLNUP)

VERSION = yyy PROJECT = "zzz" OF THIS DATA BASE IS NOT VALID FOR RESTART PURPOSES.

USER ACTION:
SUBSEQUENT RESTARTS SHOULD REFERENCE VERSION = xxx OR A PRIOR VALID VERSION

If the version is not restartable, you must restart from a previous valid version. The “xxx” and
“yyy” above denote version numbers. The “zzz” denotes a project description provided by you.
This project description is alphanumeric and can contain up to 40 characters (the default is
blank). The project description is often not used and is an optional statement.
If, for some reason, the records for the old runs are no longer available, then the DBDIR FMS
statement can be used to query the database contents to find out which versions are being stored
in the database. The following simple setup is all that is required for this purpose.
ASSIGN MASTER=’ddddd.MASTER’
DBDIR VERSION=*,PROJECT=*
ENDJOB
where ddddd.MASTER is the name of the database being used.
The Executive Control, Case Control, or Bulk Data Section is not required in this case.
Furthermore, a new version is not created in this case. Near the top of the F06 output, a
PROJECT VERSION DIRECTORY TABLE is printed listing all the versions in the database. A
“**” next to a version number indicates that this particular version was deleted from the
database. This deletion may be due to the NOKEEP option or the use of the DBCLEAN
command when performing a restart run. A version number with a “**” next to it is not
restartable. A version number without a “**” next to it is restartable if the run that created the
version did not terminate with UWM 784. For more advanced users, this DBDIR FMS statement
can also be used to check the database directory for the existence of data blocks, e.g., UG. If the
job fails very early in the run (e.g., error in the FMS section), then a new version may not be
created.
It is always good practice to back up the database on a regular basis. If the system aborts the run
(e.g., disk space exhausted or time quota exceeded on a system level), then there is a chance that
the database is corrupted and will not allow restarts. Another good practice to ensure that only
good models are retained in the database is to perform the following:

Main Index
 CHAPTER 8 227
Restarts In Dynamic Analysis

• Use RESTART VERSION = a,KEEP

where “a” is a version number.
• If a version contains errors or is no longer of interest, then use the FMS statement
DBCLEAN to remove obsolete or incorrect versions of the model from the database.
Removing these versions allows the Executive System to reuse some of this space for
new versions.

Main Index
228

8.6 Examples
The examples perform a typical series of runs starting from a normal modes run and restarting
into transient and frequency response analyses. Table 8-2 summarizes this series of nine runs
along with a brief description. Listings of the nine runs are also included (Listing 8-3 through
Listing 8-11).

Table 8-2 Typical Series of Restart Runs

Run Solution
Name of Version Version
Sequence Sequence Description of Runs
Input File Created Deleted
Number Number
1 bd08bar1.dat 103 Perform a normal modes cold start 1 None
analysis and save the database.
2 bd08bar2.dat 103 Restart from run number 1 and 2 None
request eigenvector output.
3 bd08bar3.dat 103 The first two modes of the structure 3 2
are very close to one of the forcing
frequencies. The structure is
modified in order to stay away from
resonance. This run restarts from run
number 2 to delete the old PBAR
entry and replace it with the modified
PBAR entry. The modes are
recalculated, and the eigenvector
output is requested in this run.
4 bd08bar4.dat N/A This run deletes the data blocks in the None 1
database previously occupied by
Version 1. This space can then be
reused by future restarts. Note that
the statement DBCLEAN does not
reduce the database size. It merely
removes some of its contents so that
this space can be reused. A new
version is not created as a result of
this run. This is an optional run
especially if disk space is of no
concern to you.

Main Index
 CHAPTER 8 229
Restarts In Dynamic Analysis

Table 8-2 Typical Series of Restart Runs (continued)

Run Solution
Name of Version Version
Sequence Sequence Description of Runs
Input File Created Deleted
Number Number
5 bd08bar5.dat 112 This is a transient restart run from the 4 None
modes saved in Version 3. The
applied load is a unit step function.
The modes calculated in run number
3 are also saved at the end of this run.
Since the calculation of the modes is
the most expensive operation in a
dynamic analysis, it is probably a
good idea to save Version 3 once you
have confidence in the results. This
way you can always restart from this
version. Partial output is shown at
the top of Figure 8-2. A 1% critical
damping value is applied to the
structure.
6 bd08bar6.dat 112 This is another transient restart run 5 4
using the solution from run number
5. The purpose of this run is to
request additional output. Partial
output is shown at the bottom of
Figure 8-2. Note that the maximum
displacement at grid point 11 is the
same as in run number 5 as expected
since you are asking for data recovery
on the same point due to the same
loading condition.
7 bd08bar7.dat 112 This is another restart run from 6 5
Version 6 with a different load
condition (triangular pulse). Partial
output is shown in Figure 8-3. In this
case, you can just as easily restart
from Version 3.

Main Index
230

Table 8-2 Typical Series of Restart Runs (continued)

Run Solution
Name of Version Version
Sequence Sequence Description of Runs
Input File Created Deleted
Number Number
8 bd08bar8.dat 111 This is a frequency response restart 7 6
run. Note that this can also be
restarted from Version 3 since it was
saved in the database. Partial output
is shown in Figure 8-4. A 2% critical
damping value is applied to the
structure.
9 bd08bar9.dat N/A This is a database directory printout None None
run. As shown in Figure 8-5, there
are seven versions in the database.
Run numbers 4 and 9 did not create
any new versions. Only Versions 3
and 7 are restartable. This is an
optional run.

Remarks
If the results for run number 1 are not going to be used for any future purposes, then you may
consider making run number 3 as a cold start run instead of a restart run. Model changes do not
save you much time, if any, in a non-superelement analysis. By making run 3 a cold start run,
you reduce the total amount of disk space required. In this case, run number 4 is not necessary
since you are starting with a new database. However, if you want to keep both physical models
in the database, then run number 3 should be a restart run as shown in this example. An
application of this can be a parametric study of two different configurations. This type of restart
allows you to make efficient data recovery or response analysis from two different physical
models. However, this type of restart is not used often in a non-superelement analysis since, in
general, it is not very efficient. However, in a superelement analysis (see “Advanced Dynamic
Analysis Capabilities” on page 311), this type of restart can still be very efficient since the
changes can be localized to a small region.

Main Index
 CHAPTER 8 231
Restarts In Dynamic Analysis

Listing 8-3 Input File for Normal Modes Run

$ FILE - bd08bar1.dat
$
$ NORMAL MODES RUN
$
ID CANT BEAM
SOL 103
TIME 10
CEND
TITLE = CANTILEVER BEAM - NORMAL MODES - COLD START RUN
SPC = 1
METHOD = 10
$
BEGIN BULK
$
CBAR 1 1 1 2 0. 1. 0.
CBAR 2 1 2 3 0. 1. 0.
CBAR 3 1 3 4 0. 1. 0.
CBAR 4 1 4 5 0. 1. 0.
CBAR 5 1 5 6 0. 1. 0.
CBAR 6 1 6 7 0. 1. 0.
CBAR 7 1 7 8 0. 1. 0.
CBAR 8 1 8 9 0. 1. 0.
CBAR 9 1 9 10 0. 1. 0.
CBAR 10 1 10 11 0. 1. 0.
EIGRL 10 -0.1 50.
GRID 1 0.0 0. 0.
GRID 2 0.3 0. 0.
GRID 3 0.6 0. 0.
GRID 4 0.9 0. 0.
GRID 5 1.2 0. 0.
GRID 6 1.5 0. 0.
GRID 7 1.8 0. 0.
GRID 8 2.1 0. 0.
GRID 9 2.4 0. 0.
GRID 10 2.7 0. 0.
GRID 11 3.0 0. 0.
MAT1 1 7.1+10 .33 2.65+4
PARAM AUTOSPC YES
PARAM WTMASS .102
PBAR 1 1 6.158-4 3.-8 3.-8 6.-8 2.414
SPC1 1 123456 1
$
ENDDATA

Main Index
232

Listing 8-4 Input File for Requesting Eigenvectors

$ FILE - bd08bar2.dat
$
$ NORMAL MODES RUN
$ REQUEST EIGENVECTOR PRINTOUTS FROM PREVIOUS RUN
$
RESTART VERSION=1,KEEP
ASSIGN MASTER=’bd08bar1.MASTER’
$
ID CANT BEAM
SOL 103
TIME 10
CEND
TITLE = EIGENVECTORS DATA RECOVERY RESTART RUN
SPC = 1
METHOD = 10
DISP = ALL $ PRINT EIGENVECTORS
$
BEGIN BULK
$
ENDDATA

Listing 8-5 Input File for Modifying a Bar Element

$ FILE - bd08bar3.dat
$
$ NORMAL MODES RUN
$ MODIFY PBAR
$
RESTART
ASSIGN MASTER=’bd08bar1.MASTER’
$
ID CANT BEAM
SOL 103
TIME 10
CEND
TITLE = CANTILEVER BEAM - NORMAL MODES - RESTART RUN
SPC = 1
METHOD = 10
DISP = ALL
$
BEGIN BULK
$
$ DELETE OLD PBAR ENTRY, LINE 26 OF SORTED BULK DATA COUNT
$
/,26
$
$ ADD NEW PBAR ENTRY
$
PBAR,1,1,6.158-4,2.9-8,3.1-8,6.-8,2.414
$
ENDDATA

Listing 8-6 Input File for Cleaning a Database

$ FILE - bd08bar4.dat
$
assign master=’bd08bar1.MASTER’
dbclean version=1
endjob

Main Index
 CHAPTER 8 233
Restarts In Dynamic Analysis

Listing 8-7 Input File for Transient Response

$ FILE - bd08bar5.dat
$
$ THIS IS A TRANSIENT RESTART RUN FROM THE MODES
$ CALCULATED BY THE RUN “bd08bar3.dat”
$
RESTART VERSION=3,KEEP
ASSIGN MASTER=’bd08bar1.MASTER’
ID CANT BEAM
SOL 112
TIME 10
CEND
TITLE = TRANSIENT RESTART - UNIT STEP FUNCTION INPUT
SUBTITLE = REQUEST DISPLACEMENT TIME HISTORY AT GRID POINT 11
SPC = 1
METHOD = 10
SET 1 = 11
DISP = 1
SUBCASE 1
SDAMP = 100
TSTEP = 100
DLOAD = 100
$
BEGIN BULK
$
$ ADDITIONAL ENTRIES FOR DYNAMIC LOADS
$ FOR UNIT STEP FUNCTION
$
$ SID DAREA DELAY TYPE TID
TLOAD1 100 101 102
$
DAREA,101,11,3,1.0
$
TABLED1,102,,,,,,,,+TBL1
+TBL1,0.0,0.0,.001,1.0,10.0,1.0,ENDT
$
$ TRANSIENT TIME STEPS
$
$ SID N(1) DT(1) NO(1)
TSTEP 100 600 .001 5
$
$ MODAL DAMPING TABLE
$
TABDMP1,100,CRIT,,,,,,,+TDAMP
+TDAMP,0.,.01,200.,.01,ENDT
$
ENDDATA

Main Index
234

Listing 8-8 Input File for an Additional Output Request

$ FILE - bd08bar6.dat
$
$ THIS IS ANOTHER TRANSIENT RESTART RUN. THE PURPOSE
$ OF THIS RUN IS TO REQUEST ADDITIONAL OUTPUT.
$
RESTART
ASSIGN MASTER=’bd08bar1.MASTER’
ID CANT BEAM
SOL 112
TIME 10
CEND
TITLE = T R A N S I E N T R E S T A R T
SUBTITLE = U N I T S T E P F U N C T I O N I N P U T
SPC = 1
METHOD = 10
SET 1 = 11
SET 2 = 10
ACCE = 2
SUBCASE 1
SDAMP = 100
TSTEP = 100
DLOAD = 100
$
$
$ PLOT RESULTS
$
$...X-Y plot commands ...
$
BEGIN BULK
$
$
ENDDATA

Main Index
 CHAPTER 8 235
Restarts In Dynamic Analysis

Listing 8-9 Input File for an Additional Transient Load

$ FILE - bd08bar7.dat
$
$ THIS IS ANOTHER TRANSIENT RESTART RUN USING
$ A DIFFERENT LOAD CONDITION. NOTE THAT SINCE
$ THERE ARE NO MODEL CHANGES, THE SAME MODES
$ WERE USED FROM THE DATABASE FOR THE RESPONSE
$ CALCULATIONS.
$
RESTART
ASSIGN MASTER=’bd08bar1.MASTER’
ID CANT BEAM
SOL 112
TIME 10
CEND
$
$ NOTE THAT TITLE CHANGES HAVE NO EFFECT
$ ON SOLUTION PROCESS, THEY ONLY CHANGE THE
$ PRINTOUT TITLE
$
TITLE = T R A N S I E N T R E S T A R T
SUBTITLE = TRIANGLE PULSE - 1.0 AT T=0 AND 0.0 AFTER .2 SEC
SPC = 1
METHOD = 10
SET 1 = 11
DISP = 1
SUBCASE 1
SDAMP = 100
TSTEP = 100
DLOAD = 300
$
$ PLOT RESULTS
$
$...X-Y plot commands ...
$
BEGIN BULK
$
$ SID DAREA DELAY TYPE TID
TLOAD1 300 301 302
$
DAREA,301,11,3,1.0
$
TABLED1,302,,,,,,,,+TBL3
+TBL3,0.0,0.0,.001,1.0,.20,0.0,10.0,0.0,+TBL4
+TBL4,ENDT
$
ENDDATA

Main Index
236

Listing 8-10 Input File for Frequency Response Analysis

$ FILE - bd08bar8.dat
$
$ THIS IS RESTART RUN TO PERFORM FREQUENCY RESPONSE
$
RESTART
ASSIGN MASTER=’bd08bar1.MASTER’
ID CANT BEAM
SOL 111
TIME 10
CEND
TITLE = CANTILEVER BEAM - FREQUENCY RESPONSE RESTART
SPC = 1
METHOD = 10
SET 1 = 11
DISP(PHASE) = 1
SUBCASE 1
$
$ A TWO-PERCENT CRITICAL DAMPING IS
$ APPLIED TO THIS RUN AS OPPOSED TO
$ ONE-PERCENT CRITICAL DAMPING IN
$ THE TRANSIENT ANALYSIS
$
SDAMP = 1000
DLOAD = 1000
FREQ = 1000
$
$ PLOT RESULTS
$
$...X-Y plot commands ...
$
BEGIN BULK
$
$ ADDITIONAL ENTRIES FOR FREQUENCY RESPONSE
$
$ SID DAREA M N TC TD
$
RLOAD1 1000 1001 1002
$
DAREA,1001,11,3,0.1
$
TABLED1,1002
,0.,1.,200.,1.,ENDT
$
$ FORCING FREQUENCIES
$
$ RESONANT FREQUENCIES
$
FREQ,1000,2.03174,2.100632,12.59101,13.01795
FREQ,1000,34.90217,36.08563
$
$ SPREAD THROUGHOUT
$ FREQUENCY RANGE OF INTEREST
$ WITH BIAS BETWEEN
$ HALF-POWER POINTS
$
FREQ,1000,1.437,1.556,1.675,1.794,1.913
FREQ,1000,2.046,2.059,2.073,2.087
FREQ,1000,2.224,2.347,2.47,2.593,2.716
FREQ,1000,8.903,9.641,10.378,11.116,11.853
FREQ,1000,12.676,12.762,12.847,12.933
FREQ,1000,13.781,14.543,15.306,16.068,16.831
FREQ,1000,24.680,26.724,28.769,30.813,32.858
FREQ,1000,35.139,35.376,35.612,35.849
FREQ,1000,41.189,46.292,51.395,56.499,61.602
$
FREQ1,1000,0.,.5,200
$
$ DAMPING
$
TABDMP1,1000,CRIT,,,,,,,+DAMP
+DAMP,0.,.02,200.,.02,ENDT
$
ENDDATA

Main Index
 CHAPTER 8 237
Restarts In Dynamic Analysis

Listing 8-11 Input File to Print the Database Dictionary

$ FILE - bd08bar9.dat
$
assign master=’bd08bar1.MASTER’
dbdir
endjob

***** PARTIAL OUTPUT FROM bd08bar5.f06 ***** POINT-ID = 11

D I S P L A C E M E N T V E C T O R

TIME TYPE T1 T2 T3 R1 R2 R3
.0 G .0 .0 .0 .0 .0 .0
5.000000E-03 G 9.024590E-22 1.614475E-21 1.923969E-05 .0 -3.001072E-05 3.518218E-20
1.000000E-02 G 3.367912E-21 8.536304E-21 7.739825E-05 .0 -1.118816E-04 1.271978E-19
.
.
2.300003E-01 G 3.941339E-19 8.273585E-18 7.832402E-03 .0 -3.655237E-03 5.515358E-18
2.350003E-01 G 3.956722E-19 8.583470E-18 7.862186E-03 .0 -3.686721E-03 5.464303E-18
2.400003E-01 G 3.957035E-19 9.029934E-18 7.880123E-03 .0 -3.709195E-03 5.756178E-18
2.450003E-01 G 3.944803E-19 9.498899E-18 7.886210E-03 .0 -3.726098E-03 6.049827E-18
2.500003E-01 G 3.927639E-19 9.878397E-18 7.884440E-03 .0 -3.754001E-03 6.040225E-18
2.550003E-01 G 3.909298E-19 1.020556E-17 7.873885E-03 .0 -3.796297E-03 5.837433E-18
2.600002E-01 G 3.885339E-19 1.063947E-17 7.844862E-03 .0 -3.831216E-03 5.871013E-18
2.650001E-01 G 3.848278E-19 1.128651E-17 7.785291E-03 .0 -3.828911E-03 6.419444E-18
2.700001E-01 G 3.796167E-19 1.207232E-17 7.690622E-03 .0 -3.778875E-03 7.287524E-18
.
.
5.949959E-01 G 1.986670E-19 -1.304284E-17 4.027668E-03 .0 -2.010369E-03 -4.506177E-18
5.999959E-01 G 2.106813E-19 -1.285776E-17 4.244778E-03 .0 -2.079488E-03 -4.219749E-18

----------------------------------------------------------------------------------------------------------------------------
***** PARTIAL OUTPUT FROM bd08bar6.f06 *****
X Y - O U T P U T S U M M A R Y ( R E S P O N S E )

SUBCASE CURVE FRAME XMIN-FRAME/ XMAX-FRAME/ YMIN-FRAME/ X FOR YMAX-FRAME/ X FOR

ID TYPE NO. CURVE ID. ALL DATA ALL DATA ALL DATA YMIN ALL DATA YMAX
1 DISP 1 11( 5) 0.000000E+00 5.999959E-01 0.000000E+00 0.000000E+00 7.886210E-03 2.450003E-01
0.000000E+00 5.999959E-01 0.000000E+00 0.000000E+00 7.886210E-03 2.450003E-01
1 ACCE 2 10( 5) 0.000000E+00 5.999959E-01 -8.511892E-01 2.650001E-01 9.618800E-01 5.000000E-03
0.000000E+00 5.999959E-01 -8.511892E-01 2.650001E-01 9.618800E-01 5.000000E-03

Figure 8-2 Partial Output from Transient Analysis with Unit Step Function Input

Main Index
238

***** PARTIAL OUTPUT FROM bd08bar7.f06 *****

POINT-ID = 11
D I S P L A C E M E N T V E C T O R

TIME TYPE T1 T2 T3 R1 R2 R3
.0 G .0 .0 .0 .0 .0 .0
5.000000E-03 G 8.975580E-22 1.606407E-21 1.913647E-05 .0 -2.984776E-05 3.499062E-20
1.000000E-02 G 3.317352E-21 8.430443E-21 7.628838E-05 .0 -1.102005E-04 1.252464E-19
.
.
1.650002E-01 G 2.077055E-19 2.232167E-18 4.099007E-03 .0 -1.833254E-03 1.845365E-18
1.700002E-01 G 2.104969E-19 2.275425E-18 4.180374E-03 .0 -1.908738E-03 1.502202E-18
1.750002E-01 G 2.131340E-19 2.401124E-18 4.255632E-03 .0 -2.000815E-03 1.338691E-18
1.800002E-01 G 2.146708E-19 2.724639E-18 4.308840E-03 .0 -2.072395E-03 1.631772E-18
1.850002E-01 G 2.144943E-19 3.204978E-18 4.328179E-03 .0 -2.097317E-03 2.240218E-18
1.900002E-01 G 2.129652E-19 3.684523E-18 4.313660E-03 .0 -2.081293E-03 2.749993E-18
1.950002E-01 G 2.109444E-19 4.053196E-18 4.272623E-03 .0 -2.048310E-03 2.916553E-18
.
.
5.949959E-01 G 1.315062E-19 -1.196586E-17 2.607416E-03 .0 -1.203858E-03 -4.331415E-18
5.999959E-01 G 1.410901E-19 -1.111419E-17 2.775622E-03 .0 -1.249621E-03 -3.777846E-18

X Y - O U T P U T S U M M A R Y ( R E S P O N S E )

SUBCASE CURVE FRAME XMIN-FRAME/ XMAX-FRAME/ YMIN-FRAME/ X FOR YMAX-FRAME/ X FOR

ID TYPE NO. CURVE ID. ALL DATA ALL DATA ALL DATA YMIN ALL DATA YMAX
1 DISP 1 11( 5) 0.000000E+00 6.000000E-01 -4.039956E-03 4.199981E-01 4.328179E-03 1.850002E-01
0.000000E+00 5.999959E-01 -4.039956E-03 4.199981E-01 4.328179E-03 1.850002E-01

Figure 8-3 Partial Output from Transient Analysis with a Triangular Pulse

***** PARTIAL OUTPUT FROM bd08bar8.f06 *****

POINT-ID = 11
C O M P L E X D I S P L A C E M E N T V E C T O R
(MAGNITUDE/PHASE)

FREQUENCY TYPE T1 T2 T3 R1 R2 R3
.0 G 2.016639E-20 4.838324E-19 4.085192E-04 .0 2.024146E-04 3.208819E-19
.0 .0 .0 .0 180.0000 .0
5.000000E-01 G 2.136733E-20 5.467976E-19 4.323836E-04 .0 2.134119E-04 3.549737E-19
359.4255 359.0918 359.4344 .0 179.4676 359.2066
.
.
2.073000E+00 G 4.228285E-19 1.831657E-16 8.402211E-03 .0 3.874130E-03 8.302084E-17
303.5255 169.4034 303.5720 .0 123.7533 170.2924
2.087000E+00 G 4.789151E-19 1.747242E-16 9.514135E-03 .0 4.382166E-03 7.877973E-17
288.0493 145.7774 288.0963 .0 108.2796 146.6783
2.100632E+00 G 5.003335E-19 1.561973E-16 9.936970E-03 .0 4.572148E-03 7.006467E-17
270.0177 122.0074 270.0652 .0 90.2504 122.9200
2.224000E+00 G 1.560747E-19 1.943312E-17 3.091950E-03 .0 1.408677E-03 8.308275E-18
199.3224 32.8670 199.3742 .0 19.5776 33.8900
2.347000E+00 G 7.916779E-20 5.971262E-18 1.564160E-03 .0 7.050781E-04 2.427647E-18
190.2228 19.2229 190.2792 .0 10.5024 20.3631
.
.
9.950000E+01 G .0 .0 5.272466E-07 .0 8.710860E-07 1.024487E-21
.0 .0 180.4391 .0 .6610 180.7005
1.000000E+02 G .0 .0 5.216914E-07 .0 8.615909E-07 1.013485E-21
.0 .0 180.4361 .0 .6566 180.6966

X Y - O U T P U T S U M M A R Y ( R E S P O N S E )

SUBCASE CURVE FRAME XMIN-FRAME/ XMAX-FRAME/ YMIN-FRAME/ X FOR YMAX-FRAME/ X FOR

ID TYPE NO. CURVE ID. ALL DATA ALL DATA ALL DATA YMIN ALL DATA YMAX
1 DISP 1 11( 5,--) 1.000000E-01 1.000000E+02 5.216914E-07 1.000000E+02 9.936970E-03 2.100632E+00
0.000000E+00 1.000000E+02 5.216914E-07 1.000000E+02 9.936970E-03 2.100632E+00
1 DISP 1 11(--, 11) 1.000000E-01 1.000000E+02 1.804361E+02 1.000000E+02 3.594344E+02 5.000000E-01
0.000000E+00 1.000000E+02 0.000000E+00 0.000000E+00 3.594344E+02 5.000000E-01

Figure 8-4 Partial Output from Frequency Response Analysis

Main Index
 CHAPTER 8 239
Restarts In Dynamic Analysis

***** PARTIAL OUTPUT FROM bd08bar9.f06 *****

P R O J E C T V E R S I O N D I R E C T O R Y P R I N T
PROJECT_ID ASSIGNED INT. VALUE VERSION_ID CREATION TIME
---------- ------------------- ---------- -------------
" B L A N K " 1 ** 1 9/ 2/93 13:22.20
** 2 9/ 2/93 13:23. 0
3 9/ 2/93 13:23.25
** 4 9/ 2/93 13:23.59
** 5 9/ 2/93 13:24.28
** 6 9/ 2/93 13:24.57
7 9/ 2/93 13:25.26

Figure 8-5 Partial Output from a Database Directory Run

Main Index
240

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Plotted Output
9
■ Overview
■ Structure Plotting

■ X-Y Plotting

Main Index
242

9.1 Overview
Plotted output is important in verifying your model and understanding its results. Plots show
information in a format that is much easier to interpret than printed output. Plots are especially
important for dynamic analysis because the analysis can produce voluminous output. For
example, consider a transient response analysis for which there are 1000 output time steps, 100
grid points of interest, and 10 elements of interest. Printed output is too large to interpret
efficiently and effectively, and it does not easily show the time variation. Plotted output
overcomes these problems.
There are two kinds of plotted output: structure plots and X-Y plots. Structure plots can depict
the entire structure or a portion of it. Structure plots are useful for verifying proper geometry
and connectivity. They also can be used to show the deformed shape or stress contours at a
specified time or frequency. X-Y plots, on the other hand, show how a single-response quantity,
such as a grid point displacement or element stress, varies across a portion or all of the time or
frequency range.
There are numerous commercial and in-house plotting programs that interface to MSC.Nastran
for structure and/or X-Y plotting. Popular commercial programs include MSC/ARIES, PDA
PATRAN, and SDRC I-DEAS, among others. Details about these programs can be obtained from
their companies.
The commercial programs are similar because they operate interactively. The MSC.Nastran plot
capabilities, on the other hand, are performed as a batch operation, which means that you
predefine your plots when you make your MSC.Nastran run.
This chapter briefly describes the kinds of plots available in MSC.Nastran. Detailed information
is located in the MSC.Nastran Reference Manual.

Main Index
 CHAPTER 9 243
Plotted Output

9.2 Structure Plotting

Structure plotting is performed to verify the model’s geometry and element connectivity prior
to performing a dynamic analysis.
After (or during) the analysis, structure plotting is performed to view deformed shapes and
contours. For dynamic response, deformed shape and contour plots can be made for normal
modes analysis (for which there is a plot, or set of plots, per mode), frequency response analysis
(for which there is a plot, or set of plots, per output frequency), and transient response analysis
(for which there is a plot, or set of plots, per output time).
Structure plot commands are described in the MSC.Nastran Reference Manual. In the
MSC.Nastran input file, structure plotting commands are listed in the OUTPUT(PLOT) section,
which immediately precedes the Bulk Data Section. The structure plotting commands define the
set of elements to be plotted (SET), the viewing axes (AXES), the viewing angles (VIEW) as well
as the plot type and parameters (PLOT). Optionally, the scale of the plotted deformation
(MAXIMUM DEFORMATION) can be specified; if not specified, the plotted deformation is
scaled such that the maximum deformation is 5% of the maximum dimension of the structure.
Listing 9-1 shows the structure plotting commands applied to a normal modes analysis of the
bracket model. Figure 9-1 shows the resulting structure plots. The first plot shows the
undeformed shape, and the next two plots show the undeformed shape overlaid on the
deformed shapes for modes 1 and 2. The default is chosen such that the maximum plotted
deformation is 5% of the maximum dimension of the bracket; the actual maximum deformation
is printed at the top of the plot. The plots shown in this chapter were converted to PostScript
format for printing on a PostScript-compatible printer.

Listing 9-1 Normal Modes Structure Plot Commands for the Bracket Model

$ PLOT COMMANDS FOR BRACKET NORMAL MODES ANALYSIS

... Executive, Case Control ...
$
$ OUTPUT REQUESTS
DISPLACEMENT(PLOT) = ALL
$
$ STRUCTURE PLOTS
OUTPUT(PLOT)
CSCALE = 3.0
SET 333 = ALL
AXES MX,MY,Z
VIEW 20.,20.,20.
FIND SCALE, ORIGIN 5, SET 333
$ PLOT UNDEFORMED SHAPE
PLOT SET 333, ORIGIN 5
$ PLOT DEFORMED, UNDEFORMED SHAPES
PLOT MODAL DEFORMATION 0,1
PLOT MODAL DEFORMATION 0,2
$
BEGIN BULK
$
... Bulk Data ...
$
ENDDATA

Main Index
244

Undeformed Shape

Max. Def. = 3.82

Mode 1

Max. Def. = 5.06

Mode 2

Figure 9-1 Normal Modes Structure Plots for the Bracket Model

Listing 9-2 shows the structure plotting commands applied to a modal frequency response
analysis of the cantilever beam model. The displacements are computed in magnitude/phase
form. PARAM,DDRMM,-1 and PARAM,CURVPLOT,1 are required in the Bulk Data to create
structure plots at specified frequencies. (Note that PARAM,DDRMM,-1 generally increases the
amount of computer time and is not recommended unless otherwise required.) Figure 9-2
shows the resulting plots. The first plot shows the undeformed shape, and the next plot shows
Main Index
 CHAPTER 9 245
Plotted Output

the magnitude of response at 2.05 Hz, which is overlaid on the undeformed shape. The default
is chosen such that the maximum plotted deformation is 5% of the length of the bar; the actual
maximum deformation is printed at the top of the plot.

Listing 9-2 Frequency Response Structure Plot Commands for the Bar Model – Magnitude/Phase

$ PLOT COMMANDS FOR BAR MODAL FREQUENCY RESPONSE

$ --- MAGNITUDE, PHASE ---
... Executive, Case Control ...
$
$ OUTPUT REQUESTS
$ MAGNITUDE, PHASE
DISPLACEMENT(PHASE,PLOT) = ALL
$
$ STRUCTURE PLOTS
OUTPUT(PLOT)
$ DEFINE ELEMENTS IN PLOT SET
SET 333 = ALL
$ PLOT AXES: R=-Z, S=X, T=Y
AXES MZ,X,Y
VIEW 0.,0.,0.
FIND SCALE, ORIGIN 5, SET 333
$ PLOT UNDEFORMED SHAPE
PLOT SET 333, ORIGIN 5
$ PLOT DEFORMED SHAPE AT 2.05 HZ
PLOT FREQUENCY DEFORMATION 0 RANGE 2.05,2.051 MAGNITUDE SET 333
$
BEGIN BULK
$
$ REQUIRED FOR "FREQUENCY FROZEN" STRUCTURE PLOTS
PARAM,DDRMM,-1
PARAM,CURVPLOT,1
$
... rest of Bulk Data ...
$
ENDDATA

Undeformed Shape

Max. Def. = 0.394

Frequency 2.05

Figure 9-2 Frequency Response Structure Plots for the Bar Model -- Magnitude/Phase

The same plots are regenerated except that now the displacements are in real/imaginary format
(the default). Listing 9-3 shows the plot commands. Note that the imaginary component is
selected by PHASE LAG -90. Figure 9-3 shows the resulting plots. The default is chosen such
that the maximum plotted deformation is 5% of the length of the bar; the actual maximum
deformation is printed at the top of the plot.

Main Index
246

Listing 9-3 Frequency Response Structure Plot Commands for the Bar Model -- Real/Imaginary

$ PLOT COMMANDS FOR BAR MODAL FREQUENCY RESPONSE

$ --- REAL, IMAGINARY ---
... Executive, Case Control ...
$
$ OUTPUT REQUESTS
$ REAL, IMAGINARY
DISPLACEMENT(PLOT) = ALL
$
$ STRUCTURE PLOTS
OUTPUT(PLOT)
$ DEFINE ELEMENTS IN PLOT SET
CSCALE = 1.8
SET 333 = ALL
$ PLOT AXES: R=-Z, S=X, T=Y
AXES MZ,X,Y
VIEW 0.,0.,0.
FIND SCALE, ORIGIN 5, SET 333
$ PLOT UNDEFORMED SHAPE
PLOT SET 333, ORIGIN 5
$ PLOT DEFORMED SHAPE--REAL
PLOT FREQ DEFORM 0 RANGE 2.05,2.051 SET 333
$ PLOT DEFORMED SHAPE--IMAGINARY
PLOT FREQ DEFORM 0 RANGE 2.05,2.051 PHASE LAG -90. SET 333
$
BEGIN BULK
$
$ REQUIRED FOR "FREQUENCY FROZEN" STRUCTURE PLOTS
PARAM,DDRMM,-1
PARAM,CURVPLOT,1
$
... rest of Bulk Data ...
$
ENDDATA

Undeformed Shape

Max. Def. = 0.185

Frequency 2.05 Phase 0.

Max. Def. = 0.348

Frequency 2.05 Phase -90.00

Figure 9-3 Frequency Response Structure Plots for the Bar Model -- Real/Imaginary

Listing 9-4 shows the structure plotting commands applied to a modal transient response
analysis of the cantilever beam model. PARAM,DDRMM,-1 is required in the Bulk Data in order
to create structure plots at various times and/or frequencies. (Note that PARAM,DDRMM,-1
generally increases the amount of computer time and is not recommended unless otherwise
Main Index
 CHAPTER 9 247
Plotted Output

required.) Figure 9-4 shows the resulting plots. The plots show the displacements at the
following times: 0.25, 0.50, and 0.75 seconds, which are overlaid on the undeformed shape. The
default is chosen such that the maximum plotted deformation is 5% of the length of the bar; the
actual maximum deformation is printed at the top of the plot. For this case it may be better to
specify a maximum deformation so that the plots will show relative amplitudes.

Listing 9-4 Transient Response Structure Plot Commands for the Bar Model

$ PLOT COMMANDS FOR BAR MODAL TRANSIENT RESPONSE

... Executive, Case Control ...
$
$ OUTPUT REQUESTS
DISPLACEMENT(PLOT) = ALL
$
$ STRUCTURE PLOTS
OUTPUT(PLOT)
$ DEFINE ELEMENTS IN PLOT SET
SET 333 = ALL
$ PLOT AXES: R=-Z, S=X, T=Y
AXES MZ,X,Y
VIEW O.,O.,O.
FIND SCALE, ORIGIN 5, SET 333
$ PLOT DEFORMED SHAPE AT TIMES 0.25,0.5,0.75
PLOT TRANSIENT DEFORMATION 0 RANGE 0.25,0.251 SET 333
PLOT TRANSIENT DEFORMATION 0 RANGE 0.50,0.501 SET 333
PLOT TRANSIENT DEFORMATION 0 RANGE 0.75,0.751 SET 333
$
BEGIN BULK
$
$ REQUIRED FOR "TIME FROZEN" STRUCTURE PLOTS
PARAM,DDRMM,-1
$
... rest of Bulk Data ...
$
ENDDATA

Max. Def. = 0.040

Time 0.25

Max. Def. = 0.069

Time 0.50

Max. Def. = 0.057

Time 0.75
Figure 9-4 Transient Response Structure Plots for the Bar Model

Main Index
248

Although structure plotting is best performed in an interactive environment outside of

MSC.Nastran, the batch structure plotting capability in MSC.Nastran is nevertheless a useful
tool for model verification and results processing. The batch plotting capability can save time
and effort when many plots are required for a model that is run repeatedly.

Main Index
 CHAPTER 9 249
Plotted Output

9.3 X-Y Plotting

X-Y plots are used to display frequency and transient response results where the x-axis is
frequency or time and the y-axis is any output quantity. Unlike structure plotting, which is often
performed in an interactive environment, X-Y plotting is ideal for the MSC.Nastran batch
environment due to the large volume of data.
X-Y plot commands are contained in the OUTPUT(XYPLOT) section that immediately precedes
the Bulk Data Section. You define the titles (XTITLE and YTITLE) and plots (XYPLOT). You can
specify the plots to be generated in log format (XLOG and YLOG), and you can specify different
line styles. You can also specify that the plots are to be made in pairs (with a top and a bottom
plot), which is particularly useful for frequency response results when you want to display
magnitude/phase or real/imaginary pairs. Details about X-Y plotting and its commands are
located in the MSC.Nastran Reference Manual.
Element force and stress component numbers are also described in the manual. The X-Y plot
commands use numbers to identify a single component of grid point and element data. Note
that frequency response requires the use of complex force and stress components.
Once a good set of X-Y plot commands is established, it is wise to use this set repeatedly. The
examples that follow provide a good starting point.
Listing 9-5 shows X-Y plot commands for a modal frequency response analysis of the cantilever
beam model. Plots are made in pairs in magnitude/phase format. The “t” as the second letter
in XTGRID, YTGRID, YTLOG, and YTTITLE corresponds to the top plot of each pair; the letter
“b” corresponds to the bottom plot (XBGRID, YBGRID, etc.) Plots are made for the applied loads
(OLOAD) at grid points 6 and 11, displacements (DISP) of grid points 6 and 11, bending moment
(ELFORCE) at end A in plane 1 for element 6, and modal displacements (SDISP) for modes 1 and
2. For CBAR elements, force component 2 is the bending moment component at end A in plane
1 for real or magnitude output, and force component 10 is the similar component for imaginary
or phase output. Figure 9-5 shows the resulting plots. The plots in this chapter were converted
to PostScript format for printing on a PostScript-compatible printer.
The X-Y plotter makes reasonable choices of upper and lower bounds for the axes for both the x-
and y-axes on most plots. If it does not, the bounds can be fixed with the XMIN, XMAX, YMIN,
YMAX, and their variations for half-frame curves. Instances where setting bounds explicitly
results in better plots include the following situations:

• If you expect a variable to be constant or vary only slightly but want to plot it to confirm
that it indeed does not vary, set the YMIN and YMAX to include the expected value,
but separate them by at least 10 percent of their average value.
For this case the automatic bound selection chooses bounds very close to each other to
make the data fill up the plot. This selection causes the bounds to be nearly equal and
magnifies the scale of the plot orders of magnitude larger than other plots made of
varying functions. The response appears to be erratic when, in fact, it is smooth within
engineering criteria. The extreme cases occur when the function varies only in its last
digit. Then the function appears to be oscillating between the upper and lower limits
or it can even cause a fatal error due to numerical overflow when it attempts to divide
numbers by the difference YMAX-YMIN.
Main Index
250

• The automatic bound selector tends to round up the bounds to integer multiples of 100.
When plotting phase angles, bounds that cause grid lines at 90 degrees are more
readable. For example, for a plot that traverses the range of 0 to 360 degrees, the usual
selected bounds are 0 to 400 degrees. If you prefer to have grid lines drawn at integer
multiples of 90 degrees, set YMIN to 0.0 and YMAX at 360.0 degrees.
• When plotting log plots, any bounds you input may be rounded up or down to a value
that the plotter considers more reasonable. In general, changing bounds on log plots
sometimes requires experimentation before a reasonable set can be found.
• Producing good quality plots is an interactive process whether the plot is produced by
an interactive or batch plotter. The restart feature discussed in “Restarts In Dynamic
Analysis” in Chapter 8 can reduce the computer costs for this iteration since restarts
performed to change only plot requests are made efficiently.

Main Index
 CHAPTER 9 251
Plotted Output

Listing 9-5 X-Y Plot Commands for the Bar Frequency Response Analysis

$ X-Y PLOT COMMANDS FOR BAR MODAL FREQUENCY RESPONSE

$ --- MAGNITUDE, PHASE ---
...Executive, Case Control ...
$
$ APPLIED LOAD OUTPUT
SET 15 = 6,11
OLOAD(PHASE,PLOT) = 15
$
$ PHYSICAL OUTPUT REQUEST
SET 11 = 6,11
DISPLACEMENT(PHASE,PLOT) = 11
$
$ MODAL SOLUTION SET OUTPUT
SET 12 = 1,2
SDISP(PHASE,PLOT) = 12
$
$ ELEMENT FORCE OUTPUT
SET 13 = 6
ELFORCE(PHASE,PLOT) = 13
$
OUTPUT(XYPLOT)
XTGRID = YES
YTGRID = YES
XBGRID = YES
YBGRID = YES
$
$ PLOT RESULTS
XTITLE = FREQUENCY
$
YTLOG = YES
YTTITLE = DISPL. MAG. 6
YBTITLE = DISPL. PHASE 6
XYPLOT DISP /6(T2RM,T2IP)
YTTITLE = DISPL. MAG. 11
YBTITLE = DISPL. PHASE 11
XYPLOT DISP /11(T2RM,T2IP)
$
YTTITLE = SDISP. MAG. MODE 1
YBTITLE = SDISP. PHASE MODE 1
$ XYPLOT SDISP /mode(T1)
XYPLOT SDISP /1(T1RM,T1IP)
YTTITLE = SDISP. MAG. MODE 2
YBTITLE = SDISP. PHASE MODE 2
XYPLOT SDISP /2(T1RM,T1IP)
$
YTTITLE = BEND. MOMENT A1 EL. 6 MAG.
YBTITLE = BEND. MOMENT A1 EL. 6 PHASE
XYPLOT ELFORCE /6(2,10)
$
YTLOG = NO
YBMAX = 90.0
YBMIN = 0.0
CURVELINESYMBOL = -2
YTMAX = 4.0
YTMIN = 0.0
YTTITLE = LOAD MAG. 6
YBTITLE = LOAD PHASE 6
XYPLOT OLOAD /6(T2RM,T2IP)

YTMAX = 8.0
YTMIN = 0.0
YTTITLE = LOAD MAG. 11
YBTITLE = LOAD PHASE 11
XYPLOT OLOAD /11(T2RM,T2IP)
$
BEGIN BULK
$
... Bulk Data ...
$
ENDDATA
Main Index
252

1.0

Displ. Mag. 6
1.0E-5
400
Displ. Phase 6

1.0
Displ. Mag. 11

1.0E-4
400
Displ. Phase 11

0
0 Frequency (Hz) 20
Figure 9-5 X-Y Plots for the Bar Frequency Response Analysis

Main Index
 CHAPTER 9 253
Plotted Output

1.0

Sdisp. Mag. Mode1

1.0E-4
200
Sdisp. Phase Mode1

1.0E-2
Sdisp. Mag. Mode 2

1.0E-4
360
Sdisp. Phase Mode 2

160
0 Frequency (Hz) 20
Figure 9-5 X-Y Plots for the Bar Frequency Response Analysis (continued)

Main Index
254

1.0E3

Bend. Moment A1 El. 6 Mag.

1.0E-1

400
Bend. Moment A1 El. 6 Phase

0
0 Frequency (Hz) 20
Figure 9-5 X-Y Plots for the Bar Frequency Response Analysis (continued)

Main Index
 CHAPTER 9 255
Plotted Output

4.0

Load Mag. 6
0
100
Load Phase 6

0
8.0
Load Mag. 11

0
100
LoadP hase 11

0
0 20
Frequency (Hz)

Figure 9-5 X-Y Plots for the Bar Frequency Response Analysis (continued)
Listing 9-6 shows X-Y plot commands for a modal transient response analysis of the cantilever
beam model. Plots are made for the applied loads (OLOAD) at grid points 6 and 11,
displacements (DISP) of grid points 6 and 11, accelerations (ACCE) for grid points 6 and 11,
bending moment (ELFORCE) at end A in plane 1 for element 6, and modal displacements
(SDISP) for modes 1 and 2. Figure 9-6 shows the resulting plots.

Main Index
256

Listing 9-6 X-Y Plot Commands for the Bar Transient Response Analysis

$ X-Y PLOT COMMANDS FOR BAR MODAL TRANSIENT RESPONSE

... Executive, Case Control ...
$
$ PHYSICAL OUTPUT REQUEST
SET 11 = 6,11
DISPLACEMENT(PLOT) = 11
ACCELERATION(PLOT) = 11
$
$ MODAL SOLUTION SET OUTPUT
SET 12 = 1,2
SDISP(PLOT) = 12
$
$ ELEMENT FORCE OUTPUT
SET 13 = 6
ELFORCE(PLOT) = 13
$
$ APPLIED LOAD OUTPUT
SET 15 = 6,11
OLOAD(PLOT) = 15
$
OUTPUT(XYPLOT)
XGRID = YES
YGRID = YES
$
$ PLOT RESULTS
XTITLE = TIME
$
YTITLE = DISPL. 6
XYPLOT DISP /6(T2)
YTITLE = DISPL. 11
XYPLOT DISP /11(T2)
YTITLE = ACCEL. 6
XYPLOT ACCE /6(T2)
YTITLE = ACCEL. 11
XYPLOT ACCE /11(T2)
$
YTITLE = SDISP. 1
XYPLOT SDISP /1(T1)
YTITLE = SDISP. 2
XYPLOT SDISP /2(T1)
$
YTITLE = BEND. MOMENT A1, EL. 6
XYPLOT ELFORCE /6(2)
$
YTITLE = LOAD 6
XYPLOT OLOAD /6(T2)
YTITLE = LOAD 11
XYPLOT OLOAD /11(T2)
$
BEGIN BULK
$
... Bulk Data ...
$
ENDDATA

Main Index
 CHAPTER 9 257
Plotted Output

0.03

Displ. 6
0.0

-0.02

0.08
Displ. 11

0.0

-0.06
0.0 2.0
Time (sec)

Figure 9-6 X-Y Plots for the Bar Transient Response Analysis

Main Index
258

4.0

0.0

Accel. 6

-6.0

10.0
Accel. 11

0.0

-15.0
0.0 2.0
Time (sec)
Figure 9-6 X-Y Plots for the Bar Transient Response Analysis (continued)

Main Index
 CHAPTER 9 259
Plotted Output

25.0

Bend. Moment A1, El. 6

0.0

-20.0
0.0 2.0
Time (sec)

Figure 9-6 X-Y Plots for the Bar Transient Response Analysis (continued)

Main Index
260

0.08

Sdisp. 1
0.0

-0.10

0.0015
Sdisp. 2

0.0

-0.0010
0.0 2.0
Time (sec)
Figure 9-6 X-Y Plots for the Bar Transient Response Analysis (continued)

Main Index
 CHAPTER 9 261
Plotted Output

3.0

Load 6 0.0

-3.0

6.0
Load 11

0.0

-6.0
0.0 2.0
Time (sec)
Figure 9-6 X-Y Plots for the Bar Transient Response Analysis (continued)

Main Index
262

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Guidelines for Effective Dynamic Analysis
10
■ Overview
■ Overall Analysis Strategy
■ Units
■ Mass

■ Damping
■ Boundary Conditions
■ Loads
■ Meshing
■ Eigenvalue Analysis
■ Frequency Response Analysis
■ Transient Response Analysis
■ Results Interpretation and Verification
■ Computer Resource Requirements

Main Index
288

10.1 Overview
Dynamic analysis is more complicated than static analysis because of more input (mass,
damping, and time- and frequency-varying loads) and more output (time- and frequency-
varying results). Results from static analysis are usually easier to interpret, and there are
numerous textbook solutions for static analysis that make it relatively easy to verify certain static
analyses. Nevertheless, the guidelines in this chapter help you to perform dynamic analysis in
a manner that will give you the same level of confidence in the dynamic results that you would
have with static results.
This chapter covers the following topics:

• Overall analysis strategy

• Units
• Mass
• Damping
• Boundary conditions
• Loads
• Meshing
• Eigenvalue analysis
• Frequency response analysis
• Transient response analysis
• Results interpretation
• Computer resource estimation

Main Index
 CHAPTER 10 289
Guidelines for Effective Dynamic Analysis

10.2 Overall Analysis Strategy

Part of any analysis strategy, whether it be for dynamic analysis or static analysis, is to gain
confidence with the modeling procedures first. The best way to accomplish this is to run small,
simple models, preferably models that have textbook solutions. The references (see “References
and Bibliography” on page 499) provide numerous textbook solutions. Start with a simple
model first and then gradually add complexity, verifying the results at each stage. Follow the
steps outlined below and in Figure 10-1.
Once you have confidence in a small model and are ready to analyze your actual model, again
do the analysis in steps. The following is a suggested order for performing dynamic analysis on
any structure:
1. Create the initial model only; do not apply any loads. Verify the model's connectivity,
element and material properties, and boundary conditions. Use a graphical tool such
as MSC/PATRAN to assist in this. Make sure that mass is specified for this model.
2. Perform a static analysis (SOL 101) first in order to verify proper load paths and overall
model integrity. (Note that you have to constrain the structure for static analysis even
if you were not planning to do so for dynamic analysis.) For a three-dimensional
model, you should run three load cases, each with a 1g gravity load applied in a
different direction. Compute displacements and SPC forces, and verify the results.
Check for unusually large grid point displacements and unreasonable SPC forces. The
use of a graphical postprocessor can aid you at this step.
Next, apply static loads that have the same spatial distribution that your subsequent
dynamic loads have. Verify the results for reasonableness. Do not go to dynamic
analysis until you are satisfied with the results from your static analysis.
It is recommended at this stage that the model contain PARAM,GRDPNT,n (where n
is a reference grid point or 0, the origin of the basic coordinate system). Verify the
results from the grid point weight generator in order to ensure that the model's rigid-
body mass and inertia look reasonable. This step, in conjunction with the static analysis
results, helps to ensure that the proper mass units are specified.
3. Perform an eigenvalue analysis (SOL 103) next. Compute only a few modes first, verify
their frequencies, and view their mode shapes for reasonableness. If your graphical
postprocessor can animate the mode shapes, do so because that helps you to visualize
them. Things to check at this step are local mode shapes, in which one or a few grid
points are moving a very large amount relative to the rest of the model (this can
indicate poor stiffness modeling in that region), and unwanted rigid-body modes
(which can arise due to improper specification of the boundary conditions or a
mechanism).
Once you are satisfied with these results, perform the full eigenvalue analysis (for as
many modes as you need).
4. If you have frequency-dependent loads, perform frequency response analysis (SOL 108
or SOL 111) using the dynamic load spatial distribution. If your structure is
constrained, then apply the dynamic load at only one frequency, which should be 0.0
Hz. Compare the 0.0 Hz displacement results to the static analysis displacement
Main Index
290

results. The results should be the same if direct frequency response (without structural
damping) is used. If the results are not equal, then there is probably an error in the
specification of the dynamic load, and you should check the LSEQ and DAREA entries.
If modal frequency response (without structural damping) is used, then the 0.0 Hz
results should be close to the static results; the difference is due to modal truncation.
Next apply the load across the entire frequency range of interest. If you are running
modal frequency response, then make sure that you have enough modes to ensure
accurate results for even the highest forcing frequency. Also be sure to have a small
enough ∆f in order to accurately capture the peak response. Verify these results for
reasonableness (it may be easier to look at magnitude and phase results instead of real
and imaginary results, which are the default values).
If your ultimate goal is a transient response analysis for which damping is to be
neglected, then the frequency response analysis can also omit damping. However, if
damping is to be included, then use the correct damping in your frequency response
analysis. The proper specification of damping can be verified by looking at the half-
power bandwidth.
Plots are important at this stage to assist in results interpretation. X-Y plots are
necessary in order to see the variation in response versus frequency. Deformed
structure plots at a frequency near a resonant frequency can also help to interpret the
results. If structure plots are made, look at the imaginary component because the single
degree-of-freedom (SDOF) displacement response at resonance is purely imaginary
when damping is present (this response does not occur in practice because the response
is usually due to several modes).
5. If you have time-dependent loads, perform transient response (SOL 109 or SOL 112)
analysis. If your structure is constrained, apply the load "very quickly" (over one or
two time steps) as a step function and look at the displacement results. The duration
of the analysis needs to be as long as the period of the lowest frequency mode. For an
SDOF system, a quickly applied load results in a peak displacement response that is
twice the response resulting from the same load applied statically. This peak response
does not occur in your actual model because of multiple modes and damping, but the
results should be close. If your structure is unconstrained, the displacements will grow
with time (unless the rigid-body modes are excluded in a modal transient response
analysis). However, the stresses should be roughly twice those from the static analysis.
In any event, examine the results via X-Y plots to ensure reasonableness.
Once you are satisfied, apply the correct time variation to the load and compute the
results. Again, use X-Y plots to verify the accuracy of the results.
6. Finally, perform any other dynamic analyses, such as response spectrum analysis,
random response, nonlinear transient response, or dynamic response optimization.
The confidence gained by using the previous steps first helps to ensure that you have
an accurate model at this stage.

Main Index
 CHAPTER 10 291
Guidelines for Effective Dynamic Analysis

These and other guidelines are described further in the remainder of this chapter.

Verify Model Connectivity,

Properties, Boundary Conditions

Verify Static Analysis

(with gravity load, with actual load)

Verify Normal Modes Analysis

(few modes, many modes)

Verify Frequency Response Analysis

(0.0 Hz, entire frequency range)

Verify Transient Response Analysis

(step load, actual load)

Run Other Dynamic Response Analyses

(response spectrum, random response, etc.)

Figure 10-1 Simplified Flow Chart of the Overall Analysis Strategy

Main Index
292

10.3 Units
Mistakes in units and boundary conditions (see “Boundary Conditions” on page 297) are
among the most common mistakes made when performing finite element analysis.
MSC.Nastran does not assume a particular set of units, but it does assume that they are
consistent.
It is up to you to ensure that the units that you use are both consistent and correct. Table 2-2 lists
consistent units for common variables.
There are several ways to verify units. For mass, you can print the results from the grid point
weight generator and verify that the mass is correct. For stiffness, you can apply a simple load
and verify that the resulting static displacements seem reasonable. For both, you can verify that
the natural frequencies are reasonable.
These checks assume that you have enough knowledge about your structure to know when the
results are reasonable and when a mistake has been made. In other words, running small
models and/or proceeding through dynamic analysis via the steps outlined in the previous
section are necessary in order to be confident that the correct units are specified.

Main Index
 CHAPTER 10 293
Guidelines for Effective Dynamic Analysis

10.4 Mass
Mistakes with mass primarily involve mistakes in mass units as described in the previous
section. A common mistake is to mix mass and weight units. Using PARAM,WTMASS does not
solve this problem because it scales all mass and weight input (except certain types of direct
input matrices) and still leaves the mixture of units. Also, the use of PARAM,WTMASS can have
unwanted effects as well because it also scales the large mass used in enforced motion, thereby
scaling the value of the enforced motion input. In order to reduce the chances for error, it is
recommended that only mass units be used everywhere. Therefore, avoid the use of weight
input wherever possible as well as the use of PARAM,WTMASS.
One way to verify mass input is to apply gravity in several directions and print the SPC forces.
This verification ensures that the mass is correctly applied and that the units are correct.

Main Index
294

10.5 Damping
Proper specification of damping is probably the most difficult modeling input to verify because
its verification can only be done via dynamic response analysis. In addition, its units are not
necessarily familiar because damping is normally not a modeling input that you use frequently,
unless dynamic response analysis is all you run. Also, there are several ways to specify damping
in MSC.Nastran, which increase the chances of making a mistake. Finally, even though the
damping units are correct and the damping is input correctly in MSC.Nastran, it is difficult to
know that the damping specification itself is correct from a physical standpoint. While there are
relatively easy checks for correctness in mass and stiffness input, including comparison to static
and modal test data, there are no such easy checks for damping input.
The easiest way to specify damping is to use modal damping, which is often specified as the
percentage of critical damping. The TABDMP1 Bulk Data entry is used to specify modal
damping. The following are several potential mistakes associated with using modal damping.

• Forgetting to select the TABDMP1 entry using the SDAMPING Case Control
command.
• Forgetting to specify the damping TYPE (field 3). The default damping type is
structural damping, which is twice the critical damping percentage.
• Not making the table cover an adequate range of frequencies. Like almost all of the
MSC.Nastran tables, the TABDMP1 entry extrapolates beyond the endpoints of the
table by using the first two or the last two entered values. As a rule, you should try to
provide enough input points so that the table lookup point is always an interpolated
value.
Modal damping can be used only in modal frequency response and modal transient response.
Other forms of damping have to be used for the direct methods of response. For frequency
response analysis, GE (field 9 of the MAT1 entry) and PARAM,G define structural damping.
These variables are also used to specify structural damping for transient response analysis but
are not activated unless PARAM,W3 and PARAM,W4 are set to nonzero values. A common
mistake is to forget to set these values.
Note that damping is additive, that is, the damping matrix is comprised of all of the input
damping sources. For example, damping due to CDAMPi elements adds to damping due to
PARAM,G and GE. Mixing of damping types can increase the chances for error, and you are
cautioned against the mixing of damping types unless it is necessary.
In many cases damping is not an important consideration. For example, a structure's peak
response due to an impulsive load is relatively unaffected by damping since the peak response
occurs during the first cycle of response. Damping in a long duration transient excitation, such
as an earthquake, can make a difference in the peak response on the order of 10 to 20% or so, but
this difference is small when compared to the other modeling uncertainties. Therefore, it is often
conservative to ignore damping in a transient response analysis.

Main Index
 CHAPTER 10 295
Guidelines for Effective Dynamic Analysis

For frequency response analysis, however, the value of damping is critical to the value of the
computed response at resonance since the dynamic magnification factor is inversely
proportional to the damping value. One way to verify the accuracy of the modal damping ratio
input is to run the modal frequency response across the half-power bandwidth of the modal
frequencies of interest as shown in Figure 10-2.

Peak Response

Peak
------------ = Half-Power Point
2
Response

f1 fn f2 Frequency

Half-Power
Bandwidth

Figure 10-2 Half-Power Bandwidth

For lightly damped structures ( ζ < 0.1 ), the approximate relationship between the critical
damping ratio , the half-power bandwidth ( f 2 – f 1 ), and the resonant frequency f n is

f2 – f1
ζ ≈ ---------------- Eq. 10-1
2f n

Another approximate way to verify damping is to run transient response analysis and look at
the successive peak values of the free vibration response as shown in Figure 10-3. In direct
transient response analysis, this plot can be generated by using a displacement initial condition;
in modal transient response analysis, it can be generated by applying an impulsive force that
ramps up and down quickly with respect to the dominant period of response. The logarithmic
decrement ∆ is the natural log of the amplitude ratio of two successive cycles of free vibration
response given by

x( n – 1 )
∆ = ln ------------------ Eq. 10-2
xn

The logarithmic decrement provides an approximate damping relationship for lightly damped
structures as given by

∆
ζ ≈ ------ Eq. 10-3
2π

In transient response analysis, remember to use PARAM,W3 or PARAM,W4 to include

structural damping if GE or PARAM,G is used.

Main Index
296

Figure 10-3 Damped Free Vibration Response

Both the half-power bandwidth method and the logarithmic decrement method assume an
SDOF response. These approximations are less accurate when there are multiple modes of
response; however, they are useful for verifying that the damping input is within a factor of two
or three of the desired damping.

Main Index
 CHAPTER 10 297
Guidelines for Effective Dynamic Analysis

10.6 Boundary Conditions

The proper specification of boundary conditions is just as important for dynamic analysis as it is
for static analysis. The improper specification of the boundary conditions leads to incorrect
answers. One such improper specification of boundary conditions is forgetting to fully constrain
the structure. Unlike static analysis, for which an under-constrained model does not run, an
under-constrained model does run in dynamic analysis. You should perform an eigenvalue
analysis first and verify that there are no unwanted rigid-body modes.
The large mass used for enforced motion simulates a constrained condition as well as adds a
rigid-body mode. The value of the large mass is important for obtaining accurate answers. The
large mass value must be large enough to properly simulate the constrained condition, but it
must not be so large as to create numerical difficulties when solving the equations. A range of
3 6
10 to 10 times the overall structural mass is adequate to meet both conditions. One way to
verify that a proper value is chosen is to run a normal modes analysis with the enforced DOF
constrained via SPCs. Then, run a normal modes analysis with your choice of the large mass(es)
and compare the frequencies of the flexible modes. If the frequencies compare favorably (i.e., to
within four or five significant digits), then the large mass value is accurate. If the frequencies do
not compare, then increase the value of the large mass(es). A common mistake is to use too low
a value (or omit it entirely) for the rotational components. Using the grid point weight generator
is very important to obtain the overall structural mass and inertias so that you can specify a good
value for the large mass(es).

Main Index
298

10.7 Loads
Because of their time- or frequency-varying nature, it is more complicated to apply dynamic
loads than it is to apply static loads. Therefore, it is important to verify that the dynamic loads
are correctly specified and that there are no discontinuous loads.
The best way to verify proper dynamic load specification is to plot the loads as a function of time
or frequency. Another way to verify proper dynamic load specification is to print the loads as a
function of time or frequency. Use the OLOAD Case Control command to plot or print the loads.

Main Index
 CHAPTER 10 299
Guidelines for Effective Dynamic Analysis

10.8 Meshing
An adequate element mesh is required to create an accurate model. For static analysis the mesh
density is primarily controlled by the load paths; the element mesh must be fine enough so that
there is a smooth transition of stress from one element to another in the region of interest.
Load paths are also important for dynamic analysis, but there is an additional consideration: the
mesh must be fine enough to accurately represent the deformed shape of the modes of interest.
If many modes are to be considered in the analysis, then the model must be fine enough to
accurately represent the mode shapes of the highest modes of interest. Table 10-1 shows the
frequencies resulting from several cantilever beam models; each column represents the same
structure but with a different number of elements. Note that the frequencies are closer to the
theoretical results as the model fineness in increased. In addition, note that the error is greatest
in the higher modes. This table shows the computed frequencies compared to theory; it does not
show the computed mode shapes compared to theory. However, the comparison for mode
shapes shows even more error than is shown for the frequencies.
A general rule is to use at least five to ten grid points per half-cycle of response amplitude.
Figure 10-4 shows the theoretical response shape for the fifth mode. Note that there are four
half-cycles in the mode shape, which means that 20 to 40 grid points at a minimum are
recommended to accurately represent that mode shape. This modeling guideline is also
reflected in Table 10-1, which shows that the 40-element model is much more accurate than the
10-element model.

Table 10-1 Frequencies for a Cantilever Beam Model (Lumped Mass)

Frequencies (Hz)
Mode 10-Element 40-Element 70-Element 100-Element
Theory
Model Model Model Model
1 2.08 2.07 2.08 2.08 2.08
2 13.24 12.81 13.00 13.01 13.01
3 36.47 35.50 36.37 36.41 36.42
4 71.52 68.84 71.22 71.34 71.36
5 118.21 112.53 117.65 117.90 117.95

Four Half-Cycles

Figure 10-4 Fifth Mode Shape of a Cantilever Beam

Main Index
300

Another way to verify the accuracy of the mesh density is to apply static loads that give a
deformed shape the same as the mode of interest and perform stress discontinuity calculations.
This process can be laborious and is not recommended as a general checkout procedure.

Main Index
 CHAPTER 10 301
Guidelines for Effective Dynamic Analysis

10.9 Eigenvalue Analysis

In addition to the meshing guidelines described earlier, the other primary factor in eigenvalue
analysis is the proper selection of the eigenvalue solution method. The Lanczos method is the
best overall method and is the recommended method to use. The automatic Householder
method is useful for small, dense matrices. The SINV method is useful when only a few modes
are required. The other methods should be regarded as backup methods.

Carefully examine the computed frequencies and mode shapes. Viewing only one or the other
is usually not enough to verify accuracy of your model. Modes with 0.0 Hz (or computational
zero) frequencies indicate rigid-body or mechanism modes. If these modes are unintended, then
there is a mistake in boundary conditions or connectivity. The existence and cause of
unintended zero-frequency modes can also be ascertained from the mode shapes. In addition,
mode shape plots are useful for assessing local modes, in which a group of one or a few grid
points displaces and the rest of the structure does not. Local modes may also be unintended and
are often the result of incorrect connectivity or element properties.

Main Index
302

10.10 Frequency Response Analysis

Several factors are important for computing accurate frequency response results. These factors
include the number of retained modes (for modal frequency response analysis), the frequency
increment ∆f , and damping. These guidelines are only approximate but are nevertheless useful.
Running a normal modes analysis first helps to compute accurate frequency response results.

Number of Retained Modes

Use enough modes to cover the range of excitation. For example, if the model is to be excited
from 1 to 100 Hz, you must use all of the modes with frequencies up to at least 100 Hz. This is
only a minimum requirement, however. A better guideline is to use enough modes to cover two
to three times the range of excitation in order to provide accurate answers at the high end of the
frequency range. For example, when excitation is applied to 100 Hz, modes with frequencies up
to 200 to 300 Hz should all be used.

Size of the Frequency Increment

The size of the frequency increment ∆f must be small enough to ensure that the magnitude of
the peak response is accurately computed. To ensure this, you need to choose a frequency
increment small enough so that there are at least five to ten increments within the half-power
bandwidth frequencies (illustrated in Figure 10-2).
The frequency increment is defined by ∆f on the FREQ1 Bulk Data entry. Note that FREQ and
FREQ2 entries can be used in conjunction with FREQ1 to define more solutions in the areas of
resonance (the frequencies of these solutions should have been determined by a prior normal
modes analysis). A nonuniform ∆f imposes no cost increase relative to a uniform ∆f .

Relationship of Damping to the Frequency Increment

The response at resonance is inversely proportional to the amount of damping, but the half-
power bandwidth is directly proportional to the amount of damping. For lightly-damped
structures ( ζ < 0.1 ), an approximate relationship between the half-power bandwidth ( f 2 – f 1 ),
resonant frequency f n , and critical damping ratio ζ is given by Eq. 10-1. This equation can be
rewritten to define the appropriate value of ∆f :

f 2 – f 1 = ( m – 1 )∆f Eq. 10-4

f2 – f1 ( m – 1 )∆f
ζ ≈ ---------------- = -------------------------- Eq. 10-5
2f n 2f n

2ζf n
∆f = -------------------- Eq. 10-6
(m – 1)

where m is the number of frequency points within the half-power bandwidth. For example, with
1% critical damping and six points used to define the half-power bandwidth, the maximum
frequency increment is 0.004 ⋅ f n . The frequency increment is smaller for lighter damped
structures.
Main Index
 CHAPTER 10 303
Guidelines for Effective Dynamic Analysis

Another good check is looking at the X-Y plots. If the response curves are not smooth, there is a
good chance that ∆f is too large.

Verification of the Applied Load

The applied load can be verified by exciting your model at 0.0 Hz and comparing the results to
a static solution with the same spatial load distribution. The 0.0 Hz results should match the
static results if direct frequency response analysis is used without structural damping. If modal
frequency response analysis is used without structural damping, those results should be close to
the static results; any difference is due to mode truncation. If the 0.0 Hz results do not match the
static results, check the LSEQ and DAREA entries. Also, use OLOAD to print the applied force
in order to compare the loads.

Main Index
304

10.11 Transient Response Analysis

Several factors are important in computing accurate transient response. These factors include:
the number of retained modes (for modal transient response), the integration time step ∆t , the
time duration of the computed response, and damping. The guidelines are only approximate
but are nevertheless useful. Running a normal modes analysis first helps to compute transient
response.

Number of Retained Modes

In modal transient response analysis, a larger number of modes produces a more accurate
response (although at the expense of increased run times). The number of modes must be large
enough to cover the frequency range of interest. (The term "range of interest" means the range
of frequencies whose response is to be computed as well as the range of frequencies for which
the loading is applied.) As a general rule, there should be enough modes to cover a range up to
about two times the highest frequency of interest. For example, if response needs to be
computed to 100 Hz, then enough modes should be used in the transient response to cover up
to at least 200 Hz. As another general rule, you should use enough modes to cover a range up
to two to ten times the dominant frequency of response.
A sufficient number of modes must be retained to cover the time and spatial distribution of
loading. For example, if the applied transient load has a spatial distribution resembling the fifth
mode shape, then that mode should be included in the transient response.

Size of the Integration Time Step

The value of the integration time step ∆t , denoted by DT on the TSTEP Bulk Data entry, is
important for the accurate integration of the equations of motion. Rough guidelines for the
selection of ∆t are as follows:

• ∆t must be small enough to accurately capture the magnitude of the peak response,
which means that at least ten time steps per cycle of the highest mode be used. For
example, if the highest frequency of interest is 100 Hz, then ∆t should be 0.001 second
or smaller.
• ∆t must be small enough to accurately represent the frequency content of the applied
loading. If the applied loading has a frequency content of 1000 Hz, then ∆t must be
0.001 second or less (preferably much less in order to represent the applied loading
with more than one point per cycle).
The integration time step can be changed in a transient response analysis, but it is not
recommended. Much of the cost of direct transient response occurs with the decomposition of
the dynamic matrix, which occurs only once if there is a constant ∆t . A new decomposition is
performed every time ∆t changes, thereby increasing the cost of the analysis if the time
increment is changed. Therefore, the recommendation is to determine the smallest ∆t required
based on the loading and frequency range and then use that ∆t throughout the entire transient
response analysis.

Main Index
 CHAPTER 10 305
Guidelines for Effective Dynamic Analysis

Duration of the Computed Response

The length of the time duration is important so that the lowest flexible (e.g., non-rigid body)
mode oscillates through at least one cycle. For example, if the lowest flexible mode has a
frequency of 0.2 Hz, then the time duration of the computed response should be at least
5.0second. A better guideline is to make the duration the longer of the following: twice the
period of the lowest flexible mode or one period of the lowest flexible mode after the load has
been removed or reached a constant value. The time duration is set on the TSTEP entry by
multiplying (the integration time step) by the number of time steps (N).

Value of Damping
The proper selection of the damping value is relatively unimportant for analyses that are of very
short duration, such as a crash impulse or a shock blast. The specification of the damping value
is most important for long duration loadings (such as earthquakes) and is critical for loadings
(such as sine dwells) that continually add energy into the system.

Verification of the Applied Load

The applied load can be verified by applying the load suddenly (over one or two time
increments) and comparing the results to a static solution with the same spatial load
distribution. The transient results should have a peak value of twice the static results. If the
transient results are not twice the static results, check the LSEQ and DAREA entries. Another
way to verify the applied load is to inspect it visually via the X-Y plots.

Main Index
306

10.12 Results Interpretation and Verification

Because of the time- and frequency-varying nature of dynamic analysis, results can be more
difficult to interpret than for static analysis. The key to proper results interpretation is plotting.
Structure plotting is necessary for a proper understanding of the deformed shapes of the modes,
and X-Y plotting is necessary for a proper understanding of the frequency and transient
response.

Comparing results to hand calculations, to known results from similar models, or to test data is
also very useful. Do not accept any analysis results without first performing some type of
verification.
Consider the cantilever beam model shown in Figure 10-5. This is a planar model of the
cantilever beam used in the examples in Chapter 3 through Chapter 7. Figure 10-5 shows the
loads applied to the beam.

6 11

1N 2N

Figure 10-5 Cantilever Beam Model with Static Loads

A static analysis (SOL 101) is run first. Then, modal frequency response (SOL 111) is run from
0.0 to 20.0 Hz with a damping ratio of 2% critical damping used for all modes. Modes are
computed up to 500 Hz using the Lanczos method. Finally, modal transient response (SOL112)
is run with the time variation shown in Figure 10-6. Modes are computed up to 3000Hz (using
the Lanczos method), and a damping ratio of 2% critical damping is used for all modes.

1.0

Load
Factor

0.003 0.006
Time (sec)

Figure 10-6 Time Variation of Transient Loads

Table 10-2 shows the y-displacements for grid points 6 and 11. As expected, the frequency
response results at 0.0 Hz are the same as the static analysis results (see “Frequency Response
Analysis” on page 302). The frequency response results at 2.05 Hz (very near the first natural
frequency) are approximately 25 times the static analysis results. The factor of 25 is the dynamic
amplification factor at resonance for a damping ratio of 2% critical damping.

Main Index
 CHAPTER 10 307
Guidelines for Effective Dynamic Analysis

1 1
Amplification Factor at Resonance = ------ = ---------------- = 25
2ζ 2 ( .02 )

The transient response results are approximately twice the static analysis results. The factor of
two is the amplification of response for a transient load applied suddenly (see “Transient
Response Analysis” on page 304).

Table 10-2 Comparison of Results for the Cantilever Beam Model

Frequency Response
Static Analysis Analysis Results Transient Response
Results Peak Displacement
0.0 Hz 2.05 Hz
Displacement 6 Y 3.17E-3 3.17E-3 7.61E-2 6.29E-3
Displacement 11 Y 9.77E-3 9.77E-3 2.25E-1 1.87E-2

Main Index
308

10.13 Computer Resource Requirements

The efficiency of a dynamic analysis can be measured based on computer resource requirements,
such as computer runtime, disk storage, and memory. These requirements increase as the
problem size increases, but they can also vary greatly depending on other factors.
In general, a dynamic analysis uses more computer resources than a static analysis. In normal
modes analysis, the requirements vary depending on the density of the model, the eigenvalue
extraction method used, the number of modes requested, and the type of decomposition
(symmetric or sparse).
If your model has already been built, estimate its resource requirements by using the
new ESTIMATE utility. For instructions on the use of the ESTIMATE utility, see Section
6.1 of the MSC.Nastran Configuration and Operations Guide for Version 69. If your
model has not been created, but you wish to estimate resource usage, then read the
following.
In the past, we have established benchmark runs to provide guidelines on performance for
normal modes analyses. We used a cylindrical plate model scaled to various sizes and solved
for ten modes using different solution methods. Testing has shown that this model provides a
very good guideline for most industry models.
Since the previous edition of this book established that the Sparse Lanczos method is
far superior in performance than other methods, we now only use this one method for
the benchmark runs. Using MSC.Nastran Version 69 on an SGI R10000 workstation, we
applied the Sparse Lanczos method to models of various sizes and extracted data on
memory and disk space requirement.
The memory requirements increase with problem size. Empirical formulas were developed for
these benchmark problems to estimate the memory requirement given the number of degrees-
of-freedom (DOF) contained in the model.
For the equation

Log(mem) = (.7934) ⋅ Log(DOF) – 2.3671 Eq. 10-7

mem The memory required in megabytes.

DOF The number of degrees-of-freedom contained in the model.

Note that these memory requirements are for no spill conditions. The symmetric solver can run
with less memory, but because of spill conditions, the performance is degraded in both
computer runtime and disk space usage.
Empirical formulas were developed for these benchmark problems to estimate the disk space
requirement given the number of degrees-of-freedom (DOF) contained in the model.
For the equation

Log(space) = (1.0815) ⋅ Log(DOF) – 2.5411 Eq. 10-8

Main Index
 CHAPTER 10 309
Guidelines for Effective Dynamic Analysis

space The disk space usage in megabytes.

DOF Represents the number of degrees-of-freedom contained in the model.

These values were obtained from a 32-bit computer and therefore should be doubled for a 64-bit
computer. If you wish to calculate more than ten modes, then multiply the space calculated
above by 1.18^(#modes/10 - 1). For example, for 30 modes you would multiply the space value
calculated above by 1.18^2.

Main Index
310

Main Index
 MSC.Nastran Basic Dynamics User’s GuideMSC.Nastran Basic Dynamics User’s Guide
MSC.Nastran Basic Dynamics User’s Guide

CHAPTER
Advanced Dynamic Analysis Capabilities
11
■
■ Dynamic Reduction
■ Complex Eigenvalue Analysis
■ Response Spectrum Analysis

■ Random Vibration Analysis

■ Mode Acceleration Method
■ Fluid Structure Interaction
■ Nonlinear Transient Response Analysis
■ Superelement Analysis
■ Design Optimization and Sensitivity
■ Control System Analysis
■ Aeroelastic Analysis
■ DMAP

Main Index
312

11.1 Overview
The previous chapters describe the most common types of dynamic analysis: normal modes
analysis, frequency response analysis, and transient response analysis. However, MSC.Nastran
contains numerous additional types of dynamic analyses; many of these types are documented
in the MSC.Nastran Advanced Dynamic Analysis User’s Guide.
The advanced dynamic analysis capabilities include

• Dynamic reduction
• Complex eigenanalysis
• Response spectrum analysis
• Random response analysis
• Mode acceleration method
• Fluid-structure interaction
• Nonlinear transient response analysis
• Superelement analysis
• Design sensitivity and optimization
• Control systems
• Aeroelastic analysis
• DMAP
These capabilities are described briefly in the remainder of this chapter.

Main Index
 CHAPTER 11 313
Advanced Dynamic Analysis Capabilities

11.2 Dynamic Reduction

Dynamic reduction is an optional numerical approach that can be used in MSC.Nastran to
reduce a dynamic model to one with fewer degrees-of-freedom. Typically, the intent of dynamic
reduction is to increase the efficiency of a dynamic solution by working with smaller matrices
while maintaining the dynamic characteristics of the system.
Dynamic reduction is used for a number of reasons. One possible reason may be that a particular
model may be too large from a computer resource standpoint (computer runtime and/or disk
space) to be solved without using reduction. A related issue is that the model may have more
detail than required. Many times dynamic analyses are performed on models that were initially
created for detailed static stress analyses, which typically require refined meshes to accurately
predict stresses. When these static models are used in a dynamic analysis, the detailed meshes
result in significantly more detail than is reasonably required to predict the natural frequencies
and mode shapes accurately.
Two methods are available in MSC.Nastran for the dynamic reduction of the eigenequation
prior to modal extraction: static condensation (also called Guyan reduction) and generalized
dynamic reduction (GDR). Once the natural frequencies and mode shapes are calculated for the
reduced model, these modes can be used in the transient or frequency response solution process.

Static Condensation. In the static condensation process you select a set of dynamic DOFs
called the a-set; these are the retained DOFs that form the analysis set. The complementary set
is called the o-set and is the set of DOFs that are omitted from the dynamic analysis through a
reduction process. The process distributes the o-set mass, stiffness, and damping to the a-set
DOFs by using a transformation that is based on a partition of the stiffness matrix (hence the
term static condensation). This reduction process is exact only for static analysis, which leads to
approximations in the dynamic response. The a-set DOFs are defined by the ASET or ASET1
Bulk Data entires, and the o-set DOFs are defined by the OMIT or OMIT1 Bulk Data entries.

Generalized Dynamic Reduction (GDR). Generalized dynamic reduction (GDR) offers a

more accurate solution by including an approximation of the dynamic effects when forming the
transformation. Bulk Data entries for GDR are DYNRED, QSET, and QSET1, and SPOINT. In
addition, the Case Control command DYNRED is also required.
It is emphasized that dynamic reduction is an optional technique and is best left to the dynamic
analysis specialist. Whereas dynamic reduction was required in the days of small computer
memory and disk space, now it is no longer required due to increased computer resources and
better eigenanalysis methods (in particular, the Lanczos method).

Main Index
314

11.3 Complex Eigenvalue Analysis

Complex eigenvalue analysis is used to compute the damped modes of structures and assess the
stability of systems modeled with transfer functions (including servomechanisms and rotating
systems).
Complex eigenvalue analysis solves for the eigenvalues and mode shapes based on the
following equation in operator notation:

2
[ p M + pB + K ] { u } = 0 Eq. 11-1

where: p = α + iω
This equation is similar to that for normal modes analysis (see “Real Eigenvalue Analysis” on
page 33) except that damping is added and the eigenvalue is now complex. In addition, the
mass, damping, and stiffness matrices may be unsymmetric, and they may contain complex
coefficients.
Complex eigenvalue analysis is controlled with the EIGC Bulk Data entry (similar to the EIGRL
or EIGR Bulk Data entries for normal modes analysis). There are four methods of solution:
upper Hessenberg, complex Lanczos, determinant search, and inverse power. Complex
eigenvalue analysis is available as a direct method (SOLs 107, 67, or 28), in which the equations
are of the same size as the number of physical variables. Complex eigenvalue analysis is also
available as a modal method (SOLs 110, 70, and 29), in which undamped modes are first
computed and then are used to transform the matrices from physical to modal variables.

Main Index
 CHAPTER 11 315
Advanced Dynamic Analysis Capabilities

11.4 Response Spectrum Analysis

Response spectrum analysis is an approximate method of computing the peak response of a
transient excitation applied to a simple structure or component. This method is used in civil
engineering to predict the peak response of a component on a building that is subjected to
earthquake excitation; it is also used in aerospace engineering to predict the peak response of
equipment in a spacecraft that is subjected to an impulsive load due to stage separation. Because
it is an approximate method, response spectrum analysis is often used as a design tool. Response
spectrum analysis is also called shock spectrum analysis.
There are two parts to response spectrum analysis: (1) generation of the spectrum and (2) use of
the spectrum for dynamic response such as stress analysis. Both are available in MSC.Nastran.
Figure 11-1 depicts the generation of a response spectrum.

(d)
f1 f2 f3 fmax (e)
(c)
... Peak
(b)
Response

Series of f1 f2 f3 . . . fmax
(a) Transient
Response Oscillators Resonant Frequency

Base
Transient Structure
Excitation
Transient excitation (a) is applied to a base structure (b), from which
transient response (c) is computed for each floor. This response is
applied to a series of oscillators (d), for which the peak response is
plotted (e). Steps (d) and (e) are repeated for different damping values
to form response spectra as shown below.
= 5% critical damping
= 2% critical damping
= 0% critical damping
Peak
Response

Resonant Frequency (f)

Figure 11-1 Response Spectrum Generation

Note that the peak response for one oscillator does not necessarily occur at the same time as the
peak response for another oscillator. Also note that there is no phase information since only the
magnitude of peak response is computed. It is assumed in this process that each oscillator mass
is very small relative to the base structural mass so that the oscillator does not influence the
dynamic behavior of the base structure.

Main Index
316

Once a spectrum is computed, it can be used for the dynamic response analysis of an
MSC.Nastran model of the component. For example, the spectrum generated for a floor in a
building that is subjected to an earthquake can then be applied to a complex model of a piece of
equipment attached to that floor. The peak response of each mode of the equipment model is
obtained from the spectrum, and these peak modal responses are combined to create the overall
response.
Because the peak responses do not all occur at the same time and only the magnitude of peak
responses are computed, various methods are used to combine the peak responses into the
overall response. The combination methods implemented in MSC.Nastran are SRSS (square
root of the sum of the squares), ABS (absolute values), and NRL (U.S. Navy shock design modal
summation). The typical response quantities computed are grid point displacements and
element stresses.

Main Index
 CHAPTER 11 317
Advanced Dynamic Analysis Capabilities

11.5 Random Vibration Analysis

Random vibration is vibration that can be described only in a statistical sense. The instantaneous
magnitude is not known at any given time; rather, the magnitude is expressed in terms of its
statistical properties (such as mean value, standard deviation, and probability of exceeding a
certain value).
Examples of random vibration include earthquake ground motion, ocean wave heights and
frequencies, wind pressure fluctuations on aircraft and tall buildings, and acoustic excitation
due to rocket and jet engine noise. These random excitations are usually described in terms of a
power spectral density (PSD) function.
MSC.Nastran performs random response analysis as a postprocessing step after frequency
response analysis. The frequency response analysis is used to generate the transfer function,
which is the ratio of the output to the input. The input PSD multiplies the transfer function to
form a response PSD. The input PSD can be in the form of auto- or cross-spectral densities.
Random response output consists of the response PSD, autocorrelation functions, number of
zero crossings with positive slope per unit time, and RMS (root-mean-square) values of
response.

Main Index
318

11.6 Mode Acceleration Method

The mode acceleration method is an alternate form of data recovery for modal frequency
response and modal transient response. The mode acceleration method accounts for the higher
truncated modes to give more accurate answers than either the mode displacement or matrix
methods (see “Modal Frequency Response Analysis” on page 108 and “Modal Transient
Response Analysis” on page 156), particularly if the number of retained modes is small
compared to the number of physical degrees-of-freedom.
The higher modes respond in a quasi-static manner to lower frequency excitation. Therefore, the
inertia and damping forces contain little contribution from these higher modes. In the mode
acceleration method the inertia and damping forces are first computed from the modal solution.
These forces are then combined with the applied forces and are used to obtain more accurate
displacements by performing a static analysis. The mode acceleration method provides more
accurate answers but with increased cost.
The mode acceleration method is selected by specifying PARAM,DDRMM,-1 and
PARAM,MODACC,0.

Main Index
 CHAPTER 11 319
Advanced Dynamic Analysis Capabilities

11.7 Fluid Structure Interaction

Four major methods are available in MSC.Nastran to model fluid effects. These methods are
described below.

Hydroelastic Analysis
Small motions of compressible or incompressible fluids coupled to a structure may be analyzed
with this option. The fluid is modeled with axisymmetric hydroelastic elements (CFLUIDi),
which may be connected to an arbitrary structure modeled with an axisymmetric wetted surface.
Each fluid point (RINGFL) on a cross section defines the scalar pressure, which is expanded to a
Fourier series around the circumference. Complex modes and frequency response solutions are
available for the coupled fluid-structure problems. Normal modes solutions are available for
fluid-only problems. All solutions may include gravity effects (i.e., sloshing) on a free surface.
This capability was developed specifically to analyze liquid-fueled booster rockets but may also
be useful for problems involving other types of axisymmetric storage tanks.

Virtual Fluid Mass

Small motions of incompressible fluids may be coupled directly to the structure with this option.
Fluids may be coupled to the interior and exterior surfaces (with infinite fluid boundaries).
There is no explicit fluid model; only the wetted structural elements (ELIST) have to be defined.
Although free surfaces are allowed, no gravity effects are included directly. Since the fluid is
represented by a coupled mass matrix attached directly to the structural points, this capability
is allowed in all dynamic solution sequences. This capability may be used to model a wide
variety of fluid-structure interaction problems. Some examples are fuel tanks, nuclear fluid
containers, drilling platforms, and underwater devices.

Coupled Acoustics
You may analyze the dynamics of compressible fluids contained within arbitrarily shaped
structures with the coupled fluid-structure method. You would model a three-dimensional fluid
with conventional solid elements (CHEXA, etc.) using acoustic property and material data. Each
grid point in the fluid defines the pressure at its location as its degree of freedom. The fluid is
automatically connected to the structure via the geometry and ACMODL Bulk Data inputs. You
can connect acoustic absorber elements (CHACAB) to the structural surfaces to simulate
soundproofing material. In addition, the CAABSF element is now available. This element acts
as a thin layer of acoustic absorbing material along the fluid-structure interface. Version 69
introduced several features, which are parallel to those available for structural analysis, such as
direct damping, modal damping and the ability to control the modes in a response analysis
through the use of parameter. You can define panels to provide integrated response data.
Effects of gravity, large motions, and static pressures are ignored. Complex eigenvalues,
frequency response, and transient response are the available solution sequences. Design
sensitivity and optimization processes may reference the acoustic outputs as responses with
appropriate design constraints. Applications for the coupled fluid-structure option are
automotive and truck interiors, aircraft cabins, and acoustic devices, such as loudspeakers and
microphones.
Main Index
320

Uncoupled Acoustics
Several methods are available in MSC.Nastran for the analysis of normal modes of compressible
fluids bounded by rigid containers and/or free surfaces. One method is the “acoustic cavity”
capability, which uses two-dimensional slot elements and axisymmetric ring elements to define
the fluid region. This method was specifically developed for the acoustic analysis of solid rocket
motor cavities. A better method is to use the three-dimensional fluid elements for the “coupled
acoustics” described above and provide the appropriate boundary conditions.

Main Index
 CHAPTER 11 321
Advanced Dynamic Analysis Capabilities

11.8 Nonlinear Transient Response Analysis

The analysis techniques described thus far are applicable for linear-elastic analysis for which the
mass, stiffness, and damping matrices are constant over time and constant for all values of
applied force. However, there are many cases for which the matrices are not constant, and these
cases must be solved with nonlinear analysis techniques.
Nonlinear analysis requires iterative solution methods, thereby making it far more
computationally intensive than a corresponding linear analysis. Nonlinear transient response
analysis is available in MSC.Nastran in SOLs 29 and 129 (preferred). Nonlinear problems are
classified into three broad categories: geometric nonlinearity, material nonlinearity, and contact.

Geometric Nonlinearity
Geometrically nonlinear problems involve large displacements; “large” means that the
displacements invalidate the small displacement assumptions inherent in the equations of linear
analysis. For example, consider a thin plate subject to an out-of-plane load. If the deflection of
the plate’s midplane is approximately equal to the thickness of the plate, then the displacement
is considered large, and a linear analysis is not applicable.
Another aspect of geometric nonlinear analysis involves follower forces. Consider a clamped
plate subject to a pressure load. As shown in Figure 11-2, the load has followed the plate to its
deformed position. Capturing this behavior requires the iterative update techniques of
nonlinear analysis.

Figure 11-2 Follower Forces on a Cantilever Beam

Material Nonlinearity
Material nonlinear analysis can be used to analyze problems where the stress-strain relationship
of the material is nonlinear. In addition, moderately large strain values can be analyzed.
Examples of material nonlinearities include metal plasticity, materials (such as soils and
concrete), and rubbery materials (where the stress-strain relationship is nonlinear elastic).
Various yield criteria, such as von Mises or Tresca (for metals) and Mohr-Coulomb or
Drucker-Prager (for frictional materials, such as soils or concrete), can be selected. Three
hardening rules are available in MSC.Nastran: isotropic hardening, kinematic hardening, or
combined isotropic and kinematic hardening. With such generality, most plastic material
behavior with or without the Bauschinger effect can be modeled. In addition, gaps can be used
to model the effects due to structural separation.

Main Index
322

Contact
Contact occurs when two or more structures (or portions of structures) collide. Contact can be
modeled as point-to-point contact (CGAP) or as contact along a line (BLSEG).

Nonlinear-Elastic Transient Response Analysis

There are numerous structures that contain nonlinear elastic elements. These elements possess
nonlinear force-deflection relationships, yet they are elastic since they load and unload along the
same force-deflection curve. Examples of the nonlinear elastic force-deflection curves are shown
in Figure 11-3. These types of elements are common in many mechanical and structural systems
including piping systems supported by nonlinear springs, chains (stiffness in tension only), or
constant force devices; base-mounted equipment and structure supported by nonlinear shock
isolation systems; structures with gaps (piping systems and buildings); and soil or concrete
structures (which only exhibit stiffness when in compression). Other systems exhibit
nonlinearities that are proportional to the velocity, such as seat belts and piping supports.
Force

Force
Displacement Displacement

(a) Bilinear Spring

(b) Preload Constant Force Device
with Hard Stop
Force

Force

Displacement Displacement

(c) Gap and Spring in Series (d) Stiffness in Compression Only

Force

Force

Displacement Displacement

(e) Softening System (f) Hardening System

Figure 11-3 Examples of Nonlinear Elastic Systems

Main Index
 CHAPTER 11 323
Advanced Dynamic Analysis Capabilities

There are several methods in MSC.Nastran for solving nonlinear elastic problems. A general
nonlinear solution scheme can be used, in which the material properties are specified as
nonlinear elastic (NLELAST on the MATS1 entry). Nonlinear element stiffness matrices are
generated, and equilibrium iterations and convergence checks are required at each load step.
The new CBUSH element, introduced in Version 69 adds some capabilities in this area, also. In
addition, gap elements (CGAP) can be used in conjunction with elastic elements to produce
systems with piece-wise linear force-deflection curves. This process also requires the additional
computations of nonlinear element stiffness matrix generation, equilibrium iteration, and
convergence testing. Nonlinear analyses requiring these additional computations can be
substantially more costly than an equivalent linear analysis even if the nonlinearities are
relatively few in number.
An efficient technique called the pseudoforce method exists in MSC.Nastran, in which the
localized nonlinearities are treated so that they require no additional computer time when
compared to a purely linear analysis. In this method, which is available only for transient
response solutions (SOLs 109 and 112, for example), deviations from linearity are treated as
additional applied loads. The dynamic equations of motion are written as
·· ·
[M]{u(t) } + [B]{u(t )} + [K]{u(t )} = {P(t )} + {N(t) } Eq. 11-2

where [ M ] , [ B ] , and [ K ] denote the system mass, damping, and stiffness matrices, respectively.
The vectors { P ( t ) } and { u ( t ) } denote applied nodal loads and system displacements,
respectively, as functions of time. The vector { N ( t ) } denotes the nonlinear forces, which are
added to the right-hand side of Eq. 11-2 (and hence are treated as additional applied loads).
The nonlinear forces are evaluated at the end of one time step for use in the successive time step.
The equations of motion therefore become the following:
·· ·
[ M ] { u ( t ) } + [ C ] { u ( t ) } + [ K ] { u ( t ) } = { P ( t ) } + { N ( t – ∆t ) } Eq. 11-3

Note that the nonlinear force lags the true solution by a time step, which may require using small
integration time steps (i.e., smaller than those required for a purely linear analysis). Eq. 11-3 can
be solved in physical or modal coordinates (the nonlinearity itself must be expressed in terms of
physical coordinates).
A nonlinear force can be used in conjunction with a linear elastic element to produce the desired
force-deflection curve as illustrated in Figure 11-4. The nonlinear dynamic force is formulated
using a NOLINi entry and a TABLEDi entry, which contains a force-versus-deflection table
describing the nonlinear force. For desired force-deflection curves more complicated than the
bilinear stiffness shown in the figure, the nonlinear force is made correspondingly more
complex.

Main Index
324

Force
Force
Force
= +
Displacement Displacement Displacement

Nonlinear = Elastic + Nonlinear

Element Element Force

Figure 11-4 Formulation of a Nonlinear Element

Nonlinear Normal Modes Analysis

There are times when normal modes need to be calculated for a nonlinear structure. For
inflatable, rotating, or tension structures, the stiffness—and therefore, the modes—depends
upon the applied load and deformation. A nonlinear analysis can be restarted into SOL 103 to
compute the normal modes of the deformed structure.

Main Index
 CHAPTER 11 325
Advanced Dynamic Analysis Capabilities

11.9 Superelement Analysis

Superelement analysis is a form of substructuring wherein a model is divided into pieces with
each piece represented by a reduced stiffness, mass, and damping matrix. Each superelement is
processed individually, and the solutions are then combined to solve the entire model. The final
analysis (in which all of the individual superelement solutions are combined) involves much
smaller matrices than would be required if the entire model were solved in a one-step process.
An example of a superelement model is shown in Figure 11-5.
Superelement analysis has the advantage of reducing computer resource requirements,
especially if changes are made to only one portion (superelement) of the model; in this case, only
the affected superelement needs to be reanalyzed and the final analysis repeated.
Superelement analysis presents procedural advantages as well, particularly when multiple
engineering contractors are involved in an analysis. Imagine a model of a rocket and payload:
one contractor models the booster, another contractor models the engines, and another
contractor models the satellite. Each contractor can reduce his model to its boundary
degrees-of-freedom, which is suitable for superelement analysis. The systems integrator then
combines these reduced models into one model for a liftoff analysis. Superelement analysis has
the advantage that matrices can be passed from one organization to another without revealing
proprietary modeling details or concern about whether the same superelement interior grid
point and element numbers are used by every participant.

Figure 11-5 Superelements Used to Model a Car Door

Component Mode Synthesis

Component mode synthesis (CMS) is a form of superelement dynamic reduction wherein
matrices are defined in terms of modal coordinates (corresponding to the superelement modes)
and physical coordinates (corresponding to the grid points on the superelement boundaries).
CMS is advantageous because there are fewer modal coordinates than physical coordinates—
perhaps only one percent as many. In addition, CMS can utilize modal test data, thereby
increasing the accuracy of the overall analysis.

Main Index
326

11.10 Design Optimization and Sensitivity

Design optimization is used to produce a design that possesses some optimal characteristics,
such as minimum weight, maximum first natural frequency, or minimum noise levels. Design
optimization is available in MSC.Nastran in SOL 200, in which a structure can be optimized
considering simultaneous static, normal modes, buckling, transient response, frequency
response, and aeroelastic analyses.

Possible analysis response types include:

• Weight
• Volume
• Eigenvalues
• Buckling load factor
• Static displacement, stress, strain, and element force
• Composite stress, strain, and failure criterion
• Frequency response displacement, velocity, acceleration, stress, and force
• Transient response displacement, velocity, acceleration, stress, and force
• Damping level in a flutter analysis
• Trim and stability derivative responses for static aeroelastic analysis
A frequent task is to minimize the total structural weight while ensuring that the design still
satisfies all of the design-related and performance-related constraints. The structural weight is
termed the design objective, which the optimizer attempts to minimize. The design variables are
the properties that can be changed to minimize the weight. The design variables shown in
Figure 11-6 are the height and width of the flange and web. Constraints on the design variables
may be a minimum allowable flange width w and a maximum allowable beam height h . With
the tip load shown in the figure, one performance constraint on the design may be the maximum
allowable transverse displacement at the tip, while another constraint may be the maximum
allowable bending stress. The design-related and performance-related constraints place limits
on the optimizer’s ability to reduce the weight.

P
tw

h
tf
l
w

Figure 11-6 Cantilever I-Beam Example

Main Index
 CHAPTER 11 327
Advanced Dynamic Analysis Capabilities

The optimizer in MSC.Nastran uses sensitivity analysis. The sensitivity matrix [ S ] defines the
change in response for a perturbation in design variable. A term in [ S ] is given by

∂R i
S ij = -------- Eq. 11-4
∂X j

where:

S ij = i, j term in [ S ]

∂ = partial derivative
R i = i-th response

X j = j-th design variable

The sensitivity is depicted in Figure 11-7, in which the sensitivity is the slope of the curve of
response versus the design variable value. Design optimization and design sensitivity are
described in greater depth in the MSC.Nastran Design Sensitivity and Optimization User’s
Guide.

Response R

Design Variable X

Figure 11-7 Response Sensitivity

Main Index
328

11.11 Control System Analysis

A control system provides feedback (output) to an input. In aircraft, sample control systems are
flap settings and actuator positions.
A control system involves a general input-output relationship called a transfer function.
Transfer functions are represented in MSC.Nastran by the TF Bulk Data entry. Transfer
functions can utilize grid points, extra points (EPOINT), and scalar points (SPOINT).
Nonlinearities can be simulated via the NOLINi entries with which nonlinear transient loads are
expressed as functions of displacements or velocities. Complex eigenvalue analysis can be used
to determine stability when control systems include damping and unsymmetric matrices.

Main Index
 CHAPTER 11 329
Advanced Dynamic Analysis Capabilities

11.12 Aeroelastic Analysis

MSC.Nastran provides efficient solutions of the problems of aeroelasticity, which is a branch of
applied mechanics that deals with the interaction of aerodynamic, inertial, and structural forces.
It is important in the design of airplanes, helicopters, missiles, suspension bridges, tall chimneys,
and power lines. Aeroservoelasticity is a variation in which the interaction of automatic controls
requires additional consideration.

The primary concerns of aeroelasticity include flying qualities (that is, stability and control),
flutter, and structural loads arising from maneuvers and atmospheric turbulence. Methods of
aeroelastic analysis differ according to the time dependence of the inertial and aerodynamic
forces that are involved. For the analysis of flying qualities and maneuvering loads wherein the
aerodynamic loads vary relatively slowly, quasi-static methods are applicable. The remaining
problems are dynamic, and the methods of analysis differ depending on whether the time
dependence is arbitrary (that is, transient or random) or simply oscillatory in the steady state.
MSC.Nastran considers two classes of problems in dynamic aeroelasticity: aerodynamic flutter
and dynamic aeroelastic response.

Aerodynamic Flutter
Flutter is the oscillatory aeroelastic instability that occurs at some airspeed at which the energy
extracted from the airstream during a period of oscillation is exactly dissipated by the hysteretic
damping of the structure. The motion is divergent in a range of speeds above the flutter speed.
Flutter analysis utilizes complex eigenvalue analysis to determine the combination of airspeed
and frequency for which the neutrally damped motion is sustained (see Figure 11-8).

0.2
1st Mode
2nd Mode
0.1
Damping, g

0.0

Flutter Divergence

-0.1

-0.2
0 500 1000 1500 2000
Velocity (ft/sec)
Figure 11-8 Flutter Stability Curve

Main Index
330

Three methods of flutter analysis are provided in MSC.Nastran: the American flutter method
(called the K method in MSC.Nastran), an efficient K method (called the KE method) for rapid
flutter evaluations, and the British flutter method (called the PK method) for more realistic
representation of the unsteady aerodyamic influence as frequency-dependent stiffness and
damping terms. Complex eigenvalue analysis is used with the K method, and the QR
transformation method is used with the KE and PK methods.

Dynamic Aeroelastic Response

The aeroelastic dynamic response problem determines the response of the aircraft to time- or
frequency-varying excitations. Atmospheric turbulence is the primary example of this type of
excitation, but wind shear and control surface motion can also have an aeroelastic component.
Methods of generalized harmonic (Fourier) analysis are applied to the linear system to obtain
the response to the excitation in the time domain. The gust response analysis may be regarded
either as a stationary random loading or as a discrete gust.
The gust analysis capability computes the response to random atmospheric turbulence, discrete
one-dimensional gust fields, and control surface motion and other dynamic loading. The
random response parameters calculated are the power spectral density, root mean square
response, and mean frequency of zero-crossings. The response to the discrete gust and control
surface motion is calculated by direct and inverse Fourier transform methods since the
oscillatory aerodynamics are known only in the frequency domain. The time histories of
response quantities are the output in the discrete case (see Figure 11-9).

3.0E8
Root Bending Moment (in-lbs)

-5.0E8
0 Time (sec) 8

Figure 11-9 Transient Response Resulting from a Gust

Main Index
 CHAPTER 11 331
Advanced Dynamic Analysis Capabilities

11.13 DMAP
MSC.Nastran provides a series of solution sequences (SOL 103 for normal modes analysis, for
example) written in their own language tailored to matrix manipulation. These solution
sequences consist of a series of DMAP (direct matrix abstraction program) statements.
MSC.Nastran’s DMAP capability enables you to modify these solution sequences and write your
own solution sequences.

DMAP is a high-level language with its own compiler and grammatical rules. DMAP statements
contain data blocks and parameters and operate on them in a specified manner. For example,
the DMAP statement
ADD U1,U2/U3 $
adds matrices U1 and U2 together and calls the output U3. The DMAP statement
MATPRN U3// $
prints the matrix U3.
Numerous DMAP Alters for dynamic analysis are provided (see the MSC.Nastran 2005 Release
Guide for further information). Many of these DMAP Alters are for dynamic analysis. There are
alters for model checkout, modal initial conditions, the addition of static results to transient
response results, frequency-dependent impedance, and modal test-analysis correlation, among
others. The MSC.Nastran Advanced Dynamic Analysis User’s Guide describes a DMAP Alter
for enforced motion using the Lagrange multiplier approach.
In summary, DMAP is a powerful capability that allows you to modify MSC.Nastran to satisfy
your needs.

Main Index
332

Main Index
 MSC.Nastran Basic Dynamics User’s GuideMSC.Nastran Basic Dynamics User’s Guide
MSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Glossary of Terms
A
■ Glossary of Terms

Main Index
334

A.1 Glossary of Terms

CMS Component mode synthesis.

Critical Damping The lowest value of damping for which oscillation does not occur
when the structure is displaced from its rest position. Values of
damping less than the critical damping value create an
underdamped system for which oscillatory motion occurs.
Damping Energy dissipation in vibrating structures.
DOF Degree-of-freedom.
Dynamic Amplification Ratio of dynamic response to static response, which is a function
Factor of the forcing frequency, natural frequency, and damping.
Forced Vibration Analysis Vibration response due to applied time-varying forces.
Free Vibration Analysis Vibration response when there is no applied force. Normal
modes analysis and transient response to initial conditions are
examples of free vibration analysis.
Frequency Response Computation of the steady-state response to simple harmonic
Analysis excitation.
GDR Generalized dynamic reduction.
LMT Lagrange multiplier technique.
Mechanism Mode Stress-free, zero-frequency motions of a portion of the structure.
A mechanism can be caused by an internal hinge.
Mode Shape The deformed shape of a structure when vibrating at one of its
natural frequencies.
Natural Frequency The frequency with which a structure vibrates during free
vibration response. Structures have multiple natural frequencies.
Phase Angle The angle between the applied force and the response. In
MSC.Nastran this angle is a phase lead, whereby the response
leads the force.
Positive Definite A matrix whose eigenvalues are all greater than zero.
Positive Semi-Definite A matrix whose eigenvalues are greater than or equal to zero.
Repeated Roots Two or more identical natural frequencies.
Resonance Large-amplitude vibrations that can grow without bound. At
resonance, energy is added to the system. This occurs, for
example, when the harmonic excitation frequency is equal to one
of the natural frequencies. The response at resonance is
controlled entirely by damping.

Main Index
 APPENDIX A 335
Glossary of Terms

Rigid-Body Drift Transient displacements of an unconstrained model that grow

continuously with time. This is often caused by the accumulation
of small numerical errors when integrating the equations of
motion.
Rigid-Body Mode Stress-free, zero-frequency motions of the entire structure.
SDOF Single degree-of-freedom.
Structural Damping Damping that is proportional to displacement.
Transient Response Computation of the response to general time-varying excitation.
Analysis
Viscous Damping Damping that is proportional to velocity.

Main Index
336

Main Index
 MSC.Nastran Basic Dynamics User’s GuideMSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Nomenclature for Dynamic Analysis
B
■ Nomenclature for Dynamic Analysis

Main Index
338

B.1 Nomenclature for Dynamic Analysis

The appendix provides nomenclature for terms commonly used in dynamic analysis.

General

Þ Multiplication
ª Approximately
[] Matrix
{} Vector
i –1
g Acceleration of Gravity
∆t Time Step

∆f Frequency Step

i, j, k Subscripts (Indices)

∞ Infinity

u Displacement

u(0) Initial Displacement

· Velocity
u
·
u(0) Initial Velocity
·· Acceleration
u

m Mass

m0 Large Mass

b Damping

b cr Critical Damping

k Stiffness

p Applied Force

ω Circular Frequency

ωn Circular Natural Frequency

2 Eigenvalue
ω

λ Eigenvalue

Main Index
 APPENDIX B 339
Nomenclature for Dynamic Analysis

f Frequency

fn Natural Frequency

Tn Period

ωd Damped Circular Natural Frequency

ζ Damping Ratio

Q Quality Factor

θ Phase Angle

∆ Logarithmic Decrement

Structural Properties
Geometry

L Length

A Area

Stiffness

E Young’s Modulus

G Shear Modulus

J Torsional Constant

ν Poisson’s Ratio

I Area Moment of Inertia

Mass

ρ Mass Density

ρm Mass Density

ρw Weight Density

Ip Polar Moment of Inertia

Main Index
340

Damping

G Overall Structural Damping Coefficient

GE Element Structural Damping Coefficient

Multiple Degree-of-Freedom System

{u} Displacement Vector

·
{u} Velocity Vector
··
{u} Acceleration Vector

{φ} Mode Shape Vector

{ φ rig } Rigid-Body Modes

ξi i-th Modal Displacement

· i-th Modal Velocity

ξi

·· i-th Modal Acceleration

ξi

mj j-th Generalized Mass

kj j-th Generalized Stiffness

λk k-th Eigenvalue

[M] Matrix

[ Mr ] Rigid-Body Mass Matrix

[K] Stiffness Matrix

[ KE ] Element Stiffness Matrix

[B] Damping Matrix

[P] Force Matrix

Main Index
 MSC.Nastran Basic Dynamics User’s Guidek

APPENDIX
The Set Notation System Used in Dynamic
C Analysis

■ Overview

Main Index
342

C.1 Overview
The set notation system used for dynamic analysis in MSC.Nastran continues and expands the
set notation system for static analysis. Because of the great variety of physical quantities and
displacement sets used in dynamic analysis, becoming familiar with the set notation system in
MSC.Nastran aids its efficient use for dynamic analysis.

Displacement Vector Sets

When you create a model in MSC.Nastran, equations are allocated for the purpose of assembling
the necessary structural equations. These equations are rows and columns in the matrix
equations that describe the structural behavior. Six equations are created per grid point, one
equation is created per SPOINT or EPOINT. The model details (elements, properties, loads, etc.)
are used to create the appropriate row and column entries in the matrices (e.g., stiffness
coefficients). Certain data entries (i.e., SPCs, MPCs, ASETs, etc.) cause matrix operations to be
performed in the various stages of the solution process. To organize the matrix operations,
MSC.Nastran assigns each DOF to displacement sets.
Most matrix operations used in a structural analysis involve partitioning, merging, and/or
transforming matrix arrays from one subset to another. All components of motion of a given
type form a vector set which is distinguished from other vector sets. A given component of
motion can belong to many sets.
In MSC.Nastran, there are two basic types of sets:

• Combined sets
• Mutually exclusive sets
Combined sets are formed through the union (i.e., combination) of two or more sets. Mutually
exclusive sets are important in the solution process because if a DOF is a member of a mutually
exclusive set, it cannot be a member of any of the other mutually exclusive sets. The importance
of this property is demonstrated as the sets are described below.
When MSC.Nastran starts to assemble the structural equations, it allocates six equations for each
grid point (GRID) and 1 equation for each scalar point (SPOINT). These equations are associated
with a displacement set defined as the g-set. Fundamentally, the g-set represents an
unconstrained set of structural equations.
The next step in the solution process is to partition the g-set into two subsets: the m-set and the
n-set. The dependent DOFs of all multipoint constraint relations (MPCs, RBEs, etc.) that define
the m-set are condensed into a set of independent DOFs (the n-set). The n-set represents all of
the independent DOFs that remain after the dependent DOFs were removed from the active set
of equations.
Using the n-set, SPCs are applied to the independent equations to further partition the
equations. The degrees-of-freedom defined on SPC entries define the DOFs in the s-set. When
you reduce the n-set by applying the s-set constraints, the f-set remains. The f-set is the "free"
DOF of the structure. At this stage of the solution, the f-set is comprised of the remaining
equations that represent a constrained structure. If the applied constraints are applied properly,
the f-set equations represent a statically stable solution.
Main Index
 APPENDIX C 343
The Set Notation System Used in Dynamic Analysis

If static condensation is to be performed, the f-set is partitioned into the o-set and the a-set. The
o-set degrees-of-freedom are those that are to be eliminated from the active solution through a
static condensation. The remaining DOFs reside in the a-set. The a-set is termed the "analysis"
set. The a-set is often the partition at which the solution is performed.
The a-set can be divided into physical degrees-of-freedom (the t-set) and generalized
coordinates (the q-set). This partitioning is used in dynamic analysis if generalized dynamic
reduction (GDR) is performed.
If the SUPORT entry is used, the degrees-of-freedom defined on the SUPORT entry are placed
in the r-set. When the r-set is partitioned from the t-set, the -set remains. This final set is termed
the "leftover" set and is the lowest level of partitioning performed in MSC.Nastran static
analysis. The l-set partition is the matrix on which the final solution is performed. Under special
circumstances, the l-set is divided into two types of DOFs. The DOFs that are held fixed in
component mode synthesis (CMS) or GDR are called b-set points, and those DOFs that are free
to move in CMS and GDR are called c-set points.
Table C-1 demonstrates the basic partitioning operations. When a particular set above has no
DOFs associated with it, it is a null set. Its partition is then applied, and the DOFs are moved to
the subsequent partition.

Table C-1 Basic Partitioning Operations

g-set – m-set = n-set

n-set – s-set = f-set
f-set – o-set = a-set
a-set – r-set = l-set

In the above table, the m-, s-, o-, and r-sets form the mutually exclusive sets. Physically, the
mutually exclusive set partitioning ensures that operations cannot be performed on DOFs that
are no longer active. For example, if you apply an SPC to a DOF which is a dependent degree-
of-freedom on an RBAR, a FATAL error is issued. Using an SPC entry moves a DOF to the s-set,
but this cannot occur if the same DOF is already a member of the m-set. Both the m-set and s-
set are mutually exclusive.
Table C-1 is a representation of all sets, set partitions, and set combinations used in
MSC.Nastran.
A number of additional, mutually independent sets of physical displacements (namely, q-set, c-
set, b-set and e-set) are used in dynamic analysis to supplement the sets used in static analysis.
The q-, c-, and b-sets facilitate generalized dynamic reduction and component mode synthesis.
The e-set is used to represent control systems and other nonstructural physical variables. The p-
set is created by combining the g-set with the e-set variables.
In addition to the combined sets described above, the v-set is a combined set created by
combining the c-, r- and o-sets. The DOFs in these sets are the DOFs free to vibrate during
component mode synthesis or generalized dynamic reduction.

Main Index
344

Some additional sets (sa-, k-, ps-, and pa-sets) represented in Figure C-1 are defined and used in
aeroelastic analysis.
The modal coordinate set ξ is separated into zero frequency modes ξo and elastic (finite
frequency) modes ξf . For dynamic analysis by the modal method, the extra points u e are added
to the modal coordinate set ξi to form the h-set as shown in Figure C-2.
The parameter PARAM,USETPRT can be used to print of lists of degrees-of-freedom and the sets
to which they belong. Different values of the PARAM create various tables in the printed output.

The supersets formed by the union of other sets have the following definitions.

Mutually
Exclusive Sets Combined Sets
m
sb
s
sg
o
g
q n p
ne ps
r f pa
a fe
c t d
l
b
e
sa
k

Figure C-1 Combined Sets Formed from Mutually Exclusive Sets

The set names have the definitions described in Table C-2.

Table C-2 Sets in MSC.Nastran

Mutually Exclusive Sets

Set Name Meaning
m Points eliminated by multipoint constraints.
sb* Points eliminated by single-point constraints that are included in
boundary condition changes and by the automatic SPC feature.
sg* Points eliminated by single-point constraints that are listed on
grid point Bulk Data entries.
o Points omitted by structural matrix partitioning.
q Generalized coordinates for dynamic reduction or component
mode synthesis.
r Reference points used to determine free body motion.
c The free boundary set for component mode synthesis or dynamic
reduction.
b Coordinates fixed for component mode analysis or dynamic
reduction.
Main Index
 APPENDIX C 345
The Set Notation System Used in Dynamic Analysis

Table C-2 Sets in MSC.Nastran (continued)

Mutually Exclusive Sets

Set Name Meaning
e Extra degrees-of-freedom introduced in dynamic analysis.
sa Permanently constrained aerodynamic points.
k Aerodynamic points.

* Strictly speaking, sb and sg are not exclusive with respect to one another.
Degrees-of-freedom may exist in both sets simultaneously. These sets are
exclusive, however, from the other mutually exclusive sets.

Table C-3 Sets in MSC.Nastran

Combined Sets (+ Indicates Union of Two Sets)

Set Name Meaning
s = sb + sg All points eliminated by single-point constraints.

l = b+c The structural coordinates remaining after the reference

coordinates are removed (points left over).
t = l+r The total set of physical boundary points for superelements.

a = t+q The set assembled in superelement analysis.

d = a+e The set used in dynamic analysis by the direct method.

f = a+o Unconstrained (free) structural points.

fe = f + e Free structural points plus extra points.

n = f+s All structural points not constrained by multipoint

constraints.
ne = n + e All structural points not constrained by multipoint
constraints plus extra points.
g = n+m All structural (grid) points including scalar points.

p = g+e All physical points.

ps = p + sa Physical and constrained aerodynamic points.

pa = ps + k Physical set for aerodynamics.

fr = o + l Statically independent set minus the statically determinate

supports ( o + l = f – q – r ).

v = o+c+r The set free to vibrate in dynamic reduction and component

mode synthesis.

Main Index
346

Physical Set
e

Modal Sets h

ξo
ξi
ξf

where:

ξ o = rigid-body (zero frequency) modal coordinates

ξ f = finite frequency modal coordinates

ξ i = ξ o + ξ f = the set of all modal coordinates

u h = ξ i + u e = the set used in dynamic analysis by the modal method

Figure C-2 Set Notation for Physical and Modal Sets

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Solution Sequences for Dynamic
D Analysis

■ Overview

Main Index
348

D.1 Overview
This appendix lists the solution sequences (SOLs) for the dynamic analysis types described in
this guide. The SOLs are listed for the (preferred) structured solution sequences, rigid formats,
and unstructured superelement solution sequences.
The MSC.Nastran Quick Reference Guide lists all of the solution sequences for all of the
MSC.Nastran analysis types.

Structured Solution Sequences for Basic Dynamic Analysis

SOL Number SOL Name Description
103 SEMODES Normal Modes
108 SEDFREQ Direct Frequency Response
109 SEDTRAN Direct Transient Response
111 SEMFREQ Modal Frequency Response
112 SEMTRAN Modal Transient Response

Rigid Formats for Basic Dynamic Analysis

SOL Number SOL Name Description
3 MODES Normal Modes
8 DFREQ Direct Frequency Response
9 DTRAN Direct Transient Response
11 MFREQ Modal Frequency Response
12 MTRAN Modal Transient Response

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Case Control Commands for Dynamic
E Analysis

■ Overview

Main Index
350

E.1 Overview
This appendix lists the Case Control commands that are often used for dynamic analysis.
Commands that apply to statics, such as FORCE and STRESS, are not listed. The dynamic
analysis Case Control commands are listed alphabetically. The description of each command is
similar to that found in the MSC.Nastran Quick Reference Guide. The descriptions in this guide
have been edited to apply specifically to the dynamic analysis capabilities described herein.

The MSC.Nastran Quick Reference Guide describes all of the Case Control commands.
The Case Control commands described in this appendix are summarized as follows:

Input Specification

B2GG Direct Input Damping Matrix Selection

K2GG Direct Input Stiffness Matrix Selection
M2GG Direct Input Mass Matrix Selection

Analysis Specification

BC Boundary Condition Identification

DLOAD Dynamic Load Set Selection
FREQUENCY Frequency Set Selection
IC Transient Initial Condition Set Selection
METHOD Real Eigenvalue Extraction Method Selection
SDAMPING Structural Damping Selection
SUPORT1 Fictitious Support Set Selection
TSTEP Transient Time Step Set Selection for Linear Analysis

Output Specification

ACCELERATION Acceleration Output Request

DISPLACEMENT Displacement Output Request
MODES Subcase Repeater
OFREQUENCY Output Frequency Set
OLOAD Applied Load Output Request
OTIME Output Time Set
SACCELERATION Solution Set Acceleration Output Request
SDISPLACEMENT Solution Set Displacement Output Request

Main Index
 APPENDIX E 351
Case Control Commands for Dynamic Analysis

SVECTOR Solution Set Eigenvector Output Request

SVELOCITY Solution Set Velocity Output Request
VELOCITY Velocity Output Request

Main Index
352

ACCELERATION Acceleration Output Request

Requests form and type of acceleration vector output.

Format:


ACCELERATION  SORT1 , PRINT, PUNCH , REAL or IMAG , PSDF,ATOC,CRMS ,
 SORT2 PLOT PHASE or RALL

 ALL 
RPRINT  
, RPUNCH , [ CID ] =  n 
NORPRINT  
 NONE 
Examples:
ACCELERATION=5
ACCELERTION(SORT2, PHASE)=ALL
ACCELERTION(SORT1, PRINT, PUNCH, PHASE)=17
ACCELERATION(SORT2, PRINT, PSDF, CRMS, RPUNCH)=20
ACCELERATION(PRINT, RALL, NORPRINT)=ALL

Describer Meaning
SORT1 Output will be presented as a tabular listing of grid points for
each load, frequency, eigenvalue, or time, depending on the
solution sequence.
SORT2 Output will be presented as a tabular listing of frequency or time
for each grid point.
PRINT The printer will be the output medium.
PUNCH The punch file will be the output medium.
PLOT Computes, but does not print or punch, acceleration output.
REAL or IMAG Requests rectangular format (real and imaginary) of complex
output. Use of either REAL or IMAG yields the same output.
PHASE Requests polar format (magnitude and phase) of complex output.
Phase output is in degrees.
PSDF Requests the power spectral density function be calculated for
random analysis post-processing. Request must be made above
the subcase level and RANDOM must be selected in the Case
Control.
ATOC Requests the autocorrelation function be calculated for random
analysis post-processing. Request must be made above the
subcase level and RANDOM must be selected in the Case
Control.

Main Index
 APPENDIX 353

Describer Meaning
CRMS Requests the cumulative root mean square function be calculated
for random analysis post-processing. Request must be made
above the subcase level and RANDOM must be selected in the
Case Control.
RALL Requests all of PSDF, ATOC and CRMS be calculated for random
analysis post-processing. Request must be made above the
subcase level and RANDOM must be selected in the Case
Control.
RPRINT Writes random anlaysis results in the print file (Default).
NORPRINT Disables the writing of random analysis results in the print file.
RPUNCH Writes random analysis results in the punch file.
CID Request to print output coordinate system ID in printed output
file, F06 file.
ALL Accelerations at all points will be output.
n Set identification of a previously appearing SET command. Only
accelerations of points with identification numbers that appear
on this SET command will be output. (Integer>0)
NONE Accelerations at no points will be output.

Remarks:
1. Both PRINT and PUNCH may be requested.
2. Acceleration output is only available for transient and frequency response problems.
Acceleration is only available for transient and frequency response problems and when
response spectra is requested in eigenvalue analysis.
3. See Remark 2 under “DISPLACEMENT” on page 355 for a discussion of SORT1 and
SORT2.
4. ACCELERATION=NONE allows overriding an overall output request.
5. The PLOT option is used when curve plots are desired in the magnitude/phase
representation and no printer request is present for magnitude/phase representation.
6. Acceleration results are output in the global coordinate system (see field CD on the
GRID Bulk Data entry).
7. The option of PSDF, ATOC, CRMS and RALL, or any combination of them, can be
selected for random analysis. The results can be either printed in .f06 file or punched in
the punch file, or output in both files.
8. Note that the CID keyword affects only grid point related output, such as
DISPlacement, VELOcity, ACCEleration, OLOAD, SPCForce and MPCForce. In
addition, CID keyword needs to appear only once in a grid related output request
anywhere in the Case Control Section to turn on the printing algorithm.

Main Index
354

B2GG Direct Input Damping Matrix Selection

Selects direct input damping matrices.

Format:
B2GG=name
Example:

B2GG = BDMIG
B2GG = B1, B2, B3
SET 100 = BDMIG, B1, B8
B2GG = 100

Describer Meaning
name 2
Name of [ B gg ] matrix that is input on the DMIG Bulk Data entry.
See the “Matrix Assembly Operations in SubDMAP SEMG” on
page 377 of the MSC.Nastran Reference Guide. (Character)

Remarks:
1. DMIG matrices will not be used unless selected.
2. Terms are added to the damping matrix before any constraints are applied.
3. The matrix must be symmetric and field 4 on the DMIG,name entry must contain the
integer 6.
4. A scale factor may be applied to this input via the PARAM, CB2 entry. See
“Parameters” on page 601.
5. The matrices are additive if multiple matrices are referenced on the B2GG command.

BC Boundary Condition Identification

Identifies multiple boundary conditions for normal modes, buckling, and flutter analysis in
SOLs 103, 105, 145, and 200.
Format:
BC=n
Example:
BC=23

Describer Meaning
n Identification number. (Integer > 0)

Main Index
 APPENDIX 355

Remarks:
1. In SOLs 103, 105, 145, and 200 BC is required in each subcase if multiple boundary
conditions are specified for normal modes, buckling, and flutter analysis.
2. If only one boundary condition is specified, then BC does not have to be specified and
n defaults to zero.

DISPLACEMENT Displacement Output Request

Requests the form and type of displacement or pressure vector output. Note: PRESSURE and
VECTOR are equivalent commands.
SORT1 PRINT, PUNCH REAL or IMAG PSDF, ATOC, CRMS
DISPLACEMENT  , , ,
Format:  SORT2 PLOT PHASE RALL

 ALL 
Examples: RPRINT  
, RPUNCH , [ CID ] =  n 
NOPRINT  
DISPLACEMENT=5  NONE 
DISPLACEMENTS(REAL)=ALL
DISPLACEMENT(SORT2, PUNCH, REAL)=ALL
DISPLACEMENT(SORT2, PRINT, PSDF, CRMS, RPUNCH)=20
DISPLACEMENT(PRINT, RALL, NORPRINT)=ALL

Describer Meaning
SORT1 Output will be presented as a tabular listing of grid points for
each load, frequency, eigenvalue, or time, depending on the
solution sequence.
SORT2 Output will be presented as a tabular listing of load, frequency or
time for each grid point.
PRINT The printer will be the output medium.
PUNCH The punch file will be the output medium.
PLOT Generates, but does not print, displacement data.
REAL or IMAG Requests rectangular format (real and imaginary) of complex
output. Use of either REAL or IMAG yields the same output.
PHASE Requests polar format (magnitude and phase) of complex
output. Phase output is in degrees.
PSDF Requests the power spectral density function be calculated for
random analysis post-processing. Request must be made above
the subcase level and RANDOM must be selected in the Case
Control. See Remark 8.
ATOC Requests the autocorrelation function be calculated for random
analysis post-processing. Request must be made above the
subcase level and RANDOM must be selected in the Case
Control. See Remark 8.

Main Index
356

Describer Meaning
CRMS Requests the cumulative root mean square function be calculated
for random analysis post-processing. Request must be made
above the subcase level and RANDOM must be selected in the
Case Control. See Remark 8.
RALL Request all of PSDF, ATOC and CRMS be calculated for random
analysis post-processing. Request must be made above the
subcase level and RANDOM must be selected in the Case
Control. See Remark 8.
RPRINT Writes random analysis results in the print file (Default)
NORPRINT Disables the writing of random analysis results in the print file.
RPUNCH Writes random analysis results in the punch file.
CID Request to print output coordinate system ID in printed output
file, F06 file.
ALL Displacements for all points will be output.
n Set identification of a previously appearing SET command. Only
displacements of points with identification numbers that appear
on this SET command will be output. (Integer>0)
NONE Displacement for no points will be output.

Remarks:
1. Both PRINT and PUNCH may be requested.
2. The defaults for SORT1 and SORT2 depend on the type of analysis:
• SORT1 is the default in static analysis, frequency response, steady state heat
transfer analysis, real and complex eigenvalue analysis, flutter analysis, and
buckling analysis.
• SORT2 is the default in transient response analysis (structural and heat transfer).
SORT2 is not available for real eigenvalue (including buckling), complex
eigenvalue, or flutter analysis. If SORT1 is selected in a transient solution for one
or more of the commands ACCE, DISP, ENTH, FORC, HDOT, MPCF, OLOA,
SPCF, STRA, STRE, and VELO then the remaining commands will also be
output in SORT1 format. If SORT2 is selected in a frequency response solution
for one or more of the commands ACCE, DISP, FORC, MPCF, OLOA, SPCF,
STRA, STRE, and VELO then the remaining commands will also be output in
SORT2 format.
• XY plot requests forces SORT2 and overrides SORT1 requests!
3. VECTOR and PRESSURE are alternate forms and are entirely equivalent to
DISPLACEMENT.
4. DISPLACEMENT=NONE overrides an overall output request.

Main Index
 APPENDIX 357

5. The PLOT option is used when curve plots are desired in the magnitude/phase
representation and no printer output request is present for magnitude/phase
representation.
6. The units of translation are the same as the units of length of the model. Rotations are
in units of radians.
7. Displacement results are output in the global coordinate system (see field CD on the
GRID Bulk Data entry).
8. The option of PSDF, ATOC, CRMS, and RALL, or any combination of them, can be
selected for random analysis. The results can be either printed in the .f06 file or
punched in the punch file, or output in both files.
9. Note that the CID keyword affects only grid point related output, such as
DISPlacement, VELOcity, ACCEleration, OLOAD, SPCForce and MPCForce. In
addition, CID keyword needs to appear only once in a grid related output request
anywhere in the Case Control Section to turn on the printing algorithm.

DLOAD Dynamic Load Set Selection

Selects a dynamic load or an acoustic source to be applied in a transient or frequency response

problem.
Format:
DLOAD=n
Example:
DLOAD=73

Describer Meaning
n Set identification of a DLOAD, RLOAD1, RLOAD2, TLOAD1,
TLOAD2, or ACSRCE Bulk Data entry. (Integer>0)

Remarks:
1. RLOAD1 and RLOAD2 may only be selected in a frequency response problem.
2. TLOAD1 and TLOAD2 may be selected in a transient or frequency response problem.
3. Either a RLOADi or TLOADi entry (but not both) must be selected in an aeroelastic
response problem. If RLOADi is selected, a frequency response is calculated. If
TLOADi is selected, then transient response is computed by Fourier transform. When
there are only gust loads (GUST entry), the DLOAD selects a TLOADi or RLOADi entry
with zero load along with field 3 of the GUST command.
4. The DLOAD command will be ignored if specified for upstream superelements in
dynamic analysis. To apply loads to upstream superelements, please see the
LOADSET command.

Main Index
358

FREQUENCY Frequency Set Selection

Selects the set of forcing frequencies to be solved in frequency response problems.

Format:
FREQUENCY=n
Example:
FREQUENCY=17

Describer Meaning
n Set identification number of FREQ, FREQ1, FREQ2, FREQ3,
FREQ4, and FREQ5 Bulk Data entries. (Integer>0)

Remarks:
1. A frequency set selection is required for a frequency response problem.
2. A frequency set selection is required for transient response by Fourier methods
(SOLution 146).
3. All FREQ, FREQ1, FREQ2, FREQ3, FREQ4, and FREQ5 entries with the same frequency
set identification numbers will be used. Duplicate frequencies will be ignored. f N and
f N – 1 are considered duplicated if
–5
f N – f N – 1 < 10 f MAX – f MIN

IC Transient Analysis Initial Condition Set Selection

Selects the initial conditions for transient analysis (SOLs 109, 112, 129 and 159).
Format:

PHYSICAL
IC MODAL = n
STATSUB[,DIFFK]

Examples:
IC = 10
IC(PHYSICAL) = 100
IC(MODAL) = 200
IC(STATSUB) = 1000
IC(STATSUB,DIFFK) = 2000

Main Index
 APPENDIX 359

Describer Meaning
PHYSICAL The TIC Bulk Data entries selected by set n define initial conditions
for coordinates involving grid, scalar and extra points. (Default).
MODAL The TIC Bulk Data entries selected by set n define initial conditions
for modal coordinates and extra points. See Remark 3.
STATSUB Use the solution of the static analysis subcase n as the initial
condition. See Remark 4.
DIFFK Include the effects of differential stiffness in the solution. See
Remarks 4. and 5.
n For the PHYSICAL (the default) and MODAL options, n is the set
identification number of TIC Bulk Data entries for structural analysis
(SOL 109, 112 and 129) or TEMP and TEMPD entries for heat transfer
analysis (SOL 159). For the STATSUB option, n is the ID of a static
analysis subcase. (Integer > 0)

Remarks:
1. For structural analysis, TIC entries will not be used (therefore, no initial conditions)
unless selected in the Case Control Section.
2. Only the PHYSICAL option (the default) may be specified in heat transfer analysis
(SOL 159).
3. IC(MODAL) may be specified only in modal transient analysis (SOL 112).
4. IC(STATSUB) and IC(STATSUB,DIFFK) may not both be specified in the same
execution.
5. The DIFFK keyword is meaningful only when used in conjunction with the STATSUB
keyword.
6. The following examples illustrate the usage of the various options of the IC Case
Control command.
$ SPECIFY INITIAL CONDITIONS FOR PHYSICAL COORDINATES
$ IN SOL 109 OR SOL 112
IC(PHYSICAL) = 100
or
IC = 100
$ SPECIFY INITIAL CONDITIONS FOR MODAL COORDINATES
$ IN SOL 112
IC(MODAL) = 200

Main Index
360

$ SPECIFY STATIC SOLUTION AS INITIAL CONDITION

$ IN SOL 109 OR SOL 112
$ (DIFFERENTIAL STIFFNESS EFFECT NOT INCLUDED)
SUBCASE 10 $ STATIC ANALYSIS
LOAD = 100
SUBCASE 20 $ TRANSIENT ANALYSIS
IC(STATSUB) = 10 $ POINTS TO STATIC ANALYSIS SUBCASE ID

$ SPECIFY STATIC SOLUTION AS INITIAL CONDITION

$ IN SOL 109 OR SOL 112
$ (DIFFERENTIAL STIFFNESS EFFECT INCLUDED
SUBCASE 100 $ STATIC ANALYSIS
LOAD = 1000
SUBCASE 200 $ TRANSIENT ANALYSIS
IC(STATSUB,DIFFK) = 100 $ POINTS TO STATIC ANALYSIS SUBCASE ID

K2GG Direct Input Stiffness Matrix Selection

Selects direct input stiffness matrices.

Format:
K2GG=name
Example:

K2GG = KDMIG
K2GG = K1, K2, K3
SET 100 = KDMIG, K1, K8
K2GG = 100

Describer Meaning
name Name of a [ K 2gg ] matrix that is input on the DMIG Bulk Data entry.
(Character) See “Matrix Assembly Operations in SubDMAP
SEMG” on page 377 of the MSC.Nastran Reference Guide.

Remarks:
1. DMIG matrices will not be used unless selected.
2. Terms are added to the stiffness matrix before any constraints are applied.
3. The matrix must be symmetric and field 4 on DMIG,name entry must contain the
integer 6.
4. A scale factor may be applied to this input via the PARAM, CK2 entry. See
“Parameters” on page 601.
Main Index
 APPENDIX 361

5. The matrices are additive if multiple matrices are referenced on the K2GG command.

M2GG Direct Input Mass Matrix Selection

Selects direct input mass matrices.

Format:
M2GG=name

Example:

M2GG = MDMIG
M2GG = M1, M2, M3
SET 100 = MDMIG, M1, M8
M2GG = 100

Describer Meaning
name Name of a [ M 2gg ] matrix that is input on the DMIG Bulk Data entry.
See the “Matrix Assembly Operations in SubDMAP SEMG” on
page 377 of the MSC.Nastran Reference Guide. (Character)

Remarks:
1. DMIG matrices will not be used unless selected.
2. Terms are added to the mass matrix before any constraints are applied.
3. The matrix must be symmetric and field 4 on DMIG, name entry must contain a 6.
4. M2GG input is not affected by PARAM,WTMASS. M2GG input must either be in
consistent mass units or PARAM,CM2 may be used.
5. The matrices are additive if multiple matrices are referenced on the M2GG command.

METHOD Real Eigenvalue Extraction Method Selection

Selects the real eigenvalue extraction parameters.

Format:

METHOD STRUCTURE = n
FLUID
Examples:
METHOD=33
METHOD(FLUID)=34

Main Index
362

Describer Meaning
STRUCTURE The referenced EIGR or EIGRL Bulk Data entry is applied to the
or FLUID structural or fluid portion of the model. (Default=STRUCTURE)
n Set identification number of an EIGR or EIGRL Bulk Data entry for
normal modes or modal formulation, or an EIGB or EIGRL entry for
buckling. (Integer>0)

Remarks:
1. An eigenvalue extraction method must be selected when extracting real eigenvalues
using DMAP modules READ or REIGL.
2. If the set identification number selected is present on both EIGRL and EIGR and/or
EIGB entries, the EIGRL entry will be used. This entry requests the Lanczos
eigensolution method.
3. METHOD(FLUID) permits a different request of EIGR or EIGRL for the fluid portion
of the model from the structural portion of the model in coupled fluid-structural
analysis. See “Additional Topics” on page 529 of the MSC.Nastran Reference Guide.
• If not specified, then the METHOD selection of the structure will be used for the
fluid and modal reduction will not be performed on the fluid portion of the
model in the dynamic solution sequences.
• The METHOD(FLUID) and METHOD(STRUCTURE) may be specified
simultaneously in the same subcase for the residual structure only. Do not
specify METHOD(FLUID) in a superelement subcase even if the superelement
contains fluid elements.
• The auto-omit feature (see “Real Eigenvalue Analysis in SubDMAPs SEMR3
and MODERS” on page 423 of the MSC.Nastran Reference Guide) is not
recommended. Therefore, only those methods of eigenvalue extraction that can
process a singular mass matrix should be used; e.g., EIGRL entry or MGIV and
MHOU on the EIGR entry.

MODES Subcase Repeater

Repeats a subcase.
Format:
MODES=n

Example:
MODES=3

Describer Meaning
n Number of times the subcase is to be repeated. (Integer>0)

Main Index
 APPENDIX 363

Remarks:
1. This command can be illustrated by an example. Suppose stress output is desired for
the first five modes only and displacements for the next two modes and forces for the
remaining modes. The following example would accomplish this

SUBCASE 1 $ FOR MODES 1 THRU 5

MODES = 5
STRESS = ALL
SUBCASE 6 $ FOR MODES 6 AND 7
DISPLACEMENTS =
MODES = 2
SUBCASE 8 $ FOR MODE 8 AND REMAINING MODES
FORCE = ALL

2. This command causes the results for each mode to be considered as a separate,
successively numbered subcase, beginning with the subcase number containing the
MODES command. In the example above, this means that subcases 1, 2, 3, etc. are
assigned to modes 1, 2, 3, etc., respectively.
3. If this command is not used, eigenvalue results are considered to be a part of a single
subcase. Therefore, any output requests for the single subcase will apply for all
eigenvalues.
4. All eigenvectors with mode numbers greater than the number of subcases defined in
Case Control are printed with the descriptors of the last subcase. For example, to
suppress all printout for modes beyond the first three, the following Case Control
could be used:

SUBCASE 1
MODES = 3
DISPLACEMENTS = ALL
SUBCASE 4
DISPLACEMENTS = NONE
BEGIN BULK

5. This command may be of no use in non eigenvalue analysis and may cause data
recovery processing to be repeated.

OFREQUENCY Output Frequency Set

Selects a set of frequencies for output requests.

Format:

 
OFREQUENCY =  ALL 
 n 
Examples:
OFREQUENCY=ALL
OFREQUENCY=15

Main Index
364

Describer Meaning
ALL Output for all frequencies will be computed.
n Set identification of a previously appearing SET command. Output
for frequencies closest to those given on this SET command will be
output. (Integer>0)

Remarks:
1. In real eigenvalue, buckling, and complex eigenvalue analyses, the OMODES Case
Control command allows for an alternate way of selecting the modes to be output
based on their mode numbers. In these cases, if both the OMODES and OFREQUENCY
requests appear, the OMODES request takes precedence.
2. If this command is not specified in the Case Control Section (or, in the case of real
eigenvalue, buckling, and complex eigenvalue analyses, if neither the OMODES nor
the OFREQUENCY request is specified), then output will be generated for all
frequencies.
3. The number of solutions selected will always be equal to the number of quantities in
the selected set. The closest values are used.
4. In flutter analysis (SOL 145), the selected set refers to the imaginary part of the complex
eigenvalues. The physical interpretation of this quantity depends on the method of
flutter analysis as follows:
• K- or KE-method: Velocity (input units).
• PK-method: Frequency.
5. In aeroelastic response analysis (SOL 146) with RLOAD selection, the selected set refers
to the frequency (cycles per unit time).
6. In complex eigenvalue analysis (SOLs 7, 10, 107, and 110), the selected set refers to the
imaginary part of the complex eigenvalues.
7. If this command is specified in more than one subcase, then it is recommended that the
first subcase contain OFREQ=ALL and subsequent subcases contain OFREQ=n. Also,
data recovery requests should be specified only in the subsequent subcases. For
example:

SUBCASE 1
OFREQ = ALL $ 0.0 through 0.5
SUBCASE 2
SET 10 = 0.0 0.1 0.3
OFREQ = 10
DISP = ALL
SUBCASE3
SET 20 = 0.4 0.5
OFREQ = 20
STRESS = ALL

Main Index
 APPENDIX 365

OLOAD Applied Load Output Request

Requests the form and type of applied load vector output.

Format:


OLOAD  SORT1 , PRINT, PUNCH , REAL or IMAG , PSDF, ATOC, CRMS ,
 SORT2 PLOT PHASE or RALL

 ALL 
RPRINT   
, RPUNCH , [ CID ] =  n 

NORPRINT  
 NONE 
Examples:
OLOAD=ALL‘
OLOAD(SORT1, PHASE)=5‘
OLOAD(SORT2, PRINT, PSDF, CRMS, RPUNCH=20
OLOAD(PRINT, RALL, NORPRINT)=ALL

Describer Meaning
SORT1 Output will be presented as a tabular listing of grid points for
each load, frequency, eigenvalue, or time, depending on the
solution sequence.
SORT2 Output will be presented as a tabular listing of frequency or time
for each grid point.
PRINT The printer will be the output medium.
PUNCH The punch file will be the output medium.
REAL or IMAG Requests rectangular format (real and imaginary) of complex
output. Use of either REAL or IMAG yields the same output.
PHASE Requests polar format (magnitude and phase) of complex
output. Phase output is in degrees.
PSDF Requests the power spectral density function be calculated and
stored in the database for random analysis post-processing.
Request must be made above the subcase level and RANDOM
must be selected in the Case Control.
ATOC Requests the autocorrelation function be calculated and stored in
the database for random analysis post-processing. Request must
be made above the subcase level and RANDOM must be
selected in the Case Control.

Main Index
366

Describer Meaning
CRMS Requests all of PSDF, ATOC and CRMS be calculated for random
analysis post-processing. Request must be made above the
subcase level and RANDOM must be selected in the Case
Control.
RALL Requests all of PSDF, ATOC and CRMS be calculated for random
analysis post-processing. Request must be made above the
subcase level and RANDOM must be selected in the Case
Control.
RPRINT Writes random analysis results in the print file (Default).
NORPRINT Disables the writing of random analysis results in the print file.
RPUNCH Writes random analysis results in the punch file.
CID Request to print output coordinate system ID in printed output
file, F06 file.
ALL Applied loads for all points will be output. See Remarks 2. and 8.
NONE Applied load for no points will be output.
n Set identification of a previously appearing SET command. Only
loads on points with identification numbers that appear on this
SET command will be output. (Integer > 0)

Remarks:
1. Both PRINT and PUNCH may be requested.
2. See Remark 2 under “DISPLACEMENT” on page 355 for a discussion of SORT1 and
SORT2. In the SORT1 format, only nonzero values will be output.
3. In a statics problem, a request for SORT2 causes loads at all requested points (zero and
nonzero) to be output.
4. OLOAD=NONE overrides an overall output request.
5. In the statics superelement solution sequences, and in the dynamics SOLs 107 through
112, 118, 145, 146, and 200. OLOADs are available for superelements and the residual
structure only externally applied loads are printed, and not loads transmitted from
upstream superelements. Transmitted loads can be obtained with GPFORCE requests.
• In the nonlinear transient analysis solution sequences SOLs 129 and 159,
OLOADs are available only for residual structure points and include loads
transmitted by upstream superelements.
6. In nonlinear analysis, OLOAD output will not reflect changes due to follower forces.
7. Loads generated via the SPCD Bulk Data entry do not appear in OLOAD output.

Main Index
 APPENDIX 367

8. In SORT1 format, OLOADs recovered at consecutively numbered scalar points are

printed in groups of six (sextets) per line of output. But if a scalar point is not
consecutively numbered, then it will begin a new sextet on a new line of output. If a
sextet can be formed and it is zero, then the line will not be printed. If a sextet cannot
be formed, then zero values may be output.
9. OLOAD results are output in the global coordinate system (see field CD on the GRID
Bulk Data entry).
10. In inertia relief analysis the OLOAD output is interpreted differently for SOLs 1, 101,
and 200:
• In SOL 1, the output shows only the applied loads.
• In SOLs 101 and 200, the output includes both the inertia loads and applied
loads.
11. The option of PSDF, ATOC, CRMS and RALL, or any combination of them, can be
selected for random analysis. The results can be either printed in the .f06 file or
punched n the punch file, or output in both files.
12. Note that the CID keyword affects only grid point related output, such as
DISPlacement, VELOcity, ACCEleration, OLOAD, SPCForce and MPCForce. In
addition, CID keyword needs to appear only once in a grid related output request
anywhere in the Case Control Section to turn on the printing algorithm.

OTIME Output Time Set

Selects a set of times for output requests.

Format:

 
OTIME =  ALL 
 n 
Examples:
OTIME=ALL
OTIME=15

Describer Meaning
ALL Output for all times will be computed.
n Set identification number of a previously appearing SET command.
Output for times closest to those given on this SET command will be
computed. (Integer>0)

Main Index
368

Remarks:
1. If the OTIME command is not supplied in the Case Control Section, then output for all
times will be computed.
2. This command is particularly useful for requesting a subset of the output (e.g., stresses
at only peak times, etc.).
3. This command can be used in conjunction with the MODACC module to limit the
times for which modal acceleration computations are performed.
4. If this command is specified in more than one subcase in the modal solution sequences,
then it is recommended that the first subcase contain OTIME=ALL and subsequent
subcases contain OTIME=n. Also, data recovery requests should be specified only in
the subsequent subcases. For example:

SUBCASE 1
OTIME ALL
SUBCASE 2
OTIME = 10
SET10 = . . .
DISP = ALL
SUBCASE 3
OTIME = 20
SET 20 = . . .
STRESS = ALL

5. The OTIME command is not effective in nonlinear transient analysis.

6. In superelement analysis, the set definition of an OTIME request for an upstream
superelement will not be honored unless a similar request appears for its downstream
neighbor. The downstream request should either be comprised of the union of all
upstream requests, or left blank, as the OTIME default is ALL. Note that the program
does not check to see if this condition is satisfied.

SACCELERATION Solution Set Acceleration Output Request

Requests the form and type of solution set acceleration output.

Format:

 ALL 
   
SACCELERATION  SORT1 , PRINT, PUNCH, REAL or IMAG  =  n 
 SORT2 PHASE   
 NONE 

Examples:
SACCELERATION=ALL
SACCELERATION(PUNCH,IMAG)=142

Main Index
 APPENDIX 369

Describer Meaning
SORT1 Output will be presented as a tabular listing of grid points for each
load, frequency, eigenvalue, or time, depending on the solution
sequence.
SORT2 Output will be presented as a tabular listing of frequency or time
for each grid point (or mode number).
PRINT The printer will be the output medium.
PUNCH The punch file will be the output medium.
REAL or IMAG Requests rectangular format (real and imaginary) of complex
output. Use of either REAL or IMAG yields the same output.
PHASE Requests polar format (magnitude and phase) of complex output.
Phase output is in degrees.
ALL Acceleration for all solution set points (modes) will be output.
NONE Acceleration for no solution set points (modes) will be output.
n Set identification number of a previously appearing SET
command. Only accelerations of points with identification
numbers that appear on this SET command will be output.
(Integer>0)

Remarks:
1. Acceleration output is only available for transient and frequency response problems.
2. The defaults for SORT1 and SORT2 depend on the type of analysis and is discussed in
Remark 2 under the “DISPLACEMENT” on page 355 command. If SORT1 is selected
for any of the commands SACC, SDIS, and SVEL then the remaining commands will
also be SORT1.
3. SACCELERATION=NONE allows an overall output request to be overridden.

SDAMPING Structural Damping Selection

Requests modal damping as a function of natural frequency in modal solutions or viscoelastic

materials as a function of frequency in direct frequency response analysis.
Format:

 
SDAMPING  STRUCTURE  = n
 FLUID 
Example:
SDAMPING=77

Main Index
370

Describer Meaning
STRUCTURE Modal damping is requested for the structural or fluid portion of
or FLUID the model.
n Set identification number of a TABDMP1 or TABLEDi Bulk Data
entry. (Integer>0)

Remarks:
1. In the modal solutions (e.g., SOLs 10, 11, 12, 110, 111, 112, 145, 146, and 200),
SDAMPING must reference a TABDMP1 entry.
2. In direct frequency response (e.g., SOL 108), SDAMPING must reference a TABLEDi
entry which defines viscoelastic (frequency-dependent) material properties. See
“Viscoelastic Materials in Frequency Response Analysis” on page 887 of the .
3. In frequency response analysis, changes in SDAMPING commands between subcases
must be accompanied by a new FREQUENCY command. The new FREQi entry may
be a duplicate of the previous entry except that the SID must be different.
4. SDAMPING may be requested for superelements as long as PARAM,SESDAMP,YES
is specified.

SDISPLACEMENT Solution Set Displacement Output Request

Requests the form and type of solution set displacement output.

Format:

 ALL 
   
SDISPLACEMENT  SORT1 , PRINT, PUNCH, REAL or IMAG  =  n 
 SORT2 PHASE   
 NONE 

Examples:
SDISPLACEMENT=ALL
SDISPLACEMENT(SORT2,PUNCH,PHASE)=NONE

Describer Meaning
SORT1 Output will be presented as a tabular listing of grid points for
each load, frequency, eigenvalue, or time, depending on the
solution sequence.
SORT2 Output will be presented as a tabular listing of frequency or time
for each grid point (or mode number).
PRINT The printer will be the output medium.
PUNCH The punch file will be the output medium.
REAL or IMAG Requests rectangular format (real and imaginary) of complex
output. Use of either REAL or IMAG yields the same output.

Main Index
 APPENDIX 371

Describer Meaning
PHASE Requests polar format (magnitude and phase) of complex
output. Phase output is in degrees.
ALL Displacements for all solution set points (modes) will be output.
NONE Displacements for no solution set points (modes) will be output.
n Set identification number of a previously appearing SET
command. Only displacements on points with identification
numbers that appear on this SET command will be output.
(Integer>0)

Remarks:
1. The defaults for SORT1 and SORT2 depend on the type of analysis and is discussed in
Remark 2 under the “DISPLACEMENT” on page 355 command. If SORT1 is selected
for any of the commands SACC, SDIS, and SVEL then the remaining commands will
also be SORT1.
2. SDISPLACEMENT=NONE allows an overall output request to be overridden.
3. The SDISPLACEMENT command is required to output normalized complex
eigenvectors.

SUPORT1 Fictitious Support Set Selection

Selects the fictitious support set (SUPORT1 entries only) to be applied to the model.
Format:
SUPORT1=n
Examples:
SUPORT1=15
SUPO=4

Describer Meaning
n Set identification of fictitious support set defined on the SUPORT1
Bulk Data entry. (Integer>0)

Remarks:
1. SUPORT1 entries will not be used unless selected in Case Control by the SUPORT1
command.
2. SUPORT entries will be applied in all subcases.

SVECTOR Solution Set Eigenvector Output Request

Requests the form and type of solution set eigenvector output.

Main Index
372

Format:

 ALL 
 
SVECTOR [ ( PRINT, PUNCH ) ] =  n 
 
 NONE 
Examples:
SVECTOR=ALL
SVECTOR(PUNCH)=NONE

Describer Meaning
PRINT The printer will be the output medium.
PUNCH The punch file will be the output medium.
ALL Displacements for all points (modes) will be output.
NONE Displacements for no points (modes) will be output.
n Set identification of a previously appearing SET command. Only
displacements of points with identification numbers that appear on
this SET command will be output. (Integer>0)

Remarks:
1. Both PRINT and PUNCH may be requested.
2. SVECTOR=NONE overrides an overall output request.
3. Output will be presented as a tabular listing of grid points for each eigenvector.

SVELOCITY Solution Set Velocity Output Request

Requests the form and type of solution set velocity output.

Format:

 ALL 
 REAL or IMAG   
SVELOCITY  SORT1 , PRINT, PUNCH,  =  n 
 SORT2 PHASE   
 NONE 
Examples:
SVELOCITY=5
SVELOCITY(SORT2,PUNCH,PRINT,PHASE)=ALL

Main Index
 APPENDIX 373

Describer Meaning
SORT1 Output will be presented as a tabular listing of grid points for
each load, frequency, eigenvalue, or time, depending on the
solution sequence.
SORT2 Output will be presented as a tabular listing of frequency or time
for each grid point (or mode number).
PRINT The printer will be the output medium.
PUNCH The punch file will be the output medium.
REAL or IMAG Requests rectangular format (real and imaginary) of complex
output. Use of either REAL or IMAG yields the same output.
PHASE Requests polar format (magnitude and phase) of complex
output. Phase output is in degrees.
ALL Velocity for all solution points (modes) will be output.
NONE Velocity for no solution points (modes) will be output.
n Set identification of a previously appearing SET command.
Only velocities of points with identification numbers that appear
on this SET command will be output. (Integer>0)

Remarks:
1. Both PRINT and PUNCH may be requested.
2. Velocity output is only available for transient and frequency response problems.
3. The defaults for SORT1 and SORT2 depend on the type of analysis and is discussed in
Remark 2 under the “DISPLACEMENT” on page 355 command. If SORT1 is selected
for any of the commands SACC, SDIS, and SVEL then the remaining commands will
also be SORT1.
4. SVELOCITY=NONE overrides an overall output request.

TSTEP Transient Time Step Set Selection

Selects integration and output time steps for linear or nonlinear transient analysis.
Format:
TSTEP=n
Example:
TSTEP=731

Describer Meaning
n Set identification number of a TSTEP or TSTEPNL Bulk Data entry.
(Integer>0)

Main Index
374

Remarks:
1. A TSTEP entry must be selected to execute a linear transient analysis (SOLs 9, 12, 109,
or 112) and TSTEPNL for a nonlinear transient analysis (SOLs 129 and 159).
2. A TSTEPNL entry must be selected in each subcase to execute a nonlinear transient
problem.
3. For the application of time-dependent loads in modal frequency response analysis
(SOLs 111 and 146), or TSTEP entry must be selected by the TSTEP command. The
time-dependent loads will be recomputed in frequency domain by a Fourier
Transform.

VELOCITY Velocity Output Request

Requests the form and type of velocity vector output.

Format:


VELOCITY  SORT1 , PRINT, PUNCH , REAL or IMAG PSDF, ATOC, CRMS ,
 SORT2 PLOT PHASE or RALL

 ALL 
RPRINT  
, RPUNCH , [ CID ]  =  
 n
NORPRINT  
 NONE 

Examples:
VELOCITY=5
VELOCITY(SORT2,PHASE,PUNCH)=ALL
VELOCITY(SORT2, PRINT, PSDF, CRMS, RPUNCH)=20
VELOCITY(PRINT, RALL, NORPRINT)=ALL

Describer Meaning
SORT1 Output will be presented as a tabular listing of grid points for
each load, frequency, eigenvalue, or time, depending on the
solution sequence.
SORT2 Output will be presented as a tabular listing of frequency or time
for each grid point.
PRINT The printer will be the output medium.
PUNCH The punch file will be the output medium.
PLOT Generates, but does not print, velocities.
REAL or IMAG Requests rectangular format (real and imaginary) of complex
output. Use of either REAL or IMAG yields the same output.
PHASE Requests polar format (magnitude and phase) of complex
output. Phase output is in degrees.

Main Index
 APPENDIX 375

Describer Meaning
PSDF Requests the power spectral density function be calculated and
stored in the database for random analysis post-processing.
Request must be made above the subcase level and RANDOM
must be selected in the Case Control.
ATOC Requests the autocorrelation function be calculated and stored in
the database for random analysis post-processing. Request must
be made above the subcase level and RANDOM must be
selected in the Case Control.
CRMS Requests the cumulative root mean square function be
calculated for random anlaysis post-processing. Request must be
made above the subcase level and RANDOM must be selected in
the Case Control.
RALL Requests all of PSDF, ATOC, and CRMS be calculated for
random analysis post-processing. Request must be made above
the subcase level and RANDOM must be selected in the Case
Control.
RPRINT Writes random analysis results in the print file. (Default)
NORPRINT Disables the writing of random analysis results in the print file.
RPUNCH Writes random analysis results in the punch file.
CID Request to print output coordinate system ID in printed output
file, F06 file.
ALL Velocity for all solution points will be output.
NONE Velocity for no solution points will be output.
n Set identification of a previously appearing SET command.
Only velocities of points with identification numbers that appear
on this SET command will be output. (Integer>0)

Remarks:
1. Both PRINT and PUNCH may be requested.
2. Velocity output is only available for transient and frequency response problems.
3. See Remark 2 under “DISPLACEMENT” on page 355 for a discussion of SORT1 and
SORT2.
4. VELOCITY=NONE overrides an overall output request.
5. The PLOT option is used when curve plots are desired in the magnitude/phase
representation and no printer request is present for magnitude/phase representation.
6. Velocity results are output in the global coordinate system (see field CD on the GRID
Bulk Data entry).

Main Index
376

7. The option of PSDF, ATOC, CRMS, and RALL, or any combination of them, can be
selected for random analysis. The results can be either printed in the .f06 file or
punched in the punch file, or output in both files.
8. Note that the CID keyword affects only grid point related output, such as
DISPlacement, VELOcity, ACCEleration, OLOAD, SPCForce and MPCForce. In
addition, CID keyword needs to appear only once in a grid related output request
anywhere in the Case Control Section to turn on the printing algorithm.

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Bulk Data Entries for Dynamic
F Analysis

■ Overview

Main Index
378

F.1 Overview
This appendix lists the Bulk Data entries that are often used for dynamic analysis. Entries that
apply to generic modeling or statics, such as GRID or FORCE, are not listed. Bulk Data entries
are listed alphabetically. The description of each entry is similar to that found in the
MSC.Nastran Quick Reference Guide. The descriptions in this guide have been edited to apply
specifically to the dynamic analysis capabilities described herein.

The MSC.Nastran Quick Reference Guide describes all of the Bulk Data entries.
The Bulk Data entries described in this appendix are summarized as follows:

Main Index
 APPENDIX F 379
Bulk Data Entries for Dynamic Analysis

Mass Properties
CMASS1 Scalar Mass Connection
CMASS2 Scalar Mass Property and Connection
CMASS3 Scalar Mass Connection to Scalar Points Only
CMASS4 Scalar Mass Property and Connection to Scalar Points Only
CONM1 Concentrated Mass Element Connection, General Form
CONM2 Concentrated Mass Element Connection, Rigid Body Form
PMASS Scalar Mass Property

Damping Properties

CDAMP1 Scalar Damper Connection

CDAMP2 Scalar Damper Property and Connection
CDAMP3 Scalar Damper Connection to Scalar Points Only
CDAMP4 Scalar Damper Property and Connection to Scalar Points Only
CVISC Viscous Damper Connection
PDAMP Scalar Damper Property
PVISC Viscous Element Property
TABDMP1 Modal Damping Table

Normal Modes

EIGR Real Eigenvalue Extraction Data

EIGRL Real Eigenvalue Extraction Data, Lanczos Method

Dynamic Loading

DAREA Dynamic Load Scale Factor

DELAY Dynamic Load Time Delay
DLOAD Dynamic Load Combination or Superposition
DPHASE Dynamic Load Phase Lead
LSEQ Static Load Set Definition
TABLED1 Dynamic Load Tabular Function, Form 1
TABLED2 Dynamic Load Tabular Function, Form 2
TABLED3 Dynamic Load Tabular Function, Form 3
TABLED4 Dynamic Load Tabular Function, Form 4

Main Index
380

Frequency Response
FREQ Frequency List
FREQ1 Frequency List, Alternate Form 1
FREQ2 Frequency List, Alternate Form 2
FREQ3 Frequency List, Alternate Form 3
FREQ4 Frequency List, Alternate Form 4
FREQ5 Frequency List, Alternate Form 5
RLOAD1 Frequency Response Dynamic Load, Form 1
RLOAD2 Frequency Response Dynamic Load, Form 2

Transient Response

TIC Transient Initial Condition

TLOAD1 Transient Response Dynamic Load, Form 1
TLOAD2 Transient Response Dynamic Load, Form 2
TSTEP Transient Time Step

Miscellaneous

DMIG Direct Matrix Input at Points

SUPORT Fictitious Support
SUPORT1 Fictitious Support, Alternate Form

Main Index
 APPENDIX 381

CDAMP1 Scalar Damper Connection

Defines a scalar damper element.

Format:

1 2 3 4 5 6 7 8 9 10
CDAMP1 EID PID G1 C1 G2 C2

Example:

CDAMP1 19 6 0 23 2

Field Contents
EID Unique element identification number. (0 < Integer < 100,000,000)
PID Property identification number of a PDAMP property entry.
(Integer > 0; Default = EID)
G1, G2 Geometric grid point identification number. (Integer > 0)
C1, C2 Component number. (0 < Integer < 6; 0 or up to six unique integers, 1
through 6 may be specified in the field with no embedded blanks. 0
applies to scalar points and 1 through 6 applies to grid points.)

Remarks:
1. Scalar points may be used for G1 and/or G2, in which case the corresponding C1
and/or C2 must be zero or blank. Zero or blank may be used to indicate a grounded
terminal G1 or G2 with a corresponding blank or zero C1 or C2. A grounded terminal
is a point with a displacement that is constrained to zero.
2. Element identification numbers should be unique with respect to all other element
identification numbers.
3. The two connection points (G1, C1) and (G2, C2), must be distinct.
4. For a discussion of the scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.
5. When CDAMP1 is used in heat transfer analysis, it generates a lumped heat capacity.
6. A scalar point specified on this entry need not be defined on an SPOINT entry.
7. If Gi refers to a grid point then Ci refers to degrees-of-freedom(s) in the displacement
coordinate system specified by CD on the GRID entry.

CDAMP2 Scalar Damper Property and Connection

Defines a scalar damper element without reference to a material or property entry.

Main Index
382

Format:

1 2 3 4 5 6 7 8 9 10
CDAMP2 EID B G1 C1 G2 C2

Example:

CDAMP2 16 2.98 32 1

Field Contents
EID Unique element identification number. (0 < Integer < 100,000,000)
B Value of the scalar damper. (Real)
G1, G2 Geometric grid point identification number. (Integer > 0)
C1, C2 Component number. (0 < Integer < 6; 0 or up to six unique integers, 1
through 6 may be specified in the field with no embedded blanks. 0
applies to scalar points and 1 through 6 applies to grid points.)

Remarks:
1. Scalar points may be used for G1 and/or G2, in which case the corresponding C1
and/or C2 must be zero or blank. Zero or blank may be used to indicate a grounded
terminal G1 or G2 with a corresponding blank or zero C1 or C2. A grounded terminal
is a point with a displacement that is constrained to zero.
2. Element identification numbers should be unique with respect to all other element
identification numbers.
3. The two connection points (G1, C1) and (G2, C2), must be distinct.
4. For a discussion of the scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.
5. When CDAMP2 is used in heat transfer analysis, it generates a lumped heat capacity.
6. A scalar point specified on this entry need not be defined on an SPOINT entry.
7. If Gi refers to a grid point then Ci refers to degrees-of-freedom(s) in the displacement
coordinate system specified by CD on the GRID entry.

CDAMP3 Scalar Damper Connection to Scalar Points Only

Defines a scalar damper element that is connected only to scalar points.

Format:

1 2 3 4 5 6 7 8 9 10
CDAMP3 EID PID S1 S2

Main Index
 APPENDIX 383

Example:

CDAMP3 16 978 24 36

Field Contents
EID Unique element identification number. (0 < Integer < 100,000,000)
PID Property identification number of a PDAMP entry. (Integer > 0;
Default = EID)
S1, S2 Scalar point identification numbers. (Integer > 0; S1 ≠ S2 )

Remarks:
1. S1 or S2 may be blank or zero, indicating a constrained coordinate.
2. Element identification numbers should be unique with respect to all other element
identification numbers.
3. Only one scalar damper element may be defined on a single entry.
4. For a discussion of the scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.
5. When CDAMP3 is used in heat transfer analysis, it generates a lumped heat capacity.
6. A scalar point specified on this entry need not be defined on an SPOINT entry.

CDAMP4 Scalar Damper Property and Connection to Scalar Points Only

Defines a scalar damper element that connected only to scalar points and without reference to a
material or property entry.
Format:

1 2 3 4 5 6 7 8 9 10
CDAMP4 EID B S1 S2

Example:

CDAMP4 16 -2.6 4 9

Field Contents
EID Unique element identification number. (0 < Integer < 100,000,000)
B Scalar damper value. (Real)
S1, S2 Scalar point identification numbers. (Integer > 0; S1 ≠ S2 )

Remarks:
1. S1 or S2 may be blank or zero, indicating a constrained coordinate.

Main Index
384

2. Element identification numbers should be unique with respect to all other element
identification numbers.
3. Only one scalar damper element may be defined on a single entry.
4. For a discussion of the scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.
5. If this entry is used in heat transfer analysis, it generates a lumped heat capacity.
6. A scalar point specified on this entry need not be defined on an SPOINT entry.

CMASS1 Scalar Mass Connection

Defines a scalar mass element.

Format:

1 2 3 4 5 6 7 8 9 10
CMASS1 EID PID G1 C1 G2 C2

Example:

CMASS1 32 6 2 1

Field Contents
EID Unique element identification number. (0 < Integer < 100,000,000)
PID Property identification number of a PMASS entry. (Integer > 0;
Default = EID)
G1, G2 Geometric grid or scalar point identification number. (Integer > 0)
C1, C2 Component number. (0 < Integer < 6; blank or zero if scalar point.)

Remarks:
1. Scalar points may be used for G1 and/or G2, in which case the corresponding C1
and/or C2 must be zero or blank. Zero or blank may be used to indicate a grounded
terminal G1 or G2 with a corresponding blank or zero C1 or C2. A grounded terminal
is a point with a displacement that is constrained to zero. If only scalar points and/or
ground are involved, it is more efficient to use the CMASS3 entry.
2. Element identification numbers should be unique with respect to all other element
identification numbers.
3. The two connection points (G1, C1) and (G2, C2) must not be coincident.
4. For a discussion of the scalar elements, see the MSC.Nastran Reference Manual, Section
5.6.
5. A scalar point specified on this entry need not be defined on an SPOINT entry.
6. If Gi refers to a grid point then Ci refers to degrees-of-freedom(s) in the displacement
coordinate system specified by CD on the GRID entry.
Main Index
 APPENDIX 385

7. Scalar elements input coupled mass matrices when the second pair of fields is entered.
When uncoupled point masses are desired input only the first pair of fields. When a
coupled mass matrix is requested the submatrix added has M on the diagonal, and -M
on the off-diagonal. The element is not checked for internal constraints, which is the
user’s responsibility if desired. There are instances where elements with internal
constraints are desired, although not frequently. To identify the presence of internal
constraints caused by coupled mass, inspect GPWG output, OLOAD output due to
GRAV loads, and rigid body modes of free structures. Some forms of coupled mass will
cause coupling of rigid body translational mass terms in GPWG output, and poor rigid
body modes in modal analysis.

CMASS2 Scalar Mass Property and Connection

Defines a scalar mass element without reference to a property entry.

Format:

1 2 3 4 5 6 7 8 9 10
CMASS2 EID M G1 C1 G2 C2

Example:

CMASS2 32 9.25 6 1

Field Contents
EID Unique element identification number. (0 < Integer < 100,000,000)
M Value of the scalar mass. (Real)
G1, G2 Geometric grid or scalar point identification number. (Integer > 0)
C1, C2 Component number. (0 < Integer < 6; blank or zero if scalar point.)

Remarks:
1. Scalar points may be used for G1 and/or G2, in which case the corresponding C1
and/or C2 must be zero or blank. Zero or blank may be used to indicate a grounded
terminal G1 or G2 with a corresponding blank or zero C1 or C2. A grounded terminal
is a point with a displacement that is constrained to zero. If only scalar points and/or
ground are involved, it is more efficient to use the CMASS4 entry.
2. Element identification numbers should be unique with respect to all other element
identification numbers.
3. The two connection points (G1, C1) and (G2, C2) must be distinct. Except in unusual
circumstances, one of them will be a grounded terminal with blank entries for Gi and
Ci.
4. For a discussion of the scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.
5. A scalar point specified on this entry need not be defined on an SPOINT entry.
Main Index
386

6. If Gi refers to a grid point then Ci refers to degrees-of-freedom(s) in the displacement

coordinate system specified by CD on the GRID entry.
7. See Remark 7 for “CMASS1” on page 384.

CMASS3 Scalar Mass Connection to Scalar Points Only

Defines a scalar mass element that is connected only to scalar points.

Format:

1 2 3 4 5 6 7 8 9 10
CMASS3 EID PID S1 S2

Example:

CMASS3 13 42 62

Field Contents
EID Unique element identification number. (0 < Integer < 100,000,000)
PID Property identification number of a PMASS entry. (Integer > 0;
Default = EID)
S1, S2 Scalar point identification numbers. (Integer > 0; S1 ≠ S2 )

Remarks:
1. S1 or S2 may be blank or zero, indicating a constrained coordinate.
2. Element identification numbers should be unique with respect to all other element
identification numbers.
3. Only one scalar mass element may be defined on a single entry.
4. For a discussion of the scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.
5. A scalar point specified on this entry need not be defined on an SPOINT entry.

CMASS4 Scalar Mass Property and Connection to Scalar Points Only

Defines a scalar mass element that is connected only to scalar points, without reference to a
property entry.
Format:

1 2 3 4 5 6 7 8 9 10
CMASS4 EID M S1 S2

Example:

CMASS4 23 14.92 23
Main Index
 APPENDIX 387

Field Contents
EID Unique element identification number. (0 < Integer < 100,000,000)
M Scalar mass value. (Real)
S1, S2 Scalar point identification numbers. (Integer > 0; S1 ≠ S2 )

Remarks:
1. S1 or S2 may be blank or zero, indicating a constrained coordinate. This is the usual
case.
2. Element identification numbers should be unique with respect to all other element
identification numbers.
3. Only one scalar mass element may be defined on a single entry.
4. For a discussion of the scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.
5. A scalar point specified on this entry need not be defined on an SPOINT entry.

CONM1 Concentrated Mass Element Connection, General Form

Defines a 6 x 6 symmetric mass matrix at a geometric grid point.

Format:

1 2 3 4 5 6 7 8 9 10
CONM1 EID G CID M11 M21 M22 M31 M32

M33 M41 M42 M43 M44 M51 M52 M53

M54 M55 M61 M62 M63 M64 M65 M66

Example:

CONM1 2 22 2 2.9 6.3

4.8 28.6

28.6 28.6

Field Contents
EID Unique element identification number. (Integer > 0)
G Grid point identification number. (Integer > 0)
CID Coordinate system identification number for the mass matrix.
(Integer > 0)
Mij Mass matrix values. (Real)

Main Index
388

Remarks:
1. For a less general means of defining concentrated mass at grid points, see the CONM2
entry description.
2. Element identification numbers should be unique with respect to all other element
identification numbers.

CONM2 Concentrated Mass Element Connection, Rigid Body Form

Defines a concentrated mass at a grid point.

Format:

1 2 3 4 5 6 7 8 9 10
CONM2 EID G CID M X1 X2 X3

I11 I21 I22 I31 I32 I33

Example:

CONM2 2 15 6 49.7

16.2 16.2 7.8

Field Contents
EID Element identification number. (Integer > 0)
G Grid point identification number. (Integer > 0)
CID Coordinate system identification number. For CID of -1; see X1, X2, X3
below. (Integer > -1; Default = 0)
M Mass value. (Real)
X1, X2, X3 Offset distances from the grid point to the center of gravity of the mass
in the coordinate system defined in field 4, unless CID = -1, in which
case X1, X2, X3 are the coordinates, not offsets, of the center of gravity of
the mass in the basic coordinate system. (Real)
lij Mass moments of inertia measured at the mass center of gravity in the
coordinate system defined by field 4. If CID = -1, the basic coordinate
system is implied. (For I11, I22, and I33; Real > 0.0; for I21, I31, and I32;
Real)

Remarks:
1. Element identification numbers should be unique with respect to all other element
identification numbers.
2. For a more general means of defining concentrated mass at grid points, see the CONM1
entry description.
3. The continuation is optional.

Main Index
 APPENDIX 389

4. If CID = -1, offsets are internally computed as the difference between the grid point
location and X1, X2, X3. The grid point locations may be defined in a nonbasic
coordinate system. In this case, the values of Iij must be in a coordinate system that
parallels the basic coordinate system.
5. The form of the inertia matrix about its center of gravity is taken as:

M
M symmetric
M
I11
– I21 I22
– I31 – I32 I33

where

M = ρ dV
∫
I11 = 2 2
∫ ρ ( x 2 + x 3 ) dV
I22 = 2 2
∫ ρ ( x 1 + x 3 ) dV
I33 = 2 2
∫ ρ ( x 1 + x 2 ) dV
I21 =
∫ ρx 1 x 2 dV
I31 =
∫ ρx 1 x 3 dV
I32 =
∫ ρx 2 x 3 dV
and x 1, x 2, x 3 are components of distance from the center of gravity in the coordinate
system defined in field 4. The negative signs for the off-diagonal terms are supplied
automatically. A warning message is issued if the inertia matrix is nonpositive definite,
since this may cause fatal errors in dynamic analysis modules.
6. If CID > 0, then X1, X2, and X3 are defined by a local Cartesian system, even if CID
references a spherical or cylindrical coordinate system. This is similar to the manner in
which displacement coordinate systems are defined.
7. See “Grid Point and Coordinate System Definition” on page 41 of the MSC.Nastran
Reference Guide for a definition of coordinate system terminology.

CVISC Viscous Damper Connection

Defines a viscous damper element.

Main Index
390

Format:

1 2 3 4 5 6 7 8 9 10
CVISC EID PID G1 G2

Example:

CVISC 21 6327 29 31

Field Contents
EID Element identification number. (Integer > 0)
PID Property identification number of a PVISC entry. (Integer > 0;
Default = EID)
G1, G2 Grid point identification numbers of connection points. (Integer > 0;
G1 ≠ G2 )

Remarks:
1. Element identification numbers should be unique with respect to all other element
identification numbers.
2. Only one viscous damper element may be defined on a single entry.

DAREA Load Scale Factor

Defines scale (area) factors for static and dynamic loads. In dynamic analysis, DAREA is used
in conjunction with RLOADi and TLOADi entries.
Format:

1 2 3 4 5 6 7 8 9 10
DAREA SID P1 C1 A1 P2 C2 A2

Example:

DAREA 3 6 2 8.2 15 1 10.1

Field Contents
SID Identification number. (Integer > 0)
Pi Grid, extra, or scalar point identification number. (Integer > 0)

Main Index
 APPENDIX 391

Field Contents
Ci Component number. (Integer 1 through 6 for grid point; blank or 0 for
extra or scalar point.)
Ai Scale (area) factor. (Real)

Remarks:
1. One or two scale factors may be defined on a single entry.
2. Refer to RLOAD1, RLOAD2, TLOAD1, TLOAD2, or ACSRCE entries for the formulas
that define the scale factor Ai in dynamic analysis.
3. Component numbers refer to the displacement coordinate system.
4. In dynamic analysis, DAREA entries may be used with LSEQ Bulk Data entries if
LOADSET is specified in Case Control. The LSEQ and static load entries will be used
to internally generate DAREA entries.
5. If DAREA is referenced by a GUST entry, Pi must be defined. However, it is only used
if selected through a DLOAD Case Control command. WG from the GUST entry is
used instead of Ai when requested via a GUST entry.
6. All DAREA entries corresponding to all grid and scalar points are automatically
converted internally by the program to equivalent FORCE/MOMENT/SLOAD entries
(as appropriate) if there are no LSEQ Bulk Data entries. The DAREA entries can only be
used in static analysis if there are no LSEQ Bulk Data entries
7. In superelement analysis, DAREA may be used to specify loads not only on the interior
points of the residual, but also on the interior points of upstream superelements if there
are no LSEQ Bulk Data entries.

DELAY Dynamic Load Time Delay

Defines the time delay term τ in the equations of the dynamic loading function.
Format:

1 2 3 4 5 6 7 8 9 10
DELAY SID P1 C1 T1 P2 C2 T2

Example:

DELAY 5 21 6 4.25 7 6 8.1

Field Contents
SID Identification number of the DELAY entry. (Integer > 0)
Pi Grid, extra, or scalar point identification number. (Integer > 0)

Main Index
392

Field Contents
Ci Component number. (Integer 1 through 6 for grid point, blank or 0 for
extra point or scalar point.)
Ti Time delay τ for designated point Pi and component Ci. (Real)

Remarks:
1. One or two dynamic load time delays may be defined on a single entry.
2. SID must also be specified on a RLOAD1, RLOAD2, TLOAD1, TLOAD2, or ACSRCE
entry. See those entry descriptions for the formulas that define the manner in which
the time delay τ is used.
3. A DAREA and/or LSEQ entry should be used to define a load at Pi and Ci.
4. In superelement analysis, DELAY entries may only be applied to loads on points in the
residual structure.

DLOAD Dynamic Load Combination or Superposition

Defines a dynamic loading condition for frequency response or transient response problems as
a linear combination of load sets defined via RLOAD1 or RLOAD2 entries for frequency
response or TLOAD1 or TLOAD2 entries for transient response.
Format:

1 2 3 4 5 6 7 8 9 10
DLOAD SID S S1 L1 S2 L2 S3 L3

S4 L4 -etc.- *

Examples:

DLOAD 17 1.0 2.0 6 -2.0 7 2.0 8

-2.0 9

Field Contents
SID Load set identification number. (Integer > 0)
S Scale factor. (Real)
Si Scale factors. (Real)
Li Load set identification numbers of RLOAD1, RLOAD2, TLOAD1,
TLOAD2, and ACSRC entries. (Integer > 0)

Remarks:
1. Dynamic load sets must be selected in the Case Control Section with DLOAD = SID.
2. The load vector being defined by this entry is given by

Main Index
 APPENDIX 393

{ P } = S ∑ Si { P i }
i

3. Each Li must be unique from any other Li on the same entry.

4. SID must be unique from all TLOADi and RLOADi entries.
5. Nonlinear transient load sets (NOLINi entries) may not be specified on DLOAD
entries. NOLINi entries are selected separately in the Case Control Section by the
NONLINEAR command.
6. A DLOAD entry may not reference a set identification number defined by another
DLOAD entry.
7. TLOAD1 and TLOAD2 loads may be combined only through the use of the DLOAD
entry.
8. RLOAD1 and RLOAD2 loads may be combined only through the use of the DLOAD
entry.

DMIG Direct Matrix Input at Points

Defines direct input matrices related to grid, extra, and/or scalar points. The matrix is defined
by a single header entry and one or more column entries. A column entry is required for each
column with nonzero elements.
Header Entry Format:

1 2 3 4 5 6 7 8 9 10
DMIG NAME “0" IFO TIN TOUT POLAR NCOL

Column Entry Format:

DMIG NAME GJ CJ G1 C1 A1 B1

G2 C2 A2 B2 -etc.-

Example:

DMIG STIF 0 1 3 4

DMIG STIF 27 1 2 3 3.+5 3.+3

2 4 2.5+10 0. 50 1.0 0.

Field Contents
NAME Name of the matrix. See Remark 1. (One to eight alphanumeric
characters, the first of which is alphabetic.)
IFO Form of matrix input. IFO = 6 must be specified for matrices selected by
the K2GG, M2GG, and B2GG Case Control commands. (Integer)

Main Index
394

Field Contents
1 = Square
9 or 2 = Rectangular
6 = Symmetric
TIN Type of matrix being input: (Integer)
1 = Real, single precision (One field is used per element.)
2 = Real, double precision (One field is used per element.)
3 = Complex, single precision (Two fields are used per element.)
4 = Complex, double precision (Two fields are used per element.)
TOUT Type of matrix that will be created: (Integer)
0 = Set by precision system cell (Default)
1 = Real, single precision
2 = Real, double precision
3 = Complex, single precision
4 = Complex, double precision
POLAR Input format of Ai, Bi. (Integer=blank or 0 indicates real, imaginary
format; Integer > 0 indicates amplitude, phase format.)
NCOL Number of columns in a rectangular matrix. Used only for IFO = 9. See
Remarks 5. and 6. (Integer > 0)
GJ Grid, scalar or extra point identification number for column index.
(Integer > 0)
CJ Component number for grid point GJ. (0 < Integer < 6; blank or zero if
GJ is a scalar or extra point.)
Gi Grid, scalar, or extra point identification number for row index.
(Integer > 0)
Ci Component number for Gi for a grid point. ( 0 < CJ ≤ 6 ; blank or zero
if Gi is a scalar or extra point.)
Ai, Bi Real and imaginary (or amplitude and phase) parts of a matrix element.
If the matrix is real (TIN = 1 or 2), then Bi must be blank. (Real)

Remarks:
1. Matrices defined on this entry may be used in dynamics by selection in the Case
Control with K2PP = NAME, B2PP = NAME, M2PP = NAME for [Kpp], [Bpp], or [Mpp],
respectively. Matrices may also be selected for all solution sequences by
K2GG = NAME, B2GG = NAME, and M2GG = NAME. The g-set matrices are added
to the structural matrices before constraints are applied, while p-set matrices are added
in dynamics after constraints are applied. Load matrices may be selected by
P2G = NAME for dynamic and superelement analyses.

Main Index
 APPENDIX 395

2. The header entry containing IFO, TIN and TOUT is required. Each nonnull column is
started with a GJ, CJ pair. The entries for each row of that column follows. Only
nonzero terms need be entered. The terms may be input in arbitrary order. A GJ, CJ
pair may be entered more than once, but input of an element of the matrix more than
once will produce a fatal message.
3. Field 3 of the header entry must contain an integer 0.
4. For symmetric matrices (IFO = 6), a given off-diagonal element may be input either
below or above the diagonal. While upper and lower triangle terms may be mixed, a
fatal message will be issued if an element is input both below and above the diagonal.
5. The recommended format for rectangular matrices requires the use of NCOL and
IFO = 9. The number of columns in the matrix is NCOL. (The number of rows in all
DMIG matrices is always either p-set or g-set size, depending on the context.) The GJ
term is used for the column index. The CJ term is ignored.
6. If NCOL is not used for rectangular matrices, two different conventions are available:
• If IFO = 9, GJ and CJ will determine the sorted sequence, but will otherwise be
ignored; a rectangular matrix will be generated with the columns submitted
being in the 1 to N positions, where N is the number of logical entries submitted
(not counting the header entry).
• If IFO = 2, the number of columns of the rectangular matrix will be equal to the
index of the highest numbered non-null column (in internal sort). Trailing null
columns of the g- or p-size matrix will be truncated.
7. The matrix names must be unique among all DMIGs.
8. TIN should be set consistent with the number of decimal digits required to read the
input data adequately. For a single-precision specification on a short-word machine,
the input will be truncated after about eight decimal digits, even when more digits are
present in a double-field format. If more digits are needed, a double precision
specification should be used instead. However, note that a double precision
specification requires a “D” type exponent even for terms that do not need an
exponent. For example, unity may be input as 1.0 in single precision, but the longer
form 1.0D0 is required for double precision.
9. On long-word machines, almost all matrix calculations are performed in single
precision and on short-word machines, in double precision. It is recommended that
DMIG matrices also follow these conventions for a balance of efficiency and reliability.
The recommended value for TOUT is 0, which instructs the program to inspect the
system cell that measures the machine precision at run time and sets the precision of
the matrix to the same value. TOUT = 0 allows the same DMIG input to be used on any
machine. If TOUT is contrary to the machine type specified (for example, a TOUT of 1
on a short-word machine), unreliable results may occur.
10. If any DMIG entry is changed or added on restart then a complete re-analysis is
performed. Therefore, DMIG entry changes or additions are not recommended on
restart.

Main Index
396

DPHASE Dynamic Load Phase Lead

Defines the phase lead term θ in the equation of the dynamic loading function.
Format:

1 2 3 4 5 6 7 8 9 10
DPHASE SID P1 C1 TH1 P2 C2 TH2

Example:

DPHASE 4 21 6 2.1 8 6 7.2

Field Contents
SID Identification number of DPHASE entry. (Integer > 0)
Pi Grid, extra, or scalar point identification number. (Integer > 0)
Ci Component number. (Integers 1 through 6 for grid points; zero or blank
for extra or scalar points)
THi Phase lead θ in degrees. (Real)

Remarks:
1. One or two dynamic load phase lead terms may be defined on a single entry.
2. SID must be referenced on a RLOADi entry. Refer to the RLOAD1 or RLOAD2 entry
for the formulas that define how the phase lead θ is used.
3. A DAREA and/or LSEQ entry should be used to define a load at Pi and Ci.
4. In superelement analysis, DPHASE entries may only be applied to loads on points in
the residual structure.

EIGR Real Eigenvalue Extraction Data

Defines data needed to perform real eigenvalue analysis.

Format:

1 2 3 4 5 6 7 8 9 10
EIGR SID METHOD F1 F2 NE ND

NORM G C

Example:

EIGR 13 LAN 12

Main Index
 APPENDIX 397

Field Contents
SID Set identification number. (Unique Integer > 0)
METHOD Method of eigenvalue extraction. (Character)
Modern Methods:
LAN Lanczos Method
AHOU Automatic selection of HOU or MHOU method. See
Remark 13.
Obsolete Methods:
INV Inverse Power method.
SINV Inverse Power method with enhancements.
GIV Givens method of tridiagonalization.
MGIV Modified Givens method.
HOU Householder method of tridiagonalization.
MHOU Modified Householder method.
AGIV Automatic selection of METHOD = “GIV” or “MGIV”. See
Remark 13.
NORM Method for normalizing eigenvectors. (Character: “MASS,” “MAX,” or
“POINT”; Default = “MASS”)
MASS Normalize to unit value of the generalized mass. (Default)
MAX Normalize to unit value of the largest component in the
analysis set.
POINT Normalize to a positive or negative unit value of the
component defined in fields 3 and 4. The POINT option is
not supported for METH=LAN. (Defaults to “MASS” if
defined component is zero.)
G Grid or scalar point identification number. Required only if
NORM = “POINT”. (Integer > 0)
C Component number. Required only if NORM = “POINT” and G is a
geometric grid point. (1 < Integer < 6)

Main Index
398

Table 6-1 Relationship Between METHOD Field and Other Fields for
Obsolete Methods

METHOD Field
Field
INV or SINV GIV, MGIV, HOU, or MHOU
F1, F2 Frequency range of interest. F1 Frequency range of interest. If ND is
must be input. If not blank, F1 and F2 are ignored. If ND
METHOD = “SINV” and ND, is is blank, eigenvectors are found with
blank, then F2 must be input. natural frequencies that lie in the range
(Real > 0.0) between F1 and F2. (Real > 0.0; F1 < F2)
NE Estimate of number of roots in Not used.
range (Required for
METHOD = “INV”). Not used
by “SINV” method.
(Integer > 0)
ND Desired number of roots. If this Desired number of eigenvectors. If ND
field is blank and is zero, the number of eigenvectors is
METHOD = “SINV”, then all determined from F1 and F2. If all three
roots between F1 and F2 are are blank, then ND is automatically set
searched and the limit is 600 to one more than the number of
roots. (Integer > 0, Default is degrees-of-freedom listed on SUPORTi
3 ⋅ NE for METHOD = “INV” entries. (Integer > 0; Default = 0)
only.)

Remarks:
1. The EIGR entry must be selected with the Case Control command METHOD = SID.
2. See “Real Eigenvalue Analysis” on page 33 of the MSC.Nastran Basic Dynamics User’s
Guide for a discussion of method selection.
3. The units of F1 and F2 are cycles per unit time.
4. The continuation entry is optional. If the continuation entry is not specified, then mass
normalization is performed.
5. The contemporary methods are LAN and AHOU. The other methods are in a
maintenance-only status, with no enhancements planned for them. They may be
eliminated in a future release of MSC.Nastran.
6. The LAN method is the most general-purpose method, and may be used on both small-
and large-size problems. It takes advantage of sparsity of input matrices, leading to
greater efficiency on large-size problems. Because Lanczos performance is tuned for
medium to large problems, this has caused difficulties with very small problems. Thus,
by default, on problems with fewer than 20 degrees-of-freedom when the LAN method
is selected, the method is switched to AHOU. The criteria for automatic switching is
controlled by SYSTEM(359) on the NASTRAN entry. The NE, G, and C fields are
Main Index
 APPENDIX 399

ignored for the LAN method. The NORM field may be set to MASS (the default value)
or NORM. The conventions used when both the Fi and ND fields are specified are
described in Table 1 of the EIGRL entry description. The EIGRL entry is an alternate
method to select the LAN method. It has several other input options for special cases.
When both and EIGRL and EIGR have the same SID and that SID is selected by a
METHOD command the EIGRL entry takes precedence.
7. The AHOU method is competitive with the LAN method when there are small, dense
matrices and many eigenvectors are required. This most commonly occurs when static
or dynamic reduction is performed. The AHOU method does not take advantage of
matrix sparsity, so that computation cost rises with the cube of the number of DOFs.
The AHOU method responds to all permitted values for all the other fields except NE,
which is ignored.
8. All methods require a positive semi-definite (psd) mass matrix for stable solutions. The
mass matrix may be tested for this condition for all methods of solution by setting
SYSTEM(303). A value of “-4” should be sufficient to identify problem matrices. A fatal
error exit is taken when it is not met. All MSC.Nastran metric elements are designed to
produce psd mass matrices. CMASSi elements, DMIG matrices selected by the M2GG
command, and matrices input via INPUTT4 are special methods that allow addition of
non-psd terms by use of non-metric element input. If none of this type of special input
is present and the fatal error exit is taken you may have encountered an error in a
metric element. Contact your local MSC technical representative for corrective action
in this case.
9. The LAN and AHOU methods allow singular but positive semi-definite mass matrices.
10. The tridiagonal methods include the xGIV and xHOU methods, where “x” is described
in the following comments. All tridiagonal methods compute all eigenvalues, and the
number of eigenvectors specified by the Fi and Nd fields, as described in Table 13.
11. If “x” is blank (for example, the HOU method is selected) the mass matrix must be non-
singular.
12. If “x” is M (for example, the MHOU method is selected) the mass matrix may be
singular. A modified, shifted problem is solved in an inverse basis with this method.
Some precision in the solution and longer computation time is exchanged for a more
stable solution.
13. If “x” is A (for example, the AHOU method is selected) an attempt is made to solve the
problem without shifting, in the interest of cost reduction and maximum precision. If
the mass matrix is determined to be poorly conditioned for inversion the basis is
automatically shifted with the modified method.
14. If NORM = “MAX”, components that are not in the analysis set may have values larger
than unity.
15. If NORM = “POINT”, the selected component should be in the analysis set (a-set). (The
program uses NORM = “MAX” when it is not in the analysis set.) The displacement
value at the selected component will be positive or negative unity.

Main Index
400

16. The “SINV” method is an enhanced version of the “INV” method. It uses Sturm
sequence number techniques to make it more likely that all roots in the range have been
found. It is generally more reliable and more efficient than the “INV” method.
–6
17. For the “INV” and “SINV” methods, convergence is achieved at 10 . Convergence is
tested by other criteria for the other methods.
18. For the “SINV” method only, if F2 is blank, the first shift will be made at F1, and only
one eigensolution above F1 will be calculated. If there are no modes below F1, it is
likely that the first mode will be calculated. If there are modes below F1 (including
rigid body modes defined by SUPORT entries), a mode higher than the first mode
above F1 may be calculated.
19. When F1, F2, and ND are all zero or blank, ND is reset to 1. A User Warning Message
is produced for this condition, which is interpreted as likely to be due to an inadvertent
omission by the user.

EIGRL Real Eigenvalue Extraction Data, Lanczos Method

Defines data needed to perform real eigenvalue (vibration or buckling) analysis with the
Lanczos method.
Format:

1 2 3 4 5 6 7 8 9 10
EIGRL SID V1 V2 ND MSGLVL MAXSET SHFSCL NORM

option_1 = value_1 option_2 = value_2, etc.

Example:

EIGRL 1 0.1 3.2 10

NORM=MAX NUMS=2

Field Contents
SID Set identification number. (Unique Integer > 0)
V1, V2 For vibration analysis: frequency range of interest. For buckling
analysis: eigenvalue range of interest. See Remark 4. (Real or blank,
16 16
– 5 × 10 ≤ V1 < V2 ≤ 5. × 10 )
ND Number of roots desired. See Remark 4. (Integer > 0 or blank)
MSGLVL Diagnostic level. (0 < Integer < 4; Default = 0)
MAXSET Number of vectors in block or set. Default is machine dependent. See
Remark 14.
SHFSCL Estimate of the first flexible mode natural frequency. See Remark 10.
(Real or blank)
NORM Method for normalizing eigenvectors (Character: “MASS” or “MAX”)

Main Index
 APPENDIX 401

Field Contents
MASS Normalize to unit value of the generalized mass. Not
available for buckling analysis. (Default for normal modes
analysis.)
MAX Normalize to unit value of the largest displacement in the
analysis set. Displacements not in the analysis set may be
larger than unity. (Default for buckling analysis.)
ALPH Specifies a constant for the calculation of frequencies (Fi) at the upper
boundary segments for the parallel method based on the following
formula. See Remark 13. (Integer > 0.0; Default = 1.0):
i
1 – ALPH
Fi = ( V2 – V1 ) --------------------------------------------
NUMS
1. – ALPH
NUMS Number of frequency segments for the parallel method. (Integer > 0;
Default = 1)
Fi Frequency at the upper boundary of the i-th segment. See Remark 13.
(Real or blank; V1 < F1 < F2 < …F15 < V2 )
option_i= Assigns a value to the fields above except for SID. ALPH, NUMS, and Fi
value_i must be specified in this format. V1, V2, ND, MSGLVL, MAXSET,
SHFSCL, and NORM may be specified in this format as long as their
corresponding field is blank in the parent entry.

Remarks:
1. Real eigenvalue extraction data sets must be selected with the Case Control command
METHOD = SID.
2. The units of V1 and V2 are cycles per unit time in vibration analysis, and are
eigenvalues in buckling analysis. Each eigenvalue is the factor by which the
prebuckling state of stress is multiplied to produce buckling in the shape defined by the
corresponding eigenvector.
3. NORM = “MASS” is ignored in buckling analysis and NORM = “MAX” will be
applied.
4. The roots are found in order of increasing magnitude; that is, those closest to zero are
found first. The number and type of roots to be found can be determined from Table 6-
2.

Table 6-2 Number and Type of Roots Found with EIGRL Entry

V1 V2 ND Number and Type of Roots Found

V1 V2 ND Lowest ND or all in range, whichever is smaller.
V1 V2 blank All in range

Main Index
402

Table 6-2 Number and Type of Roots Found with EIGRL Entry

V1 V2 ND Number and Type of Roots Found

V1 blank ND Lowest ND in range [V1,+∞]
V1 blank blank Lowest root in range [V1,+∞]
blank blank ND Lowest ND roots in [-∞,+∞]
blank blank blank Lowest root. See Remark 11.
blank V2 ND Lowest ND roots below V2
blank V2 blank All below V2

5. In vibration analysis, if V1 < 0.0, the negative eigenvalue range will be searched.
(Eigenvalues are proportional to Vi squared; therefore, the negative sign would be
lost.) This is a means for diagnosing improbable models. In buckling analysis, negative
V1 and/or V2 require no special logic.
6. Eigenvalues are sorted on order of magnitude for output. An eigenvector is found for
each eigenvalue.
7. MSGLVL controls the amount of diagnostic output during the eigenvalue extraction
process. The default value of zero suppresses all diagnostic output. A value of one
prints eigenvalues accepted at each shift. Higher values result in increasing levels of
diagnostic output.
8. MAXSET is used to limit the maximum block size. It is otherwise set by the region size
or by ND with a maximum size of 15. It may also be reset if there is insufficient memory
available. The default value is recommended.
9. In vibration analysis, if V1 is blank, all roots less than zero are calculated. Small
negative roots are usually computational zeroes which indicate rigid body modes.
Finite negative roots are an indication of modeling problems. If V1 is set to zero,
negative eigenvalues are not calculated.
10. A specification for SHFSCL may improve performance, especially when large mass
techniques are used in enforced motion analysis. Large mass techniques can cause a
large gap between the rigid body and flexible frequencies. If this field is blank, a value
for SHFSCL is estimated automatically.
11. On occasion, it may be necessary to compute more roots than requested to ensure that
all roots in the range have been found. However, this method will not output the
additional roots.
12. NASTRAN SYSTEM(146) provides options for I/O in sparse method only:

Main Index
 APPENDIX 403

Table 6-3 SYSTEM(146) Options

SYSTEM(146) Description
2 Increase memory reserved for sparse method
by approximately 100%.
3 Increase memory reserved for sparse method
by approximately 300%.
4 Increase memory reserved for sparse method
by approximately 400%.

13. For the distributed parallel method, the frequency range between V1 and V2 may be
subdivided into segments that can then be analyzed in parallel. V1 and V2 must be
specified for the parallel method. NUMS must be specified greater than 1 to take
advantage of the parallel method. NUMS may also be specified on the NUMSEG
keyword of the NASTRAN statement. Currently, NUMSEG must equal the number of
processors and by default NUMSEG is set to the number of processors requested by the
DMP keyword. If both are specified, then NUMS takes precedence.
The upper frequencies of each segment may be generated automatically by ALPH or
specified directly in Fi. If both are specified, then Fi takes precedence over ALPH as
long as they are consistent. ALPH if multiplied by 100 may also be specified on
FRQSEQ keyword of the NASTRAN statement.
14. Increasing MAXSET may improve performance for large problems where a large
number of eigenvalues are being found. The default is 7 on all machines except CRAY
which is 12. SYSTEM(263) may be set in an rcfile to effectively modify the default;
however the setting on the EIGRL entry always takes precedence.
15. SYSTEM(196), keyword SCRSAVE, controls reuse of scratch files when segment logic
is invoked. SYSTEM(196) is useful only when multiple frequency segments are
requested on a Lanczos run. (Multiple frequency segments can be requested via the
NUMS field in the EIGRL entry and by SYSTEM(197).) Each frequency segment
requires a minimum of three scratch files. When multiple frequency segments are used
on a single processor computer then each frequency segment is solved serially. In this
case, it makes sense to let segment #2 use the scratch files which were used by segment
#1 since work for segment #1 has been completed (otherwise it wouldn’t be working on
#2). Similarly, when work for segment #2 is finished, segment #3 should be able to use
#2’s scratch files. SYSTEM(196)=1 allows such file reuse and is considered a safe
default on Version 70 and later systems.
16. The new buckling shift logic in Version 70.5 tends to shift to 1.0 first. The logic may
have difficulty finding the lowest ND roots if a problem requests a small number of
roots (ND) when there are thousands of roots below 1. In this case either the loading
should be scaled, SHFSCL specified, or a smaller frequency range requested.

Main Index
404

17. Because Lanczos performance is tuned for medium to large problems, this has caused
difficulties with very small problems. Thus, by default, on problems with fewer than
20 degrees-of-freedom when the LAN method is selected, the method is switched to
AHOU. The criteria for automatic switching is controlled by SYSTEM(359) on the
NASTRAN entry.

FREQ Frequency List

Defines a set of frequencies to be used in the solution of frequency response problems.

Format:

1 2 3 4 5 6 7 8 9 10
FREQ SID F1 F2 F3 F4 F5 F6 F7

F8 F9 F10 -etc.-

Example:

FREQ 3 2.98 3.05 17.9 21.3 25.6 28.8 31.2

29.2 22.4 19.3

Field Contents
SID Set identification number. (Integer > 0)
Fi Frequency value in units of cycles per unit time. (Real > 0.0)

Remarks:
1. Frequency sets must be selected with the Case Control command FREQUENCY = SID.
2. All FREQi entries with the same frequency set identification numbers will be used.
Duplicate frequencies will be ignored. f N and f N – 1 are considered duplicated if

f N – f N – 1 < DFREQ ⋅ f MAX – f MIN ,

where DFREQ is a user parameter, with a default of 10 –5 . f MAX and f MIN are the
maximum and minimum excitation frequencies of the combined FREQi entries.
3. In modal analysis, solutions for modal DOFs from rigid body modes at zero excitation
frequencies may be discarded. Solutions for nonzero modes are retained.

FREQ1 Frequency List, Alternate Form 1

Defines a set of frequencies to be used in the solution of frequency response problems by

specification of a starting frequency, frequency increment, and the number of increments
desired.

Main Index
 APPENDIX 405

Format:

1 2 3 4 5 6 7 8 9 10
FREQ1 SID F1 DF NDF

Example:

FREQ1 6 2.9 0.5 13

Field Contents
SID Set identification number. (Integer > 0)
F1 First frequency in set. (Real > 0.0)
DF Frequency increment. (Real > 0.0)
NDF Number of frequency increments. (Integer > 0; Default = 1)

Remarks:
1. FREQ1 entries must be selected with the Case Control command FREQUENCY = SID.
2. The units for F1 and DF are cycles per unit time.
3. The frequencies defined by this entry are given by

f i = F1 + DF ⋅ ( i – 1 )

where i = 1 to (NDF + 1).

4. All FREQi entries with the same frequency set identification numbers will be used.
Duplicate frequencies will be ignored. f N and f N – 1 are considered duplicated if

f N – f N – 1 < DFREQ ⋅ f MAX – f MIN ,

where DFREQ is a user parameter, with a default of 10 –5 . f MAX and f MIN are the
maximum and minimum excitation frequencies of the combined FREQi entries.
5. In modal analysis, solutions for modal DOFs from rigid body modes at zero excitation
frequencies may be discarded. Solutions for nonzero modes are retained.

FREQ2 Frequency List, Alternate Form 2

Defines a set of frequencies to be used in the solution of frequency response problems by

specification of a starting frequency, final frequency, and the number of logarithmic increments
desired.
Format:

1 2 3 4 5 6 7 8 9 10
FREQ2 SID F1 F2 NF

Main Index
406

Example:

FREQ2 6 1.0 8.0 6

Field Contents
SID Set identification number. (Integer > 0)
F1 First frequency. (Real > 0.0)
F2 Last frequency. (Real > 0.0, F2 > F1)
NF Number of logarithmic intervals. (Integer > 0; Default = 1)

Remarks:
1. FREQ2 entries must be selected with the Case Control command FREQUENCY = SID.
2. The units for F1 and F2 are cycles per unit time.
3. The frequencies defined by this entry are given by

( i – 1 )d
f i = F1 ⋅ e
1 F2
- ln ------- and i = 1, 2, …, ( NF + 1 )
where d = -------
NF F1
In the example above, the list of frequencies will be 1.0, 1.4142, 2.0, 2.8284, 4.0, 5.6569
and 8.0 cycles per unit time.
4. All FREQi entries with the same frequency set identification numbers will be used.
Duplicate frequencies will be ignored. f N and f N – 1 are considered duplicated if

f N – f N – 1 < DFREQ ⋅ f MAX – f MIN ,

where DFREQ is a user parameter, with a default of 10 –5 . f MAX and f MIN are the
maximum and minimum excitation frequencies of the combined FREQi entries.
5. In modal analysis, solutions for modal DOFs from rigid body modes at zero excitation
frequencies may be discarded. Solutions for nonzero modes are retained.

FREQ3 Frequency List, Alternate 3

Defines a set of excitation frequencies for modal frequency-response solutions by specifying

number of excitation frequencies between two modal frequencies.
Format:

1 2 3 4 5 6 7 8 9 10
FREQ3 SID F1 F2 TYPE NEF CLUSTER

Example:

FREQ3 6 20.0 200.0 LINEAR 10 2.0

Main Index
 APPENDIX 407

Field Contents
SID Set identification number. (Integer > 0)
F1 Lower bound of modal frequency range in cycles per unit time.
(Real > 0.0 for TYPE = LINEAR and Real = 0.0 for TYPE = LOG)
F2 Upper bound of modal frequency range in cycles per unit time.
(Real > 0.0, F2 > F1, Default = F1)
TYPE LINEAR or LOG. Specifies linear or logarithmic interpolation between
frequencies. (Character; Default = "LINEAR")
NEF Number of excitation frequencies within each subrange including the
end points. The first subrange is between F1 and the first modal
frequency within the bounds. The second subrange is between first and
second modal frequencies between the bounds. The last subrange is
between the last modal frequency within the bounds and F2.
(Integer > 1, Default = 10)
CLUSTER Specifies clustering of the excitation frequency near the end points of the
range. See Remark 6. (Real > 0.0; Default = 1.0)

Remarks:
1. FREQ3 applies only to modal frequency-response solutions (SOLs 11, 111, 146, and 200)
and is ignored in direct frequency response solutions.
2. FREQ3 entries must be selected with the Case Control command FREQUENCY = SID.
3. In the example above, there will be 10 frequencies in the interval between each set of
modes within the bounds 20 and 2000, plus 10 frequencies between 20 and the lowest
mode in the range, plus 10 frequencies between the highest mode in the range and 2000.
4. Since the forcing frequencies are near structural resonances, it is important that some
amount of damping be specified.
5. All FREQi entries with the same set identification numbers will be used. Duplicate
frequencies will be ignored. f N and f N – 1 are considered duplicated if

f N – f N – 1 < DFREQ ⋅ f MAX – f MIN

where DFREQ is a user parameter, with a default of 10 –5 . f MAX an f MIN are the
maximum and minimum excitation frequencies of the combined FREQi entries.
6. CLUSTER is used to obtain better resolution near the modal frequencies where the
response varies the most. CLUSTER > 1.0 provides closer spacing of excitation
frequency towards the ends of the frequency range, while values of less than 1.0
provide closer spacing towards the center of the frequency range. For example, if the
frequency range is between 10 and 20, NEF = 11, TYPE = "LINEAR"; then, the
excitation frequencies for various values of CLUSTER would be as shown in Table 6-4.

1 1 1 ⁄ CLUSTER
f̂ k = --- ( f̂ 1 + f̂ 2 ) + --- ( f̂ 2 – f̂ 1 ) ξ ⋅ SIGN ( ξ )
2 2

Main Index
408

where

ξ = – 1 + 2 ( k – 1 ) ⁄ ( NEF – 1 ) is a parametric coordinate between -1

and 1
k = varies from 1 to NEF (k = 1, 2, ..., NEF)
f̂ 1 = is the lower limit of the frequency subrange

f̂ 2 = is the upper limit o the subrange

f̂ k = is the k-th excitation frequency

f̂ = is the frequency, or the logarithm of the frequency, depending on

the value specified for TYPE

Table 6-4 CLUSTER Usage Example

CLUSTER
Excitation
Frequenc ξ c=0.25 c-0.50 c-1.0 c-2.0 c-4.0
y Number
Excitation Frequencies in Hertz
1 -1.0 10.00 10.0 10.0 10.00 10.00
2 -0.8 12.95 11.8 11.0 10.53 10.27
3 -0.6 14.35 13.2 12.0 11.13 10.60
4 -0.4 14.87 14.2 13.0 11.84 11.02
5 -0.2 14.99 14.8 14.0 12.76 11.66
6 0.0 15.00 15.0 15.0 15.00 15.00
7 0.2 15.01 15.2 16.0 17.24 18.34
8 0.4 15.13 15.8 17.0 18.16 18.98
9 0.6 15.65 16.8 18.0 18.87 19.40
10 0.8 17.05 18.2 19.0 19.47 19.73
11 1.0 20.00 20.0 20.0 20.00 20.00

7. In design optimization (SOL 200), the excitation frequencies are derived from the
natural frequencies computed at each design cycle.
8. In modal analysis, solutions for modal DOFs from rigid body modes at zero excitation
frequencies may be discarded. Solutions for nonzero modes are retained.

FREQ4 Frequency List, Alternate Form 4

Defines a set of frequencies used in the solution of modal frequency-response problems by

specifying the amount of “spread” around each natural frequency and the number of equally
spaced excitation frequencies within the spread.
Main Index
 APPENDIX 409

Format:

1 2 3 4 5 6 7 8 9 10
FREQ4 SID F1 F2 FSPD NFM

Example:

FREQ4 6 20.0 200.0 0.30 21

Field Contents
SID Set identification number. (Integer > 0)
F1 Lower bound of frequency range in cycles per unit time. (Real > 0.0,
Default = 0.0)
F2 Upper bound of frequency range in cycles per unit time. (Real > 0.0,
F2 > F1, Default = 1.0E20)
FSPD Frequency spread, +/- the fractional amount specified for each mode
which occurs in the frequency range F1 to F2. (1.0 > Real > 0.0,
Default = 0.10)
NFM Number of evenly spaced frequencies per “spread” mode. (Integer > 0;
Default = 3; If NFM is even, NFM + 1 will be used.)

Remarks:
1. FREQ4 applies only to modal frequency-response solutions (SOLs 11, 111, 146, and 200
and is ignored in direct frequency-response solutions.
2. FREQ4 entries must be selected with the Case Control command FREQUENCY = SID.
3. There will be NFM excitation frequencies between ( 1 – FSPD ) ⋅ f N and ( 1 + FSPD ) ⋅ f N , for
each natural frequency in the range F1 to F2.
4. In the example above there will be 21 equally spaced frequencies across a frequency
band of 0.7 ⋅ f N to 1.3 ⋅ f N for each natural frequency that occurs between 20 and 2000.
See Figure 6-1 for the definition of frequency spread.
( 1 – FSPD ) ⋅ f N fN ( 1 + FSPD ) ⋅ fN

Figure 6-1 Frequency Spread Definition

Excitation frequencies may be based on natural frequencies that are not within the
range (F1 and F2) as long as the calculated excitation frequencies are within the range.
Similarly, an excitation frequency calculated based on natural frequencies within the
range (F1 through F2) may be excluded if it falls outside the range.
5. The frequency spread can be used also to define the half-power bandwidth. The
half-power bandwidth is given by 2 ⋅ ξ ⋅ f N , where ξ is the damping ratio. Therefore, if
FSPD is specified equal to the damping ratio for the mode, NFM specifies the number
of excitation frequency within the half-power bandwidth. See Figure 6-2 for the
definition of half-power bandwidth.
Main Index
 Peak Response

2 ⋅ ξ ⋅ fN
410

Amplitude
Half-Power Point (.707 Peak)

Figure 6-2 Half-Power Bandwidth Definition

6. Since the forcing frequencies are near structural resonances, it is important that some
amount of damping be specified.
7. All FREQi entries with the same set identification numbers will be used. Duplicate
frequencies will be ignored. f N and fN – 1 are considered duplicated if

f N – f N – 1 < DFREQ ⋅ f MAX – f MIN

fN Frequency
where DFREQ is a user parameter with a default of 10 –5 . The values f MAX and f MIN are
Half-Power
the maximum and minimumBandwidth
excitation frequencies of the combined FREQi entries.
8. In design optimization (SOL 200), the excitation frequencies are derived from the
natural frequencies computed at each design cycle.
9. In modal analysis, solutions for modal DOFs from rigid body modes at zero excitation
frequencies may be discarded. Solutions for nonzero modes are retained.

FREQ5 Frequency List, Alternate Form 5

Defines a set of frequencies used in the solution of modal frequency-response problems by

specification of a frequency range and fractions of the natural frequencies within that range.
Format:

1 2 3 4 5 6 7 8 9 10
FREQ5 SID F1 F2 FR1 FR2 FR3 FR4 FR5

FR6 FR7 -etc.-

Example:

FREQ5 6 20.0 200.0 1.0 0.6 0.8 0.9 0.95

1.05 1.1 1.2

Field Contents
SID Set identification number. (Integer > 0)
F1 Lower bound of frequency range in cycles per unit time. (Real > 0.0;
Default = 0.0)
F2 Upper bound of frequency range in cycles per unit time. (Real > 0.0,
F2 > F1, Default = 1.0E20)
FRi Fractions of the natural frequencies in the range F1 to F2. (Real > 0.0)

Remarks:
1. FREQ5 applies only to modal frequency-response solutions (SOLs 11, 111, 146, and 200)
and is ignored in direct frequency response solutions.
Main Index
 APPENDIX 411

2. FREQ5 entries must be selected with the Case Control command FREQUENCY = SID.
3. The frequencies defined by this entry are given by

f i = FRi ⋅ f N
i

where f N are the natural frequencies in the range F1 through F2.

i
4. In the example above, the list of frequencies will be 0.6, 0.8, 0.9, 0.95, 1.0, 1.05, 1.1,
and 1.2 times each natural frequency between 20 and 2000. If this computation results
in excitation frequencies less then F1 and greater than F2, those computed excitation
frequencies are ignored.
Excitation frequencies may be based on natural frequencies that are not within the
range (F1 and F2) as long as the calculated excitation frequencies are within the range.
Similarly, an excitation frequency calculated based on natural frequencies within the
range (F1 through F2) may be excluded if it falls outside the range.
5. Since the forcing frequencies are near structural resonances, it is important that some
amount of damping be specified.
6. All FREQi entries with the same set identification numbers will be used. Duplicate
frequencies will be ignored. f N and f N – 1 are considered duplicated if

f N – f N – 1 < DFREQ ⋅ f MAX – f MIN

where DFREQ is a user parameter with a default of 10 –5 . The values f MAX and f MIN are
the maximum and minimum excitation frequencies of the combined FREQi entries.
7. In design optimization (SOL 200), the excitation frequencies are derived from the
natural frequencies computed at each design cycle.
8. In modal analysis, solutions for modal DOFs from rigid body modes at zero excitation
frequencies may be discarded. Solutions for nonzero modes are retained

LSEQ Static Load Set Definition

Defines a sequence of static load sets.

Format:

1 2 3 4 5 6 7 8 9 10
LSEQ SID EXCITEID LID TID

Example:

LSEQ 100 200 1000 1001

Main Index
412

Field Contents
SID Set identification of the set of LSEQ entries. See Remark 5. (Integer > 0)
EXCITEID The EXCITEID set identification assigned to this static load vector. See
Remark 5. (Integer > 0)
LID Load set identification number of a set of static load entries such as those
referenced by the LOAD Case Control command. (Integer > 0 or blank)
TID Temperature set identification of a set of thermal load entries such as
those referenced by the TEMP(LOAD) Case Control command.
(Integer > 0 or blank)

Remarks:
1. LSEQ will not be used unless selected in the Case Control Section with the LOADSET
command.
2. The number of static load vectors created for each superelement depends upon the type
of analysis. In static analysis, the number of vectors created is equal to the number of
unique EXCITEID IDs on all LSEQ entries in the Bulk Data; in dynamic analysis, the
number of vectors created is equal to the number of unique EXCITEID IDs on all
RLOAD1, RLOAD2, TLOAD1, TLOAD2 and ACSRCE entries in the Bulk Data.
3. EXCITEID may be referenced by CLOAD, RLOAD1, RLOAD2, TLOAD1, and
TLOAD2 entries in order to apply the static load in nonlinear, static and dynamic
analysis.
4. Element data recovery for thermal loads is not currently implemented in dynamics.
5. The SID-EXCITEID number pair must be unique with respect to similiar pairs on all
other LSEQ entries in the Bulk Data.
6. In a nonsuperelement analysis, LID and TID cannot both be blank. In superelement
analysis, they may both be blank as long as static loads are prescribed in the upstream
superelements.
7. It is no longer necessary to employ LOADSET/LSEQ data to specify static loading data
for use in dynamic analysis. In the absence of LSEQ Bulk Data entries, all static loads
whose load set IDs match the EXCITEID IDs on all RLOAD1, RLOAD2, TLOAD1,
TLOAD2 and ACSRCE entries in the Bulk Data are automatically processed.

PDAMP Scalar Damper Property

Specifies the damping value of a scalar damper element using defined CDAMP1 or CDAMP3
entries.
Format:

1 2 3 4 5 6 7 8 9 10
PDAMP PID1 B1 PID2 B2 PID3 B3 PID4 B4

Main Index
 APPENDIX 413

Example:

PDAMP 14 2.3 2 6.1

Field Contents
PIDi Property identification number. (Integer > 0)
Bi Force per unit velocity. (Real)

Remarks:
1. Damping values are defined directly on the CDAMP2 and CDAMP4 entries, and
therefore do not require a PDAMP entry.
2. A structural viscous damper, CVISC, may also be used for geometric grid points.
3. Up to four damping properties may be defined on a single entry.
4. For a discussion of scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.

PMASS Scalar Mass Property

Specifies the mass value of a scalar mass element (CMASS1 or CMASS3 entries).
Format:

1 2 3 4 5 6 7 8 9 10
PMASS PID1 M1 PID2 M2 PID3 M3 PID4 M4

Example:

PMASS 7 4.29 6 13.2

Field Contents
PIDi Property identification number. (Integer > 0)
Mi Value of scalar mass. (Real)

Remarks:
1. Mass values are defined directly on the CMASS2 and CMASS4 entries, and therefore
do not require a PMASS entry.
2. Up to four mass values may be defined by this entry.
3. For a discussion of scalar elements, see “Scalar Elements (CELASi, CMASSi,
CDAMPi)” on page 193 of the MSC.Nastran Reference Guide.

PVISC Viscous Damping Element Property

Defines properties of a one-dimensional viscous damping element (CVISC entry).

Main Index
414

Format:

1 2 3 4 5 6 7 8 9 10
PVISC PID1 CE1 CR1 PID2 CE2 CR2

Example:

PVISC 3 6.2 3.94

Field Contents
PIDi Property identification number. (Integer > 0)
CE1, CE2 Viscous damping values for extension in units of force per unit velocity.
(Real)
CR1, CR2 Viscous damping values for rotation in units of moment per unit
velocity. (Real)

Remarks:
1. Viscous properties are material independent; in particular, they are temperature
independent.
2. One or two viscous element properties may be defined on a single entry.

RLOAD1 Frequency Response Dynamic Excitation, Form 1

Defines a frequency-dependent dynamic load of the form

i { θ – 2πfτ }
{ P ( f ) } = { A [ C ( f ) + iD ( f ) ]e }

for use in frequency response problems.

Format:

1 2 3 4 5 6 7 8 9 10
RLOAD1 SID EXCITEID DELAY DPHASE TC TD TYPE

Example:

RLOAD1 5 3 1

Field Contents
SID Set identification number. (Integer > 0)
EXCITEID Identification number of the DAREA or SPCD entry set that defines A.
See Remarks 5. and 6. (Integer > 0)

Main Index
 APPENDIX 415

Field Contents
DELAY Defines time delay τ. (Integer > 0, real or blank.) If it is a non-zero
integer, it represents the identification number of DELAY Bulk Data
entry that defines τ . If it is real, then it directly defines the value of τ that
will be used for all degrees of freedom that are excited by this dynamic
load entry. See also Remark 2.
DPHASE Defines phase angle θ. (Integer > 0, real or blank.) If it is a non-zero
integer, it represents the identification number of DPHASE Bulk Data
entry that defines θ (in degrees). If it is real, then it directly defines the
vlaue of θ (in degrees) that will be used for all degrees of freedom that
are excited by this dynamic load entry. See also Remark 2.
TC Set identification number of the TABLEDi entry that gives C ( f ) . See
Remark 2. (Integer > 0)
TD Set identification number of the TABLEDi entry that gives D ( f ) . See
Remark 2. (Integer > 0)
TYPE Defines the type of the dynamic excitation. See Remarks 5. and 6.
(Integer, character or blank; Default = 0)

Remarks:
1. Dynamic excitation sets must be selected with the Case Control command
DLOAD = SID.
2. If any of DELAY, DPHASE, TC, or TD fields are blank or zero, the corresponding τ,
θ, C ( f ) or D ( f ) will both be zero. Either TC or TD may be blank or zero, but not both.
3. RLOAD1 excitations may be combined with RLOAD2 excitations only by specification
on a DLOAD entry. That is, the SID on a RLOAD1 entry must not be the same as that
on a RLOAD2 entry.
4. SID must be unique for all RLOAD1, RLOAD2, TLOAD1, TLOAD2, and ACSRCE
entries.
5. The type of the dynamic excitation is specified by TYPE (field 8) according to the
following table:

TYPE TYPE of Dynamic Excitation

0, L, LO, LOA or LOAD Applied load (force or moment) (Default)
1, D, DI, DIS or DISP Enforced displacement using SPC/SPCD data
2, V, VE, VEL or VELO Enforced velocity using SPC/SPCD data
3, A, AC, ACC or ACCE Enforced acceleration SPC/SPCD data

6. TYPE (field 8) also determines the manner in which EXCITEID (field 3) is used by the
program as described below:
Main Index
416

Excitation specified by TYPE is applied load

• There is no LOADSET request in Case Control
EXCITEID may also reference DAREA, static and thermal load set entries.
• There is a LOADSET request in Case Control
The program may also reference static and thermal load set entries specified by
the LID or TID field in the selected LSEQ entries corresponding to the
EXCITEID.
Excitation specified by TYPE is enforced motion
• There is no LOADSET request in Case Control
EXCITEID will reference SPCD entries.
• There is a LOADSET request in Case Control
The program will reference SPCD entries specified by the LID field in the
selected LSEQ entries corresponding to the EXCITEID.

RLOAD2 Frequency Response Dynamic Excitation, Form 2

Defines a frequency-dependent dynamic excitation of the form.

i { φ ( f ) + θ – 2πfτ }
{ P( f) } = { A ⋅ B( f) e }

for use in frequency response problems.

Format:

1 2 3 4 5 6 7 8 9 10
RLOAD2 SID EXCITEID DELAY DPHASE TB TP TYPE

Example:

RLOAD2 5 3 7

Field Contents
SID Set identification number. (Integer > 0)
EXCITEID Identification number of the DAREA or SPCD entry set that defines A.
See Remarks 5. and 6. (Integer > 0)
DELAY .Defines time delay τ. (Integer > 0, real or blank.) If it is a non-zero
integer, it represents the identification number of DELAY Bulk Data
entry that defines τ . If it is real, then it directly defines the value of τ that
will be used for all degrees of freedom that are excited by this dynamic
load entry. See also Remark 2.

Main Index
 APPENDIX 417

Field Contents
DPHASE Defines phase angle θ. (Integer > 0, real or blank.) If it is a non-zero
integer, it represents the identification number of DPHASE Bulk Data
entry that defines θ (in degrees). If it is real, then it directly defines the
vlaue of θ (in degrees) that will be used for all degrees of freedom that
are excited by this dynamic load entry. See also Remark 2.
TB Set identification number of the TABLEDi entry that gives B ( f ) .
(Integer > 0)
TP Set identification number of the TABLEDi entry that gives φ ( f ) in
degrees. (Integer > 0)
TYPE Defines the type of the dynamic excitation. See Remarks 5. and 6.
(Integer, character or blank; Defaults = 0)

Remarks:
1. Dynamic excitation sets must be selected with the Case Control command
DLOAD = SID.
2. If any of DELAY, DPHASE, or TP fields are blank or zero, the corresponding τ, θ, or
φ ( f ) will be zero.
3. RLOAD2 excitations may be combined with RLOAD1 excitations only by specification
on a DLOAD entry. That is, the SID on a RLOAD2 entry must not be the same as that
on a RLOAD1 entry.
4. SID must be unique for all RLOAD1, RLOAD2, TLOAD1, TLOAD2, and ACSRCE
entries.
5. The type of the dynamic excitation is specified by TYPE (field 8) according to the
following table:

TYPE TYPE of Dynamic Excitation

0, L, LO, LOA or LOAD Applied load (force or moment) (Default)
1, D, DI, DIS or DISP Enforced displacement using SPC/SPCD data
2, V, VE, VEL or VELO Enforced velocity using SPC/SPCD data
3, A, AC, ACC or ACCE Enforced acceleration SPC/SPCD data

6. TYPE (field 8) also determines the manner in which EXCITEID (field 3) is used by the
program as described below:
Excitation specified by TYPE is applied load
• There is no LOADSET request in Case Control
EXCITEID may also reference DAREA, static and thermal load set entries.
• There is a LOADSET request in Case Control
Main Index
418

The program may also reference static and thermal load set entries specified by
the LID or TID field in the selected LSEQ entries corresponding to the
EXCITEID.
Excitation specified by TYPE is enforced motion
• There is no LOADSET request in Case Control
EXCITEID will reference SPCD entries.
• There is a LOADSET request in Case Control
The program will reference SPCD entries specified by the LID field in the
selected LSEQ entries corresponding to the EXCITEID.

SUPORT Fictitious Support

Defines determinate reaction degrees-of-freedom in a free body.

Format:

1 2 3 4 5 6 7 8 9 10
SUPORT ID1 C1 ID2 C2 ID3 C3 ID4 C4

Example:

SUPORT 16 215

Field Contents
IDi Grid or scalar point identification number. (Integer > 0)
Ci Component numbers. (Integer 0 or blank for scalar points. Any unique
combination of the Integers 1 through 6 for grid points with no
embedded blanks.)

Remarks:
1. The SUPORT entry specifies reference degrees-of-freedom for rigid body motion. It is
not intended to be used in place of a constraint (i.e., SPCi entry or PS on the GRID
entry).
2. SUPORT and/or SUPORT1 entries are required to perform inertia relief in static
analysis (SOLs 1 and 101).
In SOL 1, the presence of a SUPORT or selection of a SUPORT1 entry automatically
invokes inertia relief analysis.
In SOL 101, PARAM,INREL,-1 must also be specified or the SUPORTi entries will be
treated as constraints.
3. Be careful not to spell SUPORT with two Ps.

Main Index
 APPENDIX 419

4. Degrees-of-freedom specified on this entry form members of the mutually exclusive

r-set. They may not be specified on other entries that define mutually exclusive sets.
See “Degree-of-Freedom Sets” on page 843 for a list of these entries.
5. From 1 to 24 support degrees-of-freedom may be defined on a single entry.
6. See “Rigid Body Supports” on page 331 of the MSC.Nastran Reference Guide for a
discussion of supported degrees-of-freedom (members of the r-set).
7. An alternative to SUPORT is the SUPORT1 entry, which is requested by the SUPORT1
Case Control command.

SUPORT1 Fictitious Support, Alternate Form

Defines determinate reaction degrees-of-freedom (r-set) in a free body-analysis. SUPORT1 must

be requested by the SUPORT1 Case Control command.
Format:

1 2 3 4 5 6 7 8 9 10
SUPORT1 SID ID1 C1 ID2 C2 ID3 C3

Example:

SUPORT1 5 16 215

Field Contents
SID Identification number of the support set. See Remark 1. (Integer > 0)
IDi Grid or scalar point identification number. (Integer > 0)
Ci Component numbers. (Integer 0 or blank for scalar points. Any unique
combination of the Integers 1 through 6 for grid points with no
embedded blanks.)

Remarks:
1. The SUPORT entry specifies reference degrees-of-freedom for rigid body motion. It is
not intended to be used in place of a constraint; (i.e., SPCi entry or PS on the GRID
entry).
2. SUPORT and/or SUPORT1 entries are required to perform inertia relief in static
analysis; i.e., SOLs 1 and 101.
In SOL 1, the presence of a SUPORT or selection of a SUPORT1 entry automatically
invokes inertia relief analysis.
In SOL 101, PARAM,INREL,-1 must also be specified or the SUPORTi entries will be
treated as constraints.
3. SUPORT1 must be requested by the SUPORT1 Case Control command. The degrees-
of-freedom specified on SUPORT1 will be combined with those on the SUPORT entry.
4. Be careful not to spell SUPORT with two Ps.
Main Index
420

5. Degrees-of-freedom specified on this entry form members of the mutually exclusive

r-set. They may not be specified on other entries that define mutually exclusive sets.
See “Degree-of-Freedom Sets” on page 843 for a list of these entries.
6. From 1 to 18 support degrees-of-freedom may be defined on a single entry.
7. See “Rigid Body Supports” on page 331 of the MSC.Nastran Reference Guide for a
discussion of supported degrees-of-freedom (members of the r-set).
8. In superelement analysis, SUPORT1 may be specified for points belonging to the
residual structure only.

TABDMP1 Modal Damping Table

Defines modal damping as a tabular function of natural frequency.

Format:

1 2 3 4 5 6 7 8 9 10
TABDMP1 TID TYPE

f1 g1 f2 g2 f3 g3 -etc.-

Example:

TABDMP1 2

2.5 .01057 2.6 .01362 ENDT

Field Contents
TID Table identification number. (Integer > 0)
TYPE Type of damping units. (Character: “G”, “CRIT”, or “Q”; Default is
“G”)
fi Natural frequency value in cycles per unit time. (Real > 0.0)
gi Damping value. (Real)

Remarks:
1. Modal damping tables must be selected with the Case Control command
SDAMPING = TID.
2. The frequency values, fi, must be specified in either ascending or descending order, but
not both.
3. Discontinuities may be specified between any two points except the two starting points
or two end points. For example, in Figure 6-3 discontinuities are allowed only between
points f2 through f7. Also, if g is evaluated at a discontinuity, then the average value
of g is used. In Figure 6-3, the value of g at f = f3 is g = ( g3 + g4 ) ⁄ 2 .
4. At least one continuation entry must be specified.
Main Index
 APPENDIX 421

g
5. Any fi or gi entry may be ignored by placing “SKIP” in either of the two fields used for
that entry.
f value
6. The endDiscontinuity
of the table is indicatedRange
by the existence of “ENDT” in either of the two fields
of Table
following the last entry. An error is detected if any continuations
Allowed follow the entry
Discontinuity
containing the end-of-table flag “ENDT”. Not Allowed
Linear
7. The TABDMP1 uses the algorithm
Extrapolation
of Segment
f1-f2 g = gT ( f )

where f is input to the table and g is returned. The table look-up

f g T ( f ) is performed
using linear interpolation
f1 within
f2 the
f3 table
f5 and
f6 linear
f7 extrapolation outside the table
using the last two end points. Seef4Figure 6-3. No warning
f8 messages are issued if table
data is input incorrectly.
f
Extrapolated

Figure 6-3 Example of Table Extrapolation and Discontinuity

8. This form of damping is used only in modal formulations of complex eigenvalue

analysis, frequency response analysis, or transient response analysis. The type of
damping used depends on the solution sequence (structural damping is
displacement-dependent, and viscous damping is velocity-dependent). See
“Formulation of Dynamic Equations in SubDMAP GMA” on page 403 of the
MSC.Nastran Reference Guide for the equations used.
9. PARAM,KDAMP may be used in solution sequences that perform modal frequency
and modal complex analysis, to select the type of damping.

KDAMP Result
1 (Default) B Matrix
-1 ( 1 + ig )K

See “Formulation of Dynamic Equations in SubDMAP GMA” on page 403 of the

MSC.Nastran Reference Guide for a full explanation.
10. If TYPE is “G” or blank, the damping values gi, etc., are in units of equivalent viscous
dampers, as follows:

gi
b i = ----- K i
ωi

(See “Formulation of Dynamic Equations in SubDMAP GMA” on page 403 of the

MSC.Nastran Reference Guide) If TYPE is “CRIT”, the damping values gi, etc., are in the
units of fraction of critical damping C ⁄ C0 . If TYPE is “Q”, the damping values gi are in
the units of the amplification or quality factor, Q. These constants are related by the
following equations:

Main Index
422

C ⁄ C0 = g ⁄ 2

 1 ⁄ ( 2C ⁄ C 0 )
Q = 
1 ⁄ g

TABLED1 Dynamic Load Tabular Function, Form 1

Defines a tabular function for use in generating frequency-dependent and time-dependent

dynamic loads.
Format:

1 2 3 4 5 6 7 8 9 10
TABLED1 TID XAXIS YAXIS

x1 y1 x2 y2 x3 y3 -etc.- “ENDT”

Example:

TABLED1 32

-3.0 6.9 2.0 5.6 3.0 5.6 ENDT

Field Contents
TID Table identification number. (Integer > 0)
XAXIS Specifies a linear or logarithmic interpolation for the x-axis. See
Remark 6. (Character: “LINEAR” or “LOG”; Default = “LINEAR”)
YAXIS Specifies a linear or logarithmic interpolation for the y-axis. See
Remark 6. (Character: “LINEAR” or “LOG”; Default = “LINEAR”)
xi, yi Tabular values. (Real)
“ENDT” Flag indicating the end of the table.

Remarks:
1. xi must be in either ascending or descending order, but not both.
2. Discontinuities may be specified between any two points except the two starting points
or two end points. For example, in Figure 6-4 discontinuities are allowed only between
points x2 through x7. Also, if y is evaluated at a discontinuity, then the average value
of y is used. In Figure 6-4, the value of y at x=x3 is y = ( y3 + y4 ) ⁄ 2 . If the y-axis is a LOG
axis then the jump at the discontinuity is evaluated as y = y3y4 .
3. At least one continuation must be specified.
4. Any xi-yi pair may be ignored by placing the character string “SKIP” in either of the
two fields.
5. The end of the table is indicated by the existence of the character string “ENDT” in
either of the two fields following the last entry. An error is detected if any
Main Index
continuations follow the entry containing the end-of-table flag “ENDT”.
 APPENDIX 423

6. TABLED1 uses the algorithm

y = yT ( x )

where x is input to the table and y is returned. The table look-up is performed using
interpolation within the table and extrapolation outside the table using the two starting
or end points. See Figure 6-4. The algorithms used for interpolation or extrapolation
are:

XAXIS YAXIS yT(x)

LINEAR LINEAR xj – x y x – xi
---------------- yi + ---------------- yj
xj – xi xj – xi
LOG LINEAR ln ( xj ⁄ x ) ln (xxvalue ⁄ xi )
-------------------------- yi
Discontinuity -------------------------
+ Range of Table - yj
Allowed
ln ( xj ⁄ xi ) ln ( xj ⁄ xi )
Discontinuity
LINEAR LOG xj – x x – xi Not Allowed
Linear exp ---------------- ln yi + ---------------- ln yj
Extrapolation xj – xi xj – xi
of Segment
LOG LOGx1-x2 ln ( xj ⁄ x ) ln ( x ⁄ xi )
exp -------------------------- ln yi + -------------------------- ln yi
ln ( xj ⁄ xi ) ln ( xj ⁄ xi )
x
x1 x2 x3, x5 x6 x7,
where xj and yj follow xi and yi. x8
x4
No warning messages
x are issued if table data is input incorrectly.
Extrapolated

Figure 6-4 Example of Table Extrapolation and Discontinuity

7. Linear extrapolation is not used for Fourier transform methods. The function is zero
outside the range of the table.
8. For frequency-dependent loads, xi is measured in cycles per unit time.
9. Tabular values on an axis if XAXIS or YAXIS = LOG must be positive. A fatal message
will be issued if an axis has a tabular value < 0.

TABLED2 Dynamic Load Tabular Function, Form 2

Defines a tabular function for use in generating frequency-dependent and time-dependent

dynamic loads. Also contains parametric data for use with the table.
Format:

1 2 3 4 5 6 7 8 9 10
TABLED2 TID X1

x1 y1 x2 y2 x3 y3 -etc.-

Main Index
424

Example:

TABLED2 15 -10.5

1.0 -4.5 2.0 -4.2 2.0 2.8 7.0 6.5

SKIP SKIP 9.0 6.5 ENDT

Field Contents
TID Table identification number. (Integer > 0)
X1 Table parameter. See Remark 6. (Real)
xi, yi Tabular values. (Real)

Remarks:
1. xi must be in either ascending or descending order, but not both.
2. Discontinuities may be specified between any two points except the two starting points
or two end points. For example, in Figure 6-5 discontinuities are allowed only between
points x2 and x7. Also if y is evaluated at a discontinuity, then the average value of y
is used. In Figure 6-5, the value of y at x=x3 is y = ( y3 + y4 ) ⁄ 2 .
y
3. At least one continuation entry must be specified.
4. Any xi-yi pair may be ignored by placing “SKIP” in either of the two fields.
x value
5. The endDiscontinuity
of the table is indicatedRange
by the existence of “ENDT” in either of the two fields
of Table
Allowed
following the last entry. An error is detected if any continuations follow the entry
Discontinuity
containing the
Linear
end-of-table flag “ENDT”. Not Allowed
Extrapolation
6. TABLED2 uses the algorithm
of Segment
x1-x2
y = y T ( x – X1 )

where x is input to the table and y is returned. The table look-up

x is performed using
linear interpolation within
x1 the
x2 table
x3 andx5linear
x6 extrapolation
x7 outside the table using the
two starting or end points. See Figure
x4 6-5. No warning
x8 messages are issued if table
data is input incorrectly.
x
Extrapolated

Figure 6-5 Example of Table Extrapolation and Discontinuity

7. Linear extrapolation is not used for Fourier transform methods. The function is zero
outside the range of the table.
8. For frequency-dependent loads, X1 and xi are measured in cycles per unit time.

TABLED3 Dynamic Load Tabular Function, Form 3

Defines a tabular function for use in generating frequency-dependent and time-dependent

dynamic loads. Also contains parametric data for use with the table.

Main Index
 APPENDIX 425

Format:

1 2 3 4 5 6 7 8 9 10
TABLED3 TID X1 X2

x1 y1 x2 y2 x3 y3 -etc.-

Example:

TABLED3 62 126.9 30.0

2.9 2.9 3.6 4.7 5.2 5.7 ENDT

Field Contents
TID Table identification number. (Integer > 0)
X1, X2 Table parameters. (Real; X2 ≠ 0.0)
xi, yi Tabular values. (Real)

Remarks:
1. xi must be in either ascending or descending order, but not both.
2. Discontinuities may be specified between any two points except the two starting points
or two end points. For example, in Figure 6-6 discontinuities are allowed only between
points x2 and x7. Also if y is evaluated at a discontinuity, then the average value of y
is used. In Figure 6-6, the value of y at x=x3 is y = ( y3 + y4 ) ⁄ 2 .
3. At least one continuation entry must be present.
y
4. Any xi-yi pair may be ignored by placing “SKIP” in either of the two fields.
5. The end of the table is indicated by the existence of “ENDT” in either of the two fields
x value
following the last entry. An error
Discontinuity is detected
Range of Table if any continuations follow the entry
containing the end-of-table flag “ENDT”.
Allowed
Discontinuity
6. TABLED3 uses the algorithm Not Allowed
Linear
Extrapolation
of Segment
y = y T  -----------------
x – X1
x1-x2  X2 

where x is input to the table and y is returned. The table look-up

x is performed using
interpolation within the
x1 table
x2 andx3linear
x5 extrapolation
x6 x7 outside the table using the two
starting or end points. See Figurex46-6. No warningx8messages are issued if table data is
input incorrectly. x
Extrapolated

Figure 6-6 Example of Table Extrapolation and Discontinuity

7. Linear extrapolation is not used for Fourier transform methods. The function is zero
outside the range of the table.
8. For frequency-dependent loads, X1, X2, and xi are measured in cycles per unit time.
Main Index
426

TABLED4 Dynamic Load Tabular Function, Form 4

Defines the coefficients of a power series for use in generating frequency-dependent and
time-dependent dynamic loads. Also contains parametric data for use with the table.
Format:

1 2 3 4 5 6 7 8 9 10
TABLED4 TID X1 X2 X3 X4

A0 A1 A2 A3 A4 A5 -etc.-

Example:

TABLED4 28 0.0 1.0 0.0 100.

2.91 -0.0329 6.51-5 0.0 -3.4-7 ENDT

Field Contents
TID Table identification number. (Integer > 0)
Xi Table parameters. (Real; X2 ≠ 0.0; X3<X4)
Ai Coefficients. (Real)

Remarks:
1. At least one continuation entry must be specified.
2. The end of the table is indicated by the existence of “ENDT” in the field following the
last entry. An error is detected if any continuations follow the entry containing the
end-of-table flag “ENDT”.
3. TABLED4 uses the algorithm

N
i
Ai  -----------------
x – X1
y = ∑ X2
i = 0

where x is input to the table, y is returned, and N is the number of pairs. Whenever
x < X3, use X3 for x; whenever x > X4, use X4 for x. There are N + 1 entries in the table.
There are no error returns from this table look-up procedure.
4. For frequency-dependent loads, xi is measured in cycles per unit time.

TIC Transient Analysis Initial Condition

Defines values for the initial conditions of variables used in structural transient analysis. Both
displacement and velocity values may be specified at independent degrees-of-freedom. This
entry may not be used for heat transfer analysis.

Main Index
 APPENDIX 427

Format:

1 2 3 4 5 6 7 8 9 10
TIC SID G C U0 V0

Example:

TIC 100 10 3 0.1 0.5

Field Contents
SID Set identification number. (Integer > 0)
G Grid, scalar, or extra point or modal coordinate identification number.
(Integer > 0). See Remark 4.
C Component numbers. (Any one of the integers 1 through 6 for grid
points, integer zero or blank for scalar or extra points and –1 for modal
coordinates.)
U0 Initial displacement. (Real)
V0 Initial velocity. (Real)

Remarks:
1. Transient initial condition sets must be selected with the Case Control command IC =
SID. Note the use of IC in the Case Control command versus TIC on the Bulk Data
entry. For heat transfer, the IC Case Control command selects TEMP or TEMPD entries
for initial conditions and not the TIC entry.
2. If no TIC set is selected in the Case Control Section, all initial conditions are assumed
to be zero.
3. Initial conditions for coordinates not specified on TIC entries will be assumed zero.
4. In direct transient analysis (SOL 109 and 129) as well as in modal transient analysis
(SOL 112) wherein the TIC Bulk Data entry is selected by an IC or IC(PHYSICAL) Case
Control command, G may reference only grid, scalar or extra points. In modal transient
analysis (SOL 112) wherein the TIC Bulk Data entry is selected by an IC(MODAL) Case
Control command, G may reference only modal coordinates or extra points.

TLOAD1 Transient Response Dynamic Excitation, Form 1

Defines a time-dependent dynamic load or enforced motion of the form

{P(t)} = {A ⋅ F(t – τ)}

for use in transient response analysis.

Main Index
428

Format:

1 2 3 4 5 6 7 8 9 10
TLOAD1 SID EXCITEID DELAY TYPE TID

Example:

TLOAD1 5 7 LOAD 13

Field Contents
SID Set identification number. (Integer > 0)
EXCITEID Identification number of DAREA or SPCD entry set or a thermal load set
(in heat transfer analysis) that defines A. See Remarks 2. and 3. (Integer
> 0)
DELAY Defines time delay τ . (Integer > 0, real or blank) If it is a non-zero
integer, it represents the identification number of DELAY Bulk Data
entry that defines τ . If it is real, then it directly defines the value of τ that
will be used for all degrees of freedom that are excited by this dynamic
load entry. See also Remark 9.
TYPE Defines the type of the dynamic excitation. See Remarks 2. and 3.
(Integer, character or blank; Default = 0)
TID Identification number of TABLEDi entry that gives F ( t ) . (Integer > 0)

Remarks:
1. Dynamic excitation sets must be selected with the Case Control command DLOAD =
SID.
2. The type of the dynamic excitation is specified by TYPE (field 5) according to the
following table:

TYPE TYPE of Dynamic Excitation

0, L, LO, LOA or LOAD Applied load (force or moment) (Default)
1, D, DI, DIS, or DISP Enforced displacement using large mass or SPC/SPCD
data
2, V, VE, VEL or VELO Enforced velocity using large mass or SPC/SPCD data
3, A, AC, ACC or ACCE Enforced acceleration using large mass or SPC/SPCD
data

3. TYPE (field 5) also determines the manner in which EXCITEID (field 3) is used by the
program as described below
Excitation specified by TYPE is applied load
Main Index
 APPENDIX 429

• There is no LOADSET request in Case Control

EXCITEID may also reference DAREA, static, and thermal load set entries
• There is a LOADSET request in Case Control
The program may also reference static and thermal load set entries specified by
the LID or TID field in the selected LSEQ entries corresponding to the
EXCITEID.
Excitation specified by TYPE is enforced motion
• There is no LOADSET request in Case Control
EXCITEID will reference SPCD entries. If such entries indicate null enforced
motion, the program will then assume that the excitation is enforced motion
using large mass and will reference DAREA and static and thermal load set
entries just as in the case of applied load excitation.
• There is a LOADSET request in Case Control
The program will reference SPCD entries specified by the LID field in the
selected LSEQ entries corresponding to the EXCITEID. If such entries indicate
null enforced motion, the program will then assume that the excitation is
enforced motion using large mass and will reference static and thermal load set
entries corresponding to the DAREA entry in the selected LSEQ entries, just as
in the case of applied load excitation.
4. EXCITEID may reference sets containing QHBDY, QBDYi, QVECT, QVOL and
TEMPBC entries when using the heat transfer option.
5. TLOAD1 loads may be combined with TLOAD2 loads only by specification on a
DLOAD entry. That is, the SID on a TLOAD1 entry may not be the same as that on a
TLOAD2 entry.
6. SID must be unique for all TLOAD1, TLOAD2, RLOAD1, RLOAD2, and ACSRCE
entries.
7. If the heat transfer option is used, the referenced QVECT entry may also contain
references to functions of time, and therefore A may be a function of time.
8. If TLOADi entries are selected in SOL 111 or 146 then a Fourier analysis is used to
transform the time-dependent loads on the TLOADi entries to the frequency domain
and then combine them with loads from RLOADi entries. Then the analysis is
performed as a frequency response analysis but the solution and the output are
converted to and printed in the time domain. Please refer to “Fourier Transform” on
page 176 of the MSC.Nastran Advanced Dynamic Analysis User’s Guide.
9. If DELAY is blank or zero, τ will be zero.

TLOAD2 Transient Response Dynamic Excitation, Form 2

Defines a time-dependent dynamic excitation or enforced motion of the form

Main Index
430


 0 , t < ( T1 + τ ) or t > ( T2 + τ )
{P(t)} = 
 A t̃ B e C t̃ cos ( 2πF t̃ + P ) , ( T1 + τ ) ≤ t ≤ ( T2 + τ )


for use in a transient response problem, where t̃ = t – T1 – τ

Format:

1 2 3 4 5 6 7 8 9 10
TLOAD2 SID EXCITEID DELAY TYPE T1 T2 F P

C B

Example:

TLOAD2 4 10 2.1 4.7 12.0

2.0

Field Contents
SID Set identification number. (Integer > 0)
EXCITEID Identification number of DAREA or SPCD entry set or a thermal load set
(in heat transfer analysis) that defines A. See Remarks 2. and 3.
(Integer > 0)
DELAY Defines time delay τ. (Integer > 0, real or blank). If it is a non-zero
integer, it represents the identification number of DELAY Bulk Data
entry that defines τ.. If it is real, then it directly defines the value of τ that
will be used for all degrees of freedom that are excited by this dynamic
load entry. See also Remark 5.
TYPE Defines the type of the dynamic excitation. See Remarks 2. and 3.
(Integer, character or blank; Default = 0)
T1 Time constant. (Real > 0.0)
T2 Time constant. (Real; T2 > T1)
F Frequency in cycles per unit time. (Real > 0.0; Default = 0.0)
P Phase angle in degrees. (Real; Default = 0.0)
C Exponential coefficient. (Real; Default = 0.0)
B Growth coefficient. (Real; Default = 0.0)

Remarks:
1. Dynamic excitation sets must be selected with the Case Control command with
DLOAD=SID.
2. The type of the dynamic excitation is specified by TYPE (field 5) according to the
following table:
Main Index
 APPENDIX 431

TYPE TYPE of Dynamic Excitation

0, L, LO, LOA or LOAD Applied load (force or moment) (Default)
1, D, DI, DIS, or DISP Enforced displacement using large mass or SPC/SPCD
data
2, V, VE, VEL or VELO Enforced velocity using large mass or SPC/SPCD data
3, A, AC, ACC or ACCE Enforced acceleration using large mass or SPC/SPCD
data

3. TYPE (field 5) also determines the manner in which EXCITEID (field 3) is used by the
program as described below
Excitation specified by TYPE is applied load
• There is no LOADSET request in Case Control
EXCITEID may also reference DAREA, static and thermal load set entries
• There is a LOADSET request in Case Control
The program may also reference static and thermal load set entries specified by
the LID or TID field in the selected LSEQ entries corresponding to the
EXCITEID.
Excitation specified by TYPE is enforced motion
• There is no LOADSET request in Case Control
EXCITEID will reference SPCD entries. If such entries indicate null enforced
motion, the program will then assume that the excitation is enforced motion
using large mass and will reference DAREA and static and thermal load set
entries just as in the case of applied load excitation.
• There is a LOADSET request in Case Control
The program will reference SPCD entries specified by the LID field in the
selected LSEQ entries corresponding to the EXCITEID. If such entries indicate
null enforced motion, the program will then assume that the excitation is
enforced motion using large mass and will reference static and thermal load set
entries specified by the LID or TID field in the selected LSEQ entries
corresponding to the EXCITEID, just as in the case of applied load excitation.
4. EXCITEID (field 3) may reference sets containing QHBDY, QBDYi, QVECT, and QVOL
and TEMPBC entries when using the heat transfer option.
5. If DELAY is blank or zero, τ will be zero.
6. TLOAD1 loads may be combined with TLOAD2 loads only by specification on a
DLOAD entry. That is, the SID on a TLOAD1 entry may not be the same as that on a
TLOAD2 entry.

Main Index
432

7. SID must be unique for all TLOAD1, TLOAD2, RLOAD1, RLOAD2, and ACSRCE
entries.
8. If the heat transfer option is used, the referenced QVECT entry may also contain
references to functions of time, and therefore A may be a function of time.
9. If TLOADi entries are selected in SOL 111 or 146 then a Fourier analysis is used to
transform the time-dependent loads on the TLOADi entries to the frequency domain
and them combine them with loads from RLOADi entries. Then the analysis is
performed as a frequency response analysis but the solution and the output are
converted to and printed in the time domain. In this case, B will be rounded to the
nearest integer. Please refer to “Fourier Transform” on page 176 of the MSC.Nastran
Advanced Dynamic Analysis User’s Guide
10. The continuation entry is optional.

TSTEP Transient Time Step

Defines time step intervals at which a solution will be generated and output in transient analysis.
Format:

1 2 3 4 5 6 7 8 9 10
TSTEP SID N1 DT1 NO1

N2 DT2 NO2

-etc.-

Example:

TSTEP 2 10 .001 5

9 0.01 1

Field Contents
SID Set identification number. (Integer > 0)
Ni Number of time steps of value DTi. (Integer > 1)
DTi Time increment. (Real > 0.0)
NOi Skip factor for output. Every NOi-th step will be saved for output.
(Integer > 0; Default = 1)

Remarks:
1. TSTEP entries must be selected with the Case Control command TSTEP = SID.
2. Note that the entry permits changes in the size of the time step during the course of the
solution. Thus, in the example shown, there are 10 time steps of value .001 followed by
9 time steps of value .01. Also, the user has requested that output be recorded for t =
0.0, .005, .01, .02, .03, etc.

Main Index
 APPENDIX 433

3. See “Guidelines for Effective Dynamic Analysis” on page 287 of the MSC.Nastran
Basic Dynamics User’s Guide for a discussion of considerations leading to the selection
of time steps.
4. In modal frequency response analysis (SOLs 111 and 146), this entry is required only
when TLOADi is requested; i.e., when Fourier methods are selected.
5. The maximum and minimum displacement at each time step and the SIL numbers of
these variables can be printed by altering DIAGON(30) before the transient module
TRD1 and by altering DIAGOFF(30) after the module. This is useful for runs that
terminate due to overflow or excessive run times.
6. For heat transfer analysis in SOL 159, use the TSTEPNL entry.

Main Index
434

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Parameters for Dynamic Analysis
G
■ Overview

Main Index
436

G.1 Overview
This appendix lists some of the parameters that are often used for dynamic analysis. These
parameters are listed alphabetically, beginning below. See the MSC.Nastran Quick Reference
Guide for a description of all parameters.

ASlNG Default = 0
ASING specifies the action to take when singularities (null rows
and columns) exist in the dynamic matrices (or [ K ll ] in statics). If
ASING = –1, then a User Fatal Message will result.
If ASING = 0 (the default), singularities are removed by
appropriate techniques depending on the type of solution being
performed.
AUTOSPC Default = YES (in SOLutions 21, 38, 39, 61 thru 63, 65 thru 76, 81
thru 91, and 101 through 200 except 106 and 129)
Default = NO (in all other SOLutions)
AUTSPC specifies the action to take when singularities exist in the
stiffness matrix ( [ K gg ] ). AUTOSPC = YES means that singularities
will be constrained automatically. AUTOSPC = NO means that
singularities will not be constrained.
Singularity ratios smaller than PARAM,EPPRT (default = 1.E-8)
are listed as potentially singular. If PARAM,AUTOSPC has the
value YES, identified singularities with a ratio smaller than
PARAM,EPZERO (default = 1.E-8) will be automatically
constrained with single-point constraints. If PARAM,EPPRT has
the same value as PARAM,EPZERO (the default case), all
singularities are listed. If PARAM,EPPRT is larger than
PARAM,EPZERO, the printout of singularity ratios equal to
exactly zero is suppressed. If PARAM,PRGPST is set to NO
(default is YES), the printout of singularities is suppressed, except
when singularities are not going to be removed. If
PARAM,SPCGEN is set to 1 (default = 0), the automatically
generated SPCs are placed in SPCi Bulk Data entry format on the
PUNCH file.
AUTOSPC provides the correct action for superelements in all
contexts. It does not provide the correct action for the residual
structure in SOLs 64, 99, or 129. PARAM,AUTOSPCR, not
AUTOSPC, is used for the o-set (omitted set) in the residual
structure in SOLs 66 and 106.

Main Index
 APPENDIX G 437
Parameters for Dynamic Analysis

BAILOUT 0 (for superelement sequences)

Default =
-1 (for non-superelement sequences)

See MAXRATIO.
CB1, CB2 Default = 1.0
CB1 and CB2 specify factors for the total damping matrix. The
total damping matrix is

x 2
[ B jj ] = CB1 ⋅ [ B jj ] + CB2 ⋅ [ B jj ]

where [ B 2jj ] is selected via the Case Control command B2GG and
x
[ B jj ] comes from CDAMPi or CVlSC element Bulk Data entries.
These parameters are effective only if B2GG is selected in the Case
Control Section.
CK1, CK2 Default = 1.0
CK1 and CK2 specify factors for the total stiffness matrix. The
total stiffness matrix (exclusive of GENEL entries) is

x 2 2
[ K jj ] = CK1 ⋅ [ K jj ] + CK2 ⋅ [ K jj ]

where [ K 2jj ] is selected via the Case Control command K2GG and
z
[ K jj ] is generated from structural element (e.g., CBAR) entries in
the Bulk Data. These are effective only if K2GG is selected in the
Case Control Section. Note that stresses and element forces are
not factored by CK1, and must be adjusted manually.
CM1, CM2 Default = 1.0
CM1 and CM2 specify factors for the total mass matrix. The total
mass matrix is

x 2
[ M jj ] = CM1 ⋅ [ M jj ] + CM2 ⋅ [ M jj ]

where [ M 2jj ] is selected via the Case Control command M2GG and
x
[ M jj ] is derived from the mass element entries in the Bulk Data
Section. These are effective only if M2GG is selected in the Case
Control Section.
COUPMASS Default = -1

Main Index
438

COUPMASS > 0 requests the generation of coupled rather than

lumped mass matrices for elements with coupled mass capability.
This option applies to both structural and nonstructural mass for
the following elements: CBAR, CBEAM, CHEXA, CONROD,
CPENTA, CQUAD4, CQUAD8, CROD, CTETRA, CTRIA3,
CTRlA6, CTRIAX6, CTUBE. A negative value (the default) causes
the generation of lumped mass matrices (translational
components only) for all of the above elements.
CP1, CP2 Default = 1.0
The load vectors are generated from the equation

 x  2
{ P j } = CP1 ⋅  P j  + CP2 ⋅  P j 
   

where { P 2j } is selected via the Case Control command P2G, and

{ P xj } comes from Bulk Data static load entries.

CURVPLOT Default = -1
PARAM,CURVPLOT,1 requests that X-Y (or curve) plots be made
for displacements, loads, SPC forces, or grid point stresses or
strains. The y values are response values; the x values are related
to grid point locations through the parameter DOPT.
PARAM,CURVPLOT,1 suppresses SORT2-type processing; in
superelement dynamic analysis, SORT1 requests will be honored.
To obtain stress or strain plots, set the CURV parameter to +1.
DOPT controls the x spacing of curves over grid points for the
CURVPLOT module. The default for DOPT is the length between
grid points.
DDRMM Default = 0
By default, the matrix method of data recovery is used in the
modal solutions. DDRMM = –1 will force calculation of complete
g-set solution vectors by the mode displacement method, which is
needed for SORT1 output. SORT1 output is required for deformed
structure plots, grid point force balance output, the mode
acceleration technique, and postprocessing with PARAM,POST,–1
or –2.
DYNSPCF Default = NEW (Structured SOLs only)
PARAM,DYNSPCF,NEW requests that mass and damping
coupled to ground be included in the SPCForce calculations for
the linear dynamic solutions: SOLs 103, 107-112, 115, 118, 145, 146,
and 200. OLD neglects these effects and gives the same SPCForce
results obtained in versions prior to Version 68.
Main Index
 APPENDIX G 439
Parameters for Dynamic Analysis

EPZERO Default = 1.E-8

Specifies the minimum value that indicates a singularity. See
AUTOSPC.
G Default = 0.0
G specifies the uniform structural damping coefficient in the
formulation of dynamics problems. To obtain the value for the
parameter G, multiply the critical damping ratio, C/Co, by 2.0.
PARAM,G is not recommended for use in hydroelastic or heat
transfer problems. If PARAM,G is used in transient analysis,
PARAM,W3 must be greater than zero or PARAM,G will be
ignored.
HFREQ Default = 1 .+30
The parameters LFREQ and HFREQ specify the frequency range
(LFREQ is the lower limit and HFREQ is the upper limit) of the
modes to be used in the modal formulations. Note that the default
for HFREQ will usually include all vectors computed. A related
parameter is LMODES.
KDAMP Default = 1
If KDAMP is set to –1, viscous modal damping is entered into the
complex stiffness matrix as structural damping.
LFREQ Default = 0.0
See HFREQ.
LMODES Default = 0
LMODES is the number of lowest modes to use in a modal
formulation. If LMODES = 0, the retained modes are determined
by the parameters LFREQ and HFREQ.
MAXRATIO Default = 1.E5
The ratios of terms on the diagonal of the stiffness matrix to the
corresponding terms on the diagonal of the triangular factor are
computed. If, for any row, this ratio is greater than MAXRATIO,
the matrix will be considered to be nearly singular (having
mechanisms). If any diagonal terms of the factor are negative, the
stiffness matrix is considered implausible (non-positive definite).
The ratios greater than MAXRATIO and less than zero and their
associated external grid identities will be printed out. The
program will then take appropriate action as directed by the
parameter BAILOUT.

Main Index
440

By default, in the superelement solution sequences the program

will terminate processing for that superelement. A negative value
for BAILOUT directs the program to continue processing the
superelement. Although forcing the program to continue with
near-singularities is a useful modeling checkout technique, it may
lead to solutions of poor quality or fatal messages later in the run.
It is recommended that the default values be used for production
runs. A related parameter is ERROR.
In nonsuperelement solution sequences, the default value (–1) of
BAILOUT causes the program to continue processing with near
singularities and a zero value will cause the program to exit if near
singularities are detected.
In SOL 60, and SOLs 101-200 when PARAM,CHECKOUT,YES is
specified, PARAM,MAXRATIO sets the tolerance for detecting
multipoint constraint equations with poor linear independence.
MODACC Default = –1
MODACC = 0 selects the mode acceleration method for data
recovery in dynamic analysis. PARAM,DDRMM,–1 must also be
specified.
NONCUP Default = –1
NONCUP selects either a coupled or non-coupled solution
algorithm in modal transient response analysis. By default, the
noncoupled solution algorithm is selected unless the dynamic
matrices KHH, MHH, or BHH have off-diagonal terms.
NONCUP = 1 requests the coupled algorithm and –2 the
uncoupled algorithm regardless of the existence of off-diagonal
terms in the dynamic matrices. User Information Message 5222
indicates which algorithm is used in the analysis.
NOSORT1 -1 (SOL 30)
Default =
–1 (SOL 31)

NOSORT1 controls the execution of the SDR3 module, which

changes SORT1 modal solutions to SORT2. In SOL 30, where
SORT1 output is the default option, NOSORT1 is set to branch
over this operation. In SOL 31, where SORT2 output is usually
desired, PARAM,NOSORT1,1 may be used when only SORT1
output is desired.
WTMASS Default = 1.0
The terms of the structural mass matrix are multiplied by the
value of this parameter when they are generated. This parameter
is not recommended for use in hydroelastic problems.
W3, W4 Default = 0.0
Main Index
 APPENDIX G 441
Parameters for Dynamic Analysis

The damping matrix for transient analysis is assembled from the

equation:

1 2 G 1 1 4
[ B dd ] = [ B dd ] + [ B dd ] + --------- [ K dd ] + --------- [ K dd ]
W3 W4

The default values of 0.0 for W3 and W4 cause the [ K 1dd ] and [ K 4dd ]
terms to be ignored in the damping matrix, regardless of the
presence of the PARAM,G or [ K 4dd ] . [ K 1dd ] is the stiffness. [ K 4dd ] is
the structural damping and is created when GE is specified on the
MATi entries.

Main Index
442

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

APPENDIX
File Management Section
H
■ Overview
■ Definitions
■ MSC.Nastran Database
■ File Management Commands

Main Index
444

H.1 Overview
The database structure for MSC.Nastran was changed substantially with the implementation of
the Executive System first introduced in Version 66. Furthermore, a new section called the File
Management Section (FMS) was added to the MSC.Nastran input file. The goal is to make file
assignment as uniform as possible across different computer types.
This appendix is intended to give you an overview of how the database and FMS work so that
you can allocate your computer resources efficiently, especially for large models. For many
problems, due to the default values, the use or knowledge of the FMS is either transparent or not
required by you.

Main Index
 APPENDIX H 445
File Management Section

H.2 Definitions
Before presenting the details of the database description, it is helpful to define some of the basic
terms that are used throughout this appendix.

DBset Database set. This consists of an MSC.Nastran logical name, such as

MASTER, DBALL, etc., which refers to an entity within the
MSC.Nastran database. It consists of a collection of physical files.
Database Collection of all DBsets assigned to a run.
Data block Matrix or table (e.g., KAA, LAMA) that is stored in the database.
Logical name Local internal name (log-name) used by MSC.Nastran for a file or DBset.
Word For 32- and 64-bit machines, each word is equivalent to four and eight
bytes, respectively.
Buffsize Length of an MSC.Nastran buffer in terms of words (32- or 64-bit
words). It controls the physical record size for data storage/transfer that
is contained in many MSC.Nastran logical units. The default and
maximum allowable buffsize is machine dependent. The default value
is recommended except for large problems. It can be modified by using
the following NASTRAN statement:
NASTRAN BUFFSIZE = xxxxx
where xxxxx is the desired buffsize.
Block This is often referred to as a MSC.Nastran GINO block. Each GINO
block contains one MSC.Nastran buffer.
{} A brace indicates that the quantity within this bracket is mandatory. The
underlined item within { } is the default value.
[] A square bracket indicates that the quantity within this bracket is
optional.

Main Index
446

H.3 MSC.Nastran Database

When submitting a MSC.Nastran job, a series of standard output files is created (e.g., F06 file).
Conventions for the filenames are machine dependent. Furthermore, four permanent
(MASTER, DBALL, USRSOU, USROBJ) and one scratch (SCRATCH) DBsets are created during
the run.

MASTER This is the “directory” DBset that contains a list of all the DBsets used in
the job, all the physical file names assigned, and an index pointing to all
the data blocks created and where they are stored. In addition, it also
contains the MSC.Nastran Data Definition Language (NDDL) used.
NDDL is the internal MSC.Nastran language that describes the
database. You do not need to understand NDDL to use MSC.Nastran
effectively. The default maximum size for MASTER is 5000 blocks.
DBALL This is the DBset where the permanent data blocks are stored by
default. The default maximum size is 25000 blocks.
USRSOU This DBset stores the source file for user-created DMAP. The default
maximum size is 5000 blocks.
USROBJ This DBset stores the object file for user-created DMAP. The default
maximum size is 5000 blocks.
SCRATCH This DBset is used as the temporary workspace for MSC.Nastran. In
general, this DBset is deleted at the end of the run. The default
maximum size is 350100 blocks.

For most solutions, USRSOU and USROBJ are not needed and may be deleted or assigned as
temporary for the duration of the run (see “ASSIGN” on page 448).

For a typical UNIX-based workstation with an MSC.Nastran input file called “dyn1.dat”, the
following sample submittal command can be used:
nastran dyn1 scr=no
where nastran is the name of shell script for executing MSC.Nastran. The following four
physical database files are created as a result of the above command.

dyn1.MASTER
dyn1.DBALL
dyn1.USROBJ
dyn1.USRSOU
Unless otherwise stated, the input filename is assumed to be “dyn1.dat” in this appendix.

Main Index
 APPENDIX H 447
File Management Section

H.4 File Management Commands

The File Management Section is intended primarily for the attachment and initialization of
MSC.Nastran database sets (DBsets) and FORTRAN files. For many problems, due to the
default values, the use of the FMS is either transparent or not required by you. At first glance
the FMS may be overwhelming due in part to its flexibility and the many options it has to offer.
However, in its most commonly used form, it can be quite simple. Examples in the most
commonly used format are presented throughout this appendix.
There are numerous FMS statements available in MSC.Nastran. This section covers the
commonly used FMS statements in dynamic analysis. They are presented in their most
commonly used format, which in general, is a simplified format.
If an FMS statement is longer than 72 characters, it can be continued to the next line by using a
comma as long as the comma is preceded by one or more blank spaces.

INIT
Purpose
Creates/initializes permanent and/or temporary DBsets. The INIT statement has two basic
formats: one for all the DBsets and one specifically for the SCRATCH DBsets.

Format (Simplified) for All DBsets Except SCRATCH

INIT DBset-name [LOGICAL=(log-name1(max-size1) ,
log-name2(max-size2),...lognamei(max-sizei) ,
...lognamen(max-sizen)]

DBset-name This is the logical name of the DBset being used (e.g., DBALL).
log-namei This is the i-th logical name for the DBset-name referenced in this
INIT statement. You can have up to 10 logical names for each DBset
( 1 ≤ i ≤ 10 ). The i-th physical file is assigned with the “assignment”
statement.
max-sizei This is the maximum allowable number of MSC.Nastran blocks
which may be written to the i-th file.

Exampl INIT DBALL LOGICAL=(DBALL(50000))

e
This statement creates the DBALL DBset with a logical name of DBALL and the maximum size
of 50000 MSC.Nastran blocks instead of 25000 blocks, which is the default value for DBALL.
Unless an ASSIGN statement is also used, the physical file is given the name dyn1.DBALL
assuming that your input file is called “dyn1.dat”.
INIT DBALL LOGICAL=(DB1(35000),DBTWO(60000))
This statement creates the DBALL DBset with logical names of DB1 and DBTWO. Two physical
files, dyn1.DB1 and dyn1.DBTWO, are created with a maximum size of 35000 and 60000
MSC.Nastran blocks, respectively.
Main Index
448

Format (Simplified) for the SCRATCH DBset

INIT SCRATCH [LOGICAL=(log-name1(max-size1) ,
log-name2(max-size2),...log-namei(max-sizei)) ,
SCR300=(log-namei+1(max-sizei+1),...log-namen(max-sizen))]
Log-name1 through log-namei are allocated for regular scratch files as temporary workspace.
This temporary workspace is not released until the end of the job. SCR300 is a special keyword
which indicates that the log-names are members reserved for DMAP module internal scratch
files. The space occupied by these SCR300 files is for the duration of the execution of the module.
This SCR300 space is released at the end of the module execution. You can have up to a
combined total of 10 logical names for the SCRATCH DBset ( 1 ≤ i ≤ 10 ).

Example
INIT SCRATCH LOGICAL=(SCR1(150000),SCR2(100000)) ,
SCR300=(SCRA(250000),SCRB(300000))
This statement creates the SCRATCH DBset with logical names of SCR1, SCR2, SCRA, and
SCRB. Two physical files, dyn1.SCR1 and dyn1.SCR2, are created with a maximum size of
150000 and 100000 blocks, respectively. These two files are regular scratch files. Two additional
physical files, dyn1.SCRA and dyn1.SCRB, are created with a maximum size of 250000 and
300000 blocks, respectively. These two files are SCR300 type files.

ASSIGN
Purpose
Assigns physical filenames to logical filenames or special FORTRAN files that are used by other
FMS statements or DMAP modules.

Format (Simplified) to Assign Logical Files

ASSIGN log-namei=’filenamei’ [TEMP DELETE]

log-namei The i-th logical name for the DBset created by the INIT statement.
TEMP Requests that filenamei be deleted at the end of the job. This is
optional and is often used for USRSOU and USROBJ.
DELETE Requests that filenamei be deleted if it exists before the start of the
run. This is optional; however, if this option is not used and the
FORTRAN file exists prior to the current run, then the job may fail
with the following messages:

Main Index
 APPENDIX H 449
File Management Section

*** USER FATAL MESSAGE 773 (DBDEF)

THE FOLLOWING PHYSICAL FILE ALREADY EXISTS.
LOGICAL NAME = xxxx
PHYSICAL FILE = yyyy
USER INFORMATION: NO ASSOCIATED DEFAULT FILES OR ASSIGNED DBSETS CAN
EXIST PRIOR TO THE DATA BASE INITIALIZATION RUN.
USER ACTION: DELETE THIS FILE AND RESUBMIT THE JOB.

DELETE is not a suggested option if you are using RESTART since

you can delete your database inadvertently. Manual deletion of
unwanted databases is a safer approach.

Example
ASSIGN DB1=’sample.DB1’
INIT DBALL LOGICAL=(DB1(50000))
These statements create a physical file called sample.DB1 for the logical name DB1 in the current
directory. Without the ASSIGN statement, the physical file name created is called dyn1.DB1,
assuming once again that your input file is called dyn1.dat.

ASSIGN DB1=’/mydisk1/se/sample.DB1’
ASSIGN DB2=’/mydisk2/sample.DB2’
INIT DBALL LOGICAL=(DB1(50000),DB2(40000))
Two logical names DB1 and DB2 are created for the DBset DBALL. DB1 points to a physical file
called sample.DB1 that resides in the file system (directory) /mydisk1/se. DB2 points to a
physical file called sample.DB2 that resides in the file system (directory) /mydisk2.

Format (Simplified) to Assign FORTRAN Files

ASSIGN log-key=’filenamef’ [STATUS={new,old},UNIT=u ,
FORM={FORMATTED,UNFORMATTED},DELETE]

log-key This is the logical keyword for the FORTRAN file being created. The
default values depend on the keyword. Acceptable keywords are
DBC,DBMIG,INPUTT2,INPUTT4,OUTPUT2,OUTPUT4,
DBUNLOAD,DBLOAD, and USERFILE.

You should reference the MSC.Nastran Quick Reference Guide for detailed descriptions and the
defaults for these keywords.

filenamef This is the physical name of the FORTRAN file.

STATUS Specifies whether the FORTRAN file will be created (STATUS = new)
or is an existing file (STATUS = old).
UNIT Specifies the FORTRAN unit (e.g., UNIT = 12).

Main Index
450

FORM Specifies whether the file written is in ASCII (FORM = FORMATTED)

or binary (FORM = UNFORMATTED) format.
DELETE Requests that filenamef be deleted if it exists before the start of the
run.

Example
ASSIGN OUTPUT2=’sample.out’,STATUS=NEW,UNIT=11,FORM=FORMATTED,
DELETE
This example creates a new FORTRAN file to be used for OUTPUT2 operations. This file is in
ASCII format with a physical filename of sample.out and is assigned to unit 11.

EXPAND
Purpose
Concatenates files into an existing DBset in order to increase the allowable disk space. The
EXPAND statement is normally used in a restart run when you run out of disk space in your
previous run.

Format
EXPAND DBset-name LOGICAL=(log-namei(max-sizei),... )

DBset-name This is the logical name of the DBset to be expanded by the addition
of new members to this existing DBset.
log-namei This is the logical name of the i-th member of the DBset. An ASSIGN
statement should be used to point this logical name to a physical file.
max-sizei This is the maximum allowable number of MSC.Nastran blocks that
may be written to the i-th member.

Example
The original run creates a database with a name dyn1.DBALL. This database is filled and the job
fails with the following erorr messages in the F06 file:
*** USER FATAL MESSAGE 1012 (GALLOC)
DBSET DBALL IS FULL AND NEEDS TO BE EXPANDED.
For small to medium size problems, it is best to rerun the job from the beginning with a larger
file allocation. For large problems, if rerunning the job is not practical, then the database can be
expanded with the following statements:
RESTART
ASSIGN MASTER=’dyn1.MASTER’
ASSIGN DBADD=’morespace.DB’
EXPAND DBALL LOGICAL=(DBADD(50000))

Main Index
 APPENDIX H 451
File Management Section

These statements create an additional member (with a logical name of DBADD) to the existing
DBset DBALL. This member points to a new physical file called morespace.DB, which may
contain up to a maximum of 50000 MSC.Nastran blocks. In this case, you are restarting from
“dyn1.MASTER”.

RESTART

Purpose
Allows you to continue from the end of the previous run to the current run without resolving
the problem from the beginning.

Format
RESTART [PROJECT=’proj-ID’, VERSION={version-ID,LAST},
{KEEP,NOKEEP}]

proj-ID Project identifier used in the original run, which can have up to 40
characters. This is optional and is normally not used. The default
proj-ID is blank.
version-ID The version number you are restarting from. The default is the last
version.
KEEP If this option is used, then the version that you are restarting from is
also saved at the end of the current run.
NOKEEP If this option is used, then the version that you are restarting from is
deleted at the end of the current run. This is the default.

Example
RESTART
The current run uses the last version in the database for restart. At the end of the run, this last
version is deleted from the database. This statement is probably the most commonly used
format for RESTART.
RESTART VERSION=5,KEEP
The current run (version 6 or higher) uses version 5 in the database for restart. At the end of the
run, version 5 is also retained in the database. This format is used most often when you want to
ensure that a specific version is saved in the database (i.e., a large normal modes run).
RESTART PROJ=’xyz’ VERSION=3
The current run uses version 3 with a proj-ID of xyz in the database for restart. At the end of the
run, version 3 with a proj-ID of xyz is deleted from the database.

Main Index
452

INCLUDE
Purpose
Inserts an external file at the location where this “include” statement is used. This is not a pure
FMS statement because it can be used anywhere in the input file, not just in the FMS Section. The
“include” statement must not be nested; in other words, you cannot attach a file that contains an
“include” statement.

Format
Include ’filename’

filename Physical filename of the external file to be inserted at this location.

Example
Sol 101
time 10
cend
include ’sub1.dat’
begin bulk
$
include ’bulk1.dat’
include ’bulk2.dat’
$
$ rest of bulk data file
$
.
.
enddata
This run reads a file called sub1.dat with all the Case Control commands contained in it. It also
reads two additional files (bulk1.dat and bulk2.dat) in the Bulk Data Section. You may, for
example, want to include all your grid Bulk Data entries in the file bulk1.dat and all your element
connectivities in bulk2.dat. As you can see, the “include” statement can be a handy tool. For
parametric studies, you can potentially save a tremendous amount of disk space by using the
“include” statement instead of having multiple files with duplicate input data.

Summary
Due to the default values, very little knowledge of the MSC.Nastran FMS statements and
database structure is required for small to medium size problems. For large problems, however,
some knowledge of the FMS statements and database structure can help you to optimize your
computer resources.

Main Index
 MSC.Nastran Basic Dynamics User’s GuideMSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Grid Point Weight Generator
I
■ Overview
■ Commonly Used Features
■ Example with Direction Dependent Masses

Main Index
454

I.1 Overview
The grid point weight generator (GPWG) calculates the masses, centers of gravity, and inertias
of the mathematical model of the structure. The data are extracted from the mass matrix by
using a rigid-body transformation calculation. Computing the mass properties is somewhat
complex because a finite element model may have directional mass properties, that is, the mass
may differ in each of the three coordinate directions. From a mathematical point of view, the
MSC.Nastran mass may have tensor properties similar to the inertia tensor. This complexity is
reflected in the GPWG output. All of the transformations used in calculating the mass properties
are shown for the general case. Since most models have the same mass in each of the three
coordinate directions, the GPWG output provides more information than you generally need.
To avoid unnecessary confusion and at the same time provide the necessary information for the
advanced user, the discussion of the GPWG is separated into two sections. In Section I.2, a basic
discussion is given that should satisfy most users. If you need additional information, read
Section I.3.
In both sections, a simple model consisting of four concentrated masses is used to demonstrate
the GPWG output. In the first section, the mass is the same in each direction. For the second
section, the mass is different in each of the three directions.

Main Index
 APPENDIX I 455
Grid Point Weight Generator

I.2 Commonly Used Features

To demonstrate the typical output generated by the GPWG, a small model consisting of four
concentrated masses as shown in Figure I-1 is used. This model is typical of most models
because the mass is the same in each coordinate direction. The number of masses has been kept
small so you can better understand the physics.
Concentrated masses are located at four different grid points. The displacement coordinate
system for each of the grid points is the basic coordinate system.

zb
4 (5 Mass Units)
1.0

yb

1.0
3 (3 Mass Units)

2 (3 Mass Units)
1 (2 Mass Units) xb
0.0 0.5 1.0

Figure I-1 Four Concentrated Mass Model

To request the GPWG output, you must add parameter GRDPNT in either the Bulk Data Section
or the Case Control Section as follows:
PARAM,GRDPNT,x
if

x = –1 GPWG is skipped (default).

x=0 The mass properties are computed relative to the origin of the basic coordinate
system.
x>0 The mass properties are computed relative to grid point x. If grid point x does not
exist, the properties are computed relative to the basic coordinate system.
For the four masses shown in Figure I-1, the resulting GPWG output is given in Figure I-2.

Main Index
456

O U T P U T F R O M G R I D P O I N T W E I G H T G E N E R A T O R
REFERENCE POINT = 0
M O
* 1.300000E+01 0.000000E+00 0.000000E+00 0.000000E+00 5.000000E+00 -3.000000E+00 *
* 0.000000E+00 1.300000E+01 0.000000E+00 -5.000000E+00 0.000000E+00 7.000000E+00 *
* 0.000000E+00 0.000000E+00 1.300000E+01 3.000000E+00 -7.000000E+00 0.000000E+00 *
* 0.000000E+00 -5.000000E+00 3.000000E+00 8.000000E+00 -1.500000E+00 -2.500000E+00 *
* 5.000000E+00 0.000000E+00 -7.000000E+00 -1.500000E+00 1.000000E+01 0.000000E+00 *
* -3.000000E+00 7.000000E+00 0.000000E+00 -2.500000E+00 0.000000E+00 8.000000E+00 *
S
* 1.000000E+00 0.000000E+00 0.000000E+00 *
* 0.000000E+00 1.000000E+00 0.000000E+00 *
* 0.000000E+00 0.000000E+00 1.000000E+00 *
DIRECTION
MASS AXIS SYSTEM (S) MASS X-C.G. Y-C.G. Z-C.G.
X 1.300000E+01 0.000000E+00 2.307692E-01 3.846154E-01
Y 1.300000E+01 5.384616E-01 0.000000E+00 3.846154E-01
Z 1.300000E+01 5.384616E-01 2.307692E-01 0.000000E+00
I(S)
* 5.384615E+00 -1.153847E-01 -1.923079E-01 *
* -1.153847E-01 4.307692E+00 -1.153846E+00 *
* -1.923079E-01 -1.153846E+00 3.538461E+00 *
I(Q)
* 5.503882E+00 *
* 5.023013E+00 *
* 2.703873E+00 *
Q
* 8.702303E-01 4.915230E-01 3.323378E-02 *
* 3.829170E-01 -7.173043E-01 5.821075E-01 *
* 3.099580E-01 -4.938418E-01 -8.124324E-01 *

Figure I-2 GPWG Output for the Four Concentrated Mass Model

The [ MO ] matrix represents the rigid-body mass properties of the structure and is generally not
needed for model checkout. This matrix represents an intermediate step in computing the
inertia properties of the structure. The [ S ] matrix should always be equal to the identity matrix
when the mass is the same in each coordinate direction, which is the typical case. If this matrix
is not the identity matrix, inspect the model for inconsistent masses.

Following the [ S ] matrix are the mass and center of gravity locations. These are the most
commonly used information of the GPWG output. Because the mass may be different in the
three translational directions, the mass is printed for every coordinate direction. For the same
reason, the center of gravity location is given for each of the three translational masses. If the
mass is the same in all directions, a unique center of gravity exists and is located at the x-
component of the y (or z) mass, the y-component of the x (or z) mass, and the z-component of
the x (or y) mass. If the mass is not the same in all three directions, it is likely due to the CONM1,
CMASSi, or DMIG input. If the reference point is specified with PARAM,GRDPNT,0 and [ S ] is
the identity matrix, then the center of gravity location is given in the basic coordinate system.
For the example, the mass of the structure is 13.0 and the center of gravity location is (0.5384,
0.2307, 0.3846) in the basic coordinate system.

If a grid point ID is used for the reference point, and [ S ] is an identity matrix, then the center of
gravity location is in a coordinate system parallel to the basic coordinate system with an origin
located at the grid point.

Main Index
 APPENDIX I 457
Grid Point Weight Generator

If the [ S ] matrix is equal to the identity matrix, then the [ I ( S ) ] matrix represents the inertia matrix
of structure for the center of gravity with respect to the basic coordinate system, the [ I ( Q ) ] matrix
is the corresponding principal moments of inertia matrix, and [ Q ] represents the transformation
from the principal directions to the basic coordinate system.
The following additional comments on the GPWG should be noted.

• The scale factor entered with parameter WTMASS is applied to the assembled element
mass before the GPWG. The GPWG module, however, converts mass back to the
original input units that existed prior to the scaling effect of the parameter WTMASS.
(Note that the parameter WTMASS is not applied to M2GG or M2PP input, but the
M2GG mass is assembled into the mass matrix prior to GPWG. Therefore, for GPWG
output only, the M2GG mass is scaled by the same parameter as the element mass.
M2GG input may be scaled independently using the CM2 parameter.)
• The GPWG is performed on the g-size mass matrix, which is the mass matrix prior to
the processing of the rigid elements, MPCs, and SPCs.
• The mass at scalar points and fluid-related masses are not included in the GPWG
calculation.
• The GPWG for a superelement does not include the mass from upstream
superelements. Therefore, the GPWG for the residual structure includes only the mass
on the residual points. The center of gravity location is also based on the mass of the
current superelement only.
• If a large mass is used for enforced motion, the large mass dominates the structural
mass. For model checkout, it is recommended to remove the large mass and constrain
the driving point. A static analysis is a convenient way to generate a mass matrix and
obtain output from the GPWG.
• The output from the GPWG is for information purposes only and is not used in the
analysis.

Main Index
458

I.3 Example with Direction Dependent Masses

In the previous section, the mass was the same in each of the three coordinate directions thereby
producing a unique center of gravity location. However, if scalar masses are used, the total mass
may have different properties in each direction, and the center of gravity may not be a unique
location. This effect is shown in the output by providing the direction and center of gravity for
each of the three mass principal components.

When using directional mass, the axes about which the inertia matrix i [ S ] is calculated are
referred to as the principal mass axes. The principal mass axes may not necessarily intersect.
However, these axes provide uncoupled rotation and translation mass properties. If the
structural model is constructed using only real masses, the three principal mass values printed
out are equal, the center of gravity is unique, and the principal mass axes intersect at the center
of gravity.
To demonstrate all of the features of the GPWG module, the four-mass sample problem
discussed in the previous section is modified so that the mass is not equal in each of the three
translational directions (see Figure I-3). Furthermore, different displacement coordinate
systems are used for the grid points. The displacement coordinate system for grid point 1 is the
local rectangular system 1, which is oriented at an angle of 45 degrees (about the z b axis). The
displacement coordinate system for grid point 3 is the local rectangular system 3, which is
oriented at an angle of 60 degrees (about the z b axis). The grid point locations and masses are
summarized in Table I-1.

zb
4
1.0 y3
yb x3
1.0
z1 y1 60°
3
x1
2
1 45°
xb
0.0 0.5 1.0

Figure I-3 Four Concentrated Mass Model

Main Index
 APPENDIX I 459
Grid Point Weight Generator

Table I-1 Location and Size of Masses

Location Mass
Grid ID Basic System (CP Fields) Global System (CD Fields)
xb yb zb xCD yCD zCD
1 0 0 0 2 3 5
2 1 0 0 2 3 5
3 0.5 1 0 2 3 5
4 0.5 0 1 2 3 5

The GPWG output for the four mass model is shown in Figure I-4.

Listing I-1

*** USER WARNING MESSAGE 3042 MODULE = GPWG

INCONSISTENT SCALAR MASSES HAVE BEEN USED. EPSILON/DELTA = 3.8054429E-02

O U T P U T F R O M G R I D P O I N T W E I G H T G E N E R A T O R

REFERENCE POINT = 1
M O
* 9.250000E+00 -9.330128E-01 0.000000E+00 0.000000E+00 2.000000E+00 -2.966506E+00 * Rigid-Body Mass
* -9.330128E-01 1.075000E+01 0.000000E+00 -3.000000E+00 0.000000E+00 6.058013E+00 *
* 0.000000E+00 0.000000E+00 2.000000E+01 5.000000E+00 -1.000000E+01 0.000000E+00 *
Properties Matrix for
* 0.000000E+00 -3.000000E+00 5.000000E+00 8.000000E+00 -2.500000E+00 -1.500000E+00 * the Reference Point
* 2.000000E+00 0.000000E+00 -1.000000E+01 -2.500000E+00 9.500000E+00 0.000000E+00 *
* -2.966506E+00 6.058013E+00 0.000000E+00 -1.500000E+00 0.000000E+00 7.495513E+00 *

S
* 4.321332E-01 9.018098E-01 0.000000E+00 * Transformation from
* -9.018098E-01 4.321332E-01 0.000000E+00 * the Principal Mas to
* 0.000000E+00 0.000000E+00 1.000000E+00 *
the Basic Direction
DIRECTION
MASS AXIS SYSTEM (S) MASS X-C.G. Y-C.G. Z-C.G.
X 1.119709E+01 3.480388E-02 6.023980E-01 2.512494E-01
Center of Gravity
Y 8.802916E+00 -6.515555E-03 -4.426965E-02 2.484108E-01 Relative to the
Z 2.000000E+01 -9.385824E-03 5.589382E-01 0.000000E+00 Reference Point in
I(S) the Principal Mass
* 4.376953E+00 -8.768300E-01 6.624477E-01 * Axes System
* -8.768300E-01 5.623007E+00 -3.419880E-01 *
* 6.624477E-01 -3.419880E-01 3.431904E+00 * Moments of Inertia
with Respect to
I(Q)
* 4.463246E+00 * Principal Mass Axes
* 6.075616E+00 * for the Center of
* 2.893001E+00 *
Gravity
Q
* 7.201511E-01 4.586845E-01 5.205678E-01 * Principal Moments of
* -3.717336E-01 8.885992E-01 -2.687111E-01 *
* -5.858298E-01 0.000000E+00 8.104341E-01 *
Inertia

Figure I-4 Output from the Grid Point Weight Generator

Main Index
460

Before showing how each of the matrices are computed, a few items should be noted for this
model:

• User Warning Message 3042 is printed to inform you that inconsistent scalar masses
were used. This message occurs because there are different scalar masses in the three
components. In general, if you are using structural mass and/or CONM2s, you should
not get this message.
• The rigid-body mass matrix [ MO ] is computed with respect to the reference grid point1
in the basic coordinate system. Grid point 1 is used for this example because
PARAM,GNDPNT,1 is entered in the Bulk Data Section.
• The mass and center of gravity location shown are not in the basic coordinate system
for this example. The mass and center of gravity are computed and printed in the
principal mass coordinate system. This principal mass coordinate system should not
be confused with the principal axes discussed in most text books. The principal mass
axes in MSC.Nastran are the axes that have no coupling terms between the
translational masses (diagonal translational mass matrix). Also, the MSC.Nastran
principal mass axes are not the axes of the inertia ellipsoid. The [ S ] matrix is the
transformation from the principal mass direction to the basic coordinate system.
This additional step may sound confusing, but it is necessary. In real structures, the
mass of structure is generally the same in all directions, so there is no coupling between
the translational mass terms. Since text books are written to solve real structural
problems, there is no need to discuss the principal mass axes, as they are called in
MSC.Nastran. However, with MSC.Nastran you are not restricted to the same mass in
each coordinate direction—a situation that may not be physically realizable but still is
quite useful for certain modeling situations. Therefore, the additional step of
computing principal mass axes is necessary.
• If your model has the same mass in all coordinate directions, then the [ S ] matrix is the
identity matrix indicating that the principal mass axes is the basic coordinate system.
Always check the [ S ] matrix. If it is not the identity matrix, verify that the mass
distribution is correct. Do not use the directional mass and center of gravity location
blindly. Remember, these quantities are in the principal mass axes.
To fully understand how the GPWG module works, it is useful to trace the steps MSC.Nastran
follows to generate the output shown in Figure I-2. The following shows the step-by-step
procedure for the four mass example.
1. The GPWG module uses the global mass matrix, which is the mass matrix before any
constraints are applied. In this example, there are four grids, each with six degrees-of-
freedom, resulting in a total of 24 degrees-of-freedom in the mass [ MJJ ] matrix. The
matrix is shown in Figure I-5.

Main Index
 APPENDIX I 461
Grid Point Weight Generator

2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2
0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3
0 0 Grid
0 010 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6
0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7
0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8
0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9
0 0 0 0 0 0 0 0 Grid
0 02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 11
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 12
[ MJJ ] =
0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 13
0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 14
0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 15
0 0 0 0 0 0 0 0 0 0 0 0 0 0 Grid
0 03 0 0 0 0 0 0 0 0 16
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 17
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 18
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 19
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 20
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 21
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Grid
0 04 0 0 22
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 23
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 24

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Figure I-5 Global Mass Matrix

The [ MJJ ] matrix shows the mass contribution for each of the four grid points. Note
that the coordinate system associated with rows and columns 1 through 6 (grid point
1) is coordinate system 1, the coordinate system associated with rows and columns 13
through 18 (grid point 3) is coordinate system 3. The remaining rows and columns are
in the basic coordinate system.
2. To generate the 6 × 6 rigid-body mass matrix [ MO ] for the structure, it is necessary to
compute the mass matrix in the basic coordinate system relative to the reference point.
This computation requires the transformation matrix [ D ] that relates the rigid-body
displacements in the global system u g to the six unit displacements in the basic
coordinate system located at the reference grid point ( u o ) as shown in Eq. I-1.

{ u g } = [ D ] { u o } (Reference Point) Eq. I-1

The transformation matrix [ D ] is assembled from the individual transformation

matrices [ d ]i computed for each grid point. Each individual transformation matrix [ d ] i
consists of two transformations: [ Tr ] i , which relates the location of the grid point to the
reference grid point in the basic coordinate system, and [ Ti ] i , which relates the global
coordinate system at the grid point to the basic coordinate system.
The [ Tr ] i transformation matrices are first computed by constructing the location
vectors in the basic coordinate system r i for each grid point in the model relative to the
reference point as shown in Eq. I-2.

Main Index
462

 r1 
 
{ r }i = { R }i – { R }0 =  r2  Eq. I-2
 
 r3 

The location vectors for the example are as follows:

0  1   .5   .5 
       
{ r } 1 =  0  ; { r } 2 =  0 ; { r } 3 =  1 ; { r } 4 =  0  Eq. I-3
       
0  0  0  1 

Using the location vectors, the grid point transformation matrix [ Tr ] i is computed for
each grid point by expanding the location vectors to a 3 × 3 matrix as shown in Eq. I-4.

0 r3 –r2
[ Tr ] = –r3 0 r1 Eq. I-4
i
r2 –r1 0

For the example problem, the grid point transformation matrices are

0 0 0 0 0 0 0 0 –1 0 1 0
[ Tr ] 1 = 0 0 0 ; [ Tr ] 2 = 0 0 1 ; [ Tr ] 3 = 0 0 .5 ; [ Tr ]1 = – 1 0 .5 Eq. I-5
0 0 0 0 –1 0 1 – .5 0 0 – .5 0

The coordinate system transformation matrices from the global coordinate system to
the basic system coordinates are given by the direction cosine matrices as follows:

cos ( 45° ) – sin ( 45° ) 0 1 0 0

[ Ti ] 1 = sin ( 45° ) cos ( 45° ) 0 ; [ Ti ] 2 = 0 1 0 ;
0 0 1 0 0 1
Eq. I-6
cos ( 60° ) – sin ( 60° ) 0 1 0 0
[ Ti ] 3 = sin ( 60° ) cos ( 60° ) 0 ; [ Ti ] 4 = 0 1 0
0 0 1 0 0 1

The grid point transformation [ Tr ] i and the coordinate system transformation [ Ti ] i are
combined to form the individual grid point transformation matrix [ d ] i for each grid
point using Eq. I-7.

T T
Ti Ti ⋅ T r
[ d ]i = Eq. I-7
T
0 Ti

The rows of each [ d ]i form the columns of the global transformation matrix [ D ]T as
shown in Eq. I-8.

Main Index
 APPENDIX I 463
Grid Point Weight Generator

T
[D] = T
d1 d2 … dn
T T Eq. I-8

Using Eq. I-9, the global transformation matrix for the example is

.707 – .707 0 0 0 0 1 0 0 0 0 0 .5 – .866 0 0 0 0 1 0 0 0 0 0

.707 .707 0 0 0 0 0 1 0 0 0 0 .866 .5 0 0 0 0 0 1 0 0 0 0
T
[D] = 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0
0 0 0 .707 – .707 0 0 0 0 1 0 0 0 0 1 .5 – .866 0 0 –1 0 1 0 0
0 0 0 .707 .707 0 0 0 –1 0 1 0 0 0 – .5 .866 .5 0 1 0 – .5 0 1 0
0 0 0 0 0 1 0 1 0 0 0 1 – .067 1.116 0 0 0 1 0 .5 0 0 0 1 Eq. I-9

Grid 1 Grid 2 Grid 3 Grid 4

Using the global transformation matrix [ D ] , the rigid-body mass matrix about the
reference point in the basic coordinate system [ M0 ] is obtained by Eq. I-10.

T
[ M0 ] = [ D ] [ MJJ ] [ D ] Eq. I-10
For the example, [ M0 ] is determined to be

9.25 – .933 0. 0. 2. – 2.967

– .933 10.75 0. – 3. 0. 6.058
[ M0 ] = 0. 0 20. 5. – 10. 0. Eq. I-11
0. –3 5. 8. – 2.5 – 1.5
2. 0 – 10. – 2.5 9.5 0.
– 2.967 6.058 0. – 1.5 0. 7.496

Comparing the results shown in Eq. I-11 to [ M0 ] generated by the GPWG module
(Figure I-2) shows the matrices to be numerically the same.
3. The next step is to inspect the [ M0 ] to determine whether the basic coordinate system
can be used as the principal mass directions. The principal mass axes are axes that have
no coupling between the translational mass components. For real structures, there is
no coupling in the translational mass terms in the inertia matrix. However, with
inconsistent scalar masses (CONM1, CMASSi, or DMIG), you may define any type of
mass matrix you desire.
To determine whether coupling exists between the translational mass terms, [ M0 ] is
partitioned into four 3 × 3 matrices as shown in Eq. I-12.

t tr
[ M0 ] = M M Eq. I-12
rt r
M M

where the superscripts t and r refer to translation and rotation, respectively.

For this example, the translational mass partition is given by

Main Index
464

t 9.25 – .933 0
[M ] = – .933 10.75 0 Eq. I-13
0 0 20.

A check is made for coupling as follows:

t 2
δ = ∑ ( M ij ) i = j

t 2
ε = ∑ ( M ij ) i>j
Eq. I-14
If ε ⁄ δ is greater than .001, then excessive coupling exists preventing the basic
coordinate system from being used for the principal mass directions and User Warning
.5
Message 3042 is printed. For this problem, δ = ( 9.25 2 + 10.75 2 + 20 2 ) and
2 2 2 .5
ε = ( .933 + 0. + 0. ) . The ratio ε ⁄ δ = .038 agrees with the System Warning Message
3042 shown in Figure I-2.
If needed, the principal mass directions are computed by performing an eigensolution
t
with the translational mass components. The eigenvectors of [ M ] are the columns of
the transformation matrix [ S ]

[ S ] = [ { e 1 }, { e 2 }, { e 3 } ] Eq. I-15

Using this eigenvector matrix, the partitions of the rigid-body mass matrix with respect
to the principal mass direction are computed by Eq. I-16.

t T t
[M ] = [S] [M ][S]

tr T tr
[ M ] = [ S] [M ][ S]

r T r
[M ] = [S] [M ][S]
Eq. I-16
The [ S ] matrix for the example is given by

.432 .902 0.
[S] = – .902 .432 0. Eq. I-17
0. 0. 1.

The [ S ] matrix is printed after the [ M0 ] matrix as shown in Figure I-2. This represents
the transformation relating the basic coordinate system to the principal mass axes.
Again, if there is no coupling between the translational mass component, which is the
case for most problems, the eigensolution is not required, and the [ S ] matrix is set equal
to the identity matrix. This example was selected to demonstrate all of the features of
the GPWG module, but it is not a typical problem.
4. The next step is to determine the principal masses and the center of gravity location in
the principal mass axes system as shown in Eq. I-18.
Main Index
 APPENDIX I 465
Grid Point Weight Generator

t
M x = M 11

t
M y = M 22

t
M z = M 33

The “centers of gravity” are

tr tr tr
M 11 – M 13 M 12
X x = ---------- , Y x = ------------- , Z x = ---------- ,
Mx Mx Mx

tr tr tr
M 23 M 22 – M 21
X y = ---------- , Y y = ---------- , Z y = ------------- ,
My My My

tr tr tr
– M 32 M 31 M 33
X z = ------------- , Y z = ---------- , Z z = ----------
Mz Mz Mz
Eq. I-18
As can be seen, the center of gravity location is not a unique location. The center of
gravity location is computed separately for the for x-, y-, and z-directions relative to the
principal mass axes. Only if the mass is the same in each direction (which is typical) is
there a unique center of gravity location, which is relative to the reference point in the
basic coordinate system.
For this example problem, the center of gravity locations are determined to be
Table I-2 Mass Center of Gravity Locations

Center of Gravity Location in

Mass Principal Mass Axes
Component
X Y L
x - mass .035 –0.602 0.251
y - mass –0.007 –0.044 0.248
z - mass –0.009 –0.559 0

The center of gravity location given in Table I-2 is the same as shown in Figure I-2.
5. Following the center of gravity calculation is the calculation to determine the moments
of inertia for the center of gravity with respect to the principal mass axes [ I ( S ) ] as
shown in Eq. I-19.

Main Index
466

(S) r 2 2
I 11 = M 11 – M y Z y – M z Y z

(S) (S) r
I 12 = I 21 = – M 12 – M z X z Y z

(S) (S) r
I 13 = I 31 = – M 13 – M y X y Z y

(S) r 2 2
I 22 = M 22 – M z X z – M x Z x

(S) (S) r
I 32 = I 23 = – M 23 – M x Y x Z x

(S) r 2 2
I 33 = M 33 – M x Y x – M y X y
Eq. I-19
For the example, the inertia matrix [ I ( S ) ] is given by

4.377 – .877 .662

[ I( S)] = – .877 5.623 – .342 Eq. I-20
.662 – .342 3.432

6. The final step is to compute the principal moments of inertia and the principal
directions of the momental ellipsoid (commonly referred to as the principal axes in text
books). An intermediate inertia matrix [ Ĩ ] is generated by reversing the sign on the off
diagonal terms of [ I ( S ) ] . For the inertia matrix given in Eq. I-20, [ Ĩ ] is given by

4.377 .877 – .662

[ Ĩ ] = .877 5.623 .342 Eq. I-21
– .662 .342 3.432

An eigensolution is performed on the [ Ĩ ] matrix to determine the principal directions.

The resulting eigenvalues are the principal moments of inertia, which are assumed to
be the diagonal terms of the principal inertia matrix [ I ( Q ) ] . The eigenvectors form the
columns of the matrix [ Q ] , which is the transformation relating the intermediate inertia
matrix [ Ĩ ] to the principal inertia matrix [ I ( Q ) ] as shown in Eq. I-22.

T
[ I ( Q ) ] = [ Q ] [ Ĩ ] [ Q ] Eq. I-22
For the example, [ Q ] and [ I ( Q ) ] are given by

Main Index
 APPENDIX I 467
Grid Point Weight Generator

4.463 0. 0.
[I(Q)] = 0. 6.076 0.
0. 0. 2.893

.720 .459 .521

[Q] = – .372 .889 – .269
– .586 0 .810
Eq. I-23
The matrices [ S ] and [ Q ] matrices are the coordinate rotation matrices, which when
taken together, relate the principal directions of the momental ellipsoid to the basic
coordinate system. The matrices given in Eq. I-23 are in agreement with those given in
Figure I-2. The example for the four mass model is now complete.

Main Index
468

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Numerical Accuracy Considerations
J
■ Overview

Main Index
470

J.1 Overview
MSC.Nastran is an advanced finite element analysis program that can solve models in excess of
1,000,000 degrees-of-freedom. Because speed and accuracy are essential, MSC.Nastran’s
numerical analysis capabilities are continually enhanced to provide the highest level of each.
This appendix provides a brief overview for detecting and avoiding numerical ill-conditioning
problems, especially as they relate to dynamic analysis. For more information regarding
MSC.Nastran’s numerical analysis algorithms, see the MSC.Nastran Numerical Methods User’s
Guide.

Linear Equation Solution

The basic statement of the linear equation solution is

[A]{x} = {b} Eq. J-1

where [ A ] is a square matrix of known coefficients (and usually symmetric for structural
models), { b } is a known vector, and { x } is the unknown vector to be determined.
The methods used for solution in MSC.Nastran are based on a decomposition of [ A ] to triangular
matrices and are followed by forward-backward substitution to get { x } . The equations for this
solution technique are

[A] = [L][U] Eq. J-2

where [ L ] is a lower-triangular matrix and [ U ] an upper-triangular matrix, and

[L ]{ y } = { b} Eq. J-3

where { y } is an intermediate vector. Eq. J-3 is called the forward pass because the solution starts
with the first row where there is only one unknown due to the triangular form of [ L ] . The
backward pass starts with the last row and provides the solution

[U]{x} = {y} Eq. J-4

Eigenvalue Analysis
The general eigensolution equation is

2
[ K + pB + p M ] { u } = 0 Eq. J-5

where p is the complex eigenvalue. This equation can always be transformed to a special
eigenvalue problem for a matrix [ A ]

[A – λ I]{φ} = 0 Eq. J-6

where [ I ] is the identity matrix. Eq. J-6 is the basis of all the transformation methods of
MSC.Nastran (HOU, GIV, etc.). The iterative methods (INV, SINV) work directly from Eq. J-5.
The Lanczos method uses both. If [ A ] is a symmetric matrix, the eigenvectors are orthogonal,
and they can be normalized such that
Main Index
 APPENDIX J 471
Numerical Accuracy Considerations

T
[φ ][φ] = [I] Eq. J-7

where [ φ ] is a square matrix whose columns contain the eigenvectors { φ i } . With this
normalization convention, then

T
[φ ][A][φ] = [λ] Eq. J-8
and

T
[ A ] = [ φ ] [ λ ][ φ ] Eq. J-9

where [ λ ] is the eigenvalue diagonal matrix.

Matrix Conditioning
Reordering the previous equations, any matrix [ A ] can be expressed as a sum of its eigenvalues
multiplied by dyadic eigenvector products

T T  T
[ A ] = λ 1 { φ1 } { φ1 } + λ 2 { φ2 } { φ2 } + … + λ n { φn }  φn  Eq. J-10
 
 
Defining [ B i ] = { φi }  φTi  , which is a full rank 1 matrix, then
 
n
[A] = ∑ λk [ Bk ] Eq. J-11
k = 1

where n is the dimension of [ A ] . On the average, an element of [ B k ] has the same magnitude as
the corresponding element of [ B k + 1 ] . Let B max be the magnitude of the largest coefficient of all
[ B k ] matrices. Then

a ij ≤ B max ( λ 1 + λ 2 + … + λ n ) Eq. J-12

This equation shows that the terms of [ A ] are dominated by the largest eigenvalues.
Unfortunately, the smallest eigenvalues are those of greatest interest for structural models.
These small eigenvalues must be calculated by taking the differences of coefficients that are
dominated by the largest eigenvalue. For this reason, the ratio λn ⁄ λ1 is called a numerical
conditioning number. If this number is too large, numerical truncation causes a loss of accuracy
when computing the lowest eigenvalues of a system.
The assumptions that allow this simple analysis are often pessimistic in practice, that is, the
bounds predicted by the error analysis are conservative. However, the effects it predicts do
occur eventually so that models that produce acceptable results for one mesh size may produce
unacceptable results with a finer mesh size due to the higher eigenvalues included in the
larger-sized matrices occurring from the finer mesh.

Main Index
472

Definiteness of Matrices
A matrix whose eigenvalues are all greater than zero is said to be positive definite. If some
eigenvalues are zero but none are less than zero, the matrix is positive semi-definite. A stiffness
matrix assembled from elements is at least positive semi-definite. If all of the structure’s
rigid-body modes are constrained, the stiffness matrix is positive definite. Another category is
the indefinite matrices category. These matrices have zeroes or blocks of zeroes on the diagonal.
Although definiteness is most concisely defined in terms of eigenvalues, it is not a practical test
for large matrices because of the computational cost of extracting all of the eigenvalues.
However, other operations, particularly linear equation solution and dynamic reduction, may
detect nonpositive definite matrices and provide diagnostics using these terms as described later
in this appendix.

Numerical Accuracy Issues

The numerical operations of MSC.Nastran are executed in a finite 64-bit floating point
arithmetic. Depending on the specific computer’s word structure (number of bits for mantissa
versus exponent), different roundoff errors may occur. The sophistication level of the actual
hardware (or software) arithmetic units also has an influence on the numerical accuracy. To
attain the most numerical accuracy possible, the following strategies are used in MSC.Nastran.
In the decomposition of positive definite matrices, the Gaussian elimination process (which is
the basis of all decomposition algorithms in MSC.Nastran) does not require numerical pivots.
Since some of the matrices are not positive definite, sparse decomposition (both symmetric and
unsymmetric) employs a numerical pivoting (row/column interchange) algorithm. These
methods consider a pivot term suitable if

THRESH
a kk ≥ 10 ⋅ MAX a ki Eq. J-13
i

which means that a diagonal term is accepted as a pivot if it is greater than the maximum term
in that row multiplied by 10 THRESH (the default for THRESH is –6).
To ensure numerical accuracy in eigenvalue calculations, most MSC.Nastran methods use a
spectral transformation of

1
µ = ---------------- Eq. J-14
λ – λs

where λ s is an eigenvalue shift.

This transformation is shown in Figure J-1.

Main Index
 APPENDIX J 473
Numerical Accuracy Considerations

µ3

µ2
µ1
λ
λS λ1 λ2 λ 3

Figure J-1 Spectral Transformation

The spectral transformation ensures uniform accuracy throughout the frequency by “shifting”
to the area of interest. Another effect of this transformation is the welcomed µ space separation
of closely-spaced eigenvalues. When λs is close to an eigenvalue λ i , the decomposition of the
“shifted matrix”

T
[ K – λs M ] = [ L ] [ D ] [ L ] Eq. J-15

may produce high MAXRATIO messages. Automatic logic to perturb the λs value in this case
is implemented in MSC.Nastran.

Sources of Mechanisms
In all of the decomposition methods, a null row or column in [ A ] causes a fatal error message.
All other causes of singularity are not distinguishable from near-singularity because of the
effects of numerical truncation. Only warning messages are usually provided for these cases.
In standard decomposition

T
[A ] = [L ][ D][L ] Eq. J-16

the process starts to compute the first term of [ D ] with the first internal degree of freedom and
then processes each additional degree of freedom and its associated terms. It can be shown that
when processing the k-th row, the k-th row and all rows above it are in effect free, and all rows
below it are constrained to ground. The term of [ D ] at the k-th row is proportional to the amount
of coupling between that degree of freedom and the degree-of-freedom with a higher value of k.
If the terms of the k-th row and above are not connected to the remaining rows, the k-th term of
[ D ] goes to zero. Because of numerical truncation, the term may be a small positive or negative
number instead. If a term of [ D ] is calculated to be identically zero, it is reset to a small number
because of the indeterminacy of its calculation. The existence of such a small term defines a
mechanism in static analysis in MSC.Nastran. A mechanism is a group of degrees-of-freedom

Main Index
474

that may move independently from the rest of the structure as a rigid body without causing
internal loads in the structure. A hinged door, for example, is a mechanism with one rigid-body
freedom. If the hinges are disconnected, the door mechanism has six rigid-body freedoms.
Mechanisms are characterized by nondimensional numbers derived by dividing the terms of [ D ]
into the corresponding diagonal term of [ A ] . If these “matrix diagonal to factor diagonal” ratios
are large numbers, various warning and fatal messages are produced, depending on the context.

Sources of Nonpositive Definite Matrices

A negative semidefinite element stiffness matrix, which is defined as one whose eigenvalues are
all negative or zero, implies that the element has an energy source within it. MSC.Nastran
discourages using such elements by giving fatal messages when negative element thicknesses or
section properties are input. However, there are applications where such elements are useful,
such as when using an element with negative stiffness in parallel with a passive element to
model a damaged or thinned-down element. For this reason, negative material property
coefficients and negative stiffnesses for scalar elements are allowed. Also, some incorrect
modeling techniques, such as the misuse of the membrane-bending coupling term on PSHELL
entries (MID4), can lead to negative eigenvalues of the element stiffness matrix.
Stiffness matrices with negative eigenvalues cause negative terms in [ D ] . The number of such
terms is automatically output by the standard symmetric decomposition subroutine. Their
existence again causes various warning and fatal messages, depending on the context. The most
common cause of negative terms in [ D ] is true mechanisms, whose terms are small negative
numbers and are actually computational zeroes.
Indefinite matrices occur when using the Lagrange multiplier method for constraint processing.
In this method, the constraints are not eliminated from but are concatenated to the g-size system
of equations. The constraint equations have zeroes on the diagonal, resulting in indefinite
matrices. These systems can be solved using the block pivoting scheme of the sparse
decomposition.

Detection and Avoidance of Numerical Problems

Static Analysis
Models used in static analysis must be constrained to ground in at least a statically determinate
manner even for unloaded directions. For example, a model intended for only gravity loading
must be constrained in horizontal directions as well as vertical directions. The evidence of
unconstrained directions is that the entire model is a mechanism, that is, the large ratio occurs at
the last grid point in the internal sequence.
Another source of high ratios arises from connecting soft elements to stiff elements. Local
stiffness is a function of element thickness (moment per unit rotation) through element thickness
cubed (force per unit deflection) and is inversely proportional to mesh spacing, again in linear
through cubic ratios. Some relief is possible by sequencing the soft degrees-of-freedom first in
the internal sequence, although this is difficult to control in the presence of automatic
resequencing. More reliable corrections are to replace the very stiff elements with rigid elements
or to place the soft and stiff elements in different superelements.
Main Index
 APPENDIX J 475
Numerical Accuracy Considerations

A third source of high ratios is the elements omitted through oversight. The corrective action
here is to start with the grid points listed in the diagnostics and track back through the elements
connected to them through the upstream grid points. The missing elements may be anywhere
upstream. PARAM,GPECT,1 output and undeformed structure plots all provide useful data for
detecting missing elements.

At present, there are two major methods of identifying large ratios and nonpositive-definite
matrices. In some solutions, the largest matrix diagonal to factor diagonal ratio greater than 105
(MAXRATIO default) is identified by its internal sequence number, and the number of negative
factor diagonal terms is output. The best method to identify mechanisms here is to apply
checkout loads that cause internal loads in all of the elements. Then inspect the displacement
output for groups of grid points that move together with implausibly large displacements and
common values of grid point rotation. The only condition that causes a fatal error is a true null
column, and NASTRAN SYSTEM(69)=16 avoids this fatal error by placing a unit spring
coefficient on the degrees-of-freedom with null columns. This option is recommended only for
diagnostic runs because it may mask modeling errors.
In other solution sequences, all ratios greater than 10 5 are printed in a matrix format named the
MECH table. The external sequence number of each large ratio is also printed, which is the grid
point and degree-of-freedom number. If any such ratios exist, the action taken depends on the
value of PARAM,BAILOUT. In the conventional solution sequences, its default value causes the
program to continue after printing the MECH matrix. In the superelement solution sequences,
a different default causes a fatal error exit after printing the MECH matrix. For both types of
solution sequences, the opposite action may be requested by setting the value of
PARAM,BAILOUT explicitly. Also, the criterion used for identifying large ratios may be
changed using PARAM,MAXRATIO.

For static analysis, values between 10 3 and 10 6 are almost always acceptable. Values between
7 8
10 and 10 are questionable. When investigating structures after finding these values, some
types of structures may be found to be properly modeled. It is still worthwhile to investigate the
structures with questionable values.
The solutions with differential stiffness effects offer another method to obtain nonpositive
definite stiffness matrices. For example, a column undergoing compressive gravity loading has
a potential energy source in the gravity load. A lateral load that is stabilizing in the absence of
gravity (i.e., a decreased load causes a decreased deflection) is destabilizing when applied in a
postbuckled state.

Eigensolutions Using the Inverse Iteration and Lanczos Methods

The matrix [ K – λ s M ] is decomposed where λs is an eigenvalue shift and [ K ] and [ M ] are the
stiffness and mass matrices, respectively. This condition allows the solution of models with
rigid-body modes since the singularities in [ K ] are suppressed if there are compensating terms
in [ M ] . The only conditions that should cause fatal messages due to singularity are

1. The same null column in both [ K ] and [ M ] . These columns and rows are given an
uncoupled unit stiffness by the auto-omit operation if the default value (0) for
PARAM,ASING is used. If this value is set to -1, a null column in both matrices is
regarded as an undefined degree of freedom and causes a fatal error exit.
Main Index
476

2. A massless mechanism. One commonly encountered example is a straight-line, pinned

rod structure made from CBAR elements with no torsional stiffness (J) defined. The
structural mass generated for CBAR elements does not include inertia for the torsion
degrees-of-freedom. The natural frequency for this torsion mode approaches the limit
of zero torsional stiffness divided by zero mass, which is an undefined quantity. If the
elements lie along a global coordinate axis, the mass term is identically zero, which
leads to very large negative or positive eigenvalues and is usually beyond any
reasonable search region. If the elements are skewed from the global axes, the
eigenvalues may be computed at any value (including negative) because of the
indeterminacy caused by numerical truncation.

The “negative terms on factor diagonal” message generally occurs for every decomposition
performed in the iteration. It can be shown from Sturm sequence theory that the number of
negative terms is exactly equal to the number of eigenvalues below λ s . This condition is a means
of determining if all roots in the range have truly been found.

Eigensolutions Using Transformation Methods

If the GIV or HOU methods (and their variations) are selected, one of the first operations
performed is a Choleski decomposition, which is used to reduce the problem to a special
eigenvalue problem. The mass matrix is given this operation for the straight GIV or HOU
method. Columns that are identically null are eliminated by the auto-omit operation. Poor
conditioning can result from several sources. One example is a point mass input on a C0NM2
entry that uses offsets in three directions. The grid point to which it is attached has nonzero mass
coefficients for all six degrees-of-freedom. However, only three independent mass
degrees-of-freedom exist, not six. Another example results in a superelement analysis when
most of the elements in a superelement do not have mass and all interior masses are restricted
to only a few structural or mass elements. The boundary matrix produced for the superelement
is generally full, no matter what its rank, i.e., regardless of the number of independent mass
degrees-of-freedom that it contains.
The presence of a rank-deficient mass matrix when using the GIV or HOU method produces
fatal messages due to the singularity of the mass matrix or produces solutions with poor
numerical stability. Poor stability is most commonly detected when making small changes to a
model and then observing large changes in the solution. Either the MGIV and MHOU methods
or the AGIV and AHOU methods are the preferred methods to use when nearly singular mass
matrices are expected because these methods decompose the matrix [ K – λs M ] instead of the mass
matrix. The shift parameter λ s is automatically set to be large enough to control rigid-body
modes. Better modeling practices also reduce the costs and increase the reliability for the two
examples cited above. If an offset point mass is significant in any mode, it is better to attach it to
an extra grid point at its center of gravity and model the offset with a rigid element. The
auto-omit feature then eliminates the rotational degrees-of-freedom. Similarly, if only a few
interior points of a superelement have mass, it may be more economical to convert them to
exterior points, which also eliminates the singular boundary mass matrix.

Main Index
 APPENDIX J 477
Numerical Accuracy Considerations

It is possible to input negative terms into a mass matrix directly with DMIG terms or scalar mass
elements. This class of problem causes fatal errors due to nonpositive definite mass matrices for
the transformation methods, fatal errors with the Lanczos method, and wrong answers with the
INV method. The complex eigenvalue methods should be used for this type of problem, which
infers that modal or dynamic reduction methods may not be used.

A similar but quite different problem arises because Cholesky decomposition is used on the
generalized mass matrix (named MI in the diagnostics) when orthogonalizing the eigenvectors
with respect to the mass matrix. The existence of negative terms here indicates that poor
eigenvectors were computed. The row number where the negative term occurs is printed in the
diagnostic. This row number does not refer to a physical degree of freedom but refers instead to
an eigenvector number. The usual cause is computing eigenvectors for computational infinite
roots in the modified transformation methods. This problem can be avoided after the fact by
setting ND on the EIGR entry to a value less than the row number that appears in the diagnostics
or before the fact by setting F2 on the EIGR entry to a realistic value instead.

Frequency Response Analysis

Negative terms on the factor and high factor to matrix diagonal ratios can be expected when
using coupled methods and may often be safely ignored. These messages are merely an
indication that the excitation frequency is above one or some of the natural frequencies of the
system (negative terms), or is near a natural frequency (high ratio).

Transient Response Analysis

A well-conditioned model should have neither negative terms nor high ratios on its factor terms.
The causes of such messages include all of the effects described above. Negative mass terms can
be detected by rapidly diverging oscillations in the transient solution.

The Large Mass Method

One of the methods for enforcing motion is the large mass method. It may be used with direct
methods as well as all of the reduction methods. The basis of the method is to attach artificial
masses to the structure at the degrees-of-freedom where the motion is to be enforced. These
large masses should be orders of magnitude larger than the total mass or the moment of inertia
of the structure. A survey of the literature shows recommendations for mass ratios ranging from
3 8 6
10 to 10 with a value of 10 as the most common recommendation.

The mass ratio affects both the accuracy and numerical conditioning, and must be adjusted in a
compromise that meets both criteria. With regard to load accuracy, the error in the
approximation is inversely proportional to the ratio. A ratio of 10 3 causes an error of ten percent
at resonance for a mode with one-half percent of damping, which represents an extreme case.
Off-resonance excitation or higher damping ratios result in lower errors. Numerical
conditioning problems are much more difficult to predict.

Main Index
478

The Lagrange Multiplier Method

Another method to enforce motion is the Lagrange multiplier method. In this method, the input
motion function is described by a constraint equation. This method provides better accuracy
than the large mass method where the numerical error introduced is proportional to the large
mass.
The formulation introduces an indefinite system of linear equations where some numerical
problems may arise from the fact that the system matrix contains terms that have dimensions of
stiffness as well as nondimensional (constraint) terms. The decomposition of this matrix with
the sparse decomposition methods pivoting strategy is stable.
You may control the pivoting strategy with the THRESH DMAP parameter mentioned earlier.
The default value (–6) is adequate for most Lagrange multiplier solutions. In fact, higher
(ranging up to –2) values provide better accuracy, while increasing the number of pivots may
result in performance degradation.

Main Index
 MSC.Nastran Basic Dynamics User’s Guide

APPENDIX
Diagnostic Messages for Dynamic Analysis
K
■ Overview

Main Index
480

K.1 Overview
This appendix lists common diagnostic messages for dynamic analysis. The text for each
message is given in uppercase letters and is followed by additional explanatory material,
including suggestions for remedial action. These and other messages are listed and described in
the MSC/NASTRAN Reference Manual.
The messages in this section have the following format:
*** (SYSTEM/USER) (FATAL/WARNING/INFORMATION) MESSAGE ID, text
where “ID” is a unique message identification number and “text” is the message as indicated in
capital letters for each of the diagnostic messages. Four asterisks (****) in the message text
indicates information that is filled in for a specific use of the message, such as the number of a
grid point or the name of a Bulk Data entry. Some of the messages are followed by additional
explanatory material, including suggestions for remedial action.
Fatal messages cause the termination of the execution following the printing of the message text.
These messages always appear at the end of the MSC.Nastran output. Warning and information
messages appear at various places in the output stream. Such messages only convey warnings
or information to the user. Consequently, the execution continues in a normal manner following
the printing of the message text.

As an example, consider message number 2025, which appears in the printed output as follows:
*** USER FATAL MESSAGE 2025, UNDEFINED COORDINATE SYSTEM 102
The three leading asterisks (***) are always present in numbered user and system diagnostic
messages. The word USER indicates that this is a user problem rather than a system problem.
The word FATAL indicates that this is a fatal message rather than a warning or information
message. The number 2025 is the identification number for this message. The text of the
message follows the comma. The number 102 replaces the asterisks (****) in the general message
text and indicates that 102 is the identification number of the undefined coordinate system. The
abbreviation UFM refers to User Fatal Message, UWM refers to User Warning Message, UIM
refers to User Information Message, and SFM refers to System Fatal Message.

UFM 2066
*** USER FATAL MESSAGE 2066, UNDEFINED GRID POINT **** ON DAREA CARD.
A dynamic loading entry references an undefined grid point.

UFM 2069
*** USER FATAL MESSAGE 2069, UNDEFINED GRID POINT **** IN TRANSIENT INITIAL
CONDITION SET ****.
An attempt has been made to specify initial conditions for an undefined grid point. All
degrees-of-freedom with initial conditions must be in the analysis set.
Main Index
 APPENDIX K 481
Diagnostic Messages for Dynamic Analysis

UFM 2071
*** USER FATAL MESSAGE 2071, DYNAMIC LOAD SET **** REFERENCES UNDEFINED
TABLE ****.
A referenced dynamic load table was not present in the Bulk Data.

UFM 2074
*** USER FATAL MESSAGE 2074, UNDEFINED TRANSFER FUNCTION SET ****.
A transfer function set was selected in the Case Control but was not present in the Bulk Data.

UFM 2079
*** USER FATAL MESSAGE 2079, GRID OR SCALAR POINT **** HAS AN UNDEFINED
COORDINATE REFERENCED ON A DAREA, DELAY, DPHASE CARD.
The “C” or component value for scalar-type points must be zero or one.

UFM 2088
*** USER FATAL MESSAGE 2088, DUPLICATE TABLE ID ****.
All tables must have unique numbers. Check for uniqueness.

UFM 2101A
*** USER FATAL MESSAGE 2101A, GRID POINT **** COMPONENT *** ILLEGALLY
DEFINED IN SETS ****.
The above grid point and component are defined in each of the above independent subsets. A
point may belong to a maximum of one independent subset. This error occurs when a DOF is
defined as belonging to two mutually exclusive sets (see MSC.Nastran Reference Manual).
A common example of this occurs when a DOF is defined as dependent on an MPC (M-set) as
well as being constrained (s-set) on an SPC entry. The message for this states that the component
is illegally defined in the um (user-defined m-set) and us (user-defined s-set) sets. These two sets
are mutually exclusive because all MPC equations are processed before the SPCs are applied (the
exception is SOL 24) and the m-set DOFs are removed from the matrix. When the program
attempts to apply the SPC, the DOF is no longer available, and the fatal message is issued. The
normal correction for this is to modify the MPC so that the DOF in question is independent
(n-set). Then there is no conflict.

UFM 2107
*** USER FATAL MESSAGE 2107, EIGR-CARD FROM SET **** REFERENCES DEPENDENT
COORDINATE OR GRID POINT ****.
When the point option is used on an EIGR entry, the referenced point and component must be
in the analysis set (a-set) for use in normalization.

Main Index
482

UFM 2109
*** USER FATAL MESSAGE 2109, NO GRID, SCALAR OR EXTRA POINTS DEFINED.
Dynamics problems must have at least one grid, scalar, or extra point.

UFM 2133
*** USER FATAL MESSAGE 2133, INITIAL CONDITION IN SET **** SPECIFIED FOR POINT
NOT IN ANALYSIS SET.
Initial conditions can only be specified for analysis set points. Therefore the point/component
mentioned on TIC entries must belong to the analysis set.

UFM 2135
*** USER FATAL MESSAGE 2135, DLOAD CARD *** HAS A DUPLICATE SET ID FOR SET ID
***.
The Li Set IDs on a DLOAD entry are not unique. See the DLOAD Bulk Data description.

UFM 2136
*** USER FATAL MESSAGE 2136, DUPLICATE DLOAD, RLOAD, OR TLOAD SET ID
NUMBER = ****** HAS BEEN ENCOUNTERED FOR DLOAD SET = ******.
Dynamic loads may not be defined by giving multiple data entries with the same ID. Use unique
IDs.

UIM 2141
*** USER INFORMATION MESSAGE 2141, TIME ESTIMATE IS **** SECONDS. PROBLEM
SIZE IS ****, SPILL WILL OCCUR FOR THIS CORE AT A PROBLEM SIZE OF ****.
The time estimate includes the time of the tridiagonalization and eigenvalue calculation when
the GIV or HOU methods are used. If ND is given on the EIGR entry, it also includes the time
of the eigenvector generation. If F1 and F2 are used instead, the eigenvector times are not
estimated. This condition can underestimate the time when the range includes many
eigenvectors.

UFM 2200
*** USER FATAL MESSAGE 2200, INCONSISTENT RIGID BODY SYSTEM.
This message occurs if a SUPORT is used and the rigid-body mass matrix is not positive definite.
Possible causes are unconstrained mechanisms or input of negative mass terms. A diagnostic
method is to remove all SUPORT entries and inspect the resulting eigenvectors for implausible
behavior.

Main Index
 APPENDIX K 483
Diagnostic Messages for Dynamic Analysis

UFM 3031
*** USER FATAL MESSAGE 3031, UNABLE TO FIND SELECTED SET (****) IN TABLE (****)
IN SUBROUTINE (****).
A particular set used in the problem was not included in the data. Good examples are loads,
initial conditions, or frequency sets. Include the required data or change the Case Control
commands to select data already in the problem. Set zero (0) has a special meaning. A set
selection was required, but none was made. For example, no METHOD was selected for an
eigenvalue extraction problem, or no FREQ was selected for frequency response.
This message can also indicate that a LOAD entry references another LOAD entry, which is not
permitted.
This message can also occur if a DLOAD entry references a nonexisting LOAD entry, e.g.,
RLOAD1.

UWM 3034
*** USER WARNING MESSAGE 3034, ORTHOGONALITY TEST FAILED. LARGEST TERM =
****, NUMBER FAILED = ****, PAIR = ****, ****, EPSILON = ****.
This indicates that eigenvector accuracy is in doubt. This message is printed only when the
off-diagonal terms of the modal mass matrix are larger than 1.0E-10. The eigenvectors are not
orthogonal to this extent. This nonorthogonality is especially important if a modal formulation
is used. The pair of eigenvectors listed exhibit the worst behavior. The number failed is the
number of pairs above the criteria. You can improve the numerical conditioning of the problem
by reducing the range of mass ratios, stiffness ratios, and eigenvalue range.

UIM 3035
*** USER INFORMATION MESSAGE 3035, FOR DATA BLOCKS ****. SUPORT PT. NO.
EPSILON STRAIN ENERGY EPSILONS LARGER THAN .001 ARE FLAGGED WITH
ASTERISKS.
One line of output is printed for each component on a SUPORT entry. Large values of either
EPSILON or STRAIN ENERGY indicate errors in the constraints, MPCs, or offsets.

UWM 3045
*** USER WARNING MESSAGE 3045, INSUFFICIENT TIME TO COMPLETE THE
REMAINING ** SOLUTION(S) IN MODULE ***.

The estimated time for completion of the module is less than the time remaining, as specified on
the Executive Control TIME statement. The module computes one solution (for example, one
excitation frequency in frequency response analysis) and then processes all output requests. The
remaining frequencies can be obtained on restart by adding or changing a FREQ command.

Main Index
484

UFM 3046
*** USER FATAL MESSAGE 3046, YOUR SELECTED LOADING CONDITION, INITIAL
CONDITION, AND NONLINEAR FORCES ARE NULL. A ZERO SOLUTION WILL RESULT.
Transient solutions must have one of the above nonzero loading condition, initial condition, or
nonlinear forces. Also, make sure that LOADSET is spelled correctly.

UFM 3047
*** USER FATAL MESSAGE 3047, NO MODES WITHIN RANGE AND LMODES=0. A
MODAL FORMULATION CANNOT BE MADE.
The modes used for a modal formulation must be selected by a PARAM entry. Set LFREQ,
HFREQ, or LMODES to request modes.

UFM 3051
*** USER FATAL MESSAGE 3051, INITIAL CONDITION SET **** WAS SELECTED FOR A
MODAL TRANSIENT PROBLEM. INITIAL CONDITIONS ARE NOT ALLOWED IN SUCH A
PROBLEM.
The IC command is not allowed for modal transient response problems.

UWM 3053
*** USER WARNING MESSAGE 3053, THE ACCURACY OF EIGENVALUE **** IS IN DOUBT.
GIV/HOU QR FAILED TO CONVERGE IN **** ITERATIONS.
Each eigenvalue is computed to the precision limits of each machine consistent with the
maximum number of iterations allowed.

UFM 3057
*** USER FATAL MESSAGE 3057, MATRIX **** IS NOT POSITIVE DEFINITE.

A Cholesky decomposition was attempted on the above matrix, but a diagonal term of the factor
was imaginary or equal to zero such that the decomposition failed. This message is from the
regular (as opposed to sparse) decomposition method.
This message may be produced because of constraint problems. Check the output for
UWM 4698 for large factor diagonal ratios and constrain appropriately.

UWM 4193
*** USER WARNING MESSAGE 4193, A GRID AND COMPONENT SPECIFICATION ON A
(DPHASE/DELAY) SID = **, DOES NOT APPEAR ON A DAREA CARD.
The area specification is set to zero.

Main Index
 APPENDIX K 485
Diagnostic Messages for Dynamic Analysis

SFM 4276
*** SYSTEM FATAL MESSAGE 4276, **** ERROR CODE ****
This message occurs when MSC.Nastran encounters errors that are not otherwise trapped
(including system errors). There are various error codes (EC), each of which has a different
meaning.
In nearly every case, the log file contains further information about the problem, so you should
look there for further clues. In addition, because these are errors that most likely should be
caught in another manner (and with a more explicit error message), you should look at the recent
Error Reports (search for “4276").
In many cases, increasing memory, BUFFSIZE, or disk space resolves the problem, especially
when attempting to run large models on workstations.
This error often indicates machine underflow or overflow, although it is impossible to list
specific reasons for this error code. Observed problems include:
1. Modeling problems. This condition is usually accompanied by an “arithmetic fault,
floating overflow” type message in the log file on some machines. Perform a static
analysis and verify that the value for the maximum factor diagonal ratio is acceptable.
2. When SFM 4276 is followed by an “access violation”, then it is often due to a lack of
memory or disk space or due to a coding error in MSC.Nastran.

UWM 4312
*** USER WARNING MESSAGE 4312, CONM2 **** HAS NONPOSITIVE-DEFINITE INERTIA
MATRIX.
Most dynamic analysis methods require positive-definite mass matrices. Inserting inertia data
into the wrong fields can result in nonpositive-definite systems, which are not physically
realistic for normal modeling practices. Reduce the size of the off-diagonal terms to provide a
positive-definite determinant.

UFM 4346
*** USER FATAL MESSAGE 4346, FREQUENCY RESPONSE SET, ID = *** IS UNDEFINED.
Define the set of frequencies to be used for the analysis.

UFM 4391
*** USER FATAL MESSAGE 4391, NONUNIQUE DAREA SET ***** HAS BEEN SPECIFIED
FOR LSEQ DEFINED VECTOR *****.
Each LSEQ Bulk Data entry must define a unique DAREA set specification.

Main Index
486

UFM 4392
*** USER FATAL MESSAGE 4392, CONTINUATION CARD ERRORS. EXPLANATIONS
FOLLOW LIST OF CARDS IN ERROR.
ERROR
NUMBER .1..2..3...
*** Input echo
EXPLANATION OF ERROR CODES ABOVE FOLLOWS...
1. FIELD 1 IS NOT UNIQUE.
2. MORE THAN ONE CARD HAS FIELD 10 WHICH IS THE SAME AS FIELD 1 OF THIS
CARD.
3. CARD IS AN ORPHAN (I.E., NO PARENT CARD EXISTS).
Continuation mnemonics in field 10 of a parent entry and field 1 of its continuation entry must
be unique. Each continuation entry must have a parent entry. Check all continuation
mnemonics.

UFM 4405
*** USER FATAL MESSAGE 4405, NO EIGENVECTORS COMPUTED FOR COMPONENT
MODE SYNTHESIS OR SYSTEM SOLUTION.

The eigenvectors computed in component mode synthesis (CMS) are used to approximate the
motion of the component. Some vectors must be present to perform this reduction. For the
system solution, an exit is taken if eigenvalues are requested but not eigenvectors. Reset the
“range of frequency” and/or the “number desired” on the EIGR or EIGRL entry. If the
superelement does not have any eigenvalues in the range of interest, remove the CMS request
for that particular superelement.
This message is also issued when the eigenvectors calculated with a transformation method such
as GIV do not pass internal orthogonality checks. This is indicative of a modeling error.
This message can also be issued if insufficient memory is available for the Lanczos method with
sparse decomposition.
This could occur with UFM 5401 and be related to UWM 5411.

UFM 4407
*** USER FATAL MESSAGE 4407, MR MATRIX HAS NULL DIAGONAL TERM.
The MR matrix contains the rigid-body mass matrix of the structure as measured at the
degrees-of-freedom listed on the SUPORT entry. If any of these degrees-of-freedom have null
mass, they result in invalid eigenvectors. Specify enough masses to define all rigid-body modes
and check attachments between the SUPORT and the rest of the model.

Main Index
 APPENDIX K 487
Diagnostic Messages for Dynamic Analysis

UIM 4415
*** USER INFORMATION MESSAGE 4415, THE FOLLOWING A-SET
DEGREES-OF-FREEDOM HAVE EITHER NULL MASSES OR NULL MASSES AND
STIFFNESSES.
If the listed degrees-of-freedom have null mass for the GIV, HOU, MHOU, or MGIV methods,
they are automatically omitted. For the INV method, the null degrees-of-freedom are
constrained. For direct frequency or direct transient response, the null degrees-of-freedom are
given a very small mass or stiffness. Inspect the listed degrees-of-freedom to ensure that masses
or stiffnesses are not left out inadvertently.

UFM 4416
*** USER FATAL MESSAGE 4416, NO DYNAMIC LOAD TABLE AVAILABLE.
A frequency response or transient response analysis was requested, but no dynamic load data is
available. Include dynamic load data (DLOAD, RLOADi, TLOADi) in model.

UFM 4417
*** USER FATAL MESSAGE 4417, NO TRANSIENT RESPONSE LIST AVAILABLE.
A transient response dynamic analysis was requested, but no transient response list is available.
Include a TSTEP entry in the Bulk Data Section.

UFM 4418
*** USER FATAL MESSAGE 4418, NO EIGENVALUE EXTRACTION DATA IS AVAILABLE.
A dynamic analysis was requested, but no eigenvalue extraction data was available. Include
eigenvalue extraction data EIGR or EIGRL in the Bulk Data.
Possible causes are

• The METHOD command in the Case Control Section but no EIGR or EIGRL entry in
the Bulk Data Section.
• The METHOD command in the Case Control Section, EIGRL entry in Bulk Data Section
but no RF3D83 (SOL 3 only).
• No correspondence between Set IDs on the METHOD command and the EIGR/EIGRL
entries.

Main Index
488

UWM 4420
*** USER WARNING MESSAGE 4420, THE FOLLOWING DEGREES-OF-FREEDOM ARE
POTENTIALLY SINGULAR.
During decomposition, the degrees-of-freedom listed had pivot ratios greater than MAXRATIO.
Verify that the degrees-of-freedom are not part of a mechanism and that elements do not have
excessive stiffness. In SOLs 61 and higher, this condition causes run termination.
PARAM,BAILOUT may be used to continue the run. See the MSC.Nastran Numerical Methods
User’s Guide.

UFM 4421
*** USER FATAL MESSAGE 4421, NO FREQUENCY RESPONSE LIST AVAILABLE.
A frequency response dynamic analysis was requested, but no frequency data is available.
Include frequency data (FREQ, FREQ1, FREQ2) in the Bulk Data.

UFM 4501
*** USER FATAL MESSAGE 4501, RLOADi CARD SELECTED IN TRANSIENT ANALYSIS.
USE TLOADi.
RLOADi entries are used in frequency response analysis. These entries have no meaning in
transient analysis. Replace RLOADi with TLOADi entries.

UWM 4561
*** USER WARNING MESSAGE 4561, INSUFFICIENT MEMORY FOR MODE
ORTHOGONALITY CHECKS.
The amount of memory needed for eigenvector orthogonalization is 1/2 · [number of
eigenvectors · (number of eigenvectors +1)] + 2 · BUFFSIZE + number of eigenvalues. If this
equation is not met, the modes are orthogonalized, but the checking function is not performed.
However, all outputs from the module are provided. If the check is desired, you should either
increase memory or decrease the number of eigenvectors to satisfy the above equation.

UFM 4562
*** USER FATAL MESSAGE 4562, TSTEP (TIME STEPS) DATA IS MISSING.
Transient analysis requires the time step data. Add a TSTEP Bulk Data entry, and select it with
a TSTEP Case Control command.

UWM 4582
*** USER WARNING MESSAGE 4582, LSEQ CARD SID = *** REFERS TO A NONEXISTENT
STATIC LOAD MATRIX COLUMN *** (NCOLS = ***).
The most likely cause occurs when changing an LSEQ entry on a restart without regenerating
and assembling the static load matrix.

Main Index
 APPENDIX K 489
Diagnostic Messages for Dynamic Analysis

UFM 4603
*** USER FATAL MESSAGE 4603, THE LSEQ SET ID **** IS NOT UNIQUE WITH RESPECT TO
OTHER STATIC LOAD IDS.
LSEQ set IDs must be unique with respect to all other static load set IDs.

UFM 4645
*** USER FATAL MESSAGE 4645, THE SHIFTED STIFFNESS MATRIX IS NOT POSITIVE
DEFINITE.
The matrix sum K + λ ⋅ M is decomposed by the Cholesky method at the start of the MGIV method
of eigensolution. This decomposition requires that the matrix be positive-definite. A condition
that prevents this is a massless mechanism (for example, a point mass on an offset with no
rotational stiffness).

UFM 4646
*** USER FATAL MESSAGE 4646, THE MASS MATRIX IS NOT POSITIVE DEFINITE, USING
THE GIV/HOU METHOD. USE MGIV/MHOU INSTEAD.
The reduced mass matrix has columns that are not linearly independent. Common causes are
rotation degrees-of-freedom whose only inertia terms result from point masses on offsets. Use
the MGIV or MHOU method instead since it does not require a positive definite mass matrix.

UFM 4647
*** USER FATAL MESSAGE 4647, INSUFFICIENT TIME TO COMPLETE *****.
CPU ESTIMATE = **** SEC
CPU REMAINING = **** SEC
I/O ESTIMATE = **** SEC
I/O REMAINING = **** SEC
CPU and I/O limits are supplied on the Executive Control statement TIME (in minutes). The
module where the terminated program is listed. If the time to completion appears reasonable,
you should increase the estimates on the TIME statement and resubmit the run. For large
models, an increase in the system memory request should also be considered.

UWM 4648
*** USER WARNING MESSAGE 4648, THE MODAL MASS MATRIX IS NOT POSITIVE
DEFINITE.
The modal mass matrix cannot be decomposed by the Cholesky algorithm after merging elastic
and free-body modes. (Cholesky decomposition is used to orthogonalize the eigenvectors with
respect to the mass matrix.) The causes include the input of negative masses and the calculation
of eigenvectors for eigenvalues approaching machine infinity. Inspect the model or ask for
fewer eigenvectors using the F2 option. When this condition occurs, the eigenvectors are not
orthogonalized or normalized. The second parameter of the READ or REIGL module is given a
Main Index
490

negative sign. This parameter is used in the solution sequences to branch to an error exit after
printing the real eigenvalue table. You may use a DMAP Alter to print these eigenvectors if the
cause of the problem is not apparent in the eigenvalues. The solution can be forced to
completion by changing the sign of this parameter. You should be aware that a poor-quality
solution is provided for this case. This poor solution may be useful for diagnosing the problem
but should not be used for other purposes.
A possible cause of this error is when large offsets (large relative to the element length) are used
for the BEAM element and coupled mass is selected.

UFM 4671
*** USER FATAL MESSAGE 4671, LOAD COMBINATION REQUESTED BUT LSEQ CARDS
DO NOT EXIST FOR SID = ****.
Check the LSEQ entries.

UFM 4683
*** USER FATAL MESSAGE 4683, MASS/STIFFNESS MATRIX NEEDED FOR EIGENVALUE
ANALYSIS.
The eigensolution module was given a purged (that is, nonexistent) mass or stiffness matrix.
Common causes include the deletion of mass density input on MATi entries, user restart errors
in the superelement solution sequences, or when the auto-selection feature of generalized
dynamic reduction (GDR) determines that there are no natural frequencies below FMAX. It
therefore sets the number of generalized coordinates to zero. This condition can be detected
from UIM 4181. An avoidance is to increase FMAX on the DYNRED Bulk Data entry if GDR is
used. Provide mass matrix generating data by any of several means including a mass density
entry on material entries, concentrated masses, and g-type DMIG entries.

UFM 4683 (cont.)

Possibly no mass matrix is defined. Check for the following:
• RHO entry on MATi
• NSM entry on element properties (i.e., PSHELL, PBAR)
• CONMi or CMASSi
MSC.Nastran needs at least one of the above to compute the mass matrix. Incorrect
cross-sectional properties may also lead to this error but typically show up as another error.

UWM 4698
*** USER WARNING MESSAGE 4698, STATISTICS FOR DECOMPOSITION OF MATRIX ****.
THE FOLLOWING DEGREES-OF-FREEDOM HAVE FACTOR DIAGONAL RATIOS
GREATER THAN ****, OR HAVE NEGATIVE TERMS ON THE FACTOR DIAGONAL.

Main Index
 APPENDIX K 491
Diagnostic Messages for Dynamic Analysis

During decomposition, the degrees-of-freedom listed have pivot ratios that are greater than
maxratio or are negative. Verify that the degrees-of-freedom are not part of a mechanism and
that elements do not have excessive stiffness. In SOLs 61 and higher this condition causes run
termination. PARAM,BAILOUT may be used to continue the run to obtain messages issued by
subsequent modules. See the MSC.Nastran Numerical Methods User’s Guide.

UIM 5010
*** USER INFORMATION MESSAGE 5010, STURM SEQUENCE DATA FOR EIGENVALUE
EXTRACTION. TRIAL EIGENVALUE = (real), CYCLES = (real), NUMBER OF EIGENVALUES
BELOW THIS VALUE = (integer).
This message is automatic output during eigenvalue extraction using the Lanczos and SINV
methods. This message can be used, along with the list of eigenvalues, to identify the modes
found. See the MSC.Nastran Numerical Methods User’s Guide.

UFM 5025
*** USER FATAL MESSAGE 5025, LAMA PURGED. DSTA MODULE TERMINATED.
The LAMA data block contains a list of natural frequencies and may be purged because no
eigenvalues were computed or the data block was not properly recovered on restart.

UIM 5218
*** USER INFORMATION MESSAGE 5218, EIGENVALUE APPROACHING INFINITY AT ****
TH MODE. EIGENVECTORS WILL NOT BE COMPUTED BEYOND THIS POINT.
The MGIV, MHOU, AGIV, and AHOU methods substitute a very large number for eigenvalues
that approach machine infinity. If eigenvectors are computed for these artificial values, they
may be numerical noise, or they may cause overflows. Eigenvector computation is halted at the
first machine infinity instead even if you requested eigenvectors in this range.

UIM 5222
*** USER INFORMATION MESSAGE 5222, COUPLED/UNCOUPLED SOLUTION
ALGORITHM USED.

The modal methods use uncoupled solution algorithms, if possible. The uncoupled algorithms
are considerably more economical than the coupled algorithms. Coupled algorithms are
required when any of the following effects are present: transfer functions, DMIG requests of the
p-type, element damping, and PARAM,G.

Consider the use of modal damping (TABDMP1 entry) to reduce the cost of your analysis in
modal solutions.

Main Index
492

UFM 5225
*** USER FATAL MESSAGE 5225, ATTEMPT TO OPERATE ON THE SINGULAR MATRIX
**** IN SUBROUTINE DCMP.
This message is preceded by the listing of the grid point ID and degrees-of-freedom for any null
columns.

UIM 5236
*** USER INFORMATION MESSAGE 5236, THE FREQUENCY RANGE HAS BEEN SPLIT
INTO **** SUBREGIONS.
The overall frequency range for eigenanalysis is split into several smaller ranges when using the
SINV option to calculate modes and frequencies.

UFM 5238
*** USER FATAL MESSAGE 5238, THE NUMBER OF ROOTS IN THE DEFINED FREQUENCY
RANGE IS GREATER THAN 600.
More than 600 roots are in the desired frequency range, which is greater than the maximum
allowed using SINV. Decrease the size of the frequency range.

UIM 5239
*** USER INFORMATION MESSAGE 5239, BISECTIONING IN THE ***-*** INTERVAL.
The frequency subregion encompassing eigenvalues xx-yy is cut in half in order to find the
remaining roots.

UIM 5240
*** USER INFORMATION MESSAGE 5240, THE BISECTION VALUE IS: ****
The selected value is midway between the lowest and highest frequencies in the frequency
subregion.

UIM 5241
*** USER INFORMATION MESSAGE 5241, MISSING ROOT(S) IN THE ***-*** INTERVAL.
The Sturm sequence check has indicated that roots are missing in the frequency range, and they
cannot be found by further bisectioning.
If the run terminates with missing roots, decrease the frequency range.

Main Index
 APPENDIX K 493
Diagnostic Messages for Dynamic Analysis

UIM 5242
*** USER INFORMATION MESSAGE 5242, THE ROOT FOUND IS NOT THE LOWEST ONE
ABOVE FMIN.
The Sturm sequence check indicates that at least one unfounded root exists between FMIN and
the lowest frequency root found. Set FMAX close to the lowest frequency found, so that the
lower roots can be found.

UIM 5274
*** USER INFORMATION MESSAGE 5274, THE ACTUAL TIME OF (****)
TRIDIAGONALIZATION IS: ****, THE ACTUAL TIME OF EIGENVALUE ITERATION IS: ****,
THE ACTUAL TIME OF EIGENVECTOR GENERATION IS: ****.
The time spent in the major operations of the real eigensolution module are output. Note that
the number of eigenvectors requested has a large effect on solution cost.

UFM 5288
*** USER FATAL MESSAGE 5288, NO ROOT EXISTS ABOVE FMIN.
This message occurs when the eigenproblem of finding any number of roots above FMIN is
expected, but the Sturm number indicates that there is no root above FMIN. Reduce FMIN in
order to attempt to find a root.

SFM 5299
*** SYSTEM FATAL MESSAGE 5299 (This text varies depending on the reason for termination;
see the description given below.)
1. Insufficient storage for the Lanczos method.
2. Factorization error on three consecutive shifts.
3. Stack overflow in the Lanczos method.
4. Unrecoverable termination from the Lanczos method.
5. Insufficient working storage.
6. Finite interval analysis error (see UIM 6361).
This message can also occur for models with two or more widely separated groups of repeated
roots. An avoidance is to search each group separately. This error may also be caused by a
massless mechanism, which can be confirmed by performing a static analysis.
See the MSC.Nastran Numerical Methods User’s Guide for more information.

UFM 5400
*** USER FATAL MESSAGE 5400, INCORRECT RELATIONSHIP BETWEEN FREQUENCY
LIMITS.
You have incorrectly specified V1 > V2. Check V1, V2 specified on the EIGRL Bulk Data entry.
Main Index
494

SFM 5401
*** SYSTEM FATAL MESSAGE 5401 (REIGL), LANCZOS METHOD IS UNABLE TO FIND
ALL EIGENVALUES IN RANGE. ACCEPTED EIGENVALUES AND ADDITIONAL ERROR
MESSAGES MAY BE LISTED ABOVE. USER ACTION: RERUN WITH ANOTHER METHOD
OR ANOTHER SETTING ON EIGRL ENTRY.
This message can be issued if insufficient memory is available for Lanczos with sparse
decomposition. It can also be issued if UFM 5299 occurs. See the MSC.Nastran Numerical
Methods User’s Guide.
This condition can be related to the occurrence of UWM 5411.

UWM 5402
*** USER WARNING MESSAGE 5402, ---THE PROBLEM HAS NO STIFFNESS MATRIX.
The problem requires a stiffness matrix. Verify that property entries are specified correctly.

UIM 5403
*** USER INFORMATION MESSAGE 5403, CPU TIME AT START OF LANCZOS ITERATION
****.
Since several Lanczos iterations may be executed during one application of the Lanczos method
(each shift is followed by at least one iteration), this information is given to measure the time
required for the individual iterations.

UWM 5404
*** USER WARNING MESSAGE 5404, NEGATIVE MODAL MASS TERM, IS ENCOUNTERED
DURING INVERSE ITERATION. PROCESS ABORTED.
The modal mass matrix should have unit diagonal terms (for mass normalization). Negative
terms may indicate negative eigenvalues. If these negative terms are computational zeroes
(rigid-body modes, for example), then the negative terms are acceptable. If the negative terms
are finite values, there may be a modeling problem.

UWM 5405
*** USER WARNING MESSAGE 5405, ERROR OCCURRED DURING ITERATION. ERROR
NUMBER IS : Y (SEE DESCRIPTION FOR VALUES OF Y AND USER ACTION.)

This message marks the breakdown of the inverse iteration process in the Lanczos method. See
the MSC.Nastran Numerical Methods User’s Guide for additional values and actions.

Main Index
 APPENDIX K 495
Diagnostic Messages for Dynamic Analysis

Y Value User Action

-11 File open error in interface; see GINO This error should not occur; report error
error message. to MSC.
-12 File open error in postprocessing; see This error should not occur; report error
GINO error message. to MSC.
-13 File read error; see GINO error message. This error should not occur; report error
to MSC.
-21 Insufficient space for blocksize = 1. Increase memory.
-22 Three consecutive factorizations failed Possible ill-conditioning; check model.
at a shift.
-23 Lanczos internal table overflow due to Specify smaller internal; may be
enormous number of shifts. necessary to have several runs.
-31 Internal error in Lanczos (REIGL) This error should not occur; report error
module. to MSC.
-32 No convergence in solving the Possible ill-conditioning; check model.
tridiagonal problem.
-33 Too many eigenvalues were found; Check the orthogonality of the
inconsistency between the roots found eigenvectors; if it is good, then ignore
and Sturm number. this warning.

UWM 5406
*** USER WARNING MESSAGE 5406 NO CONVERGENCE IN SOLVING THE
TRIDIAGONAL PROBLEM.
This message signals eigensolution problems in the Lanczos method. There is possible
ill-conditioning; check your model.

UWM 5407
*** USER WARNING MESSAGE 5407, INERTIA (STURM SEQUENCE) COUNT DISAGREES
WITH THE NUMBER OF MODES ACTUALLY COMPUTED IN AN (SUB) INTERVAL

This message shows a serious problem. Spurious modes were found in the Lanczos method.
Check the multiplicity of the roots given in the interval. See the MSC.Nastran Numerical Methods
User’s Guide.

UWM 5408
*** USER WARNING MESSAGE 5408, FACTORIZATION FAILED. SHIFT CHANGED TO
****.
No user action to be taken. This message occurs only for the Lanczos method.
Main Index
496

UWM 5411
*** USER WARNING MESSAGE 5411, NEGATIVE TERM ON DIAGONAL OF MASS MATRIX
(VIBRATION) OR STIFFNESS (BUCKLING), ROW ****, VALUE = ****
The message is given from the REIGL module which performs a necessary (but not sufficient)
check on the positive semi-definiteness of the indicated matrix. Look for evidence of negative
mass, such as minus signs on input. Negative terms on the factor of the indicated matrix must
be removed for correct answers.

Something has caused a negative term on the diagonal of the mass or stiffness matrix. Look for
explicitly defined negative mass and/or stiffness terms. Also, check the continuation entries on
the PBEAM entry. An incorrect entry for the SO field may lead to improper mass definition. For
example, if SO is set to NO at a particular X/XB location, the continuation entry for defining four
stress locations on the cross section (C, D, E, F) is not used. If SO is NO but the C, D, E, and F
points are entered in error, negative mass terms can result if either E1 or E2 entries are entered.
The offending DOF can be traced using the USET tables.
The Lanczos method gives wrong answers for indefinite matrices. The existence of negative
diagonal terms indicates a subclass of indefinite matrix.
See the MSC.Nastran Numerical Methods User’s Guide for more information.

UFM 5429
*** USER FATAL MESSAGE 5429, INSUFFICIENT TIME TO FINISH LANCZOS ITERATION.
There is not enough time remaining to complete the analysis based on an estimate of how long
the Lanczos iteration will take. Increase the value on the TIME statement in the Executive
Control Section and rerun. In order to ensure that the run can be competed, overestimate the
TIME value using estimates or actual times from a similar model.

UIM 5458
*** USER INFORMATION MESSAGE 5458, (****) METHOD IS SELECTED. or (****) METHOD
IS (****)
The exact text of this message depends on the METHOD field on the selected EIGR Bulk Data
entry. This message indicates the eigensolution status (all eigenvalues found, not all found, etc.).

UFM 6133
*** USER FATAL MESSAGE 6133 (DFMSDD), SINGULAR MATRIX IN SPARSE
DECOMPOSITION.
USER ACTION: CHECK MODEL.
This message is often followed by UFM 4645, UFM 4646, or UWM 4648.

Main Index
 APPENDIX K 497
Diagnostic Messages for Dynamic Analysis

UFM 6134
*** USER FATAL MESSAGE 6134 (DFMSDD), MATRIX IS NOT POSITIVE DEFINITE IN
SPARSE DECOMPOSITION.
USER ACTION: CHECK MODEL.
This message is often followed by UFM 4645, UFM 4646, or UWM 4648.

SFM 6135
*** SYSTEM FATAL MESSAGE 6135, ERROR IN READING SYMBOLIC FACTOR IN SPARSE
FBS.
This message may be issued if the FBS module is using a sparse method to solve factors which
are not decomposed by the sparse method. This message can also be caused by a compatibility
or database integrity problem.

UFM 6136
*** USER FATAL MESSAGE 6136 (****), INSUFFICIENT CORE FOR (SYMBOLIC/NUMERIC)
PHASE OF SPARSE DECOMPOSITION.
USER ACTION: INCREASE CORE BY **** WORDS.
USER INFORMATION:
!!! NOW REVERTING BACK TO ACTIVE COLUMN DECOMPOSITION UPON USER
REQUEST !!!

If this message is issued in the symbolic phase, the memory estimate is not necessarily
conservative and even more memory may be required (although this estimate is fairly accurate
for Version 68). Also, the memory increase required is only for the symbolic phase. It is not
unusual for the decomposition phase to require more memory than the symbolic phase. To
increase the chances for a successful run, increase the memory even more than the amount
indicated in this message. After the run is complete, determine the amount of memory actually
used and use this as a guideline for similar runs in the future.
The user information message is written if SYSTEM(166) = 1 (that is, if there is not enough
memory for sparse decomposition, and you should switch to regular decomposition).

UWM 6137
*** USER WARNING MESSAGE 6137 (DFMSDD), INPUT MATRIX IS RANK DEFICIENT,
RANK = ****.
USER ACTION: CHECK MODEL.
One of your matrices is singular. See the MSC.Nastran Numerical Methods User’s Guide for a
discussion of singularity.

Main Index
498

UFM 6138
*** USER FATAL MESSAGE 6138 (DFMSB), INSUFFICIENT CORE FOR SPARSE FBS.
USER ACTION: INCREASE CORE BY **** WORDS.
See UFM 6136.

UIM 6214
*** USER INFORMATION MESSAGE 6214, FEWER THAN REQUESTED VECTORS
CALCULATED, DUE TO INSUFFICIENT TIME.
This information message occurs in the READ module when there is insufficient time to
compute eigenvectors. Resubmit the job with an increased time limit (TIME).

UWM 6243
*** USER WARNING MESSAGE 6243 (READ) ---- THE DEGREE OF FREEDOM (D.O.F)
REQUESTED FOR POINT NORMALIZATION HAS NOT BEEN SPECIFIED ON THE EIGR OR
EIGB ENTRY.
USER INFORMATION: THE D.O.F PRECEDING THE REQUESTED D.O.F. IN THE
INTERNAL SEQUENCE LIST WILL BE USED
The point requested was not in the a-set, so another point was chosen.

UIM 6361
*** USER INFORMATION MESSAGE 6361-LANCZOS MODULE DIAGNOSTICS
This message prints various levels of diagnostics for the Lanczos method. The amount of print
depends on the message level set on the EIGRL entry. See the MSC.Nastran Numerical Methods
User’s Guide for more information.

UIM 6480
*** USER INFORMATION MESSAGE 6480 (REIGLA) ---
EXTERNAL IDENTIFICATION TABLE FOR DECOMPOSITION MESSAGES FOR MATRIX
****
ROW NUMBER **** = GRID ID **** + COMPONENT ****
This message is output from the REIGL module when using sparse decomposition to convert the
internal (row number oriented) diagnostic messages to external (grid and component) form.

Main Index
 MSC.Nastran Basic Dynamics User’s GuideMSC.Nastran Basic Dynamics User’s Guide

APPENDIX
References and Bibliography
L
■ Overview
■ General References
■ Bibliography

Main Index
500

L.1 Overview
This appendix includes references of interest in the field of dynamic analysis. Two categories
are included. The first category, General References, lists books that cover the general range of
structural dynamic analysis. The second category, Bibliography, is an excerpt from the dynamic
analysis section of the MSC.Nastran Bibliography, Second Edition.

Main Index
 APPENDIX L 501
References and Bibliography

L.2 General References

1. Paz, M., Structural Dynamics: Theory and Computation, Van Nostrand Reinhold,
New York, N.Y., 1985.
2. Bathe, K. J. and Wilson, E. L., Numerical Methods in Finite Element Analysis,
Prentice-Hall, Englewood Cliffs, N.J., 1976.
3. Harris, C. M. and Crede, C. E., Shock and Vibration Handbook, McGraw-Hill, New York,
N.Y., 1976.
4. Clough, R. W. and Penzien, J., Dynamics of Structures, McGraw-Hill, New York, N.Y.,
1975.
5. Timoshenko, S., Young, D. H., and Weaver Jr., W., Vibration Problems in Engineering,
John Wiley and Sons, New York, N.Y., 1974.
6. Hurty, W. C. and Rubinstein, M. F., Dynamics of Structures, Prentice-Hall, Englewood
Cliffs, N.J., 1964.

Main Index
502

L.3 Bibliography
DYNAMICS – GENERAL
Abdallah, Ayman A.; Barnett, Alan R.; Widrick, Timothy W.; Manella, Richard T.; Miller, Robert
P. Stiffness-Generated Rigid-Body Mode Shapes for Lanczos Eigensolution with
Support DOF Via a MSC/NASTRAN DMAP Alter, MSC 1994 World Users’ Conf. Proc.,
Paper No. 10, June, 1994.

Anderson, William J.; Kim, Ki-Ook; Zhi, Bingchen; Bernitsas, Michael M.; Hoff, Curtis; Cho,
Kyu-Nam. Nonlinear Perturbation Methods in Dynamic Redesign, MSC/NASTRAN
Users’ Conf. Proc., Paper No. 16, March, 1983.

Barber, Pam; Arden, Kevin. Dynamic Design Analysis Method (DDAM) Using
MSC/NASTRAN, MSC 1994 World Users’ Conf. Proc., Paper No. 31, June, 1994.

Bedrossian, Herand; Veikos, Nicholas. Rotor-Disk System Gyroscopic Effect in

MSC/NASTRAN Dynamic Solutions, MSC/NASTRAN Users’ Conf. Proc., Paper No. 12,
March, 1982.

Bernstein, Murray; Mason, Philip W.; Zalesak, Joseph; Gregory, David J.; Levy, Alvin.
NASTRAN Analysis of the 1/8-Scale Space Shuttle Dynamic Model, NASTRAN: Users’
Exper., pp. 169-242, September, 1973, (NASA TM X-2893).

Berthelon, T.; Capitaine, A. Improvements for Interpretation of Structural Dynamics

Calculation Using Effective Parameters for Substructures, Proc. of the 18th MSC Eur.
Users’ Conf., Paper No. 9, June, 1991.

Birkholz, E. Dynamic Investigation of Automobile Body Parts, Proc. of the 15th

MSC/NASTRAN Eur. Users’ Conf., October, 1988.

Bishop, N. W. M.; Lack, L. W.; Li, T.; Kerr, S. C. Analytical Fatigue Life Assessment of Vibration
Induced Fatigue Damage, MSC 1995 World Users’ Conf. Proc., Paper No. 18, May, 1995.

Blakely, Ken; Howard, G. E.; Walton, W. B.; Johnson, B. A.; Chitty, D. E. Pipe Damping Studies
and Nonlinear Pipe Benchmarks from Snapback Tests at the Heissdampfreaktor,
NUREG/CR-3180, March, 1983.

Blakely, Ken. Dynamic Analysis: Application and Modeling Considerations, J. of Engineering

Computing and Applications, Fall, 1987.

Bramante, A.; Paolozzi, A; Peroni, I. Effective Mass Sensitivity: A DMAP Procedure, MSC 1995
World Users’ Conf. Proc., Paper No. 39, May, 1995.

Brutti, C.; Conte, M.; Linari, M. Reduction of Dynamic Environment to Equivalent Static Loads
by a NASTRAN DMAP Procedure, MSC 1995 European Users’ Conf. Proc., Italian
Session, September, 1995.

Butler, Thomas G. Dynamic Structural Responses to Rigid Base Acceleration, Proc. of the Conf.
on Finite Element Methods and Technology, Paper No. 8, March, 1981.
Main Index
 APPENDIX L 503
References and Bibliography

Butler, Thomas G. Telescoping Robot Arms, MSC/NASTRAN Users’ Conf. Proc., Paper No. 10,
March, 1984.

Butler, T. G. Experience with Free Bodies, Thirteenth NASTRAN Users’ Colloq., pp. 378-388,
May, 1985, (NASA CP-2373).

Butler, Thomas G. Mass Modeling for Bars, Fifteenth NASTRAN Users’ Colloq., pp. 136-165,
August, 1987, (NASA CP-2481).

Butler, T. G. Coupled Mass for Prismatical Bars, Sixteenth NASTRAN Users’ Colloq., pp. 44-63,
April, 1988, (NASA CP-2505).

Caldwell, Steve P.; Wang, B. P. An Improved Approximate Method for Computing Eigenvector
Derivatives in MSC/NASTRAN, The MSC 1992 World Users’ Conf. Proc., Vol. I, Paper
No. 22, May, 1992.

Case, William R. Dynamic Substructure Analysis of the International Ultraviolet Explorer

(IUE) Spacecraft, NASTRAN: Users’ Exper., pp. 221-248, September, 1975, (NASA TM
X-3278).

Chang, H. T.; Cao, Tim; Hua, Tuyen. SSF Flexible Multi-Body Control/Structure Interaction
Simulation, The MSC 1993 World Users’ Conf. Proc., Paper No. 15, May, 1993.

Chang, W. M.; Lai, J. S.; Chyuan, S. W.; Application of the MSC/NASTRAN Design
Optimization Capability to Identify Joint Dynamic Properties of Structure, The Sixth
Annual MSC Taiwan Users’ Conf. Proc., Paper No. 1, November, 1994.

Chargin, M.; Miura, H.; Clifford, Gregory A. Dynamic Response Optimization Using
MSC/NASTRAN, The MSC 1987 World Users Conf. Proc., Vol. I, Paper No. 14, March,
1987.

Chen, J. T.; Chyuan, S. W.; You, D. W.; Wong, H. T. A New Method for Determining the Modal
Participation Factor in Support Motion Problems Using MSC/NASTRAN, The Seventh
Annual MSC/NASTRAN Users’ Conf. Proc., Taiwan, 1995.

Chen, J. T.; Wong, H. T. Applications of Modal Reaction Method in Support Motion Problems,
Techniques in Civil Engineering, Vol. 4, pp 17 - 30, March, 1996, in Chinese.

Chen, J. T.; Hong, H. K.; Chyuan, S. W.; Yeh, C. S. A Note on the Application of Large Mass and
Large Stiffness Techniques for Multi-Support Motion, The Fifth Annual MSC Taiwan
Users’ Conf. Proc., November, 1993.

Chen, Yohchia. Improved Free-Field Analysis for Dynamic Medium-Structure Interaction

Problems, The MSC 1992 World Users’ Conf. Proc., Vol. I, Paper No. 13, May, 1992.

Chen, Yohchia. Dynamic Response of Reinforced Concrete Box-Type Structures, The MSC 1992
World Users’ Conf. Proc., Vol. I, Paper No. 24, May, 1992.

Chiu, Chi-Wai. Spacecraft Dynamics During Solar Array Panel Deployment Motion, The Fifth
Annual MSC Taiwan Users’ Conf. Proc., November, 1993.

Main Index
504

Chung, Y. T.; Kahre, L. L. A General Procedure for Finite Element Model Check and Model
Identification, MSC 1995 World Users’ Conf. Proc., Paper No. 38, May, 1995.

Cicia, C. Static, Thermal and Dynamic Analysis of the Liquid Argon Cryostat for the ICARUS
Experiment, Proc. of the 15th MSC/NASTRAN Eur. Users’ Conf., October, 1988.

Cifuentes, Arturo O. Dynamic Analysis of Railway Bridges Using MSC/NASTRAN, The MSC
1988 World Users Conf. Proc., Vol. II, Paper No. 44, March, 1988.

Cifuentes, A. O. Dynamic Response of a Beam Excited by a Moving Mass, Finite Elements in

Analysis and Design, Vol. 5, pp. 237-246, 1989.

Citerley, R. L.; Woytowitz, P. J. Ritz Procedure for COSMIC/ NASTRAN, Thirteenth NASTRAN
Users’ Colloq., pp. 225-233, May, 1985, (NASA CP-2373).

Ciuti, Gianluca. Avionic Equipment Dynamic Analysis, MSC 1995 European Users’ Conf. Proc.,
Italian Session, September, 1995.

Coates, Dr. Tim; Matthews, Peter. Transient Response in Dynamic and Thermal Behaviour, The
Second Australasian MSC Users Conf. Proc., Paper No. 4, November, 1988.

Coppolino, Robert N.; Bella, David F. Employment of MSC/STI-VAMP for Dynamic Response
Post-Processing, The MSC 1987 World Users Conf. Proc., Vol. I, Paper No. 12, March,
1987.

Corder, P. R.; Persh, R. Castigliano and Symbolic Programming in Finite Element Analysis,
Proceedings of the 16th Annual Energy - Sources Technology Conference and Exhibition,
Houston, 1993.

Coyette, J. P.; Wijker, J. J. The Combined Use of MSC/NASTRAN and Sysnoise for Evaluating
the Dynamic Behavior of Solar Array Panels, Proc. of the 20th MSC European Users’
Conf., Paper No. 16, September, 1993.

Curti, G.; Chiandussi, G.; Scarpa, F. Calculation of Eigenvalue Derivatives of

Acousto-Structural Systems with a Numerical Comparison, MSC 23rd European Users’
Conf. Proc., Italian Session, September, 1996.

Defosse, H.; Sergent, A. Vibro-Acoustic Modal Response Analysis of Aerospace Structures,

Proc. of the MSC/NASTRAN Eur. Users’ Conf., April, 1985.

Deloo, Ph.; Dunne, L.; Klein, M. Alter DMAPS for the Generation, Assembly and Recovery of
Craig-Bampton Models in Dynamic Analyses, Actes de la 2ème Confèrence Française
Utilisateurs des Logiciels MSC, Toulouse, France, September, 1995.

Denver, Richard E.; Menichello, Joseph M. Alternate Approaches to Vibration and Shock
Analysis Using NASTRAN, Sixth NASTRAN Users’ Colloq., pp. 199-212, October, 1977,
(NASA CP-2018).

Detroux, P.; Geraets, L. H. Instability at Restart or Change of Time Step with NASTRAN in the
Presence of Nonlinear Loads, Proc. of the MSC/NASTRAN Eur. Users’ Conf., June, 1983.

Main Index
 APPENDIX L 505
References and Bibliography

Deuermeyer, D. W.; Clifford, G. A.; Petesch, D. J. Traditional Finite Element Analysis:

Opportunities for Parallelism?, Computing Systems in Engineering, Vol. 2, No. 2-3,
pp. 157-165, 1991.

Dirschmid, Dr. W.; Nolte, Dr. F.; Dunne, L. W. Application of an FRF-Based Update Method to
the Model Parameter Tuning of an Hydraulic Engine Mounting, Proc. of the 18th MSC
Eur. Users’ Conf., Paper No. 10, June, 1991.

Drago, Raymond J.; Margasahayam, Ravi N. Resonant Response of Helicopter Gears Using 3-D
Finite Element Analysis, The MSC 1988 World Users Conf. Proc., Vol. I, Paper No. 20,
March, 1988.

Elchuri, V.; Smith, G. C. C.; Gallo, A. Michael. An Alternative Method of Analysis for Base
Accelerated Dynamic Response in NASTRAN, Eleventh NASTRAN Users’ Colloq.,
pp. 89-112, May, 1983.

Everstine, Gordon C.; Schroeder, Erwin A. The Dynamic Analysis of Submerged Structures,
NASTRAN: Users’ Exper., pp. 419-430, September, 1975, (NASA TM X-3278).

Everstine, Gordon C. Structural Analogies for Scalar Field Problems, Int. J. for Numerical
Methods in Engineering, Vol. 17, No. 3, pp. 471-476, March, 1981.

Everstine, G. C. Dynamic Analysis of Fluid-Filled Piping Systems Using Finite Element

Techniques, J. of Pressure Vessel Technology, Vol. 108, pp. 57-61, February, 1986.

Flanigan, Christopher C. Accurate and Efficient Mode Acceleration Data Recovery for
Superelement Models, The MSC 1988 World Users Conf. Proc., Vol. I, Paper No. 38,
March, 1988.

Flanigan, Christopher C.; Manella, Richard T. Advanced Coupled Loads Analysis Using
MSC/NASTRAN, The MSC 1991 World Users’ Conf. Proc., Vol. I, Paper No. 14, March,
1991.

Fox, Gary L. Solution of Enforced Boundary Motion in Direct Transient and Harmonic
Problems, Ninth NASTRAN Users’ Colloq., pp. 96-105, October, 1980, (NASA CP-2151).

Geyer, A.; Schweiger, W. Multiple Support Excitation for NASTRAN Piping Analysis, Proc. of
the MSC/NASTRAN Eur. Users’ Conf., April, 1982.

Geyer, A.; Schweiger, W. Aeroelastic and Stress Analysis of the CHIWEC Chinese Wind Energy
Converter Using MSC/NASTRAN, Proc. of the MSC/NASTRAN Eur. Users’ Conf., April,
1985.

Ghofranian, S.; Dimmagio, O. D. Space Station Dynamic Analysis with Active Control Systems
Using MSC/NASTRAN, The MSC 1988 World Users Conf. Proc., Vol. I, Paper No. 17,
March, 1988.

Gibson, Warren C. Experiences with Optimization Using ASD/NASOPT and MSC/NASTRAN

for Structural Dynamics, The MSC 1987 World Users Conf. Proc., Vol. I, Paper No. 13,
March, 1987.

Main Index
506

Gibson, Warren C.; Austin, Eric. Analysis and Design of Damped Structures Using
MSC/NASTRAN, The MSC 1992 World Users’ Conf. Proc., Vol. I, Paper No. 25, May,
1992.

Gielen, L.; Brughmans, M.; Petellat, C. A Stepwise Approach for Fatigue Evaluation of Engine
Accessories Prior to Prototyping Using Hybrid Modelling Technology, MSC 1996 World
Users’ Conf. Proc., Vol. III, Paper No. 29, June, 1996.

Go, James Chi-Dian. Structural Dynamic and Thermal Stress Analysis of Nuclear Reactor
Vessel Support System, NASTRAN: Users’ Exper., pp. 465-476, September, 1972, (NASA
TM X-2637).

Grasso, A.; Tomaselli, L. Whirling Speed Analysis of Multispool Systems, Proc. of the
MSC/NASTRAN Eur. Users’ Conf., May, 1984.

Grimes, Roger G.; Lewis, John G.; Simon, Horst D.; Komzsik, Louis; Scott, David S. Shifted Block
Lanczos Algorithm in MSC/NASTRAN, MSC/NASTRAN Users’ Conf. Proc., Paper
No. 12, March, 1985.

Harn, Wen-Ren; Lin, Shyang-Kuang; Chen, Jeng-Tzong. Localization of Dynamic Model

Modification Based on Constrained Minimization Method, The 2nd Annual MSC
Taiwan Users Conf., Paper No. 14, October, 1990.

Herting, David N.; Bella, David F.; Kimbrough, Patty A. Finite Element Simulation of Coupled
Automobile Engine Dynamics, The MSC 1987 World Users Conf. Proc., Vol. I, Paper
No. 10, March, 1987.

High, Gerald D. An Iterative Method for Eigenvector Derivatives, The MSC 1990 World Users
Conf. Proc., Vol. I, Paper No. 17, March, 1990.

Hill, R. G. Transient Analysis of an IVHM Grapple Impact Test, NASTRAN: Users’ Exper.,
pp. 161-178, September, 1972, (NASA TM X-2637).

Howells, R. W.; Sciarra, J. J. Finite Element Analysis Using NASTRAN Applied to Helicopter
Transmission Vibration/Noise Reduction, NASTRAN: Users’ Exper., pp. 321-340,
September, 1975, (NASA TM X-3278).

Howlett, James T. Applications of NASTRAN to Coupled Structural and Hydrodynamic

Responses in Aircraft Hydraulic Systems, NASTRAN: Users’ Exper., pp. 407-420,
September, 1971, (NASA TM X-2378).

Huang, S. L.; Rubin, H. Static and Dynamic Analysis, F-14A Boron Horizontal Stabilizer,
NASTRAN: Users’ Exper., pp. 251-264, September, 1971, (NASA TM X-2378).

Hurwitz, Myles M. New Large Deflection Analysis for NASTRAN, Sixth NASTRAN Users’
Colloq., pp. 235-256, October, 1977, (NASA CP-2018).

Hussain, M. A.; Pu, S. L.; Lorensen, W. E. Singular Plastic Element: NASTRAN Implementation
and Application, Sixth NASTRAN Users’ Colloq., pp. 257-274, October, 1977, (NASA
CP-2018).

Main Index
 APPENDIX L 507
References and Bibliography

Ishikawa, Masanori; Iwahara, Mitsuo; Nagamatsu, Akio. Dynamic Optimization Applied to

Engine Structure, The MSC 1990 World Users Conf. Proc., Vol. I, Paper No. 31, March,
1990.

Iwahara, Mitsuo. Dynamic Optimization Using Quasi Least Square Method, The Fifth
MSC/NASTRAN User’s Conf. in Japan, October, 1987, in Japanese.

Jakovich, John; Van Benschoten, John. SDRC SUPERTAB Interactive Graphics as a Front-End
to MSC/NASTRAN Dynamic Analysis, Proc. of the MSC/NASTRAN Users’ Conf.,
March, 1979.

Jones, Gary K. The Response of Shells to Distributed Random Loads Using NASTRAN,
NASTRAN: Users’ Exper., pp. 393-406, September, 1971, (NASA TM X-2378).

Kalinowski, Anthony J. Steady Solutions to Dynamically Loaded Periodic Structures, Eighth

NASTRAN Users’ Colloq., pp. 131-164, October, 1979, (NASA CP-2131).

Kalinowski, A. J. Solution Sensitivity and Accuracy Study of NASTRAN for Large Dynamic
Problems Involving Structural Damping, Ninth NASTRAN Users’ Colloq., pp. 49-62,
October, 1980, (NASA CP-2151).

Kasai, Manabu. Real Eigenvalue Analysis by Modal Synthesis Method Taking Differential
Stiffness into Account, The First MSC/NASTRAN User’s Conf. in Japan, October, 1983,
in Japanese.

Kasai, Manabu. Recovery Method for Components by DMAP of Constrained Modal Type, The
Fifth MSC/NASTRAN User’s Conf. in Japan, October, 1987, in Japanese.

Kasai, Manabu. DMAP Program for Modal Mass and Momentum, The Sixth MSC/NASTRAN
User’s Conf. in Japan, October, 1988, in Japanese.

Kienholz, Dave K.; Johnson, Conor D.; Parekh, Jatin C. Design Methods for Viscoelastically
Damped Sandwich Plates, AIAA/ASME/ASCE/AHS 24th Structures, Structural
Dynamics and Materials Conf., Part 2, pp. 334-343, May, 1983.

Lambert, Nancy; Tucchio, Michael. Ring Element Dynamic Stresses, Ninth NASTRAN Users’
Colloq., pp. 63-78, October, 1980, (NASA CP-2151).

Lee, Jyh-Chiang. Investigation for the Large Stiffness Method, The Fifth Annual MSC Taiwan
Users’ Conf. Proc., November, 1993.

Lee, Ting-Yuan; Lee, Jyh-Chiang. Modal Analysis and Structural Modification for a Harpoon
Launcher, The Fifth Annual MSC Taiwan Users’ Conf. Proc., November, 1993.

LeMaster, R. A.; Runyan, R. B. Dynamic Certification of a Thrust-Measuring System for Large

Solid Rocket Motors, Eleventh NASTRAN Users’ Colloq., pp. 207-225, May, 1983.

Lewis, J.; Komzsik, L. Symmetric Generalized Eigenproblems in Structural Engineering, SIAM

Conf. on Applied Numerical Analysis, 1985.

Lin, Chih-Kai; Harn, Wen-Ren; Lin, Shyang-Kuang. The Dynamic Response of Bridge Due to
Main Index
Passing of Vehicle, The 2nd Annual MSC Taiwan Users Conf., Paper No. 6, October, 1990.
508

Lin, S. L.; Yang, T. W.; Chen, J. T. MSC/NASTRAN Application in Inertia Relief, The 1st MSC
Taiwan Users’ Conf., Paper No. 13, October, 1989, in Chinese.

Lin, Shan. Time-Dependent Restrained Boundary Condition Simulation, The MSC 1988 World
Users Conf. Proc., Vol. I, Paper No. 9, March, 1988.

Lu, Ming-Ying; Yang, Joe-Ming. Analysis of Static and Dynamic Responses on Ship Structures
Under Wave Loadings, The Sixth Annual MSC Taiwan Users’ Conf. Proc., Paper No. 14,
November, 1994.

Magari, P. J.; Shultz, L. A.; Murthy, V. R. Dynamics of Helicopter Rotor Blades, Computers and
Structures, Vol. 29, No. 5, pp. 763-776, 1988.

Malcolm, D. J. Dynamic Response of a Darrieus Rotor Wind Turbine Subject to Turbulent Flow,
Engineering Structures, Vol. 10, No. 2, pp. 125-134, April, 1988.

Maritan, M.; Micelli, D. Dynamic Behaviour of a High-Speed Crankshaft, MSC 23rd European
Users’ Conf. Proc., Italian Session, September, 1996.

Masters, Steven G. Plant Troubleshooting with MSC/NASTRAN, Proc. of the Conf. on Finite
Element Methods and Technology, Paper No. 12, March, 1981.

Mastrorocco, David T. Predicting Dynamic Environments for Space Structure Appendages, The
MSC 1992 World Users’ Conf. Proc., Vol. II, Paper No. 56, May, 1992.

Mayer, Lee S.; Zeischka, Johann; Scherens, Marc; Maessen, Frank. Analysis of Flexible Rotating
Crankshaft with Flexible Engine Block Using MSC/NASTRAN and DADS, MSC 1995
World Users’ Conf. Proc., Paper No. 35, May, 1995.

McLaughlin, A. Finite Element Dynamic Analysis of Production Aircraft, 4th Eur. Rotorcraft
and Powered Lift Aircraft Forum, Assoc. Ital di Aeronaut ed Astronaut, pp. 20.1-20.7,
September, 1978.

Melli, R.; Rispoli, F.; Sciubba, E.; Tavani, F. Structural and Thermal Analysis of Avionic
Instruments for an Advanced Concept Helicopter, Proc. of the 15th MSC/NASTRAN Eur.
Users’ Conf., October, 1988.

Mikami, Kouichi. Dynamic Stress Analysis System for Ship’s Hull Structure Under Wave
Loads, The Second MSC/NASTRAN User’s Conf. in Japan, October, 1984, in Japanese.

Moharir, M. M. NASTRAN Nonlinear Capabilities in Dynamic Solutions, MSC/NASTRAN

Users’ Conf. Proc., Paper No. 9, March, 1985.

Moore, Gregory J.; Nagendra, Gopal K. Dynamic Response Sensitivities in MSC/NASTRAN,

The MSC 1991 World Users’ Conf. Proc., Vol. I, Paper No. 4, March, 1991.

Mulcahy, T. M.; Turula, P.; Chung, H.; Jendrzejczyk, A. Analytical and Experimental Study of
Two Concentric Cylinders Coupled by a Fluid Gap, NASTRAN: Users’ Exper.,
pp. 249-258, September, 1975, (NASA TM X-3278).

Main Index
 APPENDIX L 509
References and Bibliography

Murthy, P. L. N.; Chamis, C. C. Dynamic Stress Analysis of Smooth and Notched Fiber
Composite Flexural Specimens, National Aeronautics and Space Administration, April,
1984, (NASA TM-83694).

Murthy, P. L. N.; Chamis, C. C. Dynamic Stress Analysis of Smooth and Notched Fiber
Composite Flexural Specimens, Composite Materials: Testing and Design (Seventh
Conf.), ASTM, pp. 368-391, 1986, (ASTM STP 893).

Neal, M. Vibration Analysis of a Printed Wiring Board Assembly, Proc. of the MSC/NASTRAN
Eur. Users’ Conf., May, 1984.

Nefske, D. J.; Sung, S. H. Power Flow Finite Element Analysis of Dynamic Systems: Basic
Theory and Application to Beams, American Soc. of Mechanical Engineers, Noise Control
and Acoustics Division, Vol. 3, pp. 47-54, December, 1987.

Nowak, Bill. The Analysis of Structural Dynamic Effects on Image Motion in Laser Printers
Using MSC/NASTRAN, The MSC 1988 World Users Conf. Proc., Vol. I, Paper No. 10,
March, 1988.

Nowak, Bill. Structural Dynamics Analysis of Laser Printers, Sound and Vibration, Vol. 23,
No. 1, pp. 22-26, January, 1989.

Nowak, William J. Dynamic Analysis of Optical Scan Systems Using MSC/NASTRAN, Proc. of
the Conf. on Finite Element Methods and Technology, Paper No. 10, March, 1981.

Nowak, William; James, Courtney. Dynamic Modeling and Analysis of Spinning Polygon
Assemblies Using MSC/NASTRAN, The MSC 1993 World Users’ Conf. Proc., Paper
No. 66, May, 1993.

Oei, T. H.; Broerse, G. Reduction of Forced Vibration Levels on Ro-Ro Car Ferry-Type Ships by
Means of Minor Changes of the Inner Aft-Body Construction, Proc. of the
MSC/NASTRAN Eur. Users’ Conf., April, 1982.

Ojalvo, I. U. Extensions of MSC/NASTRAN to Solve Flexible Rotor Problems, MSC/NASTRAN

Users’ Conf. Proc., Paper No. 13, March, 1982.

Palmieri, F. Nonlinear Dynamic Analysis of STS Main Engine Heat Exchanger, Proc. of the 15th
MSC/NASTRAN Eur. Users’ Conf., October, 1988.

Palmieri, F. W. Analyzing Deployment of Spacecraft Appendages Using MSC/NASTRAN, The

MSC 1990 World Users Conf. Proc., Vol. I, Paper No. 6, March, 1990.

Pamidi, P. R.; Brown, W. K. On Eigenvectors of Multiple Eigenvalues Obtained in NASTRAN,

NASTRAN: Users’ Exper., pp. 285-300, September, 1975, (NASA TM X-3278).

Paolozzi, A. Structural Dynamics Modification with MSC/NASTRAN, Proc. of the 19th MSC
European Users’ Conf., Paper No. 14, September, 1992.

Parthasarathy, Alwar. Force-Sum Method for Dynamic Stresses in MSC/NASTRAN Aeroelastic

Analysis, The MSC 1991 World Users’ Conf. Proc., Vol. I, Paper No. 8, March, 1991.

Main Index
510

Parthasarathy, Alwar; Elzeki, Mohamed; Abramovici, Vivianne. PSDTOOL-A DMAP

Enhancement to Harmonic/Random Response Analysis in MSC/NASTRAN, The MSC
1993 World Users’ Conf. Proc., Paper No. 36, May, 1993.

Patel, Jayant S.; Seltzer, S. M. Complex Eigenvalue Solution to a Spinning Skylab Problem,
NASTRAN: Users’ Exper., pp. 439-450, September, 1971, (NASA TM X-2378).

Patel, Jayant S.; Seltzer, S. M. Complex Eigenvalue Analysis of Rotating Structures, NASTRAN:
Users’ Exper., pp. 197-234, September, 1972, (NASA TM X-2637).

Patel, Kirit V. Stress Analysis of Hybrid Pins in a Warped Printed Wiring Board Using
MSC/NASTRAN, MSC 1995 World Users’ Conf. Proc., Paper No. 20, May, 1995.

Paxson, Ernest B., Jr. Simulation of Small Structures- Optics-Controls Systems with
MSC/NASTRAN, The 1989 MSC World Users Conf. Proc., Vol. II, Paper No. 39, March,
1989.

Pinnament, Murthy. Mode Acceleration Data Recovery in MSC/NASTRAN Dynamic Analysis

with Generalized Dynamic Reduction, MSC/NASTRAN Users’ Conf. Proc., Paper
No. 24, March, 1985.

Raney, John P.; Kaszubowski, M.; Ayers, J. Kirk. Analysis of Space Station Dynamics Using
MSC/NASTRAN, The MSC 1987 World Users Conf. Proc., Vol. I, Paper No. 11, March,
1987.

Reyer, H. A Crash-Down Calculated with NASTRAN, Proc. of the MSC/NASTRAN Eur. Users’
Conf., April, 1982.

Rose, Ted L. Using Superelements to Identify the Dynamic Properties of a Structure, The MSC
1988 World Users Conf. Proc., Vol. I, Paper No. 41, March, 1988.

Rose, Ted L. Creation of and Use of ‘Craig-Bampton’ Models Using MSC/NASTRAN, The MSC
1990 World Users Conf. Proc., Vol. II, Paper No. 51, March, 1990.

Rose, Ted. Using Residual Vectors in MSC/NASTRAN Dynamic Analysis to Improve Accuracy,
The MSC 1991 World Users’ Conf. Proc., Vol. I, Paper No. 12, March, 1991.

Rose, Ted. DMAP Alters to Apply Modal Damping and Obtain Dynamic Loading Output for
Superelements, The MSC 1993 World Users’ Conf. Proc., Paper No. 24, May, 1993.

Rose, Ted.; McNamee, Martin. A DMAP Alter to Allow Amplitude-Dependent Modal Damping
in a Transient Solution, MSC 1996 World Users’ Conf. Proc., Vol. V, Paper No. 50, June,
1996.

Ross, Robert W. Prediction and Elimination of Resonance in Structural Steel Frames, The MSC
1988 World Users Conf. Proc., Vol. II, Paper No. 45, March, 1988.

Russo, A.; Mocchetti, R. Dynamic Analysis of Loaded Structures in the Helicopter Field, Proc.
of the MSC/NASTRAN Eur. Users’ Conf., May, 1984.

Main Index
 APPENDIX L 511
References and Bibliography

Salus, W. L.; Jones, R. E.; Ice, M. W. Dynamic Analysis of a Long Span, Cable-Stayed Freeway
Bridge Using NASTRAN, NASTRAN: Users’ Exper., pp. 143-168, September, 1973,
(NASA TM X-2893).

Sauer, G.; Wolf, M. Gyroscopic Effects in the Dynamic Response of Rotating Structures, Proc.
of the MSC/NASTRAN Eur. Users’ Conf., Paper No. 11, May, 1986.

Schips, C. Aero-Engine Turbine Dynamic Analysis, Proc. of the 18th MSC Eur. Users’ Conf.,
Paper No. 8, June, 1991.

Schmitz, Ronald P. Structural Dynamic Analysis of Electronic Assemblies Using NASTRAN

Restart/Format Change Capability, NASTRAN: Users’ Exper., pp. 363-392, September,
1971, (NASA TM X-2378).

Schweiger, W.; de Bruyne, F.; Dirschmid, W. Fluid Structure Interaction of Car Fuel Tanks,
Proc. of the MSC/NASTRAN Eur. Users’ Conf., May, 1984.

Shein, Shya-Ling. Generation of the Space Station Freedom On-Orbit Dynamic Loads Analysis
Model Using MSC/NASTRAN V66A Superelements, The 2nd Annual MSC Taiwan Users
Conf., Paper No. 7, October, 1990.

Shiraki, K.; Hashimoto, H.; Sato, N.; Nasu, S.; Kinno, M. Japanese Experiment Module (JEM):
On-Orbit Structural Dynamic Analysis, 1993 MSC Japan’s 11th User’s Conf. Proc., Paper
No. 10.

Shivaji, M.; Raju, V. S. N. Dynamic Analysis of R. C. C. Chimneys, MSC 1995 World Users’ Conf.
Proc., Paper No. 34, May, 1995.

Singh, Ashok K.; Nichols, Christian W. Derivation of an Equivalent Beam Model From a
Structural Finite Element Model, The MSC 1988 World Users Conf. Proc., Vol. I, Paper
No. 14, March, 1988.

Singh, Sudeep K.; Engelhardt, Charlie. Dynamic Analysis of a Large Space Structure Using
External and Internal Superelements, The MSC 1991 World Users’ Conf. Proc., Vol. I,
Paper No. 27, March, 1991.

Skattum, Knut S. Modeling Techniques of Thin-Walled Beams with Open Cross Sections,
NASTRAN: Users’ Exper., pp. 179-196, September, 1972, (NASA TM X-2637).

Smith, Michael R.; Rangacharyulu, M.; Wang, Bo P.; Chang, Y. K. Application of Optimization
Techniques to Helicopter Structural Dynamics, AIAA/ASME/ASCE/AHS/ASC 32nd
Structures, Structural Dynamics, and Materials Conf., Part 1, Paper No. 91-0924,
pp. 227-237, April, 1991.

Stockwell, Alan E.; Perez, Sharon E.; Pappa, Richard S. Integrated Modeling and Analysis of a
Space-Truss Test Article, The MSC 1990 World Users Conf. Proc., Vol. I, Paper No. 16,
March, 1990.

Struschka, M.; Goldstein, H. Approximation of Frequency Dependant Nonlinearities in Linear

FE-Models, Proc. of the 15th MSC/NASTRAN Eur. Users’ Conf., October, 1988.

Main Index
512

Subrahmanyam, K. B.; Kaza, K. R. V.; Brown, G. V.; Lawrence, C. Nonlinear Bending-Torsional

Vibration and Stability of Rotating, Pre-Twisted, Preconed Blades Including Coriolis
Effects, National Aeronautics and Space Administration, January, 1986, (NASA
TM-87207).

Tecco, T. C. Analyzing Frequency Dependent Stiffness and Damping with MSC/NASTRAN,

MSC/NASTRAN Users’ Conf. Proc., Paper No. 25, March, 1985.

Thornton, E. A. Application of NASTRAN to a Space Shuttle Dynamics Model, NASTRAN:

Users’ Exper., pp. 421-438, September, 1971, (NASA TM X-2378).

Ting, Tienko. Test/Analysis Correlation for Multiple Configurations, The MSC 1993 World
Users’ Conf. Proc., Paper No. 74, May, 1993.

Tinti, Francesco Carlo; Scaffidi, Costantino. Structural Dynamics and Acoustic Design of Engine
Component in View of Exterior Noise Reduction Using Numerical Techniques, MSC 1995
European Users’ Conf. Proc., Italian Session, September, 1995.

Turner, Patrick Ryan. Integrating Finite Element Analysis with Quasi-Static Loadings from a
Large Displacement Dynamic Analysis, The 1989 MSC World Users Conf. Proc., Vol. II,
Paper No. 37, March, 1989.

Tzong, George T. J.; Sikes, Gregory D.; Dodd, Alan J. Large Order Modal Analysis Module in
the Aeroelastic Design Optimization Program (ADOP), The MSC 1991 World Users’
Conf. Proc., Vol. II, Paper No. 36, March, 1991.

Unger, B.; Eichlseder, Wilfried; Schuch, F. Predicting the Lifetime of Dynamically Stressed
Components, Proc. of the 20th MSC European Users’ Conf., Paper No. 36, September,
1993.

Vance, Judy; Bernard, James E. Approximating Eigenvectors and Eigenvalues Across a Wide
Range of Design, The MSC 1992 World Users’ Conf. Proc., Vol. II, Paper No. 46, May,
1992.

Visintainer, Randal H.; Aslani, Farhang. Shake Test Simulation Using MSC/NASTRAN, MSC
1994 World Users’ Conf. Proc., Paper No. 32, June, 1994.

Vitiello, P.; Quaranta, V. SEA Investigation Via a FEM Based Substructuring Technique, MSC
1995 European Users’ Conf. Proc., Italian Session, September, 1995.

Walton, William B.; Blakely, Ken. Modeling of Nonlinear Elastic Structures Using
MSC/NASTRAN, MSC/NASTRAN Users’ Conf. Proc., Paper No. 11, March, 1983.

Wamsler, M.; Blanck, N.; Kern, G. On the Enforced Relative Motion Inside a Structure, Proc. of
the 20th MSC European Users’ Conf., September, 1993.

Wang, B. P.; Caldwell, S. P.; Smith, C. M. Improved Eigensolution Reanalysis Procedures in

Structural Dynamics, The MSC 1990 World Users Conf. Proc., Vol. II, Paper No. 46,
March, 1990.

Main Index
 APPENDIX L 513
References and Bibliography

Wang, B. P.; Chang, Y. K.; Lawrence, K. L.; Chen, T. Y. Optimum Design of Structures with
Multiple Configurations with Frequency and Displacement Constraints, 31st AIAA/
ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conf., Part 1,
pp. 378-384, April, 1990.

Wang, B. P.; Caldwell, S. P. Reducing Truncation Error in Structural Dynamic Modification,

The MSC 1991 World Users’ Conf. Proc., Vol. I, Paper No. 11, March, 1991.

Wang, B. P.; Caldwell, S. P. Improved Approximate Method for Computting Eigenvector

Derivatives, Finite Elements in Analysis and Design v 14 n 4 Nov 1993.

Watanabe, Masaaki. Computation of Virtual Mass to Rigid Body Structure by MSC/NASTRAN,

The First MSC/NASTRAN User’s Conf. in Japan, October, 1983, in Japanese.

Wijker, J. J. Differential Stiffness in Conjunction with Dynamics, NASTRAN User’s Conf., June,
1981.

Wijker, J. J. Acoustic Effects on the Dynamic Behaviour of Lightweight Structures, Proc. of the
MSC/NASTRAN Eur. Users’ Conf., Paper No. 3, April, 1985.

Wilhelmy, Dr. Viktor. Dynamic Analysis with Gaps, The 1989 MSC World Users Conf. Proc.,
Vol. II, Paper No. 40, March, 1989.

Woytowitz, P. J.; Jiang, K. C.; Bhat, K. P. Dynamic Analysis of Optical Beam Pointing, The MSC
1988 World Users Conf. Proc., Vol. I, Paper No. 11, March, 1988.

Yang, Jackson C. S.; Frederick, Diana L. Application of NASTRAN in Nonlinear Analysis of a

Cartridge Case Neck Separation Malfunction, NASTRAN: Users’ Exper., pp. 389-396,
September, 1975, (NASA TM X-3278).

Young, K. J.; Mitchell, L. D. On the Performance of Various Kinds of Rod and Beam Mass
Matrices on a Plane-Frame Structure, Proc. of IMAC-IX, Vol. I, pp. 1057-1065, April, 1991.

DYNAMICS – ANALYSIS / TEST CORRELATION

Allen, James J.; Martinez, David R. Techniques for Implementing Structural Model
Identification Using Test Data, Sandia National Laboratories, June, 1990,
(SAND90-1185).

Anker, J. C. Checks that Pay, Proc. of the MSC/NASTRAN Eur. Users’ Conf., May, 1984.

Blakely, Ken; Howard, G. E.; Walton, W. B.; Johnson, B. A.; Chitty, D. E. Comparison of a
Nonlinear Dynamic Model of a Piping System to Test Data, 7th Int. Conf. on Struct.
Mech. in Reactor Tech., August, 1983.

Blakely, Ken; Walton, W. B. Selection of Measurement and Parameter Uncertainties for Finite
Element Model Revision, 2nd Int. Modal Analysis Conf., February, 1984.

Blakely, Ken. Updating MSC/NASTRAN Models to Match Test Data, The MSC 1991 World
Users’ Conf. Proc., Vol. II, Paper No. 50, March, 1991.

Main Index
514

Blakely, Ken. Revising MSC/NASTRAN Models to Match Test Data, Proc. of the 9th Int. Modal
Analysis Conf., April, 1991.

Blakely, Ken. Get the Model Right, then Run the Analysis, Machine Design, October 24, 1991.

Blakely, Ken; Rose, Ted. Cross-Orthogonality Calculations for Pre-Test Planning and Model
Verification, The MSC 1993 World Users’ Conf. Proc., Paper No. 72, May, 1993.

Blakely, Ken; Bush, Richard. Using MSC/NASTRAN to Match Dynamic Test Data, Proc. of the
Int. Conf. on Structural Dynamics Modelling, July, 1993.

Blakely, Ken; Rose, Ted. Cross-Orthogonality Calculations for Pre-Test Planning and Model
Verification, Proc. of the 20th MSC European Users’ Conf., September, 1993.

Blakely, Ken. Matching Frequency Response Test Data with MSC/NASTRAN, MSC 1994 World
Users’ Conf. Proc., Paper No. 17, June 1994.

Blakely, Ken. Matching Frequency Response Test Data with MSC/NASTRAN, Proc. of the 21st
MSC European User’s Conf., Italian Session, September, 1994.

Brillhart, Ralph; Hunt, David L.; Kammer, Daniel C.; Jensen, Brent M.; Mason, Donald R. Modal
Survey and Test-Analysis Correlation of the Space Shuttle SRM, Proc. of the 6th Int.
Modal Analysis Conf., pp. 863-870, February, 1988.

Brughmans, Marc; Leuridan, Jan; Blauwkamp, Kevin. The Application of FEM-EMA

Correlation and Validation Techniques on a Body-in-White, The MSC 1993 World Users’
Conf. Proc., Paper No. 6, May, 1993.

Brughmans, M.; Lembregts, PhD. F.; Furini, PhD. F.; Storrer, O. Modal Test on the Pininfarina
Concept Car Body “ETHOS 1", Actes de la 2ème Confèrence Française Utilisateurs des
Logiciels MSC, Toulouse, France, September, 1995.

Brughmans, M.; Lembregts, F, Ph.D.; Furini, F., Ph.D. Modal Test on the Pininfarina Concept
Car Body “ETHOS 1", MSC 1995 World Users’ Conf. Proc., Paper No. 5, May, 1995.

Budynas, R.; Kolhatkar, S. Modal Analysis of a Robot Arm Using Finite Element Analysis and
Modal Testing, Proc. of the 8th Int. Modal Analysis Conf., Vol. I, pp. 67-70, January, 1990.

Budynas, R. G.; Krebs, D. Modal Correlation of Test and Finite Element Results Using Cross
Orthogonality with a Reduced Mass Matrix Obtained by Modal Reduction and
NASTRAN’s Generalized Dynamic Reduction Solution, Proc. of the 9th Int. Modal
Analysis Conf., Vol. I, pp. 549-554, April, 1991.

Butler, Thomas G. Test vs. Analysis: A Discussion of Methods, Fourteenth NASTRAN Users’
Colloq., pp. 173-186, May, 1986, (NASA CP-2419).

Call, V.; Mason, D. Space Shuttle Redesigned Solid Rocket Booster Structural Dynamic
Predictions and Correlations of Liftoff, AIAA/SAE/ASME/ASEE 26th Joint Propulsion
Conf., Paper No. AIAA 90-2081, July, 1990.

Chung, Y. T. Model Reduction and Model Correlation Using MSC/NASTRAN, MSC 1995 World
Main Index
Users’ Conf. Proc., Paper No. 8, May, 1995.
 APPENDIX L 515
References and Bibliography

Coladonato, Robert J. Development of Structural Dynamic Test Evnironments for Subsystems

and Components, Seventh NASTRAN Users’ Colloq., pp. 85-110, October, 1978, (NASA
CP-2062).

Coppolino, Robert N. Integrated Dynamic Test/Analysis Processor Overview,

MSC/NASTRAN Users’ Conf. Proc., Paper No. 5, March, 1986.

Cronkhite, J. D. Development, Documentation and Correlation of a NASTRAN Vibration Model

of the AH-1G Helicopter Airframe, NASTRAN: Users’ Exper., pp. 273-294, October, 1976,
(NASA TM X-3428).

Dascotte, E.; Von Estorff, O.; Wandinger, J. Validation and Updating of MSC/NASTRAN Finite
Element Models Using Experimental Modal Data, Proc. of the 20th MSC European Users’
Conf., Paper No. 10, September, 1993.

de Boer, A.; Kooi, B. W. A DMAP for Updating Dynamic Mathematical Models with Respect to
Measured Data, Proc. of the MSC/NASTRAN Eur. Users’ Conf., May, 1986.

Deger, Yasar. Modal Analysis of a Concrete Gravity Dam - Linking FE Analysis and Test
Results, Proc. of the 20th MSC European Users’ Conf., September, 1993.

Deutschel, Brian W.; Katnik, Richard B.; Bijlani, Mohan; Cherukuri, Ravi. Improving Vehicle
Response to Engine and Road Excitation Using Interactive Graphics and Modal
Reanalysis Methods, SAE Trans., Paper No. 900817, September, 1991.

Dirschmid, W.; Nolte, F.; Dunne, L. W. Mathematical Model Updating Using Experimentally
Determined Real Eigenvectors, Proc. of the 17th MSC Eur. Users’ Conf., Paper No. 4,
September, 1990.

Drago, Raymond J.; Margasahayam, Ravi. Stress Analysis of Planet Gears with Integral
Bearings; 3-D Finite Element Model Development and Test Validation, The MSC 1987
World Users Conf. Proc., Vol. I, Paper No. 4, March, 1987.

Ferg, D.; Foote, L.; Korkosz, G.; Straub, F.; Toossi, M.; Weisenburger, R. Plan, Execute, and
Discuss Vibration Measurements, and Correlations to Evaluate a NASTRAN Finite
Element Model of the AH-64 Helicopter Airframe, National Aeronautics and Space
Administration, January, 1990, (NASA CR-181973).

Graves, Roger W. Interfacing MSC/NASTRAN with SDRC-IDEAS to Perform Component

Mode Synthesis Combining Test, Analytical, and F. E. Data, The MSC 1988 World Users
Conf. Proc., Vol. II, Paper No. 58, March, 1988.

Hehta, Pravin K. Correlation of a NASTRAN Analysis with Test Measurements for HEAO-2
Optics, MSC/NASTRAN Users’ Conf. Proc., Paper No. 17, March, 1984.

Herbert, Andrew A.; Currie, A. O.; Wilson, W. Analysis of Automotive Axle Carrier Assembly
and Comparison with Test Data, The MSC 1987 World Users Conf. Proc., Vol. I, Paper
No. 6, March, 1987.

Herting, D. N. Parameter Estimation Using Frequency Response Tests, MSC 1994 World Users’
Conf. Proc., Paper No. 18, June, 1994.
Main Index
516

Jiang, K. C. Finite Element Model Updates Using Modal Test Data, The 1989 MSC World Users
Conf. Proc., Vol. II, Paper No. 48, March, 1989.

Kabe, Alvar M. Mode Shape Identification and Orthogonalization,

AIAA/ASME/ASCE/AHS/ASC 29th Structures, Structural Dynamics and Materials
Conf., Paper No. 88-2354, 1988.

Kammer, Daniel C.; Jensen, Brent M.; Mason, Donald R. Test-Analysis Correlation of the Space
Shuttle Solid Rocket Motor Center Segment, J. of Spacecraft, Vol. 26, No. 4, pp. 266-273,
March, 1988.

Kelley, William R.; Isley, L. D. Using MSC/NASTRAN for the Correlation of Experimental
Modal Models for Automotive Powertrain Structures, The MSC 1993 World Users’ Conf.
Proc., Paper No. 8, May, 1993.

Kelley, William R.; Isley, L. Dean; Foster, Thomas J. Dynamic Correlation Study Transfer Case
Housings, MSC 1996 World Users’ Conf. Proc., Vol. II, Paper No. 15, June, 1996.

Kientzy, Donald; Richardson, Mark; Blakely, Ken. Using Finite Element Data to Set Up Modal
Tests, Sound and Vibration, June, 1989.

Lammens, Stefan; Brughmans, Marc; Leuridan, Jan; Sas, Paul. Application of a FRF Based
Model Updating Technique for the Validation of Finite Element MOdels of Components
of the Automotive Industry, MSC 1995 World Users’ Conf. Proc., Paper No. 7, May, 1995.

Lee, John M.; Parker, Grant R. Application of Design Sensitivity Analysis to Improve
Correlations Between Analytical and Test Modes, The 1989 MSC World Users Conf.
Proc., Vol. I, Paper No. 21, March, 1989.

Linari, M.; Mancino, E. Application of the MSC/NASTRAN Program to the Study of a Simple
Reinforced Concrete Structure in Nonlinear Material Field, Proc. of the 15th
MSC/NASTRAN Eur. Users’ Conf., October, 1988.

Lowrey, Richard D. Calculating Final Mesh Size Before Mesh Completion, The MSC 1990 World
Users Conf. Proc., Vol. II, Paper No. 44, March, 1990.

Marlow, Jill M.; Lindell, Michael C. NASSTAR: An Instructional Link Between

MSC/NASTRAN and STAR, Proceedings of the 11th International Modal Analysis
Conference, Florida, 1993.

Masse, Barnard; Pastorel, Henri. Stress Calculation for the Sandia 34-Meter Wind Turbine
Using the Local Circulation Method and Turbulent Wind, The MSC 1991 World Users’
Conf. Proc., Vol. II, Paper No. 53, March, 1991.

Mindle, Wayne L.; Torvik, Peter J. A Comparison of NASTRAN (COSMIC) and Experimental
Results for the Vibration of Thick Open Cylindrical Cantilevered Shells, Fourteenth
NASTRAN Users’ Colloq., pp. 187-204, May, 1986, (NASA CP-2419).

Morton, Mark H. Application of MSC/NASTRAN for Assurance of Flight Safety and Mission
Effectiveness with Regard to Vibration upon Installation of the Stinger Missile on the
AH-64A, The MSC 1991 World Users’ Conf. Proc., Vol. II, Paper No. 52, March, 1991.
Main Index
 APPENDIX L 517
References and Bibliography

Neads, M. A.; Eustace, K. I. The Solution of Complex Structural Systems by NASTRAN within
the Building Block Approach, NASTRAN User’s Conf., May, 1979.

Nowak, William. Electro-Mechanical Response Simulation of Electrostatic Voltmeters Using

MSC/NASTRAN, The MSC 1993 World Users’ Conf. Proc., Paper No. 65, May, 1993.

O’Callahan, Dr. John; Avitabile, Peter; Reimer, Robert. An Application of New Techniques for
Integrating Analytical and Experimental Structural Dynamic Models, The 1989 MSC
World Users Conf. Proc., Vol. II, Paper No. 47, March, 1989.

Ott, Walter; Kaiser, Hans-Jurgen; Meyer, Jurgen. Finite Element Analysis of the Dynamic
Behaviour of an Engine Block and Comparison with Experimental Modal Test Results,
The MSC 1990 World Users Conf. Proc., Vol. I, Paper No. 14, March, 1990.

Paolozzi, A. Structural Dynamics Modification with MSC/NASTRAN, Proc. of the 19th MSC
European Users’ Conf., Paper No. 14, September, 1992.

Park, H. B.; Suh, J. K.; Cho, H. G.; Jung, G. S. A Study on Idle Vibration Analysis Technique
Using Total Vehicle Model, MSC 1995 World Users’ Conf. Proc., Paper No. 6, May, 1995.

Parker, Grant R.; Rose, Ted L.; Brown, John J. Kinetic Energy Calculation as an Aid to
Instrumentation Location in Modal Testing, The MSC 1990 World Users Conf. Proc.,
Vol. II, Paper No. 47, March, 1990.

Preve, A.; Meneguzzo, M.; Merlo, A.; Zimmer, H. Simulation of Vehicles’ Structural Noise:
Numerical/Experimental Correlation in the Acoustic Simulation of the Internal Noise,
Proc. of the 21st MSC European Users’ Conf., Italian Session, September, 1994.

Rabani, Hadi. Static and Dynamic FEM/Test Correlation of an Automobile Body, The 1989
MSC World Users Conf. Proc., Vol. II, Paper No. 49, March, 1989.

Rainer, I. G. MSC/NASTRAN as a Key Tool to Satisfy Increasing Demand for Numerical

Simulation Techniques, Proc. of the 20th MSC European Users’ Conf., September, 1993.

Ray, William F. The Use of MSC/NASTRAN and Empirical Data to Verify a Design,
MSC/NASTRAN Users’ Conf. Proc., Paper No. 11, March, 1984.

Scapinello, F.; Colombo, E. An Approach for Detailed Analysis of Complex Structures Avoiding
Complete Models, Proc. of the MSC/NASTRAN Eur. Users’ Conf., May, 1987.

Sok-chu, Park; Ishii, Tetsu; Honda, Shigeki; Nagamatsu, Akio. Vibration Analysis and
Optimum Design of Press Machines, 1994 MSC Japan Users’ Conf. Proc.

Stack, Charles P.; Cunningham, Timothy J. Design and Analysis of Coriolis Mass Flowmeters
Using MSC/NASTRAN, The MSC 1993 World Users’ Conf. Proc., Paper No. 54, May,
1993.

Su, Hong. Structural Analysis of Ka-BAND Gimbaled Antennas for a Communications

Satellite System, MSC 1996 World Users’ Conf. Proc., Vol. IV, Paper No. 33, June, 1996.

Tawekal, Ricky; Budiyanto, M. Agus. Finite Element Model Correlation for Structures, The
Main Index
MSC 1993 World Users’ Conf. Proc., Paper No. 73, May, 1993.
518

Ting, T.; Ojalvo, I. U. Dynamic Structural Correlation via Nonlinear Programming Techniques,
The MSC 1988 World Users Conf. Proc., Vol. II, Paper No. 57, March, 1988.

Ting, Tienko; Chen, Timothy L. C. FE Model Refinement with Actual Forced Responses of
Aerospace Structures, The MSC 1991 World Users’ Conf. Proc., Vol. II, Paper No. 51,
March, 1991.

Ting, Tienko. Test/Analysis Correlation for Multiple Configurations, The MSC 1993 World
Users’ Conf. Proc., Paper No. 74, May, 1993.

Ujihara, B. H.; Dosoky, M. M.; Tong, E. T. Improving a NASTRAN Dynamic Model with Test
Data Using Linwood, Tenth NASTRAN Users’ Colloq., pp. 74-86, May, 1982, (NASA
CP-2249).

Zeischka, H. LMS/Link Correlating and Validating F.E.A. for Dynamic Structure Behaviour
with Experimental Modal Analysis, Proc. of the MSC/NASTRAN Eur. Users’ Conf.,
May, 1987.

DYNAMICS – COMPONENT MODE SYNTHESIS

Barnett, Alan R.; Ibrahim, Omar M.; Sullivan, Timothy L.; Goodnight, Thomas W. Transient
Analysis Mode Participation for Modal Survey Target Mode Selection Using
MSC/NASTRAN DMAP, MSC 1994 World Users’ Conf. Proc., Paper No. 8, June, 1994.

Bedrossian, Herand; Rose, Ted. DMAP Alters for Nonlinear Craig-Bampton Component Modal
Synthesis, The MSC 1993 World Users’ Conf. Proc., Paper No. 25, May, 1993.

Brillhart, Ralph; Hunt, David L.; Kammer, Daniel C.; Jensen, Brent M.; Mason, Donald R. Modal
Survey and Test-Analysis Correlation of the Space Shuttle SRM, Proc. of the 6th Int.
Modal Analysis Conf., pp. 863-870, February, 1988.

Brown, J. J.; Lee, J. M.; Parker, G. R.; Zuhuruddin, K. Procedure for Processing and Displaying
Entire Physical Modes Based on Results Generated Through Component Mode Synthesis,
The MSC 1988 World Users Conf. Proc., Vol. I, Paper No. 15, March, 1988.

Call, V.; Mason, D. Space Shuttle Redesigned Solid Rocket Booster Structural Dynamic
Predictions and Correlations of Liftoff, AIAA/SAE/ASME/ASEE 26th Joint Propulsion
Conf., Paper No. AIAA 90-2081, July, 1990.

Cifuentes, A. O.; Herting, D. N. Transient Response of a Beam to a Moving Mass Using a Finite
Element Approach, Innovative Numerical Methods in Engineering, Proc. of the Fourth
Int. Symp. on Numerical Methods in Engineering, Springer-Verlag, pp. 533-539, March,
1986.

Cifuentes, A. O. Non-Linear Dynamic Problems Using a Combined Finite Element-Finite

Difference Technique, Proc. of the 6th Conf. on the Mathematics of Finite Elements and
Application, April/May, 1987.

Main Index
 APPENDIX L 519
References and Bibliography

Carney, Kelly S.; Abdallah, Ayma A.; Hucklebridge, Arthur A. Implementation of the
Block-Krylov Boundary Flexibility Method of Component Synthesis, The MSC 1993
World Users’ Conf. Proc., Paper No. 26, May, 1993.

Del Basso, Steve; Singh, Sudeep; Lindenmoyer, Alan J. Component Mode Synthesis of Space
Station Freedom Using MSC/NASTRAN Superelement Architecture, The MSC 1990
World Users Conf. Proc., Vol. II, Paper No. 48, March, 1990.

Duncan, Alan E. Application of Modal Modeling and Mount System Optimization to Light
Duty Truck Ride Analysis, 4th Int. Conf. on Veh. Struct. Mech., pp. 113-128, November,
1981, (SAE #811313).

Flanigan, Christopher C.; Abdallah, Ayman; Manella, Richard. Implementation of the

Benfield-Hruda Modal Synthesis Method in MSC/NASTRAN, The MSC 1992 World
Users’ Conf. Proc., Vol. I, Paper No. 11, May, 1992.

Garnek, Michael. Large Space Structure Analysis Using Substructure Modal Test Data,
AIAA/ASME/ASCE/AHS 25th Structures, Structural Dynamics, and Materials Conf.,
AIAA Paper 84-0942-CP, May, 1984.

Ghosh, Tarun. MSC/NASTRAN Based Component Mode Synthesis Analysis Without the Use
of DMAPS, MSC 1996 World Users’ Conf. Proc., Vol. II, Paper No. 18, June, 1996.

Gieseke, R. K. Analysis of Nonlinear Structures via Mode Synthesis, NASTRAN: Users’ Exper.,
pp. 341-360, September, 1975, (NASA TM X-3278).

Graves, Roger W. Interfacing MSC/NASTRAN with SDRC-IDEAS to Perform Component

Mode Synthesis Combining Test, Analytical, and F. E. Data, The MSC 1988 World Users
Conf. Proc., Vol. II, Paper No. 58, March, 1988.

Halcomb, J. R. Application of Component Modes to the Analysis of a Helicopter, Proc. of the

MSC/NASTRAN Users’ Conf., March, 1979.

Hambric, Stephen A. Power Flow and Mechanical Intensity Calculations in Structural Finite
Element Analysis, ASME J. of Vibration and Acoustics, Vol. 112, pp. 542-549, October,
1990.

Herting, David N.; Hoesly, R. L. Development of an Automated Multi-Stage Modal Synthesis

System for NASTRAN, Sixth NASTRAN Users’ Colloq., pp. 435-448, October, 1977,
(NASA CP-2018).

Herting, David N. Accuracy of Results with NASTRAN Modal Synthesis, Seventh NASTRAN
Users’ Colloq., pp. 389-404, October, 1978, (NASA CP-2062).

Herting, D. N. A General Purpose, Multi-Stage, Component Modal Synthesis Method, Finite

Elements in Analysis and Design, Vol. 1, No. 2, 1985.

Hill, R. G.; Merckx, K. R. Seismic Response Evaluation of a Reactor Core Using Modal
Synthesis, Proc. of the 3rd Int. Modal Analysis Conf., Vol. 2, pp. 996-1000, June, 1985.

Main Index
520

Hodgetts, P. A.; Maitimo, F. M.; Wijker, J. J. Dynamic Analysis of the Polar Platform Solar
Array Using a Multilevel Component Mode Synthesis Technique, Proc. of the 19th MSC
European Users’ Conf., Paper No. 15, September, 1992.

Ichikawa, Tetsuji; Hagiwara, Ichiro. Frequency Response Analysis of Large-Scale Damped

Structures Using Component Mode Synthesis, Nippon Kikai Gakkai Ronbunshu, C
Hen/Transactions of the Japan Society of Mechanical Engineers, Part C v 60 n 569 Jan
1994.

Jasuja, S. C.; Borowski, V. J.; Anderson, D. H. Finite Element Modeling Techniques for the
Simulation of Automotive Powertrain Dynamics, Proc. of the 5th Int. Modal Analysis
Conf., Vol. II, pp. 1520-1530, April, 1987.

Jay, Andrew; Lewis, Bryan; Stakolich, Ed. Effect of Time Dependent Flight Loads on Turbofan
Engine Performance Deterioration, ASME J. of Engineering for Power, Vol. 104, No. 3,
July, 1982.

Kammer, Daniel C.; Jensen, Brent M.; Mason, Donald R. Test- Analysis Correlation of the Space
Shuttle Solid Rocket Motor Center Segment, J. of Spacecraft, Vol. 26, No. 4, pp. 266-273,
March, 1988.

Kasai, Manabu. Generalized CMS Employing External Superelements, The Fourth

MSC/NASTRAN User’s Conf. in Japan, October, 1986, in Japanese.

Kasai, Manabu. Representation of DMAP by Generalized CMS in System Identification, The

Fourth MSC/NASTRAN User’s Conf. in Japan, October, 1986, in Japanese.

Kasai, Manabu. Approach to CMS Subjected to the Boundary Constrained at Single Point, The
Sixth MSC/NASTRAN User’s Conf. in Japan, October, 1988, in Japanese.

Kim, Hyoung M.; Bartkowicz, Theodoore J.; Van Horn, David A. Data Recovery and Model
Reduction Methods for Large Structures, The MSC 1993 World Users’ Conf. Proc., Paper
No. 23, May, 1993.

Kubota, Minoru. Efficient Use of Component Mode Sysnthesis Using Image Superelements
Applied to Dynamic Analysis of Crankshaft, MSC/NASTRAN Users’ Conf. Proc., Paper
No. 22, March, 1986.

Lee, W. M. Substructure Mode Synthesis with External Superelement, The 2nd Annual MSC
Taiwan User’s Conf., Paper No. 16, October, 1990, in Chinese.

MacNeal, R. H. A Hybrid Method of Component Mode Synthesis, Computers and Structures,

Vol. 1, No. 4, pp. 581-601, 1971.

Martinez, D. R.; Gregory, D. L. A Comparison of Free Component Mode Synthesis Techniques

Using MSC/NASTRAN, MSC/NASTRAN Users’ Conf. Proc., Paper No. 18, March, 1983.

Martinez, David R.; Gregory, Danny L. A Comparison of Free Component Mode Synthesis
Techniques Using MSC/NASTRAN, Sandia National Laboratories, June, 1984,
(SAND83-0025).

Main Index
 APPENDIX L 521
References and Bibliography

Murakawa, Osamu. Hull Vibration Analysis by Modal Synthesis Method, The First
MSC/NASTRAN User’s Conf. in Japan, October, 1983, in Japanese.

Murakawa, Osamu; Iwahashi, Yoshio; Sakato, Tsuneo. Ship Vibration Analysis Using Modal
Synthesis Technique, MSC/NASTRAN Users’ Conf. Proc., Paper No. 23, March, 1984.

Nasu, Syouichi. Modal Synthesis of Experimental Vibration Characteristics Using

MSC/NASTRAN Results as the Reference Model, The Sixth MSC/NASTRAN User’s
Conf. in Japan, October, 1988, in Japanese.

Nefske, D. J.; Sung, S. H.; Duncan, A. E. Applications of Finite Element Methods to Vehicle
Interior Acoustic Design, Proc. of the 1984 Noise and Vibration Conf., Paper No. 840743,
1984.

Ookuma, Masaaki; Nagamatsu, Akio. Comparison of Component Mode Synthesis Method with
MSC-NASTRAN, Nippon Kikai Gakkai Ronbunshu, C Hen, Vol. 49, No. 446,
pp. 1883-1889, October, 1983, in Japanese.

Ookuma, Masaaki; Nagamatsu, Akio. Comparison of Component Mode Synthesis Method with
MSC-NASTRAN, Bulletin of the JSME, Vol. 27, No. 228, pp. 1294-1298, June, 1984.

Parekh, Jatin C.; Harris, Steve G. The Application of the Ritz Procedure to Damping Prediction
Using a Modal Strain Energy Approach, Damping ’89, Paper No. CCB, November, 1989.

Philippopoulos, V. G. Dynamic Analysis of an Engine-Transmission Assembly-Superelement

and Component Mode Synthesis, Proc. of the Conf. on Finite Element Methods and
Technology, Paper No. 3, March, 1981.

Reyer, H. Modal Synthesis with External Superelements in MSC/NASTRAN, Proc. of the

MSC/NASTRAN Eur. Users’ Conf., May, 1984.

Sabahi, Dara; Rose, Ted. Special Applications of Global-Local Analysis, The MSC 1990 World
Users Conf. Proc., Vol. II, Paper No. 49, March, 1990.

Sabahi, Dara; Rose, Ted. MSC/NASTRAN Superelement Analysis of the NASA/AMES

Pressurized Wind Tunnel, The MSC 1990 World Users Conf. Proc., Vol. II, Paper No. 50,
March, 1990.

Shein, Shya-Ling; Marquette, Brian; Rose, Ted. Superelement Technology Application and
Development in Dynamic Analysis of Large Space Structures, The MSC 1991 World
Users’ Conf. Proc., Vol. I, Paper No. 26, March, 1991.

Suzukiri, Yoshihiro. Component Mode Synthesis Application of MSC/NASTRAN V66, The 2nd
Annual MSC Taiwan Users Conf., Paper No. 10, October, 1990.

Suzukiri, Yoshihiro. Component Mode Synthesis Application of MSC/NASTRAN V66, Proc. of

the First MSC/NASTRAN Users’ Conf. in Korea, Paper No. 17, October, 1990.

Tong, Edward T.; Chang, Craig C. J. An Efficient Procedure for Data Recovery of a
Craig-Bampton Component, MSC 1994 World Users’ Conf. Proc., Paper No. 26, June 1994.

Main Index
522

Wamsler, M.; Komzsik, L.; Rose, T. Combination of Quasi-Static and Dynamic System Mode
Shapes, Proc. of the 19th MSC European Users’ Conf., Paper No. 13, September, 1992.

Wang, Bo Ping. Synthesis of Structures with Multiple Frequency Constraints,

AIAA/ASME/ASCE/AHS 27th Structures, Structural Dynamics and Materials Conf.,
Part 1, pp. 394-397, May, 1986.

DYNAMICS – DAMPING
El Maddah, M.; Imbert, J. F. A Comparison of Damping Synthesis Methods for Space Vehicle
Dynamic Analysis, NASTRAN User’s Conf., May, 1979.

Everstine, Gordon C.; Marcus, Melvyn S. Finite Element Prediction of Loss Factors for
Structures with Frequency- Dependent Damping Treatments, Thirteenth NASTRAN
Users’ Colloq., pp. 419-430, May, 1985, (NASA CP-2373).

Gibson, W. C.; Johnson, C. D. Optimization Methods for Design of Viscoelastic Damping

Treatments, ASME Design Engineering Division Publication, Vol. 5, pp. 279-286,
September, 1987.

Gibson, Warren C.; Austin, Eric. Analysis and Design of Damped Structures Using
MSC/NASTRAN, The MSC 1992 World Users’ Conf. Proc., Vol. I, Paper No. 25, May,
1992.

Johnson, Conor D.; Keinholz, David A. Prediction of Damping in Structures with Viscoelastic
Materials Using MSC/NASTRAN, MSC/NASTRAN Users’ Conf. Proc., Paper No. 17,
March, 1983.

Kalinowski, A. J. Modeling Structural Damping for Solids Having Distinct Shear and
Dilational Loss Factors, Seventh NASTRAN Users’ Colloq., pp. 193-206, October, 1978,
(NASA CP-2062).

Kalinowski, A. J. Solution Sensitivity and Accuracy Study of NASTRAN for Large Dynamic
Problems Involving Structural Damping, Ninth NASTRAN Users’ Colloq., pp. 49-62,
October, 1980, (NASA CP-2151).

Kienholz, Dave K.; Johnson, Conor D.; Parekh, Jatin C. Design Methods for Viscoelastically
Damped Sandwich Plates, AIAA/ASME/ASCE/AHS 24th Structures, Structural
Dynamics and Materials Conf., Part 2, pp. 334-343, May, 1983.

Li, Tsung-hsiun; Bernard, James. Optimization of Damped Structures in the Frequency Domain,
The MSC 1993 World Users’ Conf. Proc., Paper No. 28, May, 1993.

Lu, Y. P.; Everstine, G. C. More on Finite Element Modeling of Damped Composite Systems, J.
of Sound and Vibration, Vol. 69, No. 2, pp. 199-205, 1980.

Mace, M. Damping of Beam Vibrations by Means of a Thin Constrained Viscoelastic Layer:

Evaluation of a New Theory, Journal of Sound and Vibration v 172 n 5 May 19 1994.

Main Index
 APPENDIX L 523
References and Bibliography

Merchant, D. H.; Gates, R. M.; Ice, M. W.; Van Derlinden, J. W. The Effects of Localized Damping
on Structural Response, NASTRAN: Users’ Exper., pp. 301-320, September, 1975, (NASA
TM X-3278).

Parekh, Jatin C.; Harris, Steve G. The Application of the Ritz Procedure to Damping Prediction
Using a Modal Strain Energy Approach, Damping ’89, Paper No. CCB, November, 1989.

Rose, Ted. DMAP Alters to Apply Modal Damping and Obtain Dynamic Loading Output for
Superelements, The MSC 1993 World Users’ Conf. Proc., Paper No. 24, May, 1993.

Rose, Ted.; McNamee, Martin. A DMAP Alter to Allow Amplitude-Dependent Modal Damping
in a Transient Solution, MSC 1996 World Users’ Conf. Proc., Vol. V, Paper No. 50, June,
1996.

Shieh, Rong C. A Superefficient, MSC/NASTRAN-Interfaced Computer Code System for

Dynamic Response Analysis of Nonproportionally Damped Elastic Systems, The MSC
1993 World Users’ Conf. Proc., Paper No. 14, May, 1993.

Tonin, Renzo. Vibration Isolation of Impacts in High-Rise Structures, The Second Australasian
MSC Users Conf., Paper No. 11, November, 1988.

DYNAMICS – FREQUENCY RESPONSE

Balasubramanian, B.; Wamsler, M. Identification of Contributing Modes in MSC/NASTRAN
Modal Frequency Response Analyses, Proc. of the MSC/NASTRAN Eur. Users’ Conf.,
May, 1987.

Barnett, Alan R.; Widrick, Timothy W.; Ludwiczak, Damian R. Combining Acceleration and
Displacement Dependent Modal Frequency Responses Using An MSC/NASTRAN DMAP
Alter, MSC 1996 World Users’ Conf. Proc., Vol. II, Paper No. 17, June, 1996.

Bellinger, Dean. Dynamic Analysis by the Fourier Transform Method with MSC/NASTRAN,
MSC 1995 World Users’ Conf. Proc., Paper No. 10, May, 1995.

Bianchini, Emanuele; Marulo, Francesco; Sorrentino, Assunta. MSC/NASTRAN Solution of

Structural Dynamic Problems Using Anelastic Displacement Fields, Proceedings of the
36th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials
Conference and AIAA/ASME Adpative Structures Forum, Part 5 (of 5), New Orleans,
1995.

Blakely, Ken. Matching Frequency Response Test Data with MSC/NASTRAN, MSC 1994 World
Users’ Conf. Proc., Paper No. 17, June 1994.

Carlson, David L.; Shipley, S. A.; Yantis, T. F. Procedure for FRF Model Tuning in
MSC/NASTRAN, The MSC 1993 World Users’ Conf. Proc., Paper No. 71, May, 1993.

Charron, Francois; Donato, Louis; Fontaine, Mark. Exact Calculation of Minimum Margin of
Safety for Frequency Response Analysis Stress Results Using Yielding or Failure
Theories, The MSC 1993 World Users’ Conf. Proc., Paper No. 5, May, 1993.

Main Index
524

de la Fuente, E.; San Millán, J. Calculation within MSC/NASTRAN of the Forces Transmitted
by Multipoint Constraints (MPC) and the Forces Generated in Support Constraints, MSC
1996 World Users’ Conf. Proc., Vol. II, Paper No. 20, June, 1996.

Herbert, S.; Janavicius, P. MSC/NASTRAN Frequency Response Analysis of the Raven Army
Communication Shelter, The Third Australasian MSC Users Conf. Proc., Paper No. 14,
November, 1989.

Herting, D. N. Parameter Estimation Using Frequency Response Tests, MSC 1994 World Users’
Conf. Proc., Paper No. 18, June, 1994.

Kajiwara, Itsurou; Nagamatsu, Akio; Seto, Kazuto. New Theory for Elimination of Resonance
Peak and Optimum Design of Optical Servosystem, 1994 MSC Japan Users’ Conf. Proc.,
Paper No. 1.

Liew, K. M.; Jiang, L.; Lim, M. K.; Low, S. C. Numerical Evaluation of Frequency Responses for
Delaminated Honeycomb Structures, Computers and Structures v 55 n 2 Apr 17 1995.

Parker, G. R.; Brown, J. J. Evaluating Modal Contributors in a NASTRAN Frequency Response

Analysis, MSC/NASTRAN Users’ Conf. Proc., Paper No. 14, March, 1983.

Rose, Ted. Using Optimization in MSC/NASTRAN to Minimize Response to a Rotating

Imbalance, 1994 MSC Japan Users’ Conf. Proc., Paper No. 28, December 1994.

Shieh, Rong C. A Superefficient, MSC/NASTRAN-Interfaced Computer Code System for

Dynamic Response Analysis of Nonproportionally Damped Elastic Systems, The MSC
1993 World Users’ Conf. Proc., Paper No. 14, May, 1993.

Soni, Ravi, et al. Development of a Methodology to Predict the Road Noise Performance
Characteristics, The MSC 1993 World Users’ Conf. Proc., Paper No. 9, May, 1993.

Tsutsui, Keicchiro; Nogami, Ray. Development of a Nonlinear Frequency Response Program for
Simulating Vehicle Ride Comfort, MSC 1995 World Users’ Conf. Proc., Paper No. 37,
May, 1995.

Visintainer, Randal H.; Aslani, Farhang. Shake Test Simulation Using MSC/NASTRAN, MSC
1994 World Users’ Conf. Proc., Paper No. 32, June, 1994.

Wamsler, Manfred; Krusemann, Rolf. Calculating and Interpreting Contact Forces Between
Brake Disc and Linings in Frequency Response Analysis, Proc. of the 18th MSC Eur.
Users’ Conf., Paper No. 7, June, 1991.

Yen, K. Z. Y.; Hsueh, W. C.; Hsui, T. C. Chatter Suppression of a CNC Lathe in Inside-Diameter
Cutting, The Sixth Annual MSC Taiwan Users’ Conf. Proc., Paper No. 4, November, 1994.

DYNAMICS – MODES, FREQUENCIES, AND VIBRATIONS

(Korean). Structural Analysis of Solar Array Substate, MSC 1994 Korea Users’ Conf. Proc.,
December, 1994, in Korean.

Main Index
 APPENDIX L 525
References and Bibliography

(Korean). Optimal Design of Chip Mounter Considering Dynamic Characteristics, MSC 1994
Korea Users’ Conf. Proc., December, 1994, in Korean.

(Korean). A Study on Vibration Characteristics of a Steering Wheel According to Geometric

Variations, MSC 1994 Korea Users’ Conf. Proc., December, 1994, in Korean.

(Korean). A Study on Idle Vibration Analysis Technique Using Total Vehicle Model, MSC 1994
Korea Users’ Conf. Proc., December, 1994, in Korean.

(Korean). Vibration Analysis for Outercase in Drum Washer and Floor, MSC 1994 Korea Users’
Conf. Proc., December, 1994, in Korean.

Ahmad, M. Fouad; Guile, Carl W. Analysis of Coupled Natural Frequencies of Thin-Walled

Beams with Open Cross Sections Using MSC/NASTRAN, The MSC 1990 World Users
Conf. Proc., Vol. I, Paper No. 15, March, 1990.

Allen, James J.; Martinez, David R. Techniques for Implementing Structural Model
Identification Using Test Data, Sandia National Laboratories, June, 1990,
(SAND90-1185).

Arakawa, H.; Murakami, T.; Ito, H. Vibration Analysis of the Turbine Generator Stator Frame,
The MSC 1988 World Users Conf. Proc., Vol. II, Paper No. 43, March, 1988.

Armand, Sasan; Lin, Paul. Influence of Mass Moment of Inertia on Normal Modes of Preloaded
Solar Array Mast, The MSC 1992 World Users’ Conf. Proc., Vol. I, Paper No. 12, May,
1992.

Arora, Tejbir; Birmingham, Lily. Application of MSC/NASTRAN Superelement Dynamic

Reduction Techniques for the Vertical Launching System, The MSC 1988 World Users
Conf. Proc., Vol. I, Paper No. 39, March, 1988.

Barnes, R. A.; Schmid, R.; Adrick, H. C. Rotor Dynamic Analysis with MSC/NASTRAN via the
Important Modes Method, The 1989 MSC World Users Conf. Proc., Vol. I, Paper 13,
March, 1989.

Barnett, Alan R.; Abdallah, Ayma A.; Ibrahim, Omar M.; Manella, Richard T. Solving Modal
Equations of Motion with Initial Conditions Using MSC/NASTRAN DMAP Part 1:
Implementing Exact Mode Superposition, The MSC 1993 World Users’ Conf. Proc., Paper
No. 12, May, 1993.

Barnett, Alan R.; Abdallah, Ayma A.; Ibrahim, Omar M.; Sullivan, Timothy L. Solving Modal
Equations of Motion with Initial Conditions Using MSC/NASTRAN DMAP Part 2:
Coupled vs. Uncoupled Integration, The MSC 1993 World Users’ Conf. Proc., Paper
No. 13, May, 1993.

Bella, David F.; Steinhard, E. Critical Frequency Determination of a Flexible Rotating Structure
Attached to a Flexible Support, Proc. of the 18th MSC Eur. Users’ Conf., Paper No. 28,
June, 1991.

Main Index
526

Bella, David; Hartmueller, Hans; Muehlenfeld, Karsten; Tokar, Gabriel. Identification of

Critical Speeds of Rotors Attached to Flexible Supports, The MSC 1993 World Users’
Conf. Proc., Paper No. 34, May, 1993.

Blakely, Ken; Rose, Ted. Cross-Orthogonality Calculations for Pre-Test Planning and Model
Verification, The MSC 1993 World Users’ Conf. Proc., Paper No. 72, May, 1993.

Blakely, Ken; Rose, Ted. Cross-Orthogonality Calculations for Pre-Test Planning and Model
Verification, Proc. of the 20th MSC European Users’ Conf., September, 1993.

Brughmans, M.; Lembregts, PhD. F.; Furini, PhD. F.; Storrer, O. Modal Test on the Pininfarina
Concept Car Body “ETHOS 1", Actes de la 2ème Confèrence Française Utilisateurs des
Logiciels MSC, Toulouse, France, September, 1995.

Brughmans, M.; Lembregts, F, Ph.D.; Furini, F., Ph.D. Modal Test on the Pininfarina Concept
Car Body “ETHOS 1", MSC 1995 World Users’ Conf. Proc., Paper No. 5, May, 1995.

Buchanan, Guy. Superelement Data Recovery via the Modal Acceleration Method, The MSC
1988 World Users Conf. Proc., Vol. I, Paper No. 40, March, 1988.

Budynas, R.; Kolhatkar, S. Modal Analysis of a Robot Arm Using Finite Element Analysis and
Modal Testing, Proc. of the 8th Int. Modal Analysis Conf., Vol. I, pp. 67-70, January, 1990.

Budynas, R. G.; Krebs, D. Modal Correlation of Test and Finite Element Results Using Cross
Orthogonality with a Reduced Mass Matrix Obtained by Modal Reduction and
NASTRAN’s Generalized Dynamic Reduction Solution, Proc. of the 9th Int. Modal
Analysis Conf., Vol. I, pp. 549-554, April, 1991.

Butler, Thomas G.; Muskivitch, John C. Application of Flanigan’s Mode Acceleration in

MSC/NASTRAN Version 66, The 1989 MSC World Users Conf. Proc., Vol. I, Paper No. 25,
March, 1989.

Caldwell, Steve; Wang, B. P. Application of Approximate Techniques in the Estimation of

Eigenvalue Quality, The MSC 1993 World Users’ Conf. Proc., Paper No. 11, May, 1993.

Campanile, P.; Pisino, E.; Testi, R.; Manzilli, G.; Minen, D. Flexible Structures in Adams Using
Modal Data from NASTRAN, Proc. of the 21st MSC European Users’ Conf., Italian
Session, September, 1994.

Carlson, Mark. Applications of Finite Element Analysis for an Improved Musical Instrument
Design, MSC 1996 World Users’ Conf. Proc., Vol. I, Paper No. 8, June, 1996.

Carneiro, S. H. S.; Duarte, J. A. A.; Mendonca, C. B. Theoretical and Experimental Modal

Analysis of the VLS (Satellite Launcher Vehicle) Bent, Proc. of the 2nd MSC/NASTRAN
Users’ Conf. in Brazil, Paper No. 3, March, 1993, in Portuguese.

Case, William R. Jr. NASTRAN DMAP Alter for Determining a Local Stiffness Modification to
Obtain a Specified Eigenvalue, NASTRAN: Users’ Exper., pp. 269-284, September, 1973,
(NASA TM X-2893).

Main Index
 APPENDIX L 527
References and Bibliography

Case, William R. A NASTRAN DMAP Procedure for Calculation of Base Excitation Modal
Participation Factors, Eleventh NASTRAN Users’ Colloq., pp. 113-140, May, 1983.

Cattani, E.; Micelli, D.; Sereni, L.; Cocordano, S. Cylinder Block Eigenfrequencies and
Eigenvectors Prediction with a Linear Brick and Wedge Finite Element Model, Proc. of
the 19th MSC European Users’ Conf., Paper No. 16, September, 1992.

Chang, Cuann-yeu; Chang, Yuan-bing. Using MSC/NASTRAN to Obtain Modal Parameters,

The MSC 1988 World Users Conf. Proc., Vol. I, Paper No. 21, March, 1988.

Chargin, M. L.; Dunne, L. W.; Herting, D. N. Nonlinear Dynamics of Brake Squeal, MSC 1996
World Users’ Conf. Proc., Vol. V, Paper No. 47, June, 1996.

Chen, J. T.; Chyuan, S. W.; You, D. W.; Wong, H. T. A New Method for Determining the Modal
Participation Factor in Support Motion Problems Using MSC/NASTRAN, The Seventh
Annual MSC/NASTRAN Users’ Conf. Proc., Taiwan, 1995.

Clary, Robert R. Practical Analysis of Plate Vibrations Using NASTRAN, NASTRAN: Users’
Exper., pp. 325-342, September, 1971, (NASA TM X-2378).

Cohen, Allan R.; Laurenson, Robert M. Application of a Substructure Technique for

STS/Payload Coupled Modal Analysis, Proc. of the MSC/NASTRAN Users’ Conf.,
March, 1979.

Concilio, A.; Del Gatto, S.; Lecce, L.; Miccoli, G. Simple and Cheap Noise and Vibration Active
Control System Using Collocated Piezoelectric Devices on a Panel, Proceedings of the
11th International Modal Analysis Conference, Florida, 1993.

Courtney, Roy Leon. NASTRAN Modeling Studies in the Normal-Mode Method and
Normal-Mode Synthesis, NASTRAN: Users’ Exper., pp. 181-200, September, 1971,
(NASA TM X-2378).

Cronkhite, James D.; Smith, Michael R. Experiences in NASTRAN Airframe Vibration

Prediction at Bell Helicopter Textron, American Helicopter Soc. Dynamics Specialists
Mtg., Section 6, Vibrations Session I, Paper No. 1, November, 1989.

Cross, C.; Rao, A. Comparison of Modal Performance of Alternate Compressor Bracket Design,
Seventh Australasian Users Conf. Proc., Sydney, October, 1993.

Deger, Yasar. Modal Analysis of a Concrete Gravity Dam - Linking FE Analysis and Test
Results, Proc. of the 20th MSC European Users’ Conf., September, 1993.

Deutschel, Brian William. A Systematic Approach Using Finite Elements for Improving Vehicle
Ride, CAD/CAM Robotics and Factories of the Future Integration of Design, Analysis,
and Manufacturing (Proc.), Springer-Verlag Berlin, Heidelberg, Vol. I, pp. 150-154, 1989.

Deutschel, Brian W.; Katnik, Richard B.; Bijlani, Mohan; Cherukuri, Ravi. Improving Vehicle
Response to Engine and Road Excitation Using Interactive Graphics and Modal
Reanalysis Methods, SAE Trans., Paper No. 900817, September, 1991.

Main Index
528

Egashira, Yuji. Large Scale Vibration Analysis of Car Body Using Superelement Method, The
Second MSC/NASTRAN User’s Conf. in Japan, October, 1984, in Japanese.

El-Bayoumy, Lotfi. Identification and Correction of Damaging Resonances in Gear Drives, Gear
Technology, Vol. 1, No. 2, pp. 14-19, August/September, 1984.

Ferg, D.; Foote, L.; Korkosz, G.; Straub, F.; Toossi, M.; Weisenburger, R. Plan, Execute, and
Discuss Vibration Measurements, and Correlations to Evaluate a NASTRAN Finite
Element Model of the AH-64 Helicopter Airframe, National Aeronautics and Space
Administration, January, 1990, (NASA CR-181973).

Flanigan, Chris. Methods for Calculating and Using Modal Initial Conditions in
MSC/NASTRAN, Proc. of the Conf. on Finite Element Methods and Technology, March,
1980.

Flanigan, C. An Alternate Method for Mode Acceleration Data Recovery in MSC/NASTRAN,

Proc. of the Conf. on Finite Element Methods and Technology, Paper No. 7, March, 1981.

Flanigan, Christopher C. Efficient and Accurate Procedures for Calculating Data Recovery
Matrices for Superelement Models, The 1989 MSC World Users Conf. Proc., Vol. II, Paper
No. 44, March, 1989.

Flanigan, Christopher C. Implementation of the IRS Dynamic Reduction Method in

MSC/NASTRAN, The MSC 1990 World Users Conf. Proc., Vol. I, Paper No. 13, March,
1990.

Friberg, Olof; Karlsson, Rune; Akesson, Bengt. Linking of Modal and Finite Elements in
Structural Vibration Analysis, Proc. of the 2nd Int. Modal Analysis Conf. and Exhibit,
Vol. 1, pp. 330-339, February, 1984.

Gallaher, Bruce. Determination of Structural Dynamic Response Sensitivity to Modal

Truncation, MSC/NASTRAN Users’ Conf. Proc., Paper No. 10, March, 1986.

Ghosh, Tarun; Nall, Marsha; Muniz, Ben; Cheng, Joseph. Space Station Solar Array Pointing
System Control/Structure Interaction Study Using CO-ST-IN for Modal Reduction, The
MSC 1993 World Users’ Conf. Proc., Paper No. 68, May, 1993.

Gieseke, R. K. Modal Analysis of the Mated Space Shuttle Configuration, NASTRAN: Users’
Exper., pp. 221-236, September, 1971, (NASA TM X-2378).

Girard, A.; Boullet, A.; Dardel, R. Dynamic Analysis of a Satellite Using the Normal Modes of
the Appendages, Proc. of the MSC/NASTRAN Eur. Users’ Conf., April, 1985.

Grandle, Robert E.; Rucker, Carl E. Modal Analysis of a Nine-Bay Skin-Stringer Panel,
NASTRAN: Users’ Exper., pp. 343-362, September, 1971, (NASA TM X-2378).

Gupta, Viney K.; Zillmer, Scott D.; Allison, Robert E. Solving Large-Scale Dynamic Systems
Using Band Lanczos Method in Rockwell NASTRAN on Cray X-MP, Fourteenth
NASTRAN Users’ Colloq., pp. 236-246, May, 1986, (NASA CP-2419).

Main Index
 APPENDIX L 529
References and Bibliography

Hardman, E. S. Static and Normal Modes Analysis of an Aircraft Structure Using the
NASTRAN External Superelement Method, Proc. of the MSC/NASTRAN Eur. Users’
Conf., May, 1986.

Harn, Wen-Ren; Hwang, Chi-Ching. Evaluation of Direct Model Modification Methods via
MSC/NASTRAN DMAP Procedures, The MSC 1990 World Users Conf. Proc., Vol. II,
Paper No. 43, March, 1990.

Hayashida, Mirihiro. Application of Design Sensitivity Analysis to Reduction of Vibration of

Ship’s Deck Structure, The Sixth MSC/NASTRAN User’s Conf. in Japan, October, 1988,
in Japanese.

Herting, D. N.; Joseph, J. A.; Kuusinen, L. R.; MacNeal, R. H. Acoustic Analysis of Solid Rocket
Motor Cavities by a Finite Element Method, National Aeronautics and Space
Administration, pp. 285-324, September, 1971, (NASA TM X-2378).

Herting, David N. Accuracy of Results with NASTRAN Modal Synthesis, Seventh NASTRAN
Users’ Colloq., pp. 389-404, October, 1978, (NASA CP-2062).

Hill, R. G. The Use of MSC/NASTRAN to Determine the Impact Response of a Reactor Core Due
to Seismic Loading, The MSC 1987 World Users Conf. Proc., Vol. I, Paper No. 26, March,
1987.

Hirano, Tohru. Visual Evaluation Method for the Vibration Analysis Utilizing a Super
Intelligent Color Graphic Display, The First MSC/NASTRAN User’s Conf. in Japan,
October, 1983, in Japanese.

Hsueh, W. c.; Hsui, T. C.; Yen, K. Z. Y. Modal and Frequency Response Analyses of Vertical
Machining Center Structures - VC65, The Fifth Annual MSC Taiwan Users’ Conf. Proc.,
November, 1993.

Huang, Jieh-Shan. Detect the Variety of Structural System by the Vibration Test, The 2nd
Annual MSC Taiwan Users Conf., Paper No. 11, October, 1990, in Chinese.

Igarashi, Mitsuo. Eigenvalue Analysis of Shaft Supported by Anti-Isotropic Bearing, The Fifth
MSC/NASTRAN User’s Conf. in Japan, October, 1987, in Japanese.

Ito, Hiroyuki. Application of Modal Analysis Technique for Cars, The Fourth MSC/NASTRAN
User’s Conf. in Japan, October, 1986, in Japanese.

Jabbour, K. N. Normal Mode Analysis of the Radio Astronomy Explorer (RAE) Booms and
Spacecraft, NASTRAN: Users’ Exper., pp. 237-250, September, 1971, (NASA TM X-2378).

Jasuja, S. C.; Borowski, V. J.; Anderson, D. H. Finite Element Modeling Techniques for the
Simulation of Automotive Powertrain Dynamics, Proc. of the 5th Int. Modal Analysis
Conf., Vol. II, pp. 1520-1530, April, 1987.

Jiang, K. C.; Gahart, R. Analysis and Modal Survey Test of Intelsat VIIA Deployed Solar Array,
MSC 1994 World Users’ Conf. Proc., Paper No. 27, June 1994.

Main Index
530

Jiang, L.; Liew, K.M.; Lim, M.K.; Low, S.C. Vibratory Behaviour of Delaminated Honeycomb
Structures: a 3-D Finite Element Modelling, Computers and Structures v 55 n 5 Jun 3 1995.

Ju, Yeuan Jyh; Ting, Tienko. Modelling and Analysis of an Accelerometer Using MSC/ARIES
and MSC/NASTRAN, MSC 1995 World Users’ Conf. Proc., Paper No. 22, May, 1995.

Kabe, Alvar M. Mode Shape Identification and Orthogonalization,

AIAA/ASME/ASCE/AHS/ASC 29th Structures, Structural Dynamics and Materials
Conf., Paper No. 88-2354, 1988.

Kajiwara, Itsuro; Nagamatsu, Akio. Optimum Design of Structure and Control Systems by
Modal Analysis, 1994 MSC Japan Users’ Conf. Proc.

Kam, T. Y.; Yang, C. M.; Wu, J. H. Determination of Natural Frequencies of Laminated

Composite Space Structures Via The Experiemental and Finite Element Approaches, The
Sixth Annual MSC Taiwan Users’ Conf. Proc., Paper No. 5, November, 1994.

Kang, J. M.; Kim, J. Y.; Lee, K. J.; Yum, D. J.; Seol, Y. S.; Rashed, S. Kawahara, A. Simulation of
3-D Sloshing and Structural Response in Ship’s Tanks Taking Account of Fluid-Structure
Interaction, 1994 MSC Japan Users’ Conf. Proc., Paper No. 29, December 1994.

Kasai, Manabu. Better Accuracy of Response Derived from Modal Analysis, The Second
MSC/NASTRAN User’s Conf. in Japan, October, 1984, in Japanese.

Katnik, Richard B.; Yu, Che-Hsi; Wolf, Walt. Interactive Modal Animation and Structural
Modification, Proc. of the 6th Int. Modal Analysis Conf., Vol. I, pp. 947-952, February,
1988.

Kelley, William R.; Isley, L. D. Using MSC/NASTRAN for the Correlation of Experimental
Modal Models for Automotive Powertrain Structures, The MSC 1993 World Users’ Conf.
Proc., Paper No. 8, May, 1993.

Kientzy, Donald; Richardson, Mark; Blakely, Ken. Using Finite Element Data to Set Up Modal
Tests, Sound and Vibration, June, 1989.

Knott, George; Ishin, Young; Chargin, M. A Modal Analysis of the Violin, The MSC 1988 World
Users Conf. Proc., Vol. II, Paper No. 42, March, 1988.

Kodiyalam, Srinivas; Graichen, Catherine M.; Connell, Isobel J.; Finnigan, Peter M. Design
Optimization of Satellite Structures for Frequency, Strength, and Buckling
Requirements, Aerospace Sciences Meeting and Exhibit, AIAA, January, 1993.

Krishnamurthy, Ravi S. Stress and Vibration Analysis of Radial Gas Turbine Components,
Sixteenth NASTRAN Users’ Colloq., pp. 128-137, April, 1988, (NASA CP-2505).

Lapi, M.; Grangier, H. Modal Effective Parameters: an Application to Shipboard Support

Structures to Reduce Vibrations Transmission, Proc. of the 17th MSC Eur. Users’ Conf.,
Paper No. 5, September, 1990.

Larkin, Paul A.; Miller, Michael W. MSC/NASTRAN Superelement Analysis, MSC/NASTRAN

Users’ Conf. Proc., Paper No. 10, March, 1982.
Main Index
 APPENDIX L 531
References and Bibliography

Lawrie, Geoff. The Determination of the Normal Modes of a Gliding Vehicle, The Sixth
Australasian MSC Users Conf. Proc., November, 1992.

Lee, Jyh-Chian. Using Residual Vector in MSC/NASTRAN Modal Frequency Response to

Improve Accuracy, The 4th MSC Taiwan Users’ Conf., Paper No. 8, November, 1992, in
Chinese.

Lee, Sang H. Effective Modal Mass for Characterization of Vibration Modes, Proc. of the
Second MSC/NASTRAN Users’ Conf. in Korea, Paper No. 22, October, 1991.

Liepins, Atis A.; Conaway, John H. Application of NASTRAN to Propeller-Induced Ship

Vibration, NASTRAN: Users’ Exper., pp. 361-376, September, 1975, (NASA TM X-3278).

Lim, Tae W.; Kashangaki, Thomas A. L. Structural Damage Detection of Space Truss Structures
Using Best Achievable Eigenvectors, AIAA Journal v32 n 5 May 1994.

Lipman, Robert R. Computer Animation of Modal and Transient Vibrations, Fifteenth

NASTRAN Users’ Colloq., pp. 111-117, August, 1987, (NASA CP-2481).

Liu, Dauh-Churn, Shieh, Niahn-Chung. Vibration Suppression of High Precision Grinding

Machine Using Finite Element Method (MSC/NASTRAN), The Sixth Annual MSC
Taiwan Users’ Conf. Proc., Paper No. 7, November, 1994.

Lu, Y. P.; Killian, J. W.; Everstine, G. C. Vibrations of Three Layered Damped Sandwich Plate
Composites, J. of Sound and Vibration, Vol. 64, No. 1, pp. 63-71, 1979.

Lui, C. Y.; Mason, D. R. Space Shuttle Redesigned Solid Rocket Motor Nozzle Natural
Frequency Variations with Burn Time, AIAA/SAE/ASME 27th Joint Propulsion Conf.,
Paper No. AIAA 91-2301, June, 1991.

Lundgren, Gert. MSC’s Solvers Predict the Best Attachment of the Sunbeam Tiger Fiberglass
Front-End, MSC 1996 World Users’ Conf. Proc., Vol. II, Paper No. 16, June, 1996.

Marcus, Melvyn S. A Finite-Element Method Applied to the Vibration of Submerged Plates, J.

of Ship Research, Vol. 22, No. 2, pp. 94-99, June, 1978.

Marcus, Melvyn S.; Everstine, Gordon C.; Hurwitz, Myles M. Experiences with the QUAD4
Element for Shell Vibrations, Sixteenth NASTRAN Users’ Colloq., pp. 39-43, April, 1988,
(NASA CP-2505).

Mase, M.; Saito, H. Application of FEM for Vibrational Analysis of Ground Turbine Blades,
MSC/NASTRAN Users’ Conf., March, 1978.

McMeekin, Michael; Kirchman, Paul. An Advanced Post Processing Methodology for Viewing
MSC/NASTRAN Generated Analyses Results, MSC 1994 World Users’ Conf. Proc., Paper
No. 21, June, 1994.

Mei, Chuh; Rogers, James L. Jr. NASTRAN Nonlinear Vibration Analyses of Beam and Frame
Structures, NASTRAN: Users’ Exper., pp. 259-284, September, 1975, (NASA TM X-3278).

Main Index
532

Mei, Chuh; Rogers, James L. Jr. Application of the TRPLT1 Element to Large Amplitude Free
Vibrations of Plates, Sixth NASTRAN Users’ Colloq., pp. 275-298, October, 1977, (NASA
CP-2018).

Meyer, Karl A. Normal Mode Analysis of the IUS/TDRS Payload in a Payload

Canister/Transporter Environment, Eighth NASTRAN Users’ Colloq., pp. 113-130,
October, 1979, (NASA CP-2131).

Michiue, Shinsuke. On the Accuracy in Vibration Analysis for Cylindrical Shell-Comparison

Between QUAD4/QUAD8, The First MSC/NASTRAN User’s Conf. in Japan, October,
1983, in Japanese.

Miller, R. D. Theoretical Analysis of HVAC Duct Hanger Systems, Fifteenth NASTRAN Users’
Colloq., pp. 222-249, August, 1987, (NASA CP-2481).

Nack, Wayne V.; Joshi, Arun M. Friction Induced Vibration, MSC 1995 World Users’ Conf.
Proc., Paper No. 36, May, 1995.

Nagayasu, Katsuyosi. Method for Prediction of Noise-Oriented Vibration on Pipe Lines to

Refrigerators, The Third MSC/NASTRAN User’s Conf. in Japan, October, 1985, in
Japanese.

Nagendra, Gopal K.; Herting, David N. Design Sensitivity for Modal Analysis, The 1989 MSC
World Users Conf. Proc., Vol. I, Paper No. 22, March, 1989.

Nagy, Lajos I.; Cheng, James; Hu, Yu-Kan. A New Method Development to Predict Brake Squeal
Occurence, MSC 1994 World Users’ Conf. Proc., Paper No. 14, June, 1994.

Neads, M. A.; Eustace, K. I. The Solution of Complex Structural Systems by NASTRAN within
the Building Block Approach, NASTRAN User’s Conf., May, 1979.

Newman, Malcolm; Pipano, Aaron. Fast Modal Extraction in NASTRAN via the FEER
Computer Program, NASTRAN: Users’ Exper., pp. 485-506, September, 1973, (NASA TM
X-2893).

Nishiwaki, Nobukiyo. Coupled Vibration of Rotating Disc and Blades, The Second
MSC/NASTRAN User’s Conf. in Japan, October, 1984, in Japanese.

Overbye, Vern D. MSC/NASTRAN Dynamic Analysis: Modal or Direct?, MSC/NASTRAN

Users’ Conf. Proc., Paper No. 6, March, 1986.

Pamidi, M. R.; Pamidi, P. R. Modal Seismic Analysis of a Nuclear Power Plant Control Panel
and Comparison with SAP IV, NASTRAN: Users’ Exper., pp. 515-530, October, 1976,
(NASA TM X-3428).

Pamidi, P. R. On the Append and Continue Features in NASTRAN, Seventh NASTRAN Users’
Colloq., pp. 405-418, October, 1978, (NASA CP-2062).

Paolozzi, A. Interfacing MSC/NASTRAN with a Structural Modification Code, Proc. of the 18th
MSC Eur. Users’ Conf., Paper No. 30, June, 1991.

Main Index
 APPENDIX L 533
References and Bibliography

Park, H. B.; Suh, J. K.; Cho, H. G.; Jung, G. S. A Study on Idle Vibration Analysis Technique
Using Total Vehicle Model, MSC 1995 World Users’ Conf. Proc., Paper No. 6, May, 1995.

Parker, G. R.; Brown, J. J. Kinetic Energy DMAP for Mode Identification, MSC/NASTRAN
Users’ Conf. Proc., Paper No. 8, March, 1982.

Parker, G. R.; Brown, J. J. Evaluating Modal Contributors in a NASTRAN Frequency Response

Analysis, MSC/NASTRAN Users’ Conf. Proc., Paper No. 14, March, 1983.

Pulgrano, Louis J.; Masters, Steven G. Self-Excited Oscillation of a 165 Foot Water Tower, MSC
1995 World Users’ Conf. Proc., Paper No. 32, May, 1995.

Reyer, H. Modal Synthesis with External Superelements in MSC/NASTRAN, Proc. of the

MSC/NASTRAN Eur. Users’ Conf., May, 1984.

Rose, Ted. A Method to Apply Initial Conditions in Modal Transient Solutions, The MSC 1991
World Users’ Conf. Proc., Vol. I, Paper No. 13, March, 1991.

Rose, Ted. Some Suggestions for Evaluating Modal Solutions, The MSC 1992 World Users’
Conf. Proc., Vol. I, Paper No. 10, May, 1992.

Rose, Ted. Using Optimization in MSC/NASTRAN to Minimize Response to a Rotating

Imbalance, MSC 1994 Korea Users’ Conf. Proc., December, 1994.

Rose, Ted. Using Dynamic Optimization to Minimize Driver Response to a Tire

Out-of-Balance, MSC 1994 Korea Users’ Conf. Proc., December, 1994.

Rose, Ted. Using Optimization in MSC/NASTRAN to Minimize Response to a Rotating

Imbalance, 1994 MSC Japan Users’ Conf. Proc., Paper No. 28, December 1994.

Rose, Ted L. Using Optimization in MSC/NASTRAN to Minimize Response to a Rotating

Imbalance, The Sixth Annual MSC Taiwan Users’ Conf. Proc., Paper No. D, November,
1994.

Saito, Hiroshi; Watanabe, Masaaki. Modal Analysis of Coupled Fluid-Structure Response,

MSC/NASTRAN Users’ Conf. Proc. March, 1982.

Salvestro, Livio; Sirocco, Howden; Currie, Andrew. Cyclic Symmetry Analysis of an Air Blower
Fan, Seventh Australasian Users Conf. Proc., Sydney, October, 1993.

Scanlon, Jack; Swan, Jim. A Stand-Alone DMAP Program for Modal Cross-Correlation, MSC
1995 World Users’ Conf. Proc., Paper No. 40, May, 1995.

Schiavello, D. V.; Sinkiewicz, J. E. DMAP for Determining Modal Participation,

MSC/NASTRAN Users’ Conf. Proc., Paper No. 15, March, 1983.

Schwering, W. Shulze. A DMAP for Identification of Modeshapes, Proc. of the

MSC/NASTRAN Eur. Users’ Conf., June, 1983.

Shalev, D.; Unger, A. Nonlinear Analysis Using a Modal-Based Reduction Technique,

Composite Structures v31 n 4 1995.

Main Index
534

Shalev, Doron; Unger, A. Nonlinear Analysis Using a Modal Based Reduction Technique, The
MSC 1993 World Users’ Conf. Proc., Paper No. 51, May, 1993.

Shippen, J. M. Normal Modes Analysis of Spin Stabilised Spacecraft Possessing Cable Booms,
Proc. of the 18th MSC Eur. Users’ Conf., Paper No. 29, June, 1991.

Shirai, Yujiro; Arakawa, Haruhiko; Toda, Nobuo; Taneda, Yuji; Sakura, Kiyoshi. Active
Vibration Control for Aircraft Wing, JSME International Journal, v 36 n 3 Spe 1993.

Shy, Tyson; Hsiu, T. C.; Yen, K. Z. Y. Optimization of Structure Design of a Machining Center,
The Sixth Annual MSC Taiwan Users’ Conf. Proc., Paper No. 6, November, 1994.

Somayajula, Gopichand; Stout, Joseph; Tucker, John. Eigenvalue Reanalysis Using Subspace
Iteration Techniques, The 1989 MSC World Users Conf. Proc., Vol. I, Paper No. 26, March,
1989.

Stack, Charles P.; Cunningham, Timothy J. Design and Analysis of Coriolis Mass Flowmeters
Using MSC/NASTRAN, The MSC 1993 World Users’ Conf. Proc., Paper No. 54, May,
1993.

Starnes, James H. Jr. Vibration Studies of a Flat Plate and a Built-Up Wing, NASTRAN: Users’
Exper., pp. 637-646, September, 1971, (NASA TM X-2378).

Su, Hong. Structural Analysis of Ka-BAND Gimbaled Antennas for a Communications

Satellite System, MSC 1996 World Users’ Conf. Proc., Vol. IV, Paper No. 33, June, 1996.

Subrahmanyam, K. B.; Kaza, K. R. V.; Brown, G. V.; Lawrence, C. Nonlinear Vibration and
Stability of Rotating, Pretwisted, Preconed Blades Including Coriolis Effects, J. of
Aircraft, Vol. 24, No. 5, pp. 342-352, May, 1987.

Sundaram, S. V.; Hohman, Richard L.; Richards, Timothy R. Vibration Modes of a Tire Using
MSC/NASTRAN, MSC/NASTRAN Users’ Conf. Proc., Paper No. 26, March, 1985.

Tamba, Richard; Mowbray, Graham; Rao, Ananda. An Effective Method to Increase the Natural
Frequencies of a Transmission Assembly, The Sixth Australasian MSC Users Conf. Proc.,
November, 1992.

Tawekal, Ricky; Budiyanto, M. Agus. Finite Element Model Correlation for Structures, The
MSC 1993 World Users’ Conf. Proc., Paper No. 73, May, 1993.

Tawekal, Ricky L.; Miharjana, N. P. Validation of 3650 DWT Semi Containe Ship Finite Element
Model by Full Scale Measurements, MSC 1994 World Users’ Conf. Proc., Paper No. 19,
June, 1994.

Thornton, Earl A. A NASTRAN Correlation Study for Vibrations of a Cross-Stiffened Ship’s

Deck, NASTRAN: Users’ Exper., pp. 145-160, September, 1972, (NASA TM X-2637).

Ting, Tienko; Chen, Timothy L. C.; Twomey, William. A Practical Solution to Mode Crossing
Problem in Continuous Iterative Procedure, The 1989 MSC World Users Conf. Proc.,
Vol. I, Paper No. 14, March, 1989.

Main Index
 APPENDIX L 535
References and Bibliography

Ting, Tienko; Chen, T.; Twomey, W. Correlating Mode Shapes Based on Modal Assurance
Criterion, The MSC 1992 World Users’ Conf. Proc., Vol. I, Paper No. 21, May, 1992.

Tokuda, Naoaki; Mitikami, Shinsuke; Sakata, Yoshiuki. Accuracy of Vibration Analysis for Thin
Cylindrical Shell by MSC/NASTRAN, MSC/NASTRAN Users’ Conf. Proc., Paper No. 28,
March, 1984.

Vaillette, David. Evaluation of the Modal Response of a Pressure Vessel Filled with a Fluid, The
MSC 1991 World Users’ Conf. Proc., Vol. I, Paper No. 24, March, 1991.

Vance, Judy, M.; Bernard, James E. Approximating Eigenvectors and Eigenvalues Across a Wide
Range of Design, Finite Elements in Analysis and Design v 14 n 4 Nov 1993.

Vandepitte, D.; Wijker, J. J.; Appel, S.; Spiele, H. Normal Modes Analysis of Large Models, and
Applications to Ariane 5 Engine Frame, Proc. of the 18th MSC Eur. Users’ Conf., Paper
No. 6, June, 1991.

Wamsler, M.; Komzsik, L.; Rose, T. Combination of Quasi-Static and Dynamic System Mode
Shapes, Proc. of the 19th MSC European Users’ Conf., Paper No. 13, September, 1992.

Wang, B. P.; Cheu, T. C.; Chen, T. Y. Optimal Design of Compressor Blades with Multiple
Natural Frequency Constraints, ASME Design Engineering Division Publication, Vol. 5,
pp. 113-117, September, 1987.

Wang, B. P.; Lu, C. M.; Yang, R. J. Topology Optimization Using MSC/NASTRAN, MSC 1994
World Users’ Conf. Proc., Paper No. 12, June, 1994.

Wang, Bo Ping. Minimum Weight Design of Structures with Natural Frequency Constraints
Using MSC/NASTRAN, The MSC 1988 World Users Conf. Proc., Vol. II, Paper No. 60,
March, 1988.

Welte, Y. Vibration Analysis of an 8MW Diesel Engine, Proc. of the MSC/NASTRAN Eur.
Users’ Conf., May, 1986.

West, Timothy S. Approximate Dynamic Model Sensitivity Analysis For Large, Complex Space
Structues, MSC 1996 World Users’ Conf. Proc., Vol. I, Paper No. 6, June, 1996.

Wijker, J. J. Substructuring Technique Using the Modal Constraint Method, Proc. of the
MSC/NASTRAN Eur. Users’ Conf., June, 1983.

Wijker, J. J. MSC/NASTRAN Normal Mode Analysis on CRAY Computers, Proc. of the

MSC/NASTRAN Eur. Users’ Conf., June, 1983.

Yang, Howard J. Sorted Output in MSC/NASTRAN, Proc. of the Conf. on Finite Element
Methods and Technology, Paper No. 4, March, 1981.

Yen, K. Z. Y.; Hsueh, W. C.; Hsui, T. C. Chatter Suppression of a CNC Lathe in Inside-Diameter
Cutting, The Sixth Annual MSC Taiwan Users’ Conf. Proc., Paper No. 4, November, 1994.

Zhu, H.; Knight, D. Finite Element Forced Response Analysis on the Mondeo Front End
Accessory Drive System, Proc. of the 20th MSC European Users’ Conf., September, 1993.
Main Index
536

DYNAMICS – RANDOM RESPONSE

Barnett, Alan R.; Widrick, Timothy W.; Ludwiczak, Damian R. Combining Acceleration and
Displacement Dependent Modal Frequency Responses Using An MSC/NASTRAN DMAP
Alter, MSC 1996 World Users’ Conf. Proc., Vol. II, Paper No. 17, June, 1996.

Chiang, C. K.; Robinson, J. H.; Rizzi, S. A. Equivalent Linearization Solution Sequence for
MSC/NASTRAN, Winter Annual Meeting of the American Society of Mechanical
Engineers, pp. 133-138, November, 1992.

Ciuti, Gianluca. Avionic Equipment Dynamic Analysis, MSC 1995 European Users’ Conf. Proc.,
Italian Session, September, 1995.

Coyette, J. P.; Lecomte, C.; von Estorff, O. Evaluation of the Response of a Coupled Elastic
Structure Subjected to Random Mechanical or Acoustical Excitations Using
MSC/NASTRAN and SYSNOISE, MSC European Users’ Conf., Paper No. 21, September,
1996.

Crispino, Maurizio. A 3-D Model for the Evaluation through Random Analysis of Vertical
Dynamic Overloads in High Speed Railway Lines, MSC 1995 European Users’ Conf.
Proc., Italian Session, September, 1995.

Galletly, Robert; Wagner, R. J.; Wang, G. J.; Zins, John. Random Vibration and Acoustic
Analysis Using ARI RANDOM, a NASTRAN Post Processor, MSC/NASTRAN Users’
Conf. Proc., Paper No. 26, March, 1984.

Hatheway, A. Random Vibrations in Complex Electronic Structures, MSC/NASTRAN Users’

Conf. Proc., Paper No. 13, March, 1983.

Hatheway, Alson E. Evaluation of Ceramic Substrates for Packaging of Leadless Chip Carriers,
MSC/NASTRAN Users’ Conf. Proc., Paper No. 16, March, 1982.

Michels, Gregory J. Vibroacoustics Random Response Analysis Methodology, MSC 1995 World
Users’ Conf. Proc., Paper No. 9, May, 1995.

Palmieri, F. W. Example Problems Illustrating the Effect of Multiple Cross Correlated

Excitations on the Response of Linear Systems to Gaussian Random Excitations, The
MSC 1988 World Users Conf. Proc., Vol. I, Paper No. 18, March, 1988.

Palmieri, F. W. A Method for Predicting the Output Cross Power Spectral Density Between
Selected Variables in Response to Arbitrary Random Excitations, The MSC 1988 World
Users Conf. Proc., Vol. I, Paper No. 19, March, 1988.

Parthasarathy, Alwar; Elzeki, Mohamed; Abramovici, Vivianne. PSDTOOL-A DMAP

Enhancement to Harmonic/ Random Response Analysis in MSC/NASTRAN, The MSC
1993 World Users’ Conf. Proc., Paper No. 36, May, 1993.

Robinson, J. H.; Chiang, C. K.; Rizzi, S. A. Nonlinear Random Response Prediction Using
MSC/NASTRAN, National Aeronautics and Space Administration, Hampton, VA,
Langley Research Center, October, 1993.

Main Index
 APPENDIX L 537
References and Bibliography

Robinson, Jay H.; Chiang, C. K. An Equivalent Linearization Solution Sequence for

MSC/NASTRAN, The MSC 1993 World Users’ Conf. Proc., Paper No. 35, May, 1993.

Schwab, H. L.; Caffrey, J.; Lin, J. Fatigue Analysis Using Random Vibration, MSC 1995 World
Users’ Conf. Proc., Paper No. 17, May, 1995.

Shieh, Niahn Chung. Investigation of Swept Sine on Random Load, The 4th MSC Taiwan Users’
Conf., Paper No. 18, November, 1992, in Chinese.

Zins, J. Random Vibration and Acoustic Analysis Using ARI RANDOM a NASTRAN
Postprocessor, Proc. of the MSC/NASTRAN Eur. Users’ Conf., Paper No. 4, May, 1984.

DYNAMICS – REDUCTION METHODS

Abdallah, Ayman A.; Barnett, Alan R.; Widrick, Timothy W.; Manella, Richard T.; Miller, Robert
P. Stiffness-Generated Rigid-Body Mode Shapes for Lanczos Eigensolution with
Support DOF Via a MSC/NASTRAN DMAP Alter, MSC 1994 World Users’ Conf. Proc.,
Paper No. 10, June, 1994.

Flanigan, Christopher C. Implementation of the IRS Dynamic Reduction Method in

MSC/NASTRAN, The MSC 1990 World Users Conf. Proc., Vol. I, Paper No. 13, March,
1990.

Fox, Gary L. Evaluation and Reduction of Errors Induced by the Guyan Transformation, Tenth
NASTRAN Users’ Colloq., pp. 233-248, May, 1982, (NASA CP-2249).

Komzsik, L.; Dilley, G. Practical Experiences with the Lanczos Method, Proc. of the
MSC/NASTRAN Users Conf., Paper No. 13, March, 1985.

Kuang, Jao-Hwa; Lee, Chung-Ying. On a Guyan-Reduction Recycled Eigen Solution Technique,

The 2nd Annual MSC Taiwan Users Conf., Paper No. 13, October, 1990.

Levy, Roy. Guyan Reduction Solutions Recycled for Improved Accuracy, NASTRAN: Users’
Exper., pp. 201-220, September, 1971, (NASA TM X-2378).

Maekawa, Seiyou. Effect of Guyan Reduction and Generalized Dynamic Reduction, The Second
MSC/NASTRAN User’s Conf. in Japan, October, 1984, in Japanese.

Mera, A. MSC/NASTRAN Normal Mode Analysis with GDR: An Evaluation of Limitations,

MSC/NASTRAN Users’ Conf. Proc., Paper No. 27, March, 1985.

Mera, Andrew. Static Reduction and Symmetry Transformation of Large Finite Element
Models, Proc. of the MSC/NASTRAN Users’ Conf., Paper No. 12, March, 1979.

Vandepitte, D.; Wijker, J. J.; Appel, S.; Spiele, H. Normal Modes Analysis of Large Models, and
Applications to Ariane 5 Engine Frame, Proc. of the 18th MSC Eur. Users’ Conf., Paper
No. 6, June, 1991.

Vollan, Arne; Kaporin, Igor; Babikov, Pavel. Practical Experience with Different Iterative
Solvers for Linear Static and Modal Analysis of Large Finite Element Models, Proc. of the
21st MSC European Users’ Conf., Italian Session, September, 1994.
Main Index
538

Walker, James W. Evaluation of MSC/NASTRAN Generalized Dynamic Reduction and

Response Spectrum Analysis by Comparison with STARDYNE, MSC/NASTRAN Users’
Conf., March, 1978.

DYNAMICS – RESPONSE SPECTRUM

Barber, Pam; Arden, Kevin. Dynamic Design Analysis Method (DDAM) Using
MSC/NASTRAN, MSC 1994 World Users’ Conf. Proc., Paper No. 31, June, 1994.

Cutting, Fred. Individual Modal Accelerations as the Result of a Shock Response Spectra Input
to a Complex Structure, The 1989 MSC World Users Conf. Proc., Vol. I, Paper No. 11,
March, 1989.

Gassert, W.; Wolf, M. Floor Response Spectra of a Reactor Building Under Seismic Loading
Calculated with a 3-D Building Model, Proc. of the MSC/NASTRAN Eur. Users’ Conf.,
June, 1983.

Hirata, M.; Ishikawa, K.; Korosawa, M.; Fukushima, S.; Hoshina, H. Seismic Analysis of
Plutonium Glovebox by MSC/NASTRAN, January, 1993.

Kubota, Minoru. Response Spectrum Analysis of Underground Tank Dome Roof Using Image
Superelement Method, The Fifth MSC/NASTRAN User’s Conf. in Japan, October, 1987,
in Japanese.

Moharir, M. M. Shock Wave Propagation and NASTRAN Linear Algorithms, MSC/NASTRAN

Users’ Conf. Proc., Paper No. 7, March, 1986.

Overbye, Vern D. MSC/NASTRAN Dynamic Analysis: Modal or Direct?, MSC/NASTRAN

Users’ Conf. Proc., Paper No. 6, March, 1986.

Parris, R. A. Aspects of Seismic Analysis Using MSC/NASTRAN, Proc. of the MSC/NASTRAN

Eur. Users’ Conf., Paper No. 7, June, 1983.

Parthasarathy, A. CONSPEC-A DMAP System for Conventional Response-Spectrum Analysis

in MSC/NASTRAN, MSC/NASTRAN Users’ Conf. Proc., Paper No. 8, March, 1986.

Petteno, L.; Rossetto, P. Tecnomare Experiences in DMAP Processing, Proc. of the

MSC/NASTRAN Eur. Users’ Conf., May, 1984.

Rose, Ted L. Using Superelements for Response Spectrum and Other Handy Alters, The 1989
MSC World Users Conf. Proc., Vol. II, Paper No. 45, March, 1989.

Walker, James W. Evaluation of MSC/NASTRAN Generalized Dynamic Reduction and

Response Spectrum Analysis by Comparison with STARDYNE, MSC/NASTRAN Users’
Conf., March, 1978.

DYNAMICS – SEISMIC
Bonaldi, P.; Peano, A.; Ruggeri, G.; Venturuzzo, M. Seismic and Impact Analyses of Nuclear
Island Buildings of Italian Unified Nuclear Design, Proc. of the 15th MSC/NASTRAN
Eur. Users’ Conf., October, 1988.
Main Index
 APPENDIX L 539
References and Bibliography

Burroughs, John W. An Enhancement of NASTRAN for the Seismic Analysis of Structures,

Ninth NASTRAN Users’ Colloq., pp. 79-96, October, 1980, (NASA CP-2151).

Chen, J. T.; Chyuan, S. W.; Yeh, C. S.; Hong, H. K. Comparisons of Analytical Solution and
MSC/NASTRAN Results on Multiple-Support Motion of Long Bridge During
Earthquake, The 4th MSC Taiwan Users’ Conf., Paper No. 20, November, 1992.

Chen, J. T.; Hong, H. K.; Yen, C. S.; Chyuan, S. W. Integral Representations and Regularizations
for a Divergent Series Solution of a Beam Subjected to Support Motions, Earthquake
Engineering and Structural Dynamics, Vol. 25, pp. 909-925, 1996.

Chen, Yohchia. Nonlinear Seismic Analysis of Bridges: Practical Approach and Comparative
Study, The MSC 1993 World Users’ Conf. Proc., Paper No. 57, May, 1993.

Chen, Yohchia. Refined Analysis for Soil-Pipe Systems, MSC 1994 World Users’ Conf. Proc.,
Paper No. 38, June, 1994.

Dahlgren, F.; Citrolo, J.; Knutson, D.; Kalish, M. Dynamic Analysis of the BPX Machine
Structure, Proc. of the 14th IEEE/NPSS Symp. on Fusion Engineering, Vol. 1, pp. 47-49,
1992.

Fallet, P.; Derivery, J. C. MSC/NASTRAN Earthquake Analysis, NASTRAN User’s Conf., May,
1979.

Hill, Ronald G. Nonlinear Seismic Analysis of a Reactor Structure with Impact Between Core
Components, NASTRAN: Users’ Exper., pp. 397-418, September, 1975, (NASA TM
X-3278).

Hirata, M.; Ishikawa, K.; Korosawa, M.; Fukushima, S.; Hoshina, H. Seismic Analysis of
Plutonium Glovebox by MSC/NASTRAN, January, 1993.

Liepins, Atis A.; Nazemi, Hamid. Virtual Mass of Fluid in Egg-Shaped Digesters, The MSC 1993
World Users’ Conf. Proc., Paper No. 77, May, 1993.

Nomura, Yoshio. Seismic Response Analysis by MSC/NASTRAN for Coupled Structure -

Ground and Pile, The Fifth MSC/NASTRAN User’s Conf. in Japan, October, 1987, in
Japanese.

Pamidi, M. R.; Pamidi, P. R. Modal Seismic Analysis of a Nuclear Power Plant Control Panel
and Comparison with SAP IV, NASTRAN: Users’ Exper., pp. 515-530, October, 1976,
(NASA TM X-3428).

Tsaur, D. H.; Chyuan, S. W.; Chen, J. T. On the Accuracy of MSC/NASTRAN on Response of

Two-Span Beams to Spatially Varying Seismic Excitation, The 4th MSC Taiwan Users’
Conf., Paper No. 9, November, 1992.

Yiak, K. C.; Pezeshk, S. Seismic Study of a Cable-Stayed Bridge, Proceedings of the Structures
Congrees ’94, Atlanta, 1994.

Main Index
540

Zhou, Hongye; Chen, Youping. The Influence of Phase-Difference Effects on Earthquake

Response of Cable-Stayed Bridges, MSC 1994 World Users’ Conf. Proc., Paper No. 37,
June, 1994.

DYNAMICS – TRANSIENT ANALYSIS

(Korean). Vibration Analysis for Outercase in Drum Washer and Floor, MSC 1994 Korea Users’
Conf. Proc., December, 1994, in Korean.

Aslani, Chang; Yatheendar, Manicka; Visintainer, Randal, H.; Rohweder, David S.; Lopez de
Alda, Juan. Simulation of Proving Ground Events for Heavy Truck Cabs Using Adams,
MSC/NASTRAN, and P/FATIGUE, MSC 1994 World Users’ Conf. Proc., Paper No. 5, June
1994.

Barber, Pam; Arden, Kevin. Dynamic Design Analysis Method (DDAM) Using
MSC/NASTRAN, MSC 1994 World Users’ Conf. Proc., Paper No. 31, June, 1994.

Barnett, Alan R.; Ibrahim, Omar M.; Sullivan, Timothy L.; Goodnight, Thomas W. Transient
Analysis Mode Participation for Modal Survey Target Mode Selection Using
MSC/NASTRAN DMAP, MSC 1994 World Users’ Conf. Proc., Paper No. 8, June, 1994.

Bellinger, Dean. Dynamic Analysis by the Fourier Transform Method with MSC/NASTRAN,
MSC 1995 World Users’ Conf. Proc., Paper No. 10, May, 1995.

Chargin, M. L.; Dunne, L. W.; Herting, D. N. Nonlinear Dynamics of Brake Squeal, MSC 1996
World Users’ Conf. Proc., Vol. V, Paper No. 47, June, 1996.

Cifuentes, A. O.; Herting, D. N. Transient Response of a Beam to a Moving Mass Using a Finite
Element Approach, Innovative Numerical Methods in Engineering, Proc. of the Fourth
Int. Symp. on Numerical Methods in Engineering, Springer-Verlag, pp. 533-539, March,
1986.

Cifuentes, A. O.; Lalapet, S. Transient Response of a Plate to an Orbiting Mass, Proc. of the
Second Panamerican Cong. of Applied Mechanics, January, 1991.

Cifuentes, A. O.; Lalapet, S. A General Method to Determine the Dynamic Response of a Plate
to an Orbiting Mass, Computers and Structures, Vol. 42, No. 1, pp. 31-36, 1992.

Dai, Chung C.; Yang, Jackson C. S. Direct Transient Analysis of a Fuse Assembly by
Axisymmetric Solid Elements, Thirteenth NASTRAN Users’ Colloq., pp. 431-452, May,
1985, (NASA CP-2373).

Deloo, Ph.; Klein, M. In-Orbit Disturbance Sensitivity Analysis of the Hubble Space Telescope
New Solar Arrays, Proc. of the 19th MSC European Users’ Conf., Paper No. 11,
September, 1992.

Everstine, Gordon C. A NASTRAN Implementation of the Doubly Asymptotic Approximation

for Underwater Shock Response, NASTRAN: Users’ Exper., pp. 207-228, October, 1976,
(NASA TM X-3428).

Main Index
 APPENDIX L 541
References and Bibliography

Flanigan, Christopher C. Accurate Enforced Motion Analysis Using MSC/NASTRAN

Superelements, MSC 1994 World Users’ Conf. Proc., Paper No. 25, June, 1994.

Frye, John W. Transient Analysis of Bodies with Moving Boundaries Using NASTRAN,
NASTRAN: Users’ Exper., pp. 377-388, September, 1975, (NASA TM X-3278).

Hirata, M.; Ishikawa, K.; Korosawa, M.; Fukushima, S.; Hoshina, H. Seismic Analysis of
Plutonium Glovebox by MSC/NASTRAN, January, 1993.

Katnik, Richard B.; Deutschel, Brian; Cherukuri, Ravi. Transient Response of a Vehicle Over
Road Bumps Using the Fourier Transform in a Modal Subspace, The MSC 1992 World
Users’ Conf. Proc., Vol. I, Paper No. 6, May, 1992.

Kim, Hyoung M.; Bartkowicz, Theodoore J.; Van Horn, David A. Data Recovery and Model
Reduction Methods for Large Structures, The MSC 1993 World Users’ Conf. Proc., Paper
No. 23, May, 1993.

Larkin, Paul A.; Miller, Michael W. STS Coupled Loads Analysis Using MSC/NASTRAN,
MSC/NASTRAN Users’ Conf. Proc., Paper No. 18, March, 1985.

Lee, J. H.; Tang, J. H. K. Dynamic Response of Containments Due to Shock Wave, Proc. of the
Int. Conf. on Containment Design, pp. 25-32, June, 1984.

Lee, Sang H.; Bock, Tim L.; Hsieh, Steve S. Adaptive Time Stepping Algorithm for Nonlinear
Transient Analysis, The MSC 1988 World Users Conf. Proc., Vol. II, Paper No. 54, March,
1988.

Leifer, Joel; Gross, Michael. Non-Linear Shipboard Shock Analysis of the Tomahawk Missile
Shock Isolation System, 58th Shock and Vibration Symp., Vol. 1, pp. 97-117, October,
1978, (NASA CP-2488).

Lipman, Robert R. Computer Animation of Modal and Transient Vibrations, Fifteenth

NASTRAN Users’ Colloq., pp. 111-117, August, 1987, (NASA CP-2481).

Mattana, G.; Miranda, D. MSC/NASTRAN Applications in P-180 Analysis, Proc. of the

MSC/NASTRAN First Italian Users’ Conf., October, 1987.

McMeekin, Michael; Kirchman, Paul. An Advanced Post Processing Methodology for Viewing
MSC/NASTRAN Generated Analyses Results, MSC 1994 World Users’ Conf. Proc., Paper
No. 21, June, 1994.

McNamee, Martin J.; Zavareh, Parviz. Nonlinear Transient Analysis of a Shock Isolated
Mechanical Fuse, The MSC 1990 World Users Conf. Proc., Vol. I, Paper No. 21, March,
1990.

Neilson, H. C.; Everstine, G. C.; Wang, Y. F. Transient Response of Submerged Fluid-Coupled

Double-Walled Shell Structure to a Pressure Pulse, J. of the Acoustic Soc. of America,
Vol. 70, No. 6, pp. 1776-1782, December, 1981.

Pamidi, P. R. On the Append and Continue Features in NASTRAN, Seventh NASTRAN Users’
Colloq., pp. 405-418, October, 1978, (NASA CP-2062).
Main Index
542

Rose, Ted. A Method to Apply Initial Conditions in Modal Transient Solutions, The MSC 1991
World Users’ Conf. Proc., Vol. I, Paper No. 13, March, 1991.

Rose, Ted.; McNamee, Martin. A DMAP Alter to Allow Amplitude- Dependent Modal Damping
in a Transient Solution, MSC 1996 World Users’ Conf. Proc., Vol. V, Paper No. 50, June,
1996.

Swan, Jim. A DMAP Alter for Interface Loads Across Superelements in Dynamic Analyses, The
MSC 1992 World Users’ Conf. Proc., Vol. I, Paper No. 23, May, 1992.

Tang, C. C. Space Station Freedom Solar Array Wing: Nonlinear Transient Analysis of Plume
Impingement Load, MSC 1994 World Users’ Conf. Proc., Paper No. 35, June, 1994.

Urban, Michael R.; Dobyns, Alan. MSC/NASTRAN Transient Analysis of Cannon Recoil Loads
on Composite Helicopters, The MSC 1991 World Users’ Conf. Proc., Vol. I, Paper No. 28,
March, 1991.

West, Timothy S. Approximate Dynamic Model Sensitivity Analysis For Large, Complex Space
Structues, MSC 1996 World Users’ Conf. Proc., Vol. I, Paper No. 6, June, 1996.

Wingate, Robert T.; Jones, Thomas C.; Stephens, Maria V. NASTRAN Postprocessor Program
for Transient Response to Input Accelerations, NASTRAN: Users’ Exper., pp. 707-734,
September, 1973, (NASA TM X-2893).

Yang, Jackson C. S.; Goeller, Jack E.; Messick William T. Transient Analysis Using Conical Shell
Elements, NASTRAN: Users’ Exper., pp. 125-142, September, 1973, (NASA TM X-2893).

Main Index
 543

I N D E X
MSC.Nastran Basic Dynamics User’s Guide

I N D E X
MSC.Nastran
Basic Dynamics
User’s Guide A CMASSi, 17, 21, 22, 197, 198
Complex arithmetic, 106
ACCELERATION, 131, 177 Complex eigenvalue analysis, 314
Acceleration, 3 Component mode synthesis, 325
Acoustic analysis, 319 Computer resource requirements, 308
Aeroelastic analysis, 329 CONEAX, 17
Aeroelastic flutter, 329 CONM1 Bulk Data entry
ASETi, 313 specification of, 387
Automatic Givens method, 44, 46 CONM2 Bulk Data entry
Automatic Householder method, 44, 46, 57 specification of, 388
CONMi, 17, 21, 22, 197, 198
CONROD, 17
B Consistent mass, 18, 19, 20
Control system, 328
B2GG, 30, 152
B2PP, 30, 152 Coupled mass, 17, 18, 19, 20
CPENTA, 17, 76
Beating, 181
CQUAD4, 17, 71
CQUAD8, 17
C CQUADR, 17
CRAC2D, 17
CBAR, 17, 81 CRAC3D, 17
CBEAM, 17 Critical damping, 6
CBEND, 17 CROD, 17, 18
CDAMP1 Bulk Data entry CSHEAR, 17
specification of, 381 CTETRA, 17
CDAMP2 Bulk Data entry CTRIA3, 18, 71
specification of, 381 CTRIA6, 18
CDAMP3 Bulk Data entry CTRIAR, 18
specification of, 382 CTRIAX6, 18
CDAMP4 Bulk Data entry CTUBE, 18
specification of, 383 CVISC, 26, 112, 152, 160
CDAMPi, 25, 112, 152, 160 CVISC Bulk Data entry
CELASi, 71 specification of, 389
Characteristic frequency, 34
Characteristic shape, 34
CHEXA, 17, 76
Circular natural frequency, 5, 37
D
Index
CMASS Bulk Data entry Damped circular natural frequency, 7
Damping, 23, 302
specification of, 386
CMASS2 Bulk Data entry Damping ratio, 7
DAREA, 117, 121, 122, 132, 133, 143, 145, 147,
specification of, 385
CMASS3 Bulk Data entry 166, 169, 171, 178, 179, 185, 193, 197, 201,
214
specification of, 386
Main Index
 544 INDEX

DAREA Bulk Data entry Eigenvalue extraction methods, 43

specification of, 390 comparison, 46
DELAY, 117, 122, 132, 166, 170, 178, 179 Givens, 43
DELAY Bulk Data entry Householder, 43
specification of, 391 inverse power, 43
Design optimization, 326 Lanczos, 43
Design sensitivity, 327 modified Givens, 43
Direct frequency response analysis, 106, 114 modified Householder, 43
damping, 106 Sturm modified inverse power, 43
excitation definition, 115 Eigenvalue problem, 35, 37, 470
Direct matrix input, 29 Eigenvector, 35, 37, 38
Direct transient response analysis, 151, 162 MASS normalization, 41
damping, 152 MAX normalization, 41
excitation definition, 163 orthogonality, 39
initial conditions, 154 POINT normalization, 41
DISPLACEMENT, 54, 67, 131, 141, 177 scaling, 40
DLOAD, 197, 201 strain energy, 42
DLOAD (Bulk Data), 116, 124, 132, 133, 143, EIGR, 50, 54, 55, 57, 90, 112, 132, 161, 178
164, 172, 178, 179, 185, 214 EIGR Bulk Data entry
DLOAD (Case Control), 116, 121, 122, 123, specification of, 396
130, 135, 143, 164, 165, 169, 171, 176, 185 EIGRL, 48, 54, 55, 67, 112, 132, 133, 135, 143,
DLOAD Bulk Data entry 161, 178, 179, 185
specification of, 392 EIGRL Bulk Data entry
DMAP, 331 specification of, 400
DMASS1 Bulk Data entry ELFORCE, 131, 141, 177
specification of, 384 ELSTRESS, 131, 177
DMIG, 30, 81 Enforced motion, 196
DMIG Bulk Data entry EPOINT, 328
specification of, 393 ESE, 67
DPHASE, 117, 118, 132, 133, 143
DPHASE Bulk Data entry
specification of, 396
Dynamic absorbers, 138
F
File Management Section (FMS), 443
Dynamic aeroelastic response, 330
FORCE, 123, 131, 172, 177
Dynamic amplification factor, 9, 10
Forced vibration analysis, 8
Dynamic degrees of freedom, 3
Free vibration analysis, 4
Dynamic reduction, 313
FREQ Bulk Data entry
DYNRED, 313
specification of, 404
FREQ1 Bulk Data entry
specification of, 404
E FREQ2 Bulk Data entry
EIGC, 314 specification of, 405
Eigenvalue, 35, 38 FREQ3 Bulk Data entry
specification of, 406, 408, 410
FREQi, 132, 133, 135, 143, 302
Index
FREQUENCY, 130, 135, 143
Frequency increment, 302
Frequency response analysis, 104

Main Index
 INDEX 545

Frequency response guidelines Lanczos method, 43, 46, 48, 67, 71, 76, 91, 475
damping, 302 block size, 49
frequency increment, 302 diagnostic output, 49
number of retained modes, 302 normalizes, 50
Fundamental frequency, 34 shift, 49, 91
Fundamental shape, 34 Large mass method, 197, 199, 201, 477
LOADSET, 122, 123, 130, 171, 176
Logarithmic decrement, 295
G LSEQ, 122, 123, 132, 133, 145, 147, 170, 172,
178, 179, 193
Generalized dynamic reduction, 313
LSEQ Bulk Data entry
Generalized mass, 39, 58
specification of, 411
Generalized stiffness, 39, 58
Lumped mass, 17, 18, 20
Geometric nonlinearity, 321
Givens method, 44
GPSTRESS, 177
Gravity, acceleration of, 21 M
GRDSET, 96 M2GG, 21, 30, 85
Grid point weight generator, 454 M2PP, 21, 30
Guidelines, 288 Mass, 17
Guyan reduction, 313 grid point weight generator, 454
mass density, 21, 60
PARAM,COUPMASS, 17
H PARAM,WTMASS, 21
units, 21
Halfpower bandwidth, 129
weight density, 21, 60
Half?power bandwidth, 295
Mass matrix, 37
Hertz, 5
Material nonlinearity, 321
Householder method, 44
MATi, 22, 107, 112, 152, 160
Hydroelastic analysis, 319
GE, 25
RHO, 22
Matrix method, 112, 161
I Mechanism, 39, 88, 473
IC, 154, 176, 180 Memory requirements, 308
INCLUDE, 71 METHOD, 130, 135, 143, 176, 185
Include files, 71 Modal frequency response analysis, 108, 114
Initial conditions, 5 damping, 109
Inverse power method, 45 excitation definition, 115
mode truncation, 112
Modal transient response analysis, 156, 162
K damping, 157
excitation definition, 163
K2GG, 30, 84
mode truncation, 160
K2PP, 30
Mode acceleration method, 318
Mode displacement method, 112, 161
Mode shape, 34, 37
L MODES, 67 Index
Lanczos, 308 Modified Givens method, 44
Modified Householder method, 44
Multiple boundary conditions, 78

Main Index
 546 INDEX

N PMASS Bulk Data entry

specification of, 413
Natural frequency, 5, 20, 34 PostScript, 243, 249
NOLINi, 323, 328 Principal mass axes, 458
Nonlinear transient response analysis, 321 PVISC Bulk Data entry
Nonstructural mass, 22, 60 specification of, 413
Normal frequency, 34
Normal modes analysis, 35, 37
Numerical accuracy, 469 Q
QSET/QSET1, 313
O Quality factor, 24

OFREQUENCY, 130
OLOAD, 131, 141, 177, 298
OMITi, 313
R
Orthogonality, 39 Random vibration analysis, 317
Rayleigh's equation, 39
OTIME, 176
Overdamped system, 6 RBE2, 199
Repeated roots, 42
Resonance, 9, 10
P Resonance frequency, 34
Resonant frequency, 34
P2G, 30 Response spectrum analysis, 315
PARAMs Restarts, 217
BAILOUT, 475, 488, 491 Rigidbody drift, 199
CM1, 21 Rigidbody modes, 39, 88
CM2, 21 RLOAD1, 115, 122, 123, 133, 145, 147, 193
COUPMASS, 17, 66 RLOAD1 Bulk Data entry
DDRMM, 113, 161, 244, 246, 318 specification of, 414
FREQ, 112 RLOAD2, 116, 133, 143
G, 25, 107, 152, 153 RLOAD2 Bulk Data entry
GRDPNT, 71, 289, 455 specification of, 416
HFREQ, 112, 161
KDAMP, 112
LFREQ, 161, 199, 215
LMODES, 112
S
MAXRATIO, 93, 473 SACCELERATION, 131, 177
SDAMPING, 130, 135, 143, 176, 185, 294
MODACC, 318
MODES, 161 SDISPLACEMENT, 131, 141, 177
Single degreeoffreedom (SDOF) system, 3
W3, 25, 152, 153, 295
W4, 25, 152, 153, 295
WTMASS, 21, 60, 198, 214, 293, 457
PBAR, 22
PDAMP Bulk Data entry
specification of, 412
Index Period of response, 5
Phase lead, 10
Phase shift, 104
PLOAD1, 124, 172
PLOAD4, 133, 145, 147, 179, 193
Plotting, 241

Main Index
 INDEX 547

SOLs TABLED1 Bulk Data entry

107, 314 specification of, 422
108, 130 TABLED2, 120, 168, 201
109, 176, 323 TABLED2 Bulk Data entry
11, 130 specification of, 423
110, 314 TABLED3, 120, 168, 201
111, 130 TABLED3 Bulk Data entry
112, 176, 323 specification of, 424
12, 176 TABLED4, 121, 168, 201, 209
129, 321 TABLED4 Bulk Data entry
200, 326 specification of, 426
28, 314 TABLEDi, 116, 117, 132, 165, 178
29, 314 TF, 328
3, 54 TIC, 154, 178, 179, 180
67, 314 TIC Bulk Data entry
70, 314 specification of, 426
8, 130 Time step, 152, 157, 174, 304
9, 176 TLOAD1, 163, 170, 172, 179, 198, 203, 214
99, 321 TLOAD1 Bulk Data entry
SORT1, 131, 133 specification of, 427
SORT2, 131, 133 TLOAD2, 165, 179, 185, 198, 203
SPCFORCES, 131, 177 TLOAD2 Bulk Data entry
SPOINT, 313, 328 specification of, 429
Static condensation, 313 Transient response analysis, 150
Static degrees of freedom, 3 Transient response guidelines
Stiffness matrix, 37 damping, 305
STRAIN, 131, 177 duration of computed response, 305
STRESS, 67, 131, 177 integration time step, 304
Structural damping, 23 number of retained modes, 304
Structure plotting, 243 TSTEP, 304
Sturm modified inverse power method, 45, TSTEP (Bulk Data), 174, 177, 179, 180, 185
46, 60 TSTEP (Case Control), 174, 176, 180, 185
Substructuring, 325 TSTEP Bulk Data entry
Superelement analysis, 325 specification of, 432
SUPORT, 90, 91, 92, 93, 94, 96, 199 Tuned mass dampers, 138
SUPORT Bulk Data entry
specification of, 418
SUPORT1 Bulk Data entry
specification of, 419
U
Underdamped system, 6
SVECTOR, 131, 177
Units of measure, 27
SVELOCITY, 131, 177
User messages
Symmetry, 78
3035, 90, 98, 100
784, 226
T
TABDMP1, 110, 132, 133, 135, 143, 158, 178,
179, 185, 294
V Index

VECTOR, 54, 131, 177

TABDMP1 Bulk Data entry
VELOCITY, 131, 177
specification of, 420
Velocity, 3
TABLED1, 119, 122, 124, 133, 143, 147, 167,
Virtual fluid mass, 319
170, 172, 179, 193, 201, 214
Viscous damping, 23
Main Index
 548 INDEX

X
XY plotting, 249

Index

Main Index