Nastran 2007r1 Release Guide

MSC Nastran 2007 r1
Release Guide
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: (03)-6911-1200
Fax: (03)-6911-1201
Worldwide Web
www.mscsoftware.com
Disclaimer
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  2007 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.
This software may contain certain third-party software that is protected by copyright and licensed from
MSC.Software suppliers.
MSC, MD, Dytran, Marc, MSC Nastran, MD Nastran, MSC Patran, MD Patran, the MSC.Software corporate logo,
and Simulating Reality are trademarks or registered trademarks of the MSC.Software Corporation in the United
States and/or other countries.
NASTRAN is a registered trademark of NASA. PAMCRASH 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. All other brand names, product
names or trademarks belong to their respective owners.
k^WOMMToNWwWwWwWa`Jobi
Contents
MSC Nastran 2007 r1 Release Guide
jp`=k~ëíê~å=OMMT=
oÉäÉ~ëÉ=dìáÇÉ
List of Books x Table of Contents
Technical Support xi
Internet Resources xiii
1 Overview of MSC Nastran 2007 r1

Overview 2
Implicit Nonlinear Analysis (SOL 600) 2
NVH & Acoustics 2
Numerical Enhancements 2
Elements & Connectors 3
Optimization 3
Rotor Dynamics & Aeroelasticity 3
List of MSC Nastran Documents Released with MSC Nastran 2007 r1 4
2 Implicit & Explicit Nonlinear Analysis

MSC Nastran Implicit Nonlinear - SOL 600 6
Heat Transfer 6
Heat Transfer Examples 9
Creep Simulations 9
Element Selection 9
Other SOL 600 Items 9
Fracture Mechanics J-Integral (LORENZI) 11
Delamination 11
MATEP Extensions 12
SOL 600 Failure Description – MATF 12
Element Birth and Death 16
Unglue 16
Composite Element Numerical Analysis 16
PLOAD4 Extensions 17
Large Rotation RBE 17
Streaming Input 17
CONNECTOR TECHNOLOGY 18
COMPUTATIONAL ENHANCEMENTS 19
iv MSC Nastran 2007 r1 Release Guide
==
SUPER ELEMENTS / DMIG 19

CONTACT ENHANCEMENTS 19
New SOL 600 Parameters 20
Platform Specific Notes 21
For 64-bit Windows EM64T platforms: 22
For 32-bit Windows platforms: 22
Supported Systems for SOL 600 in MSC Nastran 2007 23
3 NVH & Acoustics

Rigid Porous Absorber 26
Introduction 26
Porous Materials 26
Inputs 27
Discussion 28
References 28
Example 28
4 Numerical Enhancements
New SPARSESOLVER Executive Statement 34
Introduction 34
Benefits 34
Method and Theory 34
Inputs 34
Outputs 34
Guidelines and Limitations 34
Improved Performance with New Sparse Solvers 37

Introduction 37
The TAUCS Sparse Solver 37
User Interface 37
Use Cases 37
Limitations 38
The UMFPACK Sparse Solver 39
Improved Memory Usage in Lanczos (Pre-release) 40
New MAXRATIO Information Output (Pre-release) 41

Introduction 41
Benefits 41
Method and Theory 41
Inputs 41
Outputs 41
Contents v

Demonstration Example 42
Example Input Data 42
Example Output 44
Performance Improvements and Expanded Capabilities for ACMS 45

Introduction 45
MDACMS for Upstream Superelements 45
Automatic FASTFR Decision Logic 46
Examples 47
Limitations for the FASTFR method 48
Miscellaneous Performance Improvements 49
Relaxed Restrictions for CASI Solver Usage 52
System Dependent Performance Improvements 53

Introduction 53
Model Description 53
Performance Improvements for Intel x86_64 54
Performance Improvements for AMD due to Atlas 54
Improved Selection of Reordering Methods 55
5 Elements & Connectors

Enhancements to Connector Elements 58
Introduction 58
CSEAM Elements 58
SWLDPRM Enhancements 61
Displacement Output of GA and GB for CWELD and CFAST Elements 62
Nonhomogeneous Multipoint Constraint 66

Introduction 66
Benefits 66
Output 66
Theory 66
Examples 68
6 Optimization
Topology Optimization Enhancements 74
Introduction 74
Benefits 74
vi MSC Nastran 2007 r1 Release Guide
==
Modified Type One Responses - Fractional Mass 76

Example 1 – Wheel (wheeltop.dat) 77
Automatic External Superelement Optimization (AESO) 80

Introduction 80
Benefits 80
Methodology 80
Input 81
Outputs 83
Examples 88
Miscellaneous 91
Randomization of a User’s Input Data File (Pre-release) 95

Introduction 95
Benefits 95
Random Elimination of Element Types (Pre-release) 97

Introduction 97
Benefits 97
7 Rotor Dynamics & Aeroelasticity

Changes to Rotordynamics for MSC Nastran 2007 100
Unbalance Entry for Frequency Response 100
New Parameter and Hybrid Damping Specifications 100
Updating/Summing of Monitor Points 101

Introduction 101
Benefits 101
Examples (monsum.dat and monsum3.dat) 101
Stripwise Aerodynamic Results 103

Introduction 103
Benefits 103
Examples (moncncm and monbodi) 104
Input of an Aerodynamic Mesh 105

Introduction 105
Benefits 105
Contents vii
Example (aegridf) 105

Rigid Body Spline 109

Introduction 109
Benefits 109
Example (splinrb) 109
Wendland Spline Functions for the Spline4/5 110

Introduction 110
Benefits 110
Example (ha144c_ris) 110
Spline Blending 111

Introduction 111
Benefits 111
Example (ha145e_blnd) 111
Export of the Spline Matrix 113

Introduction 113
Benefits 113
Examples (splinopch and splinoop2) 113
viii MSC Nastran 2007 r1 Release Guide
==
The 2005 New Template
mêÉÑ~ÅÉ
å List of Books
å Technical Support
å Internet Resources
x MSC Nastran 2007 r1 Release Guide
List of Books
Below is a list of some of the MD Nastran and MSC Nastran documents. You may order any of these
documents from the MSC.Software BooksMart site at http://store.mscsoftware.com/.
fåëí~ää~íáçå=~åÇ=oÉäÉ~ëÉ=dìáÇÉë
ç Installation and Operations Guide
ç Release Guide
oÉÑÉêÉåÅÉ=_ççâë
ç Quick Reference Guide
ç DMAP Programmer’s Guide
ç Reference Manual
rëÉêÛë=dìáÇÉë
ç Getting Started
ç Linear Static Analysis
ç Basic Dynamic Analysis
ç Advanced Dynamic Analysis
ç Design Sensitivity and Optimization
ç Thermal Analysis
ç Numerical Methods
ç Aeroelastic Analysis
ç Superelement
ç User Modifiable
ç Toolkit
ç Implicit Nonlinear (SOL 600)
ç Explicit Nonlinear (SOL 700)
ç MD User’s Guide
Preface xi
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.
tÉÄ 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, training
courses, and documentation updates at the MSC.Software Training, Technical Support, and
Documentation web page.
mÜçåÉ= råáíÉÇ=pí~íÉë cêáãäÉóI=`~ãÄÉêäÉó

~åÇ= Telephone: (800) 732-7284 pìêêÉóI=råáíÉÇ=háåÖÇçã
c~ñ Fax: (714) 784-4343 Telephone: (44) (1276) 60 19 00
Fax: (44) (1276) 69 11 11
jìåáÅÜI=dÉêã~åó qçâóçI=g~é~å
Telephone: (49) (89) 43 19 87 0 Telephone: (81) (03) 6911 1200
Fax: (49) (89) 43 61 71 6 Fax: (81) (03) 6911 1201
oçãÉI=fí~äó m~êáëI=cê~åÅÉ
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
jçëÅçïI=oìëëá~ dçìÇ~I=qÜÉ=kÉíÜÉêä~åÇë
Telephone: (7) (095) 236 6177 Telephone: (31) (18) 2543700
Fax: (7) (095) 236 9762 Fax: (31) (18) 2543707
j~ÇêáÇI=pé~áå
Telephone: (34) (91) 5560919
Fax: (34) (91) 5567280
xii MSC Nastran 2007 r1 Release Guide
bã~áä 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.
m~íê~å=pìééçêí ãëÅé~íê~åKëìééçêí]ãëÅëçÑíï~êÉKÅçã
jp`=k~ëíê~å=pìééçêí ãëÅå~ëíê~åKëìééçêí]ãëÅëçÑíï~êÉKÅçã
aóíê~å=pìééçêí ãëÅÇóíê~åKëìééçêí]ãëÅëçÑíï~êÉKÅçã
jp`=c~íáÖìÉ=pìééçêí ãëÅÑ~íáÖìÉKëìééçêí]ãëÅëçÑíï~êÉKÅçã
j~êÅ=pìééçêí ãëÅã~êÅKëìééçêí]ãëÅëçÑíï~êÉKÅçã
jp`=fåëíáíìíÉ=`çìêëÉ=fåÑçêã~íáçå ãëÅíê~áåáåÖKëìééçêí]ãëÅëçÑíï~êÉKÅçã
qê~áåáåÖ
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 markets. 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.
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.
Preface xiii
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.
xiv MSC Nastran 2007 r1 Release Guide
Ch. 1: Overview of MSC Nastran 2007 r1 MSC Nastran 2007 Release Guide
1 Overview of MSC Nastran 2007 r1

Overview
2 MSC Nastran 2007 r1 Release Guide
Overview
The MSC Nastran 2007 r1 release brings powerful new features and enhancements in the areas of high
performance computing, nonlinear analysis, assembly modeling, optimization, rotor dynamics and
aeroelasticity.
Implicit Nonlinear Analysis (SOL 600)

• Heat Transfer. Advanced thermal analysis is now available in SOL 600 including an efficient
hemi-cube viewfactor calculation method and automated procedures for thermal stress analysis
based upon the heat transfer simulation. Thermal analysis of composites may include accurate
calculation of the thermal gradient through the thickness.
• Modeling Enhancements. Connector technology has been enhanced to include large deformation
formulations of CFAST, CWELD and CBUSH. Enhancements for facture mechanics include the
calculation of stress intensity factors using either the VCCT or Lorenzi method and the
prediction of delamination.
• Performance Improvements. A new streaming input option is available that eliminates transfer
files. The analysis of composite shells has been improved such that assembly time has been often
reduced by a factor of 10 and memory requirements have also substantially been reduced. Other
performance improvements include an out-of-core iterative and direct PLOAD4 support.
NVH & Acoustics

• Rigid Porous Absorber. A new capability for modeling basic rigid skeleton porous absorber
characteristics in acoustic response analysis such as vehicle seats and absorber linings.
Numerical Enhancements
• Sparse Solvers. Two new sparse solvers have been introduced; TAUCS (statics) and UMFPACK
(unsymmetric). UMFPACK solver provides scalable performance for unsymmetric frequency
response problems. In addition, Lanczos has been enhanced to take advantage of available
memory. Automatic optimal reordering selection has been implemented for solid models to
eliminate having the user set flags.
• Iterative Solvers. Restrictions to the CASI iterative solver have been relaxed for statics of large
solid models (engines). This includes an expanded list of supported element types.
• ACMS. Automated Component Modal Synthesis (ACMS) has been extended to External
Superelements to provide significant reductions in compute time, I/O and scratch space. One
typical case study demonstrates an order of magnitude improvement.
• Other HPC enhancements. MSC Nastran 2007 has been ported to Microsoft Compute Cluster.
The compute kernels for x86_64 platforms have been optimized for both Intel and AMD based
systems. Improved user diagnostics provides a pivot ratio bar chart to localize model
singularities.
CHAPTER 1 3
Overview of MSC Nastran 2007 r1
Elements & Connectors

• Connectors. A new seamweld (CSEAM) connector element is now available for assembly
modeling. It features extended capabilities for connecting higher order elements, mesh
independent connections to top/bottom shell patches defined either by property IDs or Element
IDs, tailored parts connection, and support for anisotropic material properties. For spot weld
elements (CWELD, CFAST) end point displacement output can be obtained to view the
relationship between the spot weld and the connecting shells. A new connector type RBE2GS is
introduced to optionally search and connect independent grids of the two closest RBE2 elements
with a specified search radius.
Optimization
• Topology Optimization. This release provides combined topology, sizing and shape
optimization simultaneously to find possible better designs. Different mass targets can now be
applied on multiple design parts of the structure. Symmetry constrains have been extended to
cyclical applications such as car wheels. An adjoint design sensitivity analysis method has been
implemented for inertia relief sizing optimization. Significant performance enhancement with
minimum member size control has been achieved particularly for large number of sizing design
variables.
• Automatic External Superelement Optimization (AESO). This new feature automatically
partitions the model into a designed and non-designed part (external Superelement) for efficient
optimization. Order of magnitude speedup can be achieved without requiring user knowledge of
superlements.
• Randomization (Pre-release). This randomization capability provides a way to stochastically
introduce uncertainty into a model such as tolerances in connectivity, properties and loads. The
user selects the outputs to monitor. This beta capability is a first step in developing a multi-run
environment to spawn multiple jobs, collect the results and perform statistical post-processing.
Rotor Dynamics & Aeroelasticity

• Rotor Dynamics. Unbalance loading can now be used for frequency response with the rotor
dynamics option. Frequency response case control can be used directly in SOL 146 since we can
now handle multiple RGYRO subcases in rotor dynamics. Damping specification has been
simplified and allow for new damping formulations such as hybrid damping. As a prerelease
only capability: the effects of rotor stiffness, mass, and damping effects can be included in SOL
200 optimization.
• Aeroelasticity. Monitoring points can now be updated and summed. A new type of monitoring
point (MONCNCM) has been introduced for monitoring stripwise aerodynamic results such as
lift and pitching moments. Other enhancements include various splining techniques for
aerodynamic structural applications.
List of MSC Nastran Documents Released with MSC Nastran

2007 r1
Along with this Guide, the following documents are updated for the MSC Nastran 2007 release.
• MSC Nastran 2007 Quick Reference Guide
• MSC Nastran Installation and Operations Guide
• MSC Nastran Implicit Nonlinear (SOL 600) User’s Guide
Ch. 2: Implicit & Explicit Nonlinear Analysis MSC Nastran 2007 Release Guide
2 Implicit & Explicit Nonlinear

Analysis

MSC Nastran Implicit Nonlinear - SOL 600
MSC Nastran Implicit Nonlinear - SOL 600

The MSC Nastran 2007 r1 release contains significant enhancements to functionality and performance.
This includes the addition on heat transfer capability, support of Nastran fastener technology (CBUSH,
CFAST and CWELD) in a native mode, improved support of PLOAD4, and an enhanced composite
capabilities. Fracture mechanics may also be performed to obtain the stress intensity factors at a crack.
A new mechanism has been added for the direct transfer of data that is known as streaming input.
Heat Transfer
For heat transfer, most of the capabilities in Nastran SOL 153 and 159 are supported by SOL 600 with
the exception of CHBDYP and forced convection, the equivalents of which are not currently available in
Marc. The main advantage of using SOL 600 for heat transfer over SOL 153 or 159 is that thermal contact
is available directly and that radiation view factors may possibly be calculated faster. The user needs to
weight the drawbacks of not having CHBDYP and forced convection. Because of these alternatives,
SOL 600 offers two ways to perform a heat transfer analysis. The direct (new) method uses Marc to
perform all of the calculations and can support thermal contact that varies during the run. The other
(indirect) method is to calculate the thermal contact conditions (if they are needed) at the start of the run
and perform the rest of the calculations using Nastran SOL 153/159. This option is addressed using a new
option on the SOL 600 entry TSOLVE=M or TSOLVE=N respectively. A typical SOL 600 Executive
Control statement for heat transfer using the direct method would be:
SOL 600,153 TSOLVE=M
A typical Executive Control statement for heat transfer using the indirect method would be:
SOL 600,153 TSOLVE=N
To use the “Thermal Contact” capability released with the Nastran 2005 version either TSOLVE=N
should be used or the TSOLVE option should be left blank (which will support most existing input
decks).
SOL 600 heat transfer addresses conduction, free convection, radiation to space, cavity radiation, thermal
contact and latent heat. Steady state or transient heat transfer calculation may be obtained. All material
properties may be temperature dependent, and the material may be isotropic, orthotropic or anisotropic.
For the direct method, Marc’s table input is used for all applicable input items. The direct method requires
postprocessing using the Marc t16 file. All standard output forms (op2, xdb, f06 and/or punch) are
available using the indirect method.
The temperature history obtained may then be used in a subsequent thermal stress simulation by using
the MINSTAT and MCHSTAT bulk data options to read the temperatures off the t16 file. When used in
conjunction with the MTHERM bulk data option the time steps will be either subdivided or merged to
satisfy the accuracy and convergence requirements of the nonlinear mechanical analysis.
For the directly solutions, when CTRIA3 or CQUAD4 elements are used, the thermal conduction can be
based upon either two methods which is selected on the PSHELL option. Similar to conventional
Nastran, the thermal behavior may be membrane like only, in which case there is no thermal gradient
CHAPTER 2 7
Implicit & Explicit Nonlinear Analysis
through the thickness. To support this, new heat transfer elements have been added and are used when
appropriate. These new elements are:
Element 196 — Three-node, Bilinear Heat Transfer Membrane), 904
Element 197 — Six-node, Biquadratic Heat Transfer Membrane), 907
Element 198 — Four-node, Isoparametric Heat Transfer Element), 911
Element 199 — Eight-node, Biquadratic Heat Transfer Membrane, 915
The second method is that the element has a thermal gradient through the thickness, which may be
required for composite simulation or thermal shock type problems. This is activated by specifying a
nonzero MID2 entry. The MPHEAT options is used to specify, whether the temperature gradient is linear
or quadratic through the total thickness of the shell, or linear or quadratic variation is specified per layer
basis. In the later case if a composite shell has n layers the number of degrees of freedom per grid is n+1
or 2*n+1 for the quadratic case.
The MHEATSHL parameter may also be used to control this behavior.
Additions/changes to a standard Nastran SOL 153 or SOL 159 heat transfer input file are as follows:
Executive Control
Change SOL statement as described above.
Case Control
No changes
Bulk Data
BCBODY, BCTABLE (BCPARA if necessary to change defaults) if there is thermal contact
MPHEAT – New entry that maps to Marc’s HEAT “parameter”. See, “MPHEAT (SOL 600)” on
page 1738 of the MSC Nastran Quick Reference Guide for more information.
NLHEATC - Defines numerical analysis parameters for SOL 600 Heat Transfer Analysis. For more
information please see, “NLHEATC (SOL 600)” on page 1756 of the MSC Nastran Quick Reference
Guide.
MCHSTAT - Option to change state variables for SOL 600 – Used in SOL 600 only. For more
information please see, “MCHSTAT (SOL 600)” on page 1685 of the MSC Nastran Quick Reference
Guide.
MINSTAT - Option to define initial state variables for SOL 600 – Used in SOL 600 only. Please see,
“MINSTAT (SOL 600)” on page 1700 of the MSC Nastran Quick Reference Guide for more information.
Bulk Data Parameters

PARAM,MARCHEAT is obsolete starting with this release.
PARAM,MARHTPRT (Integer) Control heat transfer output in the Marc .out file
0 = Do not print any output except for summary tables
1 = Print the nodal temperatures
2 = Print all possible nodal heat transfer output
PARAM,MRADUNIT (Integer) Controls the units used in radiation heat transfer for SOL 600
1 = Degrees Celsius
2 = Degrees Kelvin (default if parameter not entered)
3 = Degrees Fahrenheit
Remark: Degrees Rankin are not available
PARAM,MHEMIPIX (Integer) Controls the number of pixels used in radiation heat transfer for
SOL 600 using the hemi-cube method. The default, if this parameter is not
entered is 500.
PARAM,MARVFCUT (Real) Controls the fraction of the maximum view factor that is to be used
as a cutoff. View factors calculated below this cutoff are ignored. Default
is 0.0001 if this parameter is not entered (Used in SOL 600 radiation heat
transfer only)
PARAM,MRVFIMPL (Real) Controls the fraction of the maximum view factor that is to be
treated implicitly (contribute to operator matrix). View factor values
smaller than this cutoff are treated explicitly. Default is 0.01 if this
parameter is not entered using this parameter reduces the size of the heat
transfer operator matrix, which reduces the computational costs associated
with decomposition. (Used in SOL 600 radiation heat transfer only)
PARAM,MRSTEADY (Integer) Controls the solution method for SOL 600 steady state heat
transfer
1 Marc STEADY STATE is used with TIME STEP of 1.0 (default if
parameter not entered) The specific heat matrix is not formed.
2 AUTO STEP is used.
Remark: Requires that a sufficiently large time period to be simulated for the
solution to reach steady state.
Bulk Data
BCBODY, BCTABLE (BCPARA if necessary to change defaults) if there is thermal contact. Please see
the MSC Nastran Quick Reference Guide for more details on entries, DMIGOUT, MCHSTAT,
MINSTAT, MTHERM, MPHEAT, NLHEATC.
CHAPTER 2 9
Heat Transfer Examples

The following heat transfer examples are located in the tpl1 directory:
Conduction mhqbd1, mhqbd1a, mhqbd1c, mhqbd1s, mhqbd2, mhqbd2c, mhqbd2s,

mhbc01, mhbc02, mhtepe
Free Convection mhcbv1, mhcbv1a, mhcbv1b, mhcbv1c, mhcbv1d, mhcbv1e
Radiation to Space mhrad1, mhrad2, mhrad3
Cavity Radiation mhrcv1, mhrcv1a, mhrcv2, mhrhx0, mhrdhx, mhrhx4, mhrc1t
Thermal Contact mhcnoc, mhtc07, mhtc7a
Latent Heat mtlat1, mtlat2
Creep Simulations
Creep is an important phenomena in high temperature applications. To facilitate this type of analysis
several new options have been added into SOL 600 in the MSC Nastran 2007 r1 release.
The MPCREEP allows you to select the procedure used to perform the time integration.If the explicit
method is used, then the time step must be small, but is the material is elastic and small deformation, no
reassembly of the stiffness matrix is required. If the implicit method is chosen, then larger time steps may
be used, but reassembly occurs at every increment.
In thermal creep simulations it is necessary that the time step be chosen to satisfy accuracy of both the
rate independent thermal stress problem and the rate dependent creep problem. To insure that this occurs
the MTCREEP bulk data option has been introduced.
Element Selection
To increase the flexibility in performing simulations the MRALIAS param has been augmented by the
ALIASM bulk data entry. This allows one to map the Marc element type to be used for a selection of
elements. All Marc element types may be used if they are topologically similar to the Nastran element
type. See Marc Volume B for greater details.
Other SOL 600 Items

Membrane Elements
SOL 600 now directly supports membrane-only shells without the user having to add any alias bulk data
entries or parameters. This was made possible due to a new element which was added in Marc:
Element 200 — Six-node, Biquadratic Isoparametric Membrane), 920
This new element completes the set of shell elements necessary to support membranes. Users should be
careful when using membranes with nonlinear analyses as they are unstable under certain conditions
because of a lack of bending stiffness.
Fracture Mechanics
New capabilities have been added to SOL 600 to allow greater insight in determining the fracture
resistance of your designs. These methods complement the material damage models that existed
previously and are entered through the MATHED option. The new capabilities include:
• Calculation of Energy Release Rates and Stress Intensity Factors – VCCT and LORENZI
• Option
• Crack Propagation – VCCT
• Delamination – COHESIV
• New Failure Criteria – MATF
• Birth and death of elements
Virtual Crack Closure Technique (VCCT)

Marc’s revised VCCT capability is fully supported by this release of SOL 600. It involves a anew VCCT
Case Control entry and a new VCCT Bulk Data entry (see the MSC Nastran Quick Reference Guide for
details). This option defines that the virtual crack closure technique is to be used for evaluating energy
release rates. The user defines the node (in 2-D or for shells) or nodes (in 3-D) that define each crack tip.
The supported elements are lower- and higher-order 2-D solids and 3-D shells, lower- and higher-order
3-D hexahedral solids and lower order 3-D tetrahedral solids. For 3-D solids it is important that a regular
mesh around the crack front is used.
Multiple cracks can be defined and results obtained for each crack separately. Each crack consists of a
crack tip node in 2-D and for shells and a list of nodes along the crack front for 3-D solids. Shell elements
can be used for defining a 2-D style line crack and also be connected to the face of another shell or 3-D
solid to form a 3-D style surface crack. The different cases are automatically identified.
The VCCT method is advantageous because it may be used with any material model including
orthotropic or anisotropic behavior, and because it automatically obtains the mode I, II, and III stress
intensity factors. This makes is applicable to composite structures.
For crack propagation, there are two modes of growth: fatigue and direct growth. For fatigue style, the
user specifies a load sequence time period. During the load sequence, the largest energy release rate and
the corresponding estimated crack growth direction is recorded. At the end of the load sequence, the
crack is grown using the specified method. For direct growth, the crack grows as soon as the calculated
energy release rate is larger than the user-specified Gc. Note that Gc can be made a function of the
accumulated crack growth length to model a crack growth resistance behavior. This release does not
support large crack propagation which requires remeshing.
CHAPTER 2 11
Fracture Mechanics J-Integral (LORENZI)

This option gives an estimation of the J-Integral for a crack configuration using the domain integration
method. The domain integration method has the advantage that it can also be used for problems with
thermal behavior and for dynamic analysis. This procedure is only available for continuum elements.
Only the nodes defining the crack front (crack tip in two dimensions) need to be defined. The program
automatically finds integrations paths according to the format below. The complete J-Integral is
evaluated and output. For the case of linear elastic material with no external loads on the crack faces, the
program automatically separates mode I, mode II, and mode III (3-D only) stress intensity factors from
the J-Integral. for isotropic materials.
A new Bulk Data entry, LORENZI, is necessary to activate this capability and if entered applies to all
subcases in the analysis. For more details the MSC Nastran Quick Reference Guide.
Delamination
An alternative method to model failure is to use the COHESIV bulk data option in conjunction with
special delamination or interface elements. Three different models are available along with a user
subroutine. The user defines the traction versus the relative separation. The area under the curve is the
cohesive energy, often known as the critical energy release rate.
Bilinear model Exponential Model Linear-Exponential Model
Element Type Number of Nodes Characteristic

186 4 Planar
187 8 Planar
188 8 3-D
189 20 3-D
Element Type Number of Nodes Characteristic

190 4 Axisymmetric
191 8 Axisymmetric
192 6 3-D
193 15 3-D
MATEP Extensions
Material description MATEP was extended to add Chaboche, Power Law, Kumar, Johnson Cook and
other options. For more details the MSC Nastran Quick Reference Guide.
SOL 600 Failure Description – MATF

For SOL 600 failure indices or actual material failure is only described using the MATF entry. For this
release, MATF has been totally revised to accommodate additional types of failure and improve the input
and user understanding of the input. The user should be aware that other Nastran solutions can specify
failure index calculation on various MAT entries. These specifications are not available in SOL 600 –
only MATF may be used. To activate the new entries, the PARAM,MRMATFSB, 1 must also be
included. For more details on the MATF entry, see the MSC Nastran Quick Reference Guide.
CHAPTER 2 13
Primary Format (SOL 600)
1 2 3 4 5 6 7 8 9 10
MATF MID ITYPE SB
"CRI" Criteria Xt Xc Yt Yc Zt Zc 1st
Sxy Syz Szx Find Fxy Fyz Fzx Ext
Exc Eyt Eyc Ezt Ezc Gxy Gyz Gzx
"CRI" Criteria Xt Xc Yt Yc Zt Zc 2nd
"CRI" Criteria Xt Xc Yt Yc Zt Zc 3rd
Primary Format Example:
1 2 3 4 5 6 7 8 9 10
MATF 100 0
+ CRI 1 2500. 4000. 2500. 4000. 2000. 3000. 1st
+ 4500. 4500. 4500.
+
+ CRI 2 2nd
+
+ .11 .06 .1 .05 .075 .03 .03 .03
+ CRI 4 2500. 4000. 2500. 4000. 2000. 3000. 3rd
+ 4500. 4500. 4500. 0.90
(Note: The 4th and 6th lines cannot be entirely blank and the last line of the
3rd criteria has been omitted)
Alternate Format (SOL 600 Only)
1 2 3 4 5 6 7 8 9 10
MATF MID CRI67 Xt Xc Yt Yc Zt Zc
Example Alternate Format
1 2 3 4 5 6 7 8 9 10
1 2 3 4 5 6 7 8 9 10
MATF 100 1 2500. 4000. 2500. 4000. 2000. 3000.
4500. 4500. 4500.
Field Contents
MID Identification number of a MAT1, MAT2, MAT8, MATORT or MAT9 entry
(Integer>0, no default). See Remarks 1 and 2.
ITYPE Flag to invoke progressive failure (Integer, default=1 for SOL 600)
0 No progressive failure compute failure indices only (Default)
2 Progressive failure (presently available only for Hashin and Puck methods)
SB Allowable shear stress of the bonding material between layers (composites only) (Real,
No default)
"CRI" For the alternate format, enter the character string "CRI" to start each criteria
(Character, Required)
CRI67 Used with the alternate format only (Integer, no default, required). It is highly
recommended that only one criterion be used. However, up to three criteria from the
list under Criteria below can be specified in a packed list as follows:
1000000*ITYPE+10000*C3+100*C2+C1
Where C1, C2, C3 are integer values for the various Criteria listed below.
Criteria Select an integer corresponding to the failure criteria to be applied (integer, no default)
Up to three failure criteria may be specified for each MID for 600. Only one failure
criteria may be used for the primary format.
CHAPTER 2 15
Field Contents
1. Maximum stress criterion. (SOL 600 only, See Remark 3)
2. Maximum strain criterion. (SOL 600 only, See Remark 4)
3. Hill failure criterion. (SOL 600 only, See Remark 5)
4. Hoffman failure criterion. (SOL 600 only, See Remark 6)Tsai-Wu failure
criterion. (SOL 600 only, See Remark 7)
5. Hashin failure criteria (SOL 600 only, remark 8)
6. Puck failure criteria (SOL 600 only, remark 11) - must not be combined with
any other method
7. User defined failure criteria (SOL 600 only, remark 12)
8. Hashin-Tape (SOL 600 only, remark 13)
9. Hashin-Fabric (SOL 600 only, remark 14)
Xt Maximum tensile stress in x-direction (Real>0. or blank)
Xc Maximum compressive stress (absolute value) in x-direction (Real>0., default=Xt)
Yt Maximum tensile stress in y-direction (Real>0., default=Xt)
Yc Maximum compressive stress (absolute value) in y-direction (Real>0., default=Yt)
Zt Maximum tensile stress in z-direction (Real>0., default=Xt)
Zc Maximum compressive stress (absolute value) in z-direction (Real>0., default=Zt)
Sxy Maximum shear stress in xy-plane (Real>0. or blank)
Syz Maximum shear stress in yz-plane (Real>0., default=Sxy)
Szx Maximum shear stress in zx-plane (Real>0., default=Sxy)
Find Failure index (Real>0., default=1.) See Remarks 5-7.
Fxy Interactive strength constant for xy-plane (Real<0., default= ).
1 1 1
−
2 X t X c Yt Yc
Fyz Interactive strength constant for yz-plane (Real<0., default= ).
1 1 1
−
2 Yt Yc Z t Z c
Fzx Interactive strength constant for zx-plane (Real<0., default= ).
1 1 1
−
2 Zt Zc X t X c
Ext Maximum tensile strain in x-direction (Real>0. or blank)
Exc Maximum compressive strain (absolute value) in x-direction (Real>0., default=Ext)
Field Contents
Eyt Maximum tensile strain in y-direction (Real>0., default=Ext)
Eyc Maximum compressive strain (absolute value) in y-direction (Real>0., default=Eyt)
Ezt Maximum tensile strain in z-direction (Real>0., default=Ext)
Ezc Maximum compressive strain (absolute value) in z-direction (Real>0., default=Ezt)
Gxy Maximum shear strain in xy-plane (Real>0.)
Gyz Maximum shear strain in yz-plane (Real>0., default=Gxy)
Gzx Maximum shear strain in zx-plane (Real>0., default=Gxy)
Element Birth and Death

Starting with this release it is possible to deactivate and re-activate elements in the model that have failed
or for some other reason needs to be deactivated or re-activated. This is accomplished using Case Control
commands DEACTEL and ACTIVAT as well as matching Bulk Data entries DEACTEL and ACTIVAT.
Once an element is deactivated or activated it stays that way during the entire subcase case unless it fails
due to a MATF criteria. Please see the MSC Nastran Quick Reference Guide for further details.
Unglue
Frequently in contact analysis it is known beforehand that two surfaces will never separate once they
contact. To prevent numerical chattering contact between these surfaces is frequently described using
glued contact. In order to perform VCCT analysis of such surfaces it might be necessary to unglue those
nodes near a crack. A new Bulk Data entry, UNGLUE, is available for such purposes. Please see the MSC
Nastran Quick Reference Guide for further details.
Composite Element Numerical Analysis

In previous versions, SOL 600 provided two options for composite analyses (1) complete through the
thickness integration at every iteration and (2) the “smeared” approach as used in other Nastran solution
sequences. The first approach is more accurate particularly for nonlinear analyses where local buckling
takes place and the analysis needs to extend well into the post-bucking regime. The second approach is
usually satisfactory for small deformation linear static and dynamic analyses. Method 1, complete
integration through the thickness has been modified such that the accuracy has been retained, but the
computational times and memory requirements have been significantly reduced. These are known as
“fast integration” techniques and are described by the new Bulk Data entry PCOMPF. The limitation is
that using these fast integration procedures the material may not exhibit any nonlinear behavior. Large
deformation and buckling is supported using these procedures.
CHAPTER 2 17
The following table indicates typical performance improvements with this release
when using this option.
Maximum Number of Improvement in CPU Memory

Model Number of Element Layers for Stiffness Matrix Reduction
1 10000 46 1425 % 1550 %
2 88854 182 2013 % 957 %
3 48858 33 1085 % 795 %
4 420015 48 1263 % 886 %
Please see the MSC Nastran Quick Reference Guide for further details.
PLOAD4 Extensions
SOL 600 versions previous to this one did not support the PLOAD4 continuation line. In addition, if
corner pressures with different values were entered, they were averaged. SOL 600 now fully supports
different corner pressures, pressures specified by the CID, N1, N2, and N3 fields on the continuation line
and line loads specified by the CID, N1, N2, and N3 fields. The SOLR field is fully supported. The LDIR
field is not supported. Line loads must be specified using the CID, N1, N2, and N3 fields rather than
LDIR. For SOL 600, the SORL field applies to CQUAD4 and CTRIA3 as well as CQUAD4R and
CTRIAR elements. The CID field may reference an ID of any CORD1R, CORD1C, CORD1S,
CORD2R, CORD2C or CORD2S entry but not CORD3G.
All of these new PLOAD4 extensions are activated by entering PARAM,MRPLOAD4,2 in the bulk data
or by placing this parameter in one of the RC files. For this release these capabilities are not the default
and must be activated using this parameter. For the next release, it is anticipated that MRPLOAD4=2
will become the SOL 600 default.
Large Rotation RBE

Improved large motion RBE capabilities have been added to SOL 600, however it was decided to retain
the small rotation formulation as the default. This will allow models used with previous versions to
obtain the same results. To activate the large RBE rotation capability, add the following parameter to the
bulk data:
PARAM,marc7601,1
Streaming Input
A new capability available in this release is known as “streaming input”. Normally SOL 600 will form
a Marc input file from the Nastran input file, then execute Marc to compute the results. With “streaming
input” Marc is not executed. Instead, the Marc subroutines which are now in Nastran are used directly
and called by the main SOL 600 routine. The Marc input file is still formed and saved on disk for possible
future use by the customers, but the same information is passed in memory from the main SOL 600
routine to the Marc initiation routine, thus saving computer time because a physical Marc input file on
disk does not need to be opened and read. Streaming input is activated using the Bulk Data entry:
PARAM,MRSTREAM,1
Streaming input is not available with DDM (parallel processing) or if user subroutines are necessary. For
those cases, do not include param,mrstream and run SOL 600 as before. Streaming input should also not
be used if a special version of Marc is necessary for your particular application. In that case, specify one
of the PATH options on the SOL 600 Executive statement in conjunction with a file to point the analysis
to the location of the version of Marc that is desired.
CONNECTOR TECHNOLOGY
CBUSH, CWELD and CFAST have been added to Marc as nonlinear (large deformation and rotation)
elements. These formulations are now available in SOL 600 using PARAM,MARCWELD (see the QRG
for additional details). The user should beware that CBUSH, CWELD, CFAST, RBE2, RBE3 in SOL 600
and Marc are truly nonlinear elements while in Nastran they are linear small deformation elements. Thus
different results will sometimes be obtained. In addition, solutions that converge with SOL 106, or 129
may not converge with SOL 600 (the opposite may also occur). The user should be careful when using
such elements to make sure they are applicable to both linear and nonlinear solutions when constructing
the model.
The new formulation of the CBUSH is activated using the MARCBUSH,-1, param. If CBUSH entries
are used with a nonzero CID, then the MRCOORDS,0, param should also be included.
The new formulation of the CFAST and CWELD is activated using the MARCWELD,1 param.
The other new parameters that control CWELD behavior are:
• MARIPROJ
• MRCWANGL
• MRFACEA
• MRFACEB
• MRHERRMN
• MRITTYPE
The following CWELD parameters are not supported in SOL 600

• CWDIAGP
• CWRANDEL
The following CWELD parameters are supported in SOL 600

• CWLDIGNR
If the new CWELD is used, then MSPEEDCW is ignored

CHAPTER 2 19
The following CFAST parameters are not supported in SOL 600

• CFDIAGP
• CFRANDELT
COMPUTATIONAL ENHANCEMENTS
The Direct Iterative Solver (MARCSOLV=2) can now be used with out-of-core assembly to allow larger
models to be analyzed. It should be noted that there is a decrease in performance. The program will
automatically use the out-of-core option if necessary or you can select it by using the MARCOOCC
param.
SUPER ELEMENTS / DMIG

The use of super elements in conjunction with SOL 600 has increased over the last year, and several
changed have been made in this area. In addition to the MDMIOUT that may be used to create a reduced
stiffness matrix (super element) or an Adams MNF file, the DMIGOUT option has been added.
The DMIGOUT option may be used to output either the complete global matrices or individual element
matrices at the element levels. These global matrices include:
• Stiffness matrix – This is the total stiffness matrix, including geometric stiffness, follower force
and friction contributions.
• Differential stiffness – This is only available in a buckle sub case
• Mass matrix – Available in dynamics
• Damping matrix – Available in dynamics
• Conductivity matrix – Available in heat transfer
• Specific Heat matrix – Available in transient heat transfer
The element matrices correspond to the above and can be output in either the basic system or in the
transformed system. The global stiffness matrices are always output in the transformed system.
It should be noted that the output of these matrices may be huge. To reduce the size of the DMIG files it
is possible to filter out small values. It should be noted that doing this may influence subsequent
calculations.
CONTACT ENHANCEMENTS
The spline option used with deformable bodies via the BCBODY option can now be used with higher
order elements. This improves the calculation of when contact is to occur and the normal to the surface.
New SOL 600 Parameters

The following new SOL 600 parameters have been introduced as described below. Please see the MSC
Nastran Quick Reference Guide for more details:
PARAM,MARMTLCK Determines whether a check of various property-material combinations

for SOL 600 will be made or not. This slightly slows down the input
processing.
PARAM,MARLDCMB Determines whether extraneous loads in the input file will be combined
to save computer time.
PARAM,MARLDRMV Determines whether extraneous FORCE, MOMENT and/or PLOAD4
entries in the input file will be filtered out at an early stage to save
computer time.
PARAM,MARNOCID SOL 600 by default does not support MCID defined by cylindrical or
spherical coordinate systems. This parameter determines whether MCID
defined by cylindrical or spherical coordinate systems will be ignored or
“fataled out” for shell and solid elements depending on this parameter.
Inclusion of this parameter overcomes this problem.
PARAM,MRCPENTA This entry determines how CPENTA will be mapped to Marc degenerate
solid elements. Marc does not presently have wedge elements, so
CPENTA elements must be mapped to degenerate hexa elements such as
type 7.
PARAM,MRHERRMN This entry controls whether extra grids created for such items as
hyperelastic Herrmann elements, CWELD, etc. are output or not in the
op2, f06, punch and/or xdb files. When Herrmann grids are output, the
displacement value is actually pressure which might be confusing when
looking at an f06 file.
PARAM,MARCMID3 This entry controls whether MID3 will be set to the same value as MID2
when the Marc PSHELL option is used (designated by
PARAM,MRPSHELL,1 or when the SMEAR option is used on the SOL
600 Executive Control statement)
PARAM,MFORDUPE This entry controls how duplicate forces encountered for the same load
case are handled in SOL 600.
PARAM,MARBK105 This entry controls whether linear buckling or nonlinear buckling
eigenvalues are calculated for SOL 600,105.
PARAM,MRCTRIA3 This entry controls the “default” Marc element type for CTRIA3
elements (75 or 138) in SOL 600
PARAM,MRCQUAD4 This entry controls the “default” Marc element type for CQUAD4
elements (75 or 139) in SOL 600
CHAPTER 2 21
PARAM,MARCDUPE This entry controls whether SOL 600 will check for duplicate entries for
most every type of bulk data card. SOL 600 does not allow duplicate
entries, but the portion of IFP that runs prior to spawning Marc does not
usually check for duplicate entries.
PARAM,MARCSTOP This entry controls whether a check model run will be performed, no
actual simulation will occur, and the analysis will stop with a Marc Exit 7.
PARAM,MRRSTOP2 Normally op2, xdb, punch and f06 output is not available for SOL 600
restart analyses. Setting this parameter to 1 will allow the program to
attempt to create one or more of these files. Only in limited cases will the
job be successful.
PARAM,MRBDYCVT Determines if CHBDYG is converted to CHBDYE for SOL 600 heat
transfer
PARAM,MRMATFSB Determines if the version of Marc being used supports the new MATF SB
field or not
PARAM,MRDELTTT Determines how delta time is set for each "step" of a SOL 600 transient
nonlinear analysis.
PARAM,MARCFEAT,N If entered will add FEATURE,N to the Marc input file in the parameters
section
PARAM,MRCOORDS Determines whether Marc COORD SYSTEM will be added if any
CORD1i or CORD2i entries are in the model and if CBUSH elements are
present in the model
PARAM,MARCWELD Determines how CWELD/PWELD and CFAST/PFAST elements will be
translated to Marc
PARAM,MRITTYPE Type of “constraint” used to connect the auxiliary nodes in all CWELDs
PARAM,MRCWANGL Angle in degrees over which to rotate the CWELD cross-section about
the beam axis to obtain its final orientation.
PARAM,MARIPROJ Flag to determine if auxiliary nodes of a CWELD will be projected on the
model
PARAM,MARFACEA Face number for “A” side of weld if welds are made of solid elements
PARAM,MARFACEB Face number for “B” side of weld if welds are made of solid elements
Platform Specific Notes

For Linux IA64 and EM64T platforms:
The default MPI for these platforms is HP MPI.
Intel MPI is also supported and can be used by switching to it using the maintain script under the tools
directory.
To use the Intel MPI (iMPI), please observe the following.
a. Create a .mpd.conf file in your home directory that contains the following line.
secretword=<your mpd password>
where <your mpd password> can be any arbitrary string.
Change mode of the .mpd.conf to 600, i.e. do a
chmod 600 $HOME/.mpd.conf
b. Setup a mpd.hosts file in your home directory consists of the names of nodes in your cluster
(it can have only 1 node, i.e. 1 line):
clusternode1
clusternode2
clusternode3
The rest is taken care of by the run_marc script.
However, if your cluster requires password to perform ssh or rsh between nodes, you may need to enter
your password every time you are running a parallel job. You can disable the password requirement in
the use of ssh or rsh. Please consult your system administrator.
For 64-bit Windows EM64T platforms:

The default MPI for this platform is MPICH2. MS MPI is not supported for this release of MSC Nastran
for SOL 600.
To install MPICH2, go into the mpichx64\bin directory and type
smpd -install
For 32-bit Windows platforms:

The default MPI for this platform is Argonne National Lab’s MPICH2.
To use the MPICH2 please observe the following:
a. Go into the mpich2\bin directory and type
smpd –install
If smpd could not be installed, you may need to reboot your system.
b. The first time your run a parallel job, you will be required to enter your login ID and password.
If your system is rebooted or your command prompt is closed, you will need to repeat step b.
CHAPTER 2 23
Supported Systems for SOL 600 in MSC Nastran 2007

FORTRAN C Parallel Also Works
Vendor OS Hardware Version Version Enabled Default MPI On
HP-Alpha Tru64 5.1 Alpha f90 5.5 cc 6.4 yes HP MPI 2.01
(DEC)4 Server 4100
HP (64-bit) 2,4 HPUX 11.0 PA2.0 f90 2.9.2 C.03.50 yes HP MPI 2.0
HP (64-bit) 2,4 HPUX 11.23 Itanium 2 f90 2.8.7 A.06.02 yes HP MPI 2.2
IBM (64-bit) 4 AIX 5.2 RS/6000 & xlf 8.1.1 cc 6.0.0 yes MPICH1 IBM POE 4.1
RS/6000 SP
SGI (mips4 IRIX 6.5 R12000 f90 7.4 cc 7.4 yes MPICH1
64-bit) 2,3,4
Linux 2.4.21- Itanium 2 Intel 8.1 Intel 8.1 yes SGI MPT 1.10.1 Propack 4.0
SGI (Altix 64- sgi303r2 (Propack
bit) 2, 4 3.0)
Sun (64-bit) 4 Solaris 2.8 UltraSPARC f90 8.1 cc 5.7 yes MPICH1
III
Linux (32-bit) RedHat 9 Intel Intel 8.1 Intel 8.1 yes HP MPI 2.2.5 RedHat AS 3.0
Pentium or
equiv.
Linux (64-bit) RedHat AS 3.0 Itanium 2 Intel 8.1 Intel 8.1 yes HP MPI 2.2.5 5
4,5
Linux (64-bit) RedHat WS 3.0 Intel EM64T Intel 8.1 Intel 8.1 yes HP MPI 2.2.5 5 AMD Opteron,
4,5 RedHat WS 4.0
Intel (32-bit) Windows 2000 Intel Intel 8.1 Intel 8.1 yes MPICH2 Windows XP,
Pentium or Intel 9.1
Intel (64-bit) 4 Windows equivalent Intel 8.1 Intel 8.1 yes MPICH2
Server 2003 MPICH2,
x64 Intel EM64T Windows XP 64
1 Hardware MPI version also available (via maintain in /tools directory).

2 Supports Solver 6.
3 Supports multi-threading.
4 Supports true 64-bit version.
5 Supports the Intel MPI 3.0
Ch. 3: NVH & Acoustics MSC Nastran 2007 Release Guide
3 NVH & Acoustics

Rigid Porous Absorber
Rigid Porous Absorber
Introduction
A new capability to model basic rigid skeleton porous absorber characteristics in acoustic response
analysis is now available. The capability allows some types of absorbent material to be modelled, such
as vehicle seat structures or lining materials which exhibit stiff carcasses. The absorber material is
considered using an equivalent fluid analogy and so is modelled in the same manner as a typical fluid,
using solid CHEXA, CPENTA or CTETRA elements, the GRID points for which have their CD field set
to -1. The porous absorber elements reference a PSOLID property entry with field 8 set to PFLUID in
the usual way, with field 3 of the PSOLID entry referencing a MAT10 entry which has been modified
with an additional field 7 dedicated to porous absorber materials.
Porous Materials
If a material is not completely solid, but contains voids or air pockets, then it is said to be porous. There
are a great many materials which exhibit porosity, the term given to the degree of openness of the
material, including materials generally considered “solid” like brick or stone. If the voids in the
substance are large enough, they may form an interconnected maze of passage ways allowing air (or any
other fluid) to pass through the material. However, depending on the degree of convolution in the
passage ways (known as tortuosity), the air will encounter some resistance as it passes through the
substance, requiring pressure to be exerted to push the air against the resistance. Sound waves striking
the material do not cause air to flow through the material, but they do cause local perturbations that exert
pressure and cause the air to move in the vicinity of the material; the oscillating movement of air caused
by the sound waves encounters resistance (called impedance) which uses up some of the sound energy
and damps the level of sound. This energy is eventually dissipated as heat.
If the porous material is enclosed in a frame which is considered as rigid, as will be the case for example
for a porous medium which has a high skeleton density or very large elastic modulus or weak fluid-
structure coupling, the porous material can be considered as an equivalent fluid with both density and
bulk modulus being complex frequency dependent parameters. It is possible to obtain values for these
parameters by empirical methods as introduced, for example, by Delany and Bazley1, methods which
have been widely used to describe sound propagation in fibrous materials.
In the equivalent fluid approach, the equation of motion reads
2
1 2 ω
JJJJJ ∇ P H JJJJJJ P Z 0
ρe Be
where ρe is the equivalent density, Be the equivalent bulk modulus, P the complex pressure amplitude
and ω the circular excitation frequency, in which it can be shown
1 1
JJJJJ Z JJJ ( 1 H iGE )
ρe ρ
CHAPTER 3 27
NVH & Acoustics
and
JJJJJJ Z JJJ Ó i JJJJ Z JJJ  1 Ó JJJJJ

1 1 A 1 iα
Be B ω B ω
if it is assumed that the parameters are not frequency dependent, a reasonable assumption for the study
of frequencies in a narrow band. Here, ρ , B and GE are the values of RHO, BULK and GE
respectively of the MAT10 entry for the porous absorber material. A new field 7 has been added to the
MAT10 entry to allow the value of α , the normalized admittance coefficient2, to be entered.
Inputs
MAT10 Fluid Material Property Definition
Defines material properties for fluid elements in coupled fluid-structural analysis.
Format:
1 2 3 4 5 6 7 8 9 10
MAT10 MID BULK RHO C GE ALPHA
The following may be used to calculate the equivalent fluid property values to be entered on the MAT10
entry starting from the complex density and complex bulk modulus describing the rigid porous absorber.
If a complex density and complex speed of sound are determined for the porous material, the complex
bulk modulus must first be calculated.
MAT10 density ρ
2 2 ρ r complex density, real part
ρr H ρi
ρ Z JJJJJJJJJJJJJJJJJJ
J ρ i complex density, imaginary part
ρr
MAT10 damping coefficient GE

ρ
GE Z Ó JJJJJi
ρr
MAT10 bulk modulus B

2 2
Br H Bi B r complex bulk modulus, real part
B Z JJJJJJJJJJJJJJJJJJJ
J
Br B i complex bulk modulus, imaginary part
MAT10 normalized admittance coefficient

Bi
α Z ω JJJJJ
Br
Discussion
The implementation implies that if the complex density and bulk modulus are constant, the normalized
admittance coefficient is a function of frequency. However, frequency dependent α is not supported, so
a reference frequency must be selected. Typically, this frequency will be either in the mid-range of the
desired frequency range to be studied, or will correspond to the frequency at a response peak of interest.
As frequencies progressively further away from the reference frequency are considered, the response
becomes increasingly subject to the limitations of the frequency independent formulation; the extent will
depend somewhat on the nature of the problem, and it may be necessary to study discrete frequency bands
in order to mitigate against this effect.
The use of a non-zero value in field 7 of the MAT10 entry causes the generation of a damping matrix
because the normalized admittance coefficient is multiplied by the imaginary operator i. Consequently,
the use of modal methods on the fluid are not appropriate and frequency response analysis must be carried
out using the direct method, at least for the fluid.
References
1. M.E. Delany & E.N. Bazley, Acoustical Characteristics of Fibrous Absorbent Materials,
National Physics Laboratory, Aerodynamics Division, NPL Aero Report Ac 37, March 1969.
2. J. Wandinger, Possible Implementations of Porous Absorbers in Nastran, MSC internal memo,
April 2006.
Example
Consider the following unbounded fluid (air) and porous absorber medium domains as in Figure 3-1. An
acoustic source is placed at the location indicated and the acoustic response (pressure) at the centre of the
fluid is monitored.
CHAPTER 3 29
NVH & Acoustics
Porous medium
Air
Response
Acoustic source
Figure 3-1 Porous medium
The following properties were determined using experimental methods.
Air
Density 1.225 + i0
Speed of sound 340.0 + i3.4
Porous Absorber
Density 3.8663 + i14.2204

Speed of sound 92.7076 + i70.2854
From which the following equivalent bulk moduli were obtained
Air
Bulk modulus 141595.8 + i2832.2
Porous Absorber
Bulk modulus -171190.0 + i102356.3
A frequency of 250 Hz was selected to calculate the values of alpha for air and the porous absorber.
Using the equations illustrated above, the following data is entered on the MAT10 entries.
MAT10 for air
MID BULK RHO C GE ALPHA

MAT10 1 141652.5 1.225 0.0 31.41907
MAT10 for the porous absorber material
MID BULK RHO C GE ALPHA

MAT10 2 -232389. 56.16948 -3.67804 -939.196
Notice that the values of bulk modulus, GE damping coefficient and alpha are all negative; this is a
normal characteristic of the implementation.
The response at the centre of the air domain is calculated using the new capability and the results
compared with the same model run in Actran. Both HEXA-20 (Figure 3-2) and HEXA-8 (Figure 3-3)
elements are compared.
CHAPTER 3 31
NVH & Acoustics
H EXA -20
180
160
140
Pressure (real)
120
100
80
60
0 100 200 300 400 500
Fr e que nc y ( Hz )
A ctran Nastran
Figure 3-2 HEXA-20
H EXA -8
180
160
140
Pressure (real)
120
100
80
60
0 100 200 300 400 500
Fr e que nc y ( Hz )
A ctran Nastran
Figure 3-3 HEXA-8

The results compare very favorably with those from Actran. Notice the gradual departure from the
expected Actran response for frequencies progressively further away from the reference frequency of 250
Hz, as the response becomes increasingly subject to the limitations of the frequency independent
formulation.
Ch. 4: Numerical Enhancements MSC Nastran 2007 Release Guide
4 Numerical Enhancements

New SPARSESOLVER Executive Statement

Improved Performance with New Sparse Solvers
Improved Memory Usage in Lanczos (Pre-release)

New MAXRATIO Information Output (Pre-release)

Performance Improvements and Expanded Capabilities for ACMS
Relaxed Restrictions for CASI Solver Usage

System Dependent Performance Improvements

Improved Selection of Reordering Methods
New SPARSESOLVER Executive Statement
Introduction
A new interface, the SPARSESOLVER Executive Statement, is now available for analysts to better
control the options and processes associated with sparse matrix solution methods. The interface permits
specification of the solver to be used as well as matrix reordering methods and compression techniques
to be used. Furthermore, improved control of matrix diagonal term ratio output is provided.
Benefits
The new interface provides analysts more succinct control over the process than the existing method
selection process that usually involves defining values for some SYSTEM cells via the NASTRAN
statement. In addition, a new output data option is available for matrix diagonal term ratios in the form
of a simple bar chart that provides a more comprehensive view of the ratio data.
Method and Theory

No new theory is involved. The method involves simply the specification of solver options to be used by
various DMAP modules during the solution process. The specified options are checked against feature
availability tables to ensure that they do not conflict with any limitations posed by the specification of
feature combinations. For example, the specification of a particular ordering method may not be available
for a particular solver specification. The options are available only for the DCMP, DECOMP, FRRD1,
READ, SOLVE and TRD1 modules.
Inputs
The sparse solution options are controlled by keywords specified on the SPARSESOLVER Executive
statement. See the MSC Nastran Quick Reference Guide for a complete description of this statement.
Outputs
There are no new outputs associated with this feature other than informational and diagnostic messages.
Guidelines and Limitations

The ability to specify particular sparse matrix solution options is sometimes useful in determining
whether one method is more effective than another in obtaining the solution. Other features can also be
useful as in obtaining diagnostic data output. For example, one might be interested in reviewing matrix
diagonal term ratios. In general, high ratios indicate a loss of accuracy. The feature can be used by taking
all of the program defaults for the various control variables. These defaults produce both the table and
bar outputs. The table is limited to 25 ratios that exceed 1.0E+05. The bar chart uses powers of ten for
segment widths. This can be done by simply adding:
CHAPTER 4 35
SPARSESOLVER DCMP ( MDTRATIO )

to the Executive Section of the input data file. The use of this new feature is currently limited to sparse
symmetric matrix operations in the DCMP module.
Several different sparse matrix factorization methods are available. Specification of a particular method
should be done only after thoughtful consideration. The following table summarizes the advantages and
disadvantages of the various factorization methods.
Method Advantages Limitation

MSCLDL Small memory requirements; handles None
indefinite matrices
MSCLU Small memory requirements None
TAUCSCHL Generally superior performance High memory requirements; real positive
compared to MSCLDL definite matrices only
UMFLU Generally superior performance High memory requirements
compared to MSCLU
Similarly, re-ordering methods can also be specified, but only should be done after consideration of the
potential effects. The following table summarizes the advantages and disadvantages of the various re-
ordering methods.
Method Advantage Limitation

AMF Similar to BEND but with smaller Produces less optimal reordering
memory requirements compared to BEND
BEND Optimal for small matrices and for large None
matrices from models dominated by 1-
and 2-dimensional finite elements
MD Very low memory requirements Produces sub-optimal reordering
METIS Good for large matrices dominated by 3- Large memory requirements
dimensional finite elements
MMD Low memory requirements Produces sub-optimal reordering
Similarly, compression methods can also be specified, but only should be done after consideration of the
potential effects. The following table summarizes the advantages and disadvantages of the various
compression methods.
Method Advantage Limitation

GRID Utilizes USET and SIL information from the Relies on USET and SIL
Nastran Database
SUPER Does not rely on USET and SIL tables; can None
produce better orderings for models
dominated by 3-dimensional elements
GRDSUPER Can produce better orderings for models Requires USET and SIL tables
dominated by 3-dimensional finite elements
CHAPTER 4 37
Improved Performance with New Sparse Solvers
Introduction
As part of MSC Software HPC development, surveys of solver technology from industry and academia
are conducted. This has lead to the integration of two solvers developed outside of MSC Software:
TAUCS and UMFPACK.
In addition, MSC has enhanced the Lanczos eigensolver to take better advantage of available memory.
The TAUCS Sparse Solver

A new sparse Cholesky solver is available. The solver is derived from the TAUCS package of linear
solvers.
User Interface
The new solver can be selected for the DCMP, DECOMP, and SOLVE modules through the
SPARSESOLVER Executive statement:
SPARSESOLVE DCMP (FACTMETH=TAUCSCHL)
It may also be selected by setting system cell 166 to 8192.
Use Cases
The new solver keeps the matrix data in memory, and may show improved performance over the default
solver for models which are dominated by three dimensional elements.
The four models described below demonstrate the potential performance enhancement with the new
solver.
Model Name DOF Elements

zzz3 537876 108675 TETRAs
vl0sst 408999 710768 TETRAs
xl0rst 739815 113217 HEXAs
Solidcyl 604800 176400 HEXAs
Each model was run on one node with 2 dual core 2.4GHz Opteron processors, 8Gb of memory and 56Gb
of scratch space. Each job was run with mem=7gb, and the CPU time of the DCMP module is displayed
in the following chart.
New Cholesky Solver Perform ance

800
750
700
DCMP t im e (cpu sec)
650
600
550
500
450
400 MSCLDL
350 TAUCSCHL
300
250
200
1 50
1 00
50
0
zzz3 vl0sst x l0rst solid cy l
Model
Limitations
The new solver only works for real, symmetric matrices which are positive definite. Modeling techniques
which lead to indefinite matrices, such as Lagrange multipliers, are not supported by the new solver. If
the new solver detects an indefinite matrix, or has insufficient memory to perform the factorization, or
encounters any other error, the out-of-core LDLT solver is used.
The new solver is not recommended for multiple superelement analysis.
The TAUCS code is used by MSC under the following license agreement. Please note that in the
agreement, “this program” and “this software” refers only to the TAUCS code, available at
http://www.tau.ac.il/~stoledo/taucs.
TAUCS version 2.0, November 29, 2001. Copyright © 2001 by Sivan Toledo, Tel-Aviv University,
stoledo@tau.ac.il. All Rights Reserved.
Used by permission.
THIS MATERIAL IS PROVIDED AS IS, WITH ABSOLUTELY NO WARRANTY EXPRESSED OR
IMPLIED. ANY USE IS AT YOUR OWN RISK.
Permission is hereby granted to use or copy this program, provided that the Copyright, this License, and
the Availability of the original version is retained on all copies. User documentation of any code that uses
this code or any derivative code must site the Copyright, this License, the Availability note, and “Used
by permission”. If this code or any derivative code is accessible from within MATLAB, then typing “help
taucs” must cite the Copyright, and the “type taucs” must also cite this License and the Availability note.
Permission to modify the code and to distribute modified code is granted, provided that the Copyright,
this License, and the Availability note are retained, and a notice that the code was modified is included.
This software is provided free of charge.
CHAPTER 4 39
The UMFPACK Sparse Solver

UMFPACK is a set of routines for solving unsymmetric sparse linear systems using an unsymmetric
multi-frontal method. The UMFPACK factorization has been implemented in Nastran as a licensed
software product from the University of Florida. It operates on both real and complex matrices.
UMFPACK has been implemented for solving unsymmetric frequency response problems. In addition,
it is implemented in the SOLVE and DECOMP modules. It is activated by an explicit user request via
the SPARSESOLVER Executive statement. For example:
SPARSESOLVER FRRD1 (FACTMETH=UMFLU)
This command specifies that UMFPACK is to be run in the FRRD1 module, which is used for frequency
response analysis problems. UMFPACK may also be selected by setting system cell 209 to 16.
The UMFPACK solver has shown to be very useful for problems in Exterior Acoustics, which generates
UMFPACK Memory Guidelines
The UMFPACK solver operates entirely in memory. Therefore, memory requirements for UMFPACK
may be considerably higher than for the default Nastran sparse direct unsymmetric solver.
Additionally, it is important to note that UMFPACK operates entirely outside the control of Nastran
memory management. Each time it is invoked, UMFPACK will obtain memory from the operating
system, use the new memory, and then return the new memory back to the operating system. If there is
insufficient memory to complete the UMFPACK factorization, Nastran terminates with a Fatal Error
message, and it attempts to inform the user of memory requirements. It is not possible to allocate
additional memory for UMFPACK via the nastran command line option "mem=".
It is the user's responsibility to ensure that sufficient unsubscribed memory is available from the
operating system for UMFPACK to complete successfully. This may mean lowering the amount of
memory requested at job submittal with the "mem=" option. For example, suppose there is 8GB of
memory on your computer, and you want to use Nastran with UMFPACK. If the memory requirement
for UMFPACK to complete is 4GB, then you should not submit your Nastran job with more than 4GB.
In fact, UMFPACK will be the most memory intensive portion of your Nastran analysis, so that the actual
Nastran memory could be set much lower by comparison.
This memory limitation will be removed in a future release, so that all UMFPACK memory operations
will be controlled by Nastran memory management.
Improved Memory Usage in Lanczos (Pre-release)

The Lanczos method has been improved to better take advantage of the given memory.
This beta feature may be selected by setting system cell 146 to -1.
To illustrate the performance improvements, the following models were run on an SGI Altix with four
1.4GHz itanium2 processors, 20Gb of ram and 470Gb of disk space. Each model was run with
mem=4200mb.
Model Name DOF Elements

xl0u 595303 55925 HEXAs, 18483 PENTAs
xl0h 335680 54509 QUAD4s, 15523 TRIA3s
xl0j 590474 94979 QUAD4s, 13406 TRIA3s
xl0m 529199 96185 QUAD4s
Im proved Lanczos Perform ance

700
650
600
READ t im e (cpu sec)
550
500
450
400
350 2006R1
300 2006R2
250
200
1 50
1 00
50
0
x l0u x l0j x l0h x l0m
Model
CHAPTER 4 41
New MAXRATIO Information Output (Pre-release)
Introduction
A new interface is now available for analysts to better control the generation of matrix diagonal term
ratio statistics produced by the sparse symmetric matrix decomposition process in the DCMP module.
The matrix diagonal term ratio statistics are sometimes useful in determining the quality of the matrix
decomposition process. In general, for linear static analysis, high or negative ratios indicate a loss of
accuracy and could be indicative of a modeling error.
Benefits
The new interface provides analysts more control over the process than the existing method of supplying
a value for the MAXRATIO DMAP parameter. In addition, a new output data option is available in the
form of a simple bar chart that provides a more comprehensive view of the ratio data.
Method and Theory

No new theory is involved. The method involves simply the computation of a ratio defined to be the
original matrix diagonal term divided by the decomposed matrix diagonal term. These ratios are placed
in a table together with the external identifier associated with the row/column of the term. This table is
then processed according to the options requested by the user.
Inputs
The matrix diagonal term ratio output options are controlled by keywords specified on the
SPARSESOLVER Executive statement. See “New SPARSESOLVER Executive Statement” on page 34
for a complete description of this statement.
Outputs
The matrix diagonal term ratios can be presented in two different views. The first view is the table view
in which each ratio is listed together with the external identifier of the row/column of the matrix as well
as the original input matrix diagonal term. This format is virtually identical to that produced now when
any ratio exceeds the value of the MAXRATIO input parameter. The second view of the ratios is
statistical in nature. It is similar to a bar chart. A series of bar segments are generated. There are two
options for specifying the segment widths of the bars. The default option uses powers of 10 as the widths
(e.g. 10.0 to 100.0 and 100.0 to 1000.0). The second option allows the user to specify how many
segments are desired. The program will compute the segment width using the maximum and minimum
ratios. For each bar in the chart, the total number of terms in the range is tabulated together with a visual
indication of the percentage number of terms in that particular bar.
Note that when negative matrix diagonal term ratios are detected, they will always be output if the
TABLE option is specified.
These new views of the ratios do not replace any existing diagnostics generated by the DCMP module if
a problem is detected. Under these conditions, output from the table view may duplicate previous output
generated by DCMP module error processing.

The matrix diagonal term ratio statistics are sometimes useful in determining the quality of the matrix
decomposition process. In general, high ratios indicate a loss of accuracy. The feature can be used by
taking all of the program defaults for the various control variables. These defaults produce both the table
and bar outputs. The table is limited to 25 ratios that exceed 1.0E+05. The bar chart uses powers of ten
for segment widths. This can be done by simply adding
SPARSESOLVER DCMP (MDTRATIO)
to the Executive Section of the input data file.
The use of this new feature is currently limited to sparse symmetric matrix operations in the DCMP
module.
If there are scalar-type points present in the problem, the degrees of freedom associated with these points
will be grouped into the results for the translational degrees of freedom output.
Demonstration Example
A simple example is presented that demonstrates the use of some of the new features available for output
of the matrix diagonal term ratios. The SPARSESOLVER Executive statement is used to specify the
desired features. The example problem is used for demonstration purposes only and is not representative
of anything in particular. The model data consists of a simple plate structure subject to an end load.
Example Input Data

$
$ Example problem to demonstrate matrix diagonal term ratios
$
id test,case
sol 101
SPARSESOLVER DCMP (MDTRATIO) $
cend
spc=100
load=1000
disp=all
begin bulk
grdset,,,,,,,6
cquad4,101,101,1,2,52,51
cquad4,102,101,2,3,53,52
cquad4,103,101,3,4,54,53
cquad4,104,101,4,5,55,54
cquad4,105,101,5,6,56,55
CHAPTER 4 43
cquad4,106,101,6,7,57,56
cquad4,107,101,7,8,58,57
cquad4,108,101,8,9,59,58
cquad4,109,101,9,10,60,59
cquadr,1101,101,1,2,52,51
cquadr,1102,101,2,3,53,52
cquadr,1103,101,3,4,54,53
cquadr,1104,101,4,5,55,54
cquadr,1105,101,5,6,56,55
cquadr,1106,101,6,7,57,56
cquadr,1107,101,7,8,58,57
cquadr,1108,101,8,9,59,58
cquadr,1109,101,9,10,60,59
grid, 1,, 0.0,0.0,0.0
grid, 2,, 1.0,0.0,0.0
grid, 3,, 2.0,0.0,0.0
grid, 4,, 3.0,0.0,0.0
grid, 5,, 4.0,0.0,0.0
grid, 6,, 5.0,0.0,0.0
grid, 7,, 6.0,0.0,0.0
grid, 8,, 7.0,0.0,0.0
grid, 9,, 8.0,0.0,0.0
grid,10,, 9.0,0.0,0.0
grid,51,, 0.0,1.0,0.0
grid,52,, 2.4,1.0,0.0
grid,53,, 3.5,1.0,0.0
grid,54,, 4.6,1.0,0.0
grid,55,, 5.7,1.0,0.0
grid,56,, 6.8,1.0,0.0
grid,57,, 7.9,1.0,0.0
grid,58,, 9.0,1.0,0.0
grid,59,,10.1,1.0,0.0
grid,60,,11.2,1.0,0.0
$
ctria3,201,101,101,102,151
ctria3,202,101,102,152,151
ctria3,203,101,102,103,152
ctria3,204,101,103,153,152
ctria3,205,101,103,104,153
ctria3,206,101,104,154,153
ctria3,207,101,104,105,154
ctria3,208,101,105,155,154
ctriar,1201,101,101,102,151
ctriar,1202,101,102,152,151
ctriar,1203,101,102,103,152
ctriar,1204,101,103,153,152
ctriar,1205,101,103,104,153
ctriar,1206,101,104,154,153
ctriar,1207,101,104,105,154
ctriar,1208,101,105,155,154
grid,101,, 0.0,0.0,0.0
grid,102,, 1.0,0.0,0.0
grid,103,, 2.0,0.0,0.0
grid,104,, 3.0,0.0,0.0
grid,105,, 4.0,0.0,0.0
grid,151,, 0.0,1.0,0.0
grid,152,, 3.4,1.0,0.0
grid,153,, 4.5,1.0,0.0
grid,154,, 5.6,1.0,0.0
grid,155,, 6.7,1.0,0.0
$
pshell,101,1,0.05,1
mat1,1,10.+6,,0.33
spc1,100,123,1,101
spc1,100,3,5,55,105,155
spc1,100,1,55,155
spc1,100,2,1,101
force,1000,10,,1000.0,1.0,0.0,0.0
force,1000,60,,1000.0,1.0,0.0,0.0
force,1000,105,,1000.0,1.0,0.0,0.0
force,1000,155,,1000.0,1.0,0.0,0.0
enddata
Example Output
The output generated by the previous example is shown following. Notice that there are two separate
sections of output: one for translational degrees of freedom and one for rotational. Within each section,
both a bar chart and table of matrix diagonal term ratios are output.
TRANSLATIONAL DOF DIAGONAL TERM RATIO STATISTICS CHART FOLLOWS FOR THE DECOMPOSITION OF MATRIX KLL
------------------------------------------------|---------------------------------------------------------------------------
DIAGONAL TERM RATIO RANGE #TERMS % TOT |MAXIMUM RATIO = 6.90963E+02 MINIMUM RATIO = 1.00000E+00
------------------------------------------------|---------------------------------------------------------------------------
1.0000E+00 TO 1.0000E+01 62 79.49 |**************************************************************************>
1.0000E+01 TO 1.0000E+02 12 15.38 |***************
1.0000E+02 TO 1.0000E+03 4 5.13 |*****
0
0 MATRIX/FACTOR DIAGONAL TERMS RATIO SUMMARY TABLE FOR TRANSLATIONAL DOF SORTED ON DIAGONAL RATIO
GRID POINT ID DEGREE OF FREEDOM MATRIX/FACTOR DIAGONAL RATIO MATRIX DIAGONAL
(TOP 1 RATIOS>MAXRAT= 6.90963E+02)
58 T3 6.90963E+02 5.65535E+04
ROTATIONAL DOF DIAGONAL TERM RATIO STATISTICS CHART FOLLOWS FOR THE DECOMPOSITION OF MATRIX KLL
------------------------------------------------|---------------------------------------------------------------------------
DIAGONAL TERM RATIO RANGE #TERMS % TOT |MAXIMUM RATIO = 3.35974E+02 MINIMUM RATIO = 1.00000E+00
------------------------------------------------|---------------------------------------------------------------------------
1.0000E+00 TO 1.0000E+01 38 63.33 |***************************************************************
1.0000E+01 TO 1.0000E+02 18 30.00 |******************************
1.0000E+02 TO 1.0000E+03 4 6.67 |*******
0
0 MATRIX/FACTOR DIAGONAL TERMS RATIO SUMMARY TABLE FOR ROTATIONAL DOF SORTED ON DIAGONAL RATIO
GRID POINT ID DEGREE OF FREEDOM MATRIX/FACTOR DIAGONAL RATIO MATRIX DIAGONAL
(TOP 1 RATIOS>MAXRAT= 3.35974E+02)
58 R2 3.35974E+02 2.14135E+04
CHAPTER 4 45
Performance Improvements and Expanded

Capabilities for ACMS
Introduction
Automated Component Modal Synthesis (ACMS) is a powerful tool for a variety of large modal based
analyses. ACMS functions in two domains, the Matrix Domain and the Geometric Domain. Since its
introduction in 2005, Matrix Domain ACMS (MDACMS) has become the default ACMS method.
Developments described in this section apply to MDACMS.
MDACMS has been extended to more thoroughly interact with the Nastran External Superelement
capability. See "MDACMS for Upstream Superelements" below.
An automatic decision logic has been implemented to select which path to take for frequency response
calculations. See "Automatic FASTFR Decision Logic" below.
For other enhancements, see "Miscellaneous Performance Improvements" below.
MDACMS for Upstream Superelements

MDACMS has been extended to more thoroughly interact with the Nastran External Superelement
capability. Specifically, MDACMS reduction calculations have been expanded to include all quantities
being reduced to the residual, namely stiffness, mass, and damping. (Previously, only the component
modes were computed by MDACMS.) By reducing these quantities while simultaneously computing
component modes, a significant reduction in compute time, disk I/O, and scratch disk space is realized.
One typical case study is presented here. Note that in this example, the resource requirements for MSC
Nastran R1 exceeded availability and the job did not complete, even though it had run five times longer
than the current release at the time of its failure.
No. of Grid No. of global No. of O-set No. of A-set

Points DOF DOF DOF Analysis Description
375,502 2.25 million 1.9 million 3526 Normal modes, frequency range
0-550Hz (1627 modes)
The above jobs were run on an IBM Power4 computer running the AIX operating system.
Automatic FASTFR Decision Logic

In MSC.Nastran 2004, the FASTFR method was introduced for modal frequency response analysis. It
can be selected via the Bulk Data entry PARAM,FASTFR,YES and shows significant performance
improvement for certain models in the mid-frequency range. However, the user has to make the decision
as to whether to use the new method or whether to run with the standard FRRD1 method with or without
the iterative solver before starting to run the job.
With this release, automatic decision logic has been implemented which eliminates the need for the user
to make that decision. The program will decide automatically which solution method will be most
efficient for the frequency response portion in a SOL 111 analysis. Based on the size of the modal space
and some other heuristic criteria, either the FASTFR solution method will be run, or the FRRD1 module
with or without the iterative solver will be used.
Currently, the FASTFR method is selected or deselected in the Bulk Data Section using param,fastfr,yes
or param,fastfr,no. To activate the automatic decision logic, specify
PARAM, FASTFR, AUTO
in the Bulk Data Section. This will cause the program to run as if the FASTFR method was selected until
it reaches the point where the decision will be made. If it is decided that it would be faster to not run with
the FASTFR method, the following message will be printed in the F06 file, and the program will continue
with the standard frequency response method.
Also, if the FASTFR method is deselected for a different reason, a similar message will be printed. If the
^^^ SYSTEM INFORMATION MESSAGE 9157 (GMA)

^^^ FASTFR OPTION REQUESTED, BUT THE MODEL DID NOT MEET THE FOLLOWING CRITERIA:
^^^ THE FASTFR OPTION MAY BE TOO EXPENSIVE.
^^^ USER INFORMATION: STANDARD FREQUENCY RESPONSE METHOD WILL BE USED
CHAPTER 4 47
FASTFR method will actually be used, then no message will be printed.

This new option for PARAM,FASTFR is helpful for all modal frequency response analysis jobs,
especially in the mid to upper frequency range.
Examples
The following charts show the benefit of the automatic decision logic for three examples. Please note
that the times given are the elapsed minutes for the complete job.
Example 1
For this automotive example, the automatic decision logic decided that the FASTFR method might be
too expensive. So it switched to the FRRD1 module with the iterative solver.
Job Statistics: 1,208,135 grid points;

matrix size in FRRD1 (i.e. H-size) is 9,974;
modes below 1,000 Hz;
2 frequencies.
1400
1200
1000
800
600
400
200
0
FASTFR FRRD1 (iter) FRRD1 Decision
(direct) Tim e
Example 2
For this automotive example, the automatic decision logic decided to go with the FASTFR method, and
rightly so. Both, the FRRD1 module with and without the iterative solver proved to be much slower.
Job Statistics: 233,094 grid points;

matrix size in FRRD1 is 4,350;
modes below 600 Hz;
260 frequencies.
350
300
250
200
150
100
50
0
FASTFR FRRD1 (iter) FRRD1 Decision
(direct) Tim e
Example 3
For this automotive example, the automatic decision logic decided to go with the FASTFR method which
proves to be significantly faster than the conventional method (i.e. FRRD1 with the direct method). The
iterative method inside FRRD1 is not practical for this job because there are 147 loads which
significantly slow down the iterative solver.
Job Statistics: 321,597 grid points;

matrix size in FRRD1 is 5,424;
modes below 600 Hz;
501 frequencies.
2500
2000
1500
1000
500
0
FASTFR FRRD1(direct) Decision
Tim e
Limitations for the FASTFR method

1. Fluid Damping can only be specified via the param,gfl entry and/or the sdamping(fluid)
command.
a. The CAABSF element (acoustic absorbers) is not supported;
b. The FASTFR method will be turned off automatically for fluid K4 and for fluid viscous
damping.
CHAPTER 4 49
2. All matrices must be symmetric

a. Unsymmetric formulation for acoustic coupling is not supported;
b. EPOINT Bulk Data entries are not supported.
3. The FASTFR method work only for SOL 111 and for SOL 200 with ANAL=MFREQ.
4. SESDAMP and FASTFR are not allowed in the same run.
Miscellaneous Performance Improvements

MDACMS has been enhanced to reduce the amount of disk I/O required for a successful analysis. This
happens automatically and there is no user action required to realize the reduction in I/O. Typically,
reduced I/O requirements results in reduced elapsed time and greater CPU utilization, which increases
the effectiveness of Shared Memory Parallel (SMP) computations.
To demonstrate some of these performance improvements, three example jobs were run and the
performance results were compared to the previous version. The results are displayed below.
Model Description:
Model No. of grid No. of global

No. Points DOF Analysis Description
1 44,314 265884 Auto body normal modes run w/ACMS. Frequency range 0-400Hz
(360 modes)
2 44,314 265884 Model (1) run in SOL 111 w/ACMS; 75 load cases, 250 forcing
frequencies
3 1,243,651 7.4 million Mid frequency acoustic analysis of automotive trimmed body. 9389
structure modes below 750Hz; 1486 fluid modes below 1200Hz; 4
load cases, 290 forcing frequencies
Example 1 shows time and I/O resource requirements for model one run on an IBM Power4 system under
AIX.
CHAPTER 4 51
Example 2 shows DMP scalability improvement for model two. These jobs were run on a Altix machine
with IA64 processors under Linux.
Example 3 shows I/O and elapsed time improvement for model three. These jobs were run on an IBM
Power4 system under AIX.
Relaxed Restrictions for CASI Solver Usage

The use of the element-based CASI iterative solver is limited by several restrictions noted in the remarks
for the ITER Bulk Data entry. Two of these restrictions have been somewhat relaxed so that the solver is
usable over a wider range of problems.
The first is a reduction in the number of element types that the solver does not recognize. The solver
recognizes only a sub-set of all of the element types available, but the list has been expanded to include
the FAST, SEAM and WELD element types as supported element types.
The second is a relaxation of the restriction associated with the processing of direct input matrices via the
K2GG Case Control command. K2GG matrix input is now allowed by the CASI solver interface.
However, the matrix size is limited to 100 grid/scalar points, since it must be treated as an unknown finite
element type. This should be sufficient for most cases where the input K2GG matrix represents some
form of attachment to, or base for, the structural model being analyzed.
CHAPTER 4 53
System Dependent Performance Improvements
Introduction
MSC has implemented new versions of the Basic Linear Algebra Subroutine library (BLAS) for Nastran,
on platforms supported by the "x86-64" architecture.
Version 9 of Intel's Math Kernel Library (MKL) provides optimal performance for analysis solutions that
make good use of the BLAS. This includes the ACMS solution for automotive NVH and acoustic
analysis, and the Exterior Acoustics capability in Nastran. In addition, MKL provides automatic
multithreading support. Shared memory parallel speedup is available via the smp=n command supplied
to the nastran command processor.
The ATLAS project (Automatically Tuned Linear Algebra Software) provides a portable, efficient
version of the BLAS. An ATLAS version of the BLAS library has been implemented for use on AMD
Opteron processors. ATLAS is public domain software available on the internet at
http://math-atlas.sourceforge.net/.
For demonstration purposes, two analysis model were selected and run with the current and prior release
of MSC Nastran. Performance results are shown on the following page.
Model Description
Model No of Grid No. of
No. Points Global DOF Analysis Description
1 268,486 1.6 million Auto body normal modes run w/ACMS. Frequency range
0-200Hz (1043 modes)
2 603,266 2.8 million Low frequency acoustic analysis of automotive trimmed
body
Performance Improvements for Intel x86_64

Processor Clock Speed OS Real Memory Scratch Filesystem
Woodcrest 2666MHz Linux RHWS4 16GB ext2 – not striped
Performance Improvements for AMD due to Atlas

Processor Clock Speed OS Real Memory Scratch Filesystem
Opteron 2400MHz Linux SuSE 10 8GB xfs – not striped
CHAPTER 4 55
Improved Selection of Reordering Methods

Nastran contains five methods for reordering (permuting) sparse matrices in preparation for a sparse
symmetric factorization. The five methods are minimum degree (MD), multiple minimum degree
(MMD), and three nested dissection/minimum degree hybrid methods, called BEND, AMF, and METIS.
A judicious choice of a reordering method can lead to a dramatic improvement in the performance of the
sparse factorization.
The hybrid methods (BEND, AMF, and METIS) are generally superior to MD and MMD. The previous
default behavior was to use BEND unless the USET and SIL tables were not available, in which case
MD was used. If the BEND algorithm failed, MD was also used in this case. The new default behavior
is to use BEND except for large (> 50,000 DOF) models, which are dominated by three-dimensional
elements; for those models METIS is the default. If the default method fails, the program will attempt to
use another hybrid method (for example, if BEND fails, the program tries METIS). The minimum degree
algorithms are selected only as a last resort.
To illustrate the performance improvements, the following models were run on an SGI Altix with four
1.4GHz itanium2 processors, 20Gb of ram and 470Gb of disk space.
Model Name Solution DOF Elements

xx0e Normal Modes 3308298 656570 TETRAs
xx0d Normal Modes 1920855 402441 TETRAs
xx0f Statics 1237999 246066 TETRAs
lg0q Direct Freq. Response 93375 28140 HEXAs
The chart below shows the factorization time for each job. The improvement for the xx0e, xx0d, and xx0f
jobs stems from the automatic selection of METIS rather than BEND reordering. The improvement in
lg0q is due to the selection of METIS rather than MD.
Reordering Selection Perform ance Im provem ent

2750
Fact orizat ion t im e (cpu sec)
2500
2250
2000
1 750
1 500
2 0 0 6 R1
1 250 2 0 0 6 R2
1 000
750
500
250
0
x x 0e x x 0f x x 0d lg 0q
Model
Ch. 5: Elements & Connectors MSC Nastran 2007 Release Guide
5 Elements & Connectors

Enhancements to Connector Elements

Nonhomogeneous Multipoint Constraint
Enhancements to Connector Elements
Introduction
In MSC Nastran 2007, a new seam weld element is introduced to allow for the definition of a seam line.
This new CSEAM element replaces the existing CWSEAM element and is modeled by the new CSEAM
and PSEAM Bulk Data entries. In addition, the SWLDPRM Bulk Data entry is enhanced to support
element type specific control parameters and different diagnostic output format.
For spot weld elements CWELD and CFAST, the displacements of the projected grids GA and GB are
now computed and output in this release. These data will facilitate users to view the relationship between
the spot weld and the connecting shells.
CSEAM Elements
The new CSEAM element provides the following key features to extend the analysis capabilities, provide
more flexibility in modeling, and improve the robustness of the analysis results.
• A seam line is considered continuous between two CSEAM elements that have a common face
based on either common GS/GE points or XYZ coordinates. Note that the SMLN label on a
CSEAM element does not determine the definition of a seam line. It is only intended for ease of
seam line visualization.
• This element can connect up to 64 shell grids, which allows the connection of higher order shell
elements.
• Besides selecting the connected surface patches by property IDs, the user may define the
connection by specifying shell element IDs directly.
• Each of the eight auxiliary points must have a projection onto the connected shell element. This
requirement prevents the generation of ill conditioned stiffness matrix in down stream
processing.
• The user can model tailored blank tapering by specifying different property IDs at the start and
end points.
• This element type supports the MAT9 anisotropic material properties.
Inputs
The seam connection is modeled by the new CSEAM and PSEAM Bulk Data entries and the modified
SWLDPRM Bulk Date entry. The details of these entries are described in the MSC Nastran Quick
Reference Guide.
Outputs
The connecting record based on seam line label is written as a SEAMLN record in GEOM2 data block,
which is included in OP2 file for post-processing. The contents of the SEAMLN record for each seam
line label are listed:
CHAPTER 5 59
Word Type Description

1-2 C Seam line ID
3 I Total number of elements for this seam line
4 I CSEAM element ID
5 I GS grid ID for the start point
6 I GE grid ID for the end point
: I Repeat words 4-6 for each CSEAM element
: I End the seam line data with -1
Limitations
• Each CSEAM element can connect a maximum of three shell elements on patch A and three
shell elements on patch B.
• Only line type of seam is supported.
• Super-element modeling is not supported.
• FORCE, STRESS and STRAIN output requests are not supported.
Example – A Symmetric Hat Profile (cseam_hut.dat)

This example demonstrates the application of CSEAM elements to analyze an unconstrained symmetric
hat profile model, see Figure 5-1. Each edge of the hat is connected by 27 CSEAM elements. The grids
with identification numbers 6000 to 6028 and 8000 to 8028 are used as the piercing points to define the
seams, see Figure 5-2.
Hat Profile
Figure 5-1 Hat Profile

Piercing Poin ts and Seams
Figure 5-2 Piecing Points and Seams
The input for the seam welds is listed:

CSEAM 7000 500 SMLN_A PSHELL 100 200
6000 6001
6001 6002
:
6026 6027
6027 6028
$
CSEAM 9000 500 SMLN_B PSHELL 100 200
8000 8001
8001 8002
:
CHAPTER 5 61
8026 8027
8027 8028
$
PSEAM 500 1 LINE 1.0
The normal mode analysis results are shown:
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 -5.927181E-05 7.698819E-03 1.225305E-03 1.000000E+00 -5.927181E-05
2 2 -2.941204E-05 5.423287E-03 8.631429E-04 1.000000E+00 -2.941204E-05
3 3 2.051645E-05 4.529509E-03 7.208938E-04 1.000000E+00 2.051645E-05
4 4 2.372748E-05 4.871086E-03 7.752574E-04 1.000000E+00 2.372748E-05
5 5 5.330599E-05 7.301095E-03 1.162005E-03 1.000000E+00 5.330599E-05
6 6 1.118197E-04 1.057448E-02 1.682981E-03 1.000000E+00 1.118197E-04
7 7 1.154947E+07 3.398451E+03 5.408803E+02 1.000000E+00 1.154947E+07
8 8 1.585229E+07 3.981493E+03 6.336742E+02 1.000000E+00 1.585229E+07
9 9 2.653947E+07 5.151647E+03 8.199102E+02 1.000000E+00 2.653947E+07
10 10 2.959784E+07 5.440389E+03 8.658648E+02 1.000000E+00 2.959784E+07
11 11 3.002028E+07 5.479077E+03 8.720222E+02 1.000000E+00 3.002028E+07
12 12 3.041736E+07 5.515193E+03 8.777703E+02 1.000000E+00 3.041736E+07
13 13 3.328896E+07 5.769658E+03 9.182696E+02 1.000000E+00 3.328896E+07
14 14 3.757468E+07 6.129819E+03 9.755909E+02 1.000000E+00 3.757468E+07
15 15 4.259932E+07 6.526814E+03 1.038775E+03 1.000000E+00 4.259932E+07
16 16 4.738684E+07 6.883810E+03 1.095592E+03 1.000000E+00 4.738684E+07
SWLDPRM Enhancements
Element Specific Control Parameters
The SWLDPRM Bulk Data entry is enhanced to support element type specific control parameters. Two
new parameters, CNRAGLI and CNRAGLO, are introduced to define the angle limits for checking the
geometry of seam elements. These parameters replace the GSPROJ parameter to define the allowable
angles for corner check. The GSPROJ parameter specified in the previous seam weld models must be
replaced by CNRAGLI and CNRAGLO to obtain identical results.
For example, if the SWLDPRM Bulk Data entry is defined as
SWLDPRM,PROJTOL,0.0,GSMOVE,0,GSPROJ,-1.,NREDIA,4,
,GMCHK,1,CHKRUN,1
then this entry must be changed into
SWLDPRM,PROJTOL,0.0,GSMOVE,0,GSPROJ,-1.,NREDIA,4,
,GMCHK,1,CHKRUN,1,cnraglo,-1.
to get the same results as the results running from previous version.
Real Format of Diagnostic Output

For the PRTSW parameter of SWLDPRM Bulk Data entry, two options are added to support real format
of diagnostic output so that more significant digits will show (quality 1-25861871). This parameter now
has five options.
PRTSW Output
PRTSW = 0 No Diagnostic Output
PRTSW = 1 Print diagnostic output in exponential format to .f06 file
PRTSW = 2 Punch diagnostic output in exponential format to .pch file
PRTSW = 11 Print diagnostic output in real format to .f06 file
PRTSW = 12 Punch diagnostic output in real format to .pch file
Displacement Output of GA and GB for CWELD and CFAST

Elements
The displacements of the projected grids GA and GB for CWELD elements with GRIDID, ELEMID,
ELPAT or PARTPAT format and CFAST elements are computed to display the relationship between these
elements and their connecting shell elements. As a result, the displacement output of GA and GB for
CWELD element with GRIDID or ELEMID format and MSET=OFF are no longer dummy zero values.
These displacements are calculated from the constraint equations described in the Basic Theory and
Methods section.
If GA or GB is not specified for CFAST elements or CWELD elements with ELPAT or PARTPAT option,
the program will create a grid internally, with the grid ID number starting from OSWPPT+1 (OSWPPT
is a parameter specified by PARAM Bulk Data entry). The user may request a positive PRTSW parameter
(1, 2, 11, or 12) in SWLDPRM Bulk Data entry to view the grid ID of GA or GB in the diagnostic output.
Input
The displacements of GA and GB are requested using the CONNECTOR keyword of the
DISPLACEMENT Case Control command. See MSC Nastran Quick Reference Guide for a detail
description of this command.
Outputs
The output is integrated with the displacements of the general grid points.
Basic Theory and Methods

For CWELD elements with GRIDID or ELEMID format, the displacements of the projected grids GA
and GB in basic coordinate are computed by the following equations:
CHAPTER 5 63
dA Z ∑ AI uI
I
dB Z ∑ BJ uJ
J
Where d A and d B are displacements at GA and GB. AI and BJ are constraint matrices. uI and uJ are
displacements of the connected shell grids.
For CWELD elements with ELPAT or PARTPAT format, the displacements of the auxiliary points in
basic coordinate are calculated first.
uI Z ∑ G I K uK
K
uJ Z ∑ GJ L uL
L
Where u I and u J are displacements of the auxiliary points. G IK and GJ L are RBE3 type constraint
matrices. u K and u L are displacements of the connected shell grids.
Then the displacements of the projected grids GA and GB in basic coordinate are computed by the same
constraint equations used for GRIDID and ELEMID options.
dA Z ∑ AI uI
I
dB Z ∑ BJ uJ
J
Where d A and d B are displacements at GA and GB. AI and BJ are constraint matrices. uI and uJ are
displacements of the connected auxiliary points.
Current Limitation
The displacement output of GA and GB are only available in Solution Sequences 101 and 103.
Example
This example demonstrates the various displacement output requests and their results for a small model
with two CWELD elements.
The input file follows:
nastran mesh
SOL 101
TIME 60
CEND
TITLE= two elements, identical location for GA, GB, GS
OLOAD= ALL
FORCE = ALL
SUBCASE 1
SUBTITLE= shear the weld
SPC= 1
LOAD= 1
DISP= ALL
SUBCASE 2
SUBTITLE= in plane twist
set 21 = 1002,1003,2011,thru,2014
spc= 1
LOAD= 2
DISP(CONN=ALL)=21
SUBCASE 3
SUBTITLE= upper bending
set 32 = 4
set 33 = 1012,1013,2001,thru,2004
spc= 1
LOAD= 3
DISP(CONN=32)=33
BEGIN BULK
$
$ Grids of lower shell
grid, 1001, , 0., 0., 0.
grid, 1002, , 20., 0., 0.
:
grid, 1013, , 20., 10., 5.
grid, 1014, , 0., 10., 5.
$ Grids of upper shell
grid, 2001, , 0., 0., 0.0
grid, 2002, , 20., 0., 0.0
:
grid, 2013, , 20., 10., 6.0
grid, 2014, , 0., 10., 6.0
$ spot weld grid
grid, 3001, , 10.0, 5.0, 0.0
grid, 3011, , 10.0, 5.0, 10.0
$ quad4s
cquad4, 4001, 10, 1001, 1002, 1003, 1004
cquad4, 5001, 10, 2001, 2002, 2003, 2004
cquad4, 4011, 10, 1011, 1012, 1013, 1014
cquad4, 5011, 10, 2011, 2012, 2013, 2014
:
$ property and material
pshell, 10, 10, 1.0, 10
mat1, 10, 2.e+5 , , 0.3, 0.785e-8
$
$ spot welds
$
cweld, 4, 4, 3001, elemid, , , , , +cw4
+cw4, 4001, 5001
cweld, 5, 4, 3011, elemid, , , , , +cw5
+cw5, 4011, 5011
pweld, 4, 10, 5.0
$
enddata
CHAPTER 5 65
The displacement results are shown as follows:
0 SUBCASE 1
D I S P L A C E M E N T V E C T O R
POINT ID. TYPE T1 T2 T3 R1 R2 R3

1001 G 0.0 0.0 0.0 0.0 0.0 0.0
1002 G 4.887498E-03 5.624979E-04 0.0 0.0 0.0 1.406245E-05
1003 G 4.887498E-03 -5.624979E-04 0.0 0.0 0.0 -1.406245E-05
1004 G 0.0 0.0 0.0 0.0 0.0 0.0
1011 G 0.0 0.0 0.0 0.0 0.0 0.0
1012 G 4.887498E-03 5.624979E-04 -7.685000E-01 -6.750000E-03 5.864999E-02 1.406245E-05
1013 G 4.887498E-03 -5.624979E-04 -7.685000E-01 6.750000E-03 5.864999E-02 -1.406245E-05
1014 G 0.0 0.0 0.0 0.0 0.0 0.0
2001 G 6.058339E-04 3.750000E-04 0.0 0.0 0.0 7.733938E-19
2002 G 5.605834E-03 3.750000E-04 0.0 0.0 0.0 1.084202E-1
2003 G 5.605834E-03 -3.750000E-04 0.0 0.0 0.0 1.191827E-18
2004 G 6.058339E-04 -3.750000E-04 0.0 0.0 0.0 7.623363E-19
2011 G 3.907157E-02 3.750000E-04 8.148733E-04 -2.562901E-16 3.850649E-02 2.836059E-17
2012 G 4.407157E-02 3.750000E-04 -7.693148E-01 -5.232793E-15 3.850649E-02 2.851452E-17
2013 G 4.407157E-02 -3.750000E-04 -7.693148E-01 -5.356609E-15 3.850649E-02 2.956961E-17
2014 G 3.907157E-02 -3.750000E-04 8.148733E-04 -2.636780E-16 3.850649E-02 2.803595E-17
3001 G 0.0 0.0 0.0 0.0 0.0 0.0
3011 G 0.0 0.0 0.0 0.0 0.0 0.0
101000001 G 2.443749E-03 8.917563E-18 0.0 0.0 0.0 8.402567E-19
101000002 G 3.105834E-03 8.944668E-18 0.0 0.0 0.0 9.774760E-19
101000003 G 2.443749E-03 2.517788E-16 -3.842500E-01 -2.775558E-15 3.842500E-02 2.303930E-17
101000004 G 4.157158E-02 3.029749E-15 -3.842500E-01 -2.789435E-15 3.850649E-02 2.786400E-17
0 SUBCASE 2
1002 G 1.499616E-02 3.324066E-02 0.0 0.0 0.0 2.330633E-03
1003 G -1.499616E-02 3.324066E-02 0.0 0.0 0.0 2.330633E-03
2011 G 6.107051E-03 -8.968769E-03 -1.542189E-16 3.303767E-17 1.479144E-16 1.890160E-03
2012 G 2.110705E-02 4.220943E-02 -3.132043E-15 5.117434E-17 1.498367E-16 3.390160E-03
2013 G -2.110705E-02 4.220943E-02 -2.615099E-15 5.202362E-17 1.383761E-16 3.390160E-03
2014 G -6.107051E-03 -8.968769E-03 1.645747E-16 3.118436E-17 1.398878E-16 1.890160E-03
101000001 G 6.505213E-18 1.662033E-02 0.0 0.0 0.0 1.580825E-03
101000002 G 6.722053E-18 1.662033E-02 0.0 0.0 0.0 2.640160E-03
101000003 G 3.035766E-18 1.662033E-02 -1.434168E-15 4.168935E-17 1.434168E-16 1.580825E-03
101000004 G 1.461505E-16 1.662033E-02 -1.434197E-15 4.178690E-17 1.439374E-16 2.640160E-03
0 SUBCASE 3

1012 G 3.460202E-16 9.473882E-16 -1.537000E+00 -1.350000E-02 1.173000E-01 6.294140E-17
1013 G -4.478185E-16 9.592263E-16 -1.537000E+00 1.350000E-02 1.173000E-01 6.130147E-17
2001 G 0.0 0.0 8.021409E-04 -1.965454E-14 7.693148E-02 0.0
2002 G 0.0 0.0 -1.537827E+00 6.397660E-15 7.693148E-02 0.0
2003 G 0.0 0.0 -1.537827E+00 6.758049E-15 7.693148E-02 0.0
2004 G 0.0 0.0 8.021409E-04 -2.021495E-14 7.693148E-02 0.0
101000001 G 0.0 0.0 -7.685000E-01 -6.591949E-15 7.685000E-02 0.0
101000002 G 0.0 0.0 -7.685127E-01 -6.664110E-15 7.693148E-02 0.0
Nonhomogeneous Multipoint Constraint
Introduction
In the past, if a nonhomogeneous multipoint constraint was desired then the method of “SLACK”
variable was required. That is, the nonhomogeneous right hand side of the MPC equation was written
using a scalar or grid point and an SPC or SPCD specifying the right hand side value. In this release a
MPCY Bulk Data entry is introduced that allows the user to enter in a right hand side value directly into
the MPC equation. The entry defines an equation of the form
A m um H ∑ Ai ui Z Ym
i
Associated with the MPCY entry is another new Bulk Data entry MPCD used to define a load selectable
value for Y m of nonhomogeneous multipoint constraint.
Benefits
The user can define a nonhomogeneous multipoint constraint directly.
Input
1. The MPCY Bulk Data entry is used to define a nonhomogeneous multipoint constraint directly.
This entry can also be used to define a standard homogenous multipoint constraint as well.
2. The MPCD Bulk Data entry is used to define a load selectable value for Ym of nonhomogeneous
multipoint constraint.
Output
Standard MSC Nastran MPCF output is available.

Currently this method is not supported in the Dynamic solution sequences because the dynamic load
generation modules have not yet been taught to automatically expand Y m into the required number of load
columns.
Theory
The basic relations are as follows:
The unconstrained stiffness matrix equation in MSC Nastran is
CHAPTER 5 67
[ Kg g ]{ Ug } Z { Pg } (5-1)
If we apply a nonhomogeneous MPC constraint
[ Rµ g ]{ Ug } Z { Yµ } (5-2)
Then Eq. (5-1) becomes
   
KN N KN M  UN   PN H QN 
  Z   (5-3)
KM N KM M  UM   PM H QM 
   
where QN and QM are the forces of constraint necessary to impose Eq. (5-2).
Partition Eq. (5-2) as
 
 UN 
Rµ N Rµ M   Z { Yµ } (5-4)
 UM 
 
solve for UM to get
{ UM } Z { θM H GM N UN } (5-5)
Ó1 Ó1
where θ M Z R µM Y µ and G MN Z Ó R µM R µN
Substitute Eq. (5-5) into Eq. (5-3) to get
[ KM N H KM M GM N ]{ UN } Z { PM H QM Ó KM M θM } (5-6)
or
{ QM } Z [ KM N H KM M GM N ]{ UN } Ó { PM } H { KM M θM }




















current new
(5-7)
Since the θ of Eq. (5-5) are prescribed, any virtual variation of Eq. (5-5) the results by definition of
δ θ M Z 0 . Hence the standard conjugate force transformation holds. Or
 T 
{ QN } Z  ÓGM N QM  (5-8)
 
where the minus sign comes from the fact we are imposing constraints and that QN forces are reactive to
the Q M constraint forces.
Then the upper Eq. (5-3) with Eq. (5-8) becomes
 T 
{ KN N UN H KN M [ GM N UN H θM ] } Z  PN Ó GM N QM 
 
and using Eq. (5-7) to expand the result to get

T T T T
[ K NN H K NM G MN H G MN K MN H G MN K M M G MN ] { U N } Z { P N H G MN P M } Ó [ K NM Ó G M N K MM ] { θ M }
or
T
[ KN N ]{ UN } Z { PN } Ó[ KN M H GM N KM M ]{ θM } (5-9)
























current new
Examples
As a simple example consider a horizontal rod structure. The left and right ends are clamped. However
as Figure 5-3 shows, there is a break in the structure of 0.1 units of length. If the grid id at the left end of
the break is A and the grid id at the right end of the break is B , we can tie the structure by the MPC
equation
U B Ó U A Z 0.1
In the example, we show how to write the MPC equation in two ways. The first way uses the standard
“SLACK” variable method. The second way uses the new MPCY entry.
Figure 5-3 Example Rod Structure for MPCY

SOL 101
CEND
TITLE = DEMONSTRATE USE OF MPCY
SPC = 300
MPC = 300
LOAD= 300
DISPL= ALL
OLOAD= ALL
MPCFO = ALL
SPCFO= ALL
CHAPTER 5 69
ELFOR = ALL
BEGIN BULK
$
PROD 1 1 1.
MAT1 1 1.+7 0.
GRDSET 3456
$
$
$ STANDARD “SLACK” VARIABLE METHOD
$
$ |1 2 3 4 5|
$ |O---------------O---------------O O---------------O|
$ | 1 2 3 |
$
$ GRIDs 3 and 4 have a 0.1 unit gap between them
$ We wish to impose relationship in x-direction
$ U4 = U3 + 0.1
$ When GRID 3 has a x-direction load of 1.+5
$
GRID 1 0. 0. 0. 123456
GRID 1 100. 0. 0. 23456
GRID 1 200. 0. 0. 23456
GRID 1 200.1 0. 0. 23456
GRID 1 300.1 0. 0. 123456
$
SPOINT 6
$
CROD 1 1 1 2
CROD 2 1 2 3
CROD 3 1 4 5
$
FORCE 300 3 1.+5 1.
$
SPC 300 6 1 0.1
MPC 300 4 1 -1. 3 1 1.
6 1 1.
$ ^
$ |
$ |
$ |
$ SPOINT with 0.1 in displacement gap
$
$
$ ----------------> NEW MPCY METHOD <----------------
$
$ |11 12 13 14 15|
$ |O---------------O---------------O O---------------O|
$ | 11 12 13 |
$
$ GRIDs 13 and 14 have a 0.1 unit gap between them
$ We wish to impose relationship in x-direction
$ U14 = U13 + 0.1
$ When GRID 13 has a x-direction load of 1.+5
$
GRID 11 0. 1. 0. 123456
GRID 12 100. 1. 0. 23456
GRID 13 200. 1. 0. 23456
GRID 14 200.1 1. 0. 23456
GRID 15 300.1 1. 0. 123456
$
CROD 11 1 11 12
CROD 12 1 12 13
CROD 13 1 14 15
$
FORCE 300 13 1.+5 1.
$
MPCY 300 14 1 1. 0.1
13 1 -1.
ENDDATA
The results are:

1 G 0.0 0.0 0.0 0.0 0.0 0.0
2 G 3.000000E-01 0.0 0.0 0.0 0.0 0.0
3 G 6.000000E-01 0.0 0.0 0.0 0.0 0.0
4 G 7.000000E-01 0.0 0.0 0.0 0.0 0.0
5 G 0.0 0.0 0.0 0.0 0.0 0.0
6 S 1.000000E-01
11 G 0.0 0.0 0.0 0.0 0.0 0.0
12 G 3.000000E-01 0.0 0.0 0.0 0.0 0.0
13 G 6.000000E-01 0.0 0.0 0.0 0.0 0.0
14 G 7.000000E-01 0.0 0.0 0.0 0.0 0.0
15 G 0.0 0.0 0.0 0.0 0.0 0.0
L O A D V E C T O R
3 G 1.000000E+05 0.0 0.0 0.0 0.0 0.0
6 S 0.0
13 G 1.000000E+05 0.0 0.0 0.0 0.0 0.0
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
1 G -3.000000E+04 0.0 0.0 0.0 0.0 0.0
5 G -7.000000E+04 0.0 0.0 0.0 0.0 0.0
6 S 7.000000E+04
11 G -3.000000E+04 0.0 0.0 0.0 0.0 0.0
15 G -7.000000E+04 0.0 0.0 0.0 0.0 0.0
F O R C E S I N R O D E L E M E N T S ( C R O D )
ELEMENT AXIAL ELEMENT AXIAL
ID. FORCE TORQUE ID. FORCE TORQUE
1 3.000000E+04 0.0 2 3.000000E+04 0.0
3 -7.000000E+04 0.0 11 3.000000E+04 0.0
12 3.000000E+04 0.0 13 -7.000000E+04 0.0
F O R C E S O F M U L T I P O I N T C O N S T R A I N T

3 G -7.000000E+04 0.0 0.0 0.0 0.0 0.0
4 G 7.000000E+04 0.0 0.0 0.0 0.0 0.0
6 S -7.000000E+04
13 G -7.000000E+04 0.0 0.0 0.0 0.0 0.0
14 G 7.000000E+04 0.0 0.0 0.0 0.0 0.0
* * * END OF JOB * * *
CHAPTER 5 71
Ch. 6: Optimization MSC Nastran 2007 Release Guide
6 Optimization

Topology Optimization Enhancements

Automatic External Superelement Optimization (AESO)
Randomization of a User’s Input Data File (Pre-release)

Random Elimination of Element Types (Pre-release)
Topology Optimization Enhancements
Introduction
Topology optimization capability was first released in 2005 and a number of manufacturability
constraints were added to MSC Nastran 2005 r3. New features have been added to this release based on
feedback from clients. With these enhancements, MSC Nastran 2007 SOL200 is able to support:
combined topology, sizing, and shape optimization, multiple mass reduction targets, cyclical symmetry
constraints, and adjoint design sensitivity analysis for inertia relief of static analyses. A major
performance enhancement of module DOPR1 has been made to speed up minimum member size control
and sizing optimization with many thousands variables.
Benefits
Combined Topology, Sizing, and Shape Optimization
It is often recommend that topology optimization is first used to find efficient design concepts at the early
design stage whereas sizing and/or shape optimization is used for detail design based on the topology
design proposals at a later design stage. The use of topology, sizing, and shape optimization
simultaneously may find possible better design since the interaction of sizing and/or shape variables with
topology optimization is considered during a single design optimization process. Another benefit of this
feature is that the DRESP2 BETA function (minimize the maximum responses) is now available to
topology optimization.
Cyclical Symmetry Constraints

A mirror symmetry constraint was added to MSC Nastran 2005 r3. It is also desirable to design a
rotational symmetric component or system. By using cyclical symmetry constraints in topology
optimization, a rotational symmetric design can be obtained regardless of the boundary conditions or
loads. This cyclical symmetric constraint capability can be used for irregular finite element meshes.
Adjoint Design Sensitivity Support for Inertia Relief

A direct design sensitivity analysis method is only available for inertia relief of static analyses in previous
versions. The direct method is not affordable computationally for topology optimization since many
thousands of variables are often involved. The adjoint design sensitivity analysis method has been
developed to benefit not only inertia relief topology optimization but also inertia relief sizing
optimization with many design variables. The adjoint method is automatically selected when it is more
efficient computationally.
Multiple Mass Target

Type one response FRMASS (DRESP1=FRMASS) used to be total fraction mass of topological designed
properties. Feedback from industrial users have shown that it is desirable to set up different mass
CHAPTER 6 75
Optimization
reduction targets on multiple designed parts for a built-up structures. This enhancement reflects this
requirement.
Module DOPR1 Performance Enhancements

When a large value is given for minimum member size TDMIN, a significant CPU time is spent on
module DOPR1 in previous versions. With this enhancement, a substantial performance speedup (3-100
times) is achieved for minimum member size control. Module DOPR1 is also enhanced to efficiently
support optimization problems with many thousands sizing design variables (for example, a sizing
optimization deck with 320,000 design variables, the DOPR1 module in previous MSC Nastran versions
and other Nastran products requires 16,000 seconds, MSC Nastran 2007 DOPR1 module requires only
10 seconds).
Input
The TOPVAR Bulk Data entry has been enhanced to provide cyclical symmetry constraints. To select a
topologically designable region, the user needs to specify a group of elements by using a Bulk Data entry,
TOPVAR. The cyclical symmetry constraints are then applied on all elements referencing the given
property on TOPVAR entry.
The enhanced TOPVAR format is:
N 2 3 4 5 6 7 8 9 10
TOPVAR ID LABEL PTYPE XINIT XLB DELXV POWER PID
“SYM” CID MSi MSi MSi CS NCS
“CAST” CID DDi DIE
“EXT: CID EDi
“TDMIN” TV
Field Contents
PID Property entry identifier (Integer > 0). This PID must be unique for PIDs
referenced by other TOPVAR, DVPREL1 and DVPREL2 entries. Topology
and sizing variables cannot share the same properties. (Integer > 0)
“SYM” Indicates that this line defines symmetry constraints.
CID Rectangular coordinate system ID used for specifying manufacturing
constraints. See Remark 2. (Blank or Integer > 0; Default = 0)
CS Cyclical symmetry axis (character X, Y, Z). See Remark 3.
NCS Number of cyclical symmetric segments in 360 degrees. (Integer > 0). The
angle for one segment is calculated by 360/NCS. See Remark 3..
New Remarks:
1. The topologically designable element properties include PROD, PBAR, PBARL, PBEND,
PBEAM, PBEAML, PSHELL, PSHEAR, PSOLID, and PWELD. Multiple TOPVARs are
allowed in a single file. Combined topology, sizing, and shape optimization is supported in a
single file. However, TOPVAR cannot be used with DVMREL1 and DVMREL2 entries.
2. Only CORD1R and CORD2R can be used as a referenced coordinate system to specify topology
manufacturing constraints. Only one reference coordinate system CID is allowed for each
TOPVAR entry.
3. The first cyclical symmetry segment starts at the X-axis when CS=Z (at Z-axis when CS = Y, and
at the Y-axis when CS = X). One cyclical symmetry constraint can be combined with one mirror
symmetry constraint as long as the axis of cyclic symmetry is normal to the plane of mirror
symmetry. For example, MSi = YZ and CS = X, MSi = XZ and CS = Y, and MSi = XY and CS
= Z. This feature can also be used for rotational parts with < 360 degrees but NCS must be given
in 360 degrees.
4. For “EXT” constraints, possible combinations are (ED=X, MSi=XY, and/or ZX or CS=X),
(ED=Y, MSi=YZ, and/or XY or CS=Y), (ED=Z, MSi=ZX, and/or YZ or CS=Z).
5. For “CAST” constraints, possible combinations are (DD=X or X-, MSi=XY and/or ZX or CS=X),
(DD=Y or Y-, MSi=YZ and/or XY or CS=Y), (DD=Z or Z-, MSi=ZX and/or YZ or CS=Z).
Modified Type One Responses - Fractional Mass

To allow the user to put different mass reduction constraints on multiple designed properties,
RTYPE=FRMASS entry has been extended to provide property ID (PID) at attribute field ATTi. The
format for this extension is shown in Table 6-1.
Table 6-1 Modified RTYPE=FRMASS

Response Attributes
Response Type ATTA (Integer > ATTB (Integer > 0 or
(RTYPE) 0) Real > 0.0) ATTI (Integer > 0)
coj^pp= _i^kh _i^kh _i^kh=çê=mêçéÉêíó=
oÉã~êâ=N fa=EmfaF
Remark:
1. RTYPE= FRMASS (mass fraction of topological designed elements) entries are used for
topology optimization or combined topology, sizing and or shape optimization. ATTi=Blank is
for total mass fraction of all topological designed properties. ATTi=PID is the mass fraction of
topological designed property PID.
CHAPTER 6 77
Optimization

• Although combined topology and sizing optimization is supported, TOPVAR and DVPREL1/2
entries cannot reference the same property ID (PID).
• While FRMASS is calculated for topological designed properties only, RTYPE=WEIGHT
computes total weight including all designed and non-designed parts. For combined topology
and sizing/shape optimization problems, it is recommended that RTYPE=FRMASS is used for
topological designed property mass reduction constraints and RTYPE=WEIGHT is used for
total mass reduction constraints.
Example 1 – Wheel (wheeltop.dat)

A wheel model shown in Figure 6-1 is used to demonstrate MSC Nastran topology optimization cyclical
symmetry capabilities. The wheel is modeled with six-sided solid elements (CHEXA). The wheel outer
layers and bolts are non-designable. One load case is considered. The structural compliance is minimized
(i.e., minimize the total strain energy of the structure) with a mass target 0.1 (i.e., remove 90% of the
material). Although the load is not cyclically symmetric about the Y-axis, the design is required to be
cyclically symmetric about the Y-axis with five segments.
Figure 6-1 Wheel

Input
The input data for this example related to topology optimization model is given in Listing 6-1. The
coordinate system CORD2R = 1 is created to be used to specify cyclical symmetric constraints. The field
CS on the SYM line is Y-axis with NCS=5.
Listing 6-1 Input File for Example 1
DESOBJ = 10
DESGLB = 1
ANALYSIS = STATICS
SMETHOD = ELEMENT
SUBCASE 1
SPC = 2
LOAD = 2
BEGIN BULK
CORD2R 1 10.512 33.3312 12.9921 -22.209833.3312 4.88385
28.388 33.3313 -19.7297
DCONSTR 1 2 .1
TOPVAR 1 PSOLID PSOLID .1 2
SYM 1 Y 5
DRESP1 2 FRM FRMASS
DRESP1 10 COMP COMP
Output
Figure 6-2 shows the topology optimized result that is smoothed by using Patran. It is noticed that cyclical
symmetry is obtained even though the loading is not cyclically symmetric.
CHAPTER 6 79
Optimization
Figure 6-2 Wheel Topology Design

Automatic External Superelement Optimization

(AESO)
Introduction
The Manual External Superelement Optimization capability (MESO) has been available since the
MSC.Nastran 2004. In this technique, the user manually partitions the analysis model into two parts: a
designed part and a non-designed part. The latter is treated as one external part superelement while the
former is defined as a residual structure. A creation run is performed which applies Component Mode
Synthesis (CMS) or Static Condensation to the part superelement and stores the resulting boundary
matrices in a database or a punch file. The original optimization task is performed on the assembly run
that assembles those boundary matrices into the residual model for solving system solutions. The strategy
is most efficient when the size of the design model (or residual model) is much smaller than the size of
the original analysis model. However, although the feature is efficient in CPU time, since both files of
both creation and assembly runs must be created by the user, significant effort in manual partitioning the
model might outweigh the performance gain.
The AESO capability presented here extends the MESO in an important way: rather than requiring the
user to segregate his large model into a designed and non-designed part, the process does this
automatically by identifying which parts of the finite element model are affected by the design task. In
essence, the new AESO capability provides an efficient and accurate solution in a user friendly way.
Benefits
Several major benefits are:
1. Removes a tedious and error prone task from the user in preparing the user input data.
2. Does not require the user to be knowledgeable in the specialized area of superelements in general
and external superelements in particular.
3. Provides an efficient and accurate approach for large-scale design optimization tasks.
4. Enable the performance of various design studies rapidly once the model has been divided into a
designed and non-designed part. Examples of this are the setting up of different design constraints
and objective in the studies to gain insight into the design and the available trade-offs or the
applying of various frequency excitation loadings in the frequency response analyses.
Methodology
A complete AESO task involves two separate MSC Nastran job runs: 1) the first run is an AESO creation
run (or simply creation run) whose logical flow is described in Figure 7-3 and 2) the second run is an
AESO assembly run (or simply assembly run) whose logical flow is described in Figure 7-5.
As shown in Figure 7-3, the creation run automatically partitions the original analysis model into the
residual (the designed part) and external SE (the non-designed part). This automatic partition procedure
will assign the following grid points to the residual:
CHAPTER 6 81
Optimization
1. All grid points that belong to a design model consisting of DRESP1, DVGRID, DVPRELi,
DVMRELi and DVCRELi entries;
2. All grid points that are referenced on all static or dynamic loading entries such as DAREA,
DPHASE, FORCE, MOMENT, PLOADi, TEMP entries;
3. All grid points for a rigid element that has one or more connecting grid belonging to the residual;
4. Any grid point that is a dependent grid on an MPC entry
After the automatic partition procedure, a new user input file is created from the residual (Figure 7-4).
Then, the remainder of the creation run is the application of Static Condensation and/or Component
Modes Synthesis procedures to produce stiffness, mass, damping boundary matrices. After the creation
run is complete, a Nastran database is saved to store the boundary matrices and a .asm file is also created
to include superelement boundary connection information.
The assembly run is similar to a conventional SOL 200 task as shown in Figure 7-5 by utilizing all three
types of data generated from the creation run (Figure 7-4). The original optimization problem is solved
by assembling the boundary matrices into the residual for the system solutions.
fåáíá~ä=rëÉê=
fåéìí=cáäÉ
kÉï=fåéìí=cáäÉ=
fåáíá~ä=rëÉê=
Automatic oÉëáÇì~ä= Ñçê======== fåéìí=cáäÉ
Model Partition ^ëëÉãÄäó=oìå=
bñíÉêå~ä=pb
jp`Kk~ëíê~å=
`jp=íç=ÅêÉ~íÉ= buqpb=
Optimizer
ÄçìåÇ~êó=ã~íêáÅÉë= a~í~Ä~ëÉ=
.asm file
Figure 7-3 AESO Creation Run Figure 7-4 Output of Figure 7-5 AESO Assembly Run
Creation Run and
Input of Assembly
Run
Input
1. Two parameters are added on the DOPTPRM entry for an AESO job:
AUTOSE - flag to request an AESO job (integer 0, 1, Default = 0). AUTOSE = 1 activates an
AESO creation run.
DRATIO - the threshold value that is used to turn off an active AESO job if the ratio of the size
of the design model to that of the analysis model is greater than DRATIO (Real > 0.; Default =
0.1).
2. An ASSIGN statement with a logical key name ‘AESO’ is placed in the FMS to specify an input
file name for the assembly run.
To illustrate ideas behind the input and output for an AESO task, a test problem (aeso1.dat) is used here.
Figure 6-6 shows a sample model whose upper left portion covering elements 18 to 42 (SE 1) is the non-
designed part while the rest of structure is the residual structure.
The listing below is a condensed version of the creation run file (aeso1.dat) that only shows the required
user input to invoke an AESO creation run. The assign aeso='aeso1_2.dat' statement is specified in the
FMS section and autose 1 and dratio 0.9 are requested on the DOPTPRM entry. The DRATIO=0.9 here
overrides the default.
SE1
Figure 6-6 A Sample Model

CHAPTER 6 83
Optimization
assign aeso='aeso1_2.dat'
SOL 200
CEND
desobj(max) = 1
analysis = modes
…..
Begin Bulk
……
doptprm desmax 10 autose 1 dratio 0.9
enddata
Figure 6-7 Condensed Version of the Creation Run File, dseoptl.dat
Outputs
As shown in Figure 7-4, three types of data generated from the creation run are saved in the working
directory: a Nastran database, a .asm file and a new input file for the assembly run (or an assembly file).
This section describes each of these items and explains how they are used in the assembly run. In
addition, some special print outputs from the creation run are shown that display the model partition
information.
Nastran Database Files

Two Nastran database files: aeso1.MASTER and aeso1.DBALL are automatically saved when the
creation run is submitted with SCR=NO option. Notice the size of the database is much smaller than the
regular Nastran database because only boundary matrices are stored. The Master file will be referenced
by an ASSIGN statement in the new input file described below.
The .asm File

The ‘aeso1.asm’ file that includes the boundary connection information follows. As shown in Figure 6-6,
this problem has only two boundary points, 35 and 36 between the residual and the external superelement
1. This file is accessed through an INCLUDE command in the assembly file.
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
$
$ ASSEMBLY PUNCH (.ASM) FILE FOR EXTERNAL SUPERELEMENT 1
$
$ THIS FILE CONTAINING BULK DATA ENTRIES PERTAINING TO
$ EXTERNAL SUPERELEMENT 1 IS MEANT FOR INCLUSION
$ ANYWHERE IN THE MAIN BULK DATA PORTION OF THE ASSEMBLY RUN
$
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
SEBULK 1EXTERNAL MANUAL
$
SECONCT 1 0 5.0E-05
35 35 36 36
$
$ BOUNDARY GRID DATA
$
GRID 35 -3.6 6. 0.
GRID 36 -2.8 6. 0.
$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$
New Input File for the Assembly Run

Notice that the name of this file, aeso1_2.dat is specified on the ASSIGN AESO statement in the creation
run (Figure 6-7). It is a standard Nastran input file. The AESO specific contents in the Executive Control,
Case Control and Bulk Data Sections are listed in Figure 6-8 and are described below.
}
nastran rseqcont=1
assign se1= './aeso1.MASTER'
dblocate datablk(EXTDB) logical=se1,
CONVERT(SEID=1) Executive Control Section
SOL 200
CEND
}
desobj(max) = 1
analysis = modes
…….
subcase 10 Case Control Section
method = 1
spc = 10
$
}
begin bulk
include './aeso1.asm'
…….
doptprm desmax 10 Bulk Data Section
enddata
Figure 6-8 Highlights of the Assembly Run File, aeso1_2.dat

1. Executive Control Section
CHAPTER 6 85
Optimization
The Nastran rseqcont=1 statement instructs the input file processor to ignore all continuation
fields. This statement is automatically created in this file regardless of whether the creation run
has it or not.
The next two statements assign the Nastran Master database file and locate the EXTDB datablock
that stores various boundary matrices.
Notice that the other statements in Executive Control Section of the creation run are not retained.
2. Case Control Section
The whole Case Control section of the creation run is retained in the assembly file.
3. Bulk Data Section
This section completely defines the residual structure. The include ‘./aeso1.asm’ command
allows the assembly run to access the .asm file created from the creation run. In addition, the
autose 1 and dratio 0.9 have been removed from the doptprm entry of the creation run.
Special Print Outputs from Creation Run

The following output is taken from the aeso1.f06 file. It displays detailed information about the model
partition. You may use Figure 6-6 to help read the printout here. Notice that the Superelement 1 covers
the non-designed part while the residual (or Superelement 0) covers the designed part.
Listing 6-2 Printout Showing Model Partition of Designed and Non-Designed Parts
BOUNDARY SEQUENCE ASSIGNMENT TABLE

BOUNDARY
SEQUENCE ID -------------------------- ASSIGNED TO POINT ID (SUPERELEMENT) ---------------------------------
1B 35 ( 0) 35 ( 1)
2B 36 ( 0) 36 ( 1)
SUPERELEMENT 0
LIST OF INTERIOR POINTS ( TOTAL NO. OF INTERIOR POINT = 70 )
INDEX -1- -2- -3- -4- -5- -6- -7- -8- -9- -10-
1 1 2 3 4 5 6 7 8 9 10
11 11 12 13 14 15 16 17 18 19 20
21 21 22 23 24 25 26 27 28 29 30
31 31 32 39 40 41 42 43 44 51 52
41 53 54 55 56 63 64 65 66 67 68
51 75 76 77 78 79 80 87 88 89 90
61 91 92 99 100 101 102 103 104 1B 2B
SUPERELEMENT 0
LIST OF ELEMENTS ( TOTAL NO. OF ELEMENTS = 42 )
INDEX -1- -2- -3- -4- -5- -6- -7- -8- -9- -10-
1 1 2 3 4 5 6 7 8 9 10
11 11 12 13 14 15 16 17 43 44 45
21 46 47 48 49 50 51 52 53 54 55
31 56 57 58 59 60 61 62 63 64 65
41 66 67
SUPERELEMENT 1
LIST OF EXTERIOR POINTS ( TOTAL NO. OF EXTERIOR POINT = 2 )
INDEX -1- -2- -3- -4- -5- -6- -7- -8- -9- -10-
1 1B 2B
SUPERELEMENT 1
LIST OF INTERIOR POINTS ( TOTAL NO. OF INTERIOR POINT = 34 )
INDEX -1- -2- -3- -4- -5- -6- -7- -8- -9- -10-
1 33 34 37 38 45 46 47 48 49 50
11 57 58 59 60 61 62 69 70 71 72
21 73 74 81 82 83 84 85 86 93 94
31 95 96 97 98
SUPERELEMENT 1
LIST OF ELEMENTS ( TOTAL NO. OF ELEMENTS = 25 )
INDEX -1- -2- -3- -4- -5- -6- -7- -8- -9- -10-
1 18 19 20 21 22 23 24 25 26 27
11 28 29 30 31 32 33 34 35 36 37
21 38 39 40 41 42

• You may adjust DRATIO to allow an assembly run with larger or smaller residual model. The
UIM 7824 provides brief information about the sizes of your analysis model and design model in
terms of number of the grid points.
*** USER INFORMATION MESSAGE 7824 (DSGRDM)

THE NUMBER OF GRID POINTS IN THE ANALYSIS MODEL = 104.
THE NUMBER OF GRID POINTS IN THE DESIGN MODEL = 70.
THE DESIGN MODEL COMPRISES 67.3 PERCENT OF THE ANALYSIS MODEL.
CHAPTER 6 87
Optimization
• For an AESO job with Analysis=MODES or MFREQ, it is recommended to activate the matrix
domain based decomposition with ‘domainsolver acms(partopt=dof)’ in the Executive Control
section to speed up the CMS procedure.
• Always specify the ASSIGN AESO=’filename.ext’ statement in the creation run to define the
name of the assembly file. Directly assigning the original job name to filename should be
avoided and will cause the assembly run to fail with User Fatal Message 713. A good practice is
to add some suffix to the original file name such as myjob_2nd.dat where myjob is the original
file name.
*** USER FATAL MESSAGE 732 (OPFUNT)

LOGICAL NAMES 'INPUT ' AND 'AESO ' ARE ASSIGNED TO THE SAME PHYSICAL FILE.
USER INFORMATION: PHYSICAL FILE NAME 1: ./abc.aeso
PHYSICAL FILE NAME 2: ./abc.aeso
USER ACTION: CHANGE FILE NAME ON ONE OF THE ASSOCIATED ASSIGN STATEMENTS.
• When submitting the AESO creation run, use SCR=NO option. Otherwise, the Nastran database
will not be retained after the creation run is done. However, it is optional for submitting an
assembly run.
• After the creation run is complete, check the following user information message in the f06 file
to ensure the job is terminated successfully.
^^^
^^^ USER INFORMATION MESSAGE 9181 (FEA)
^^^ THE JOB IS TERMINATED FOR AN AUTO EXTERNAL CREATION RUN
^^^
• The assembly input file may be modified to perform various design studies as long as the
changes do not affect the boundary matrices stored in the database.
• If the AESO creation run includes a GRAV entry, it will be terminated with the following
message. The same applies to TEMPD entry. They should be removed from the file if they are
temporary inactive for the current task. Gravity and TEMPD loads are not supported with AESO
since they apply loads to the entire structure and therefore block any partitioning.
*** USER FATAL MESSAGE 7699 (DSGRDM)

A GRAV Bulk Data entry is specified in an AESO creation run.
USER INFORMATION: The AESO run does not support the GRAV entry.
• Duplicate GRID entries are allowed in the .asm file and in the assembly input file. They will be
automatically removed during the assembly run within the location tolerance specified by the
TOL field on the SECONCT entry. For the AESO jobs, the default of the location tolerance has
been increased to 5.E-5 from the original 1.E-5. However, due to numerical imperfection, this
tolerance may need to be adjusted particularly for the cases in which the boundary grid points
are defined in one or multi-levels of coordinate systems.
• Bulk Data parameter SEMAPPRT can be used to control the printout of the model partition
information shown in Listing 6-2. For example, Setting SEMAPPRT to 0 will turn off the
printout.
• If the AESO task includes DVGRID entries, make sure that the grid points referenced by
DVGRID entries are inside the residual. Since the grid points on DVGRID entries vary during
the design process, including them as part of boundary grid points will invalidate the invariance
of those boundary matrices. Currently, the grid points on the DVGRID entries will not be
automatically assigned to be inside the residual. ASET and ASET1 entries can be used to create
an enclosure or a barrier to ensure the grid points referenced on the DVGRID entries are always
placed inside the residual.
• Since all the CORD1i entries are automatically converted to CORD2i entries during the AESO
creation run. The DVGRID entry should not reference the grid points that define the CORD1i
entry.
• The AESO tasks do not support acoustics response.
Examples
Road Response Optimization with CAMARO Model with Analysis=MFREQ
Figure 6-9 CAMARO Model
A CAMARO model provided by GM as shown in Figure 6-9 is modeled with 23K grid points, 37K
elements and 137K DOFs, random inputs applied on left and right suspension including cross-
correlation, to simulate road conditions.
The design task is to vary 9 spring constants of engine mounts modeled by elastic elements in two design
cases:
Case A: minimize the sum of RMS acceleration at Driver’s seat and passenger’s seat and limiting the
PSD response at steering column and
Case B: minimize the RMS acceleration at Driver’s seat and maintain frequency dependent limits on
driver’s seat.
CHAPTER 6 89
Optimization
Because these design variables and responses are limited to a small part of the total vehicle, the size ratio
of design model vs. analysis model is <1.% (the residual structure has 298 elements, 163 interior points
and 130 boundary points while the external superelement has 33345 elements and 22761 interior points).
To demonstrate the efficiency of the AESO capability, the tables below present the results from a regular
run and the AESO job. First, the same final designs are achieved by both jobs. However, the AESO job
takes 1/5 of the time to complete Case A. Furthermore, since Case B just changes the design objective
and constraint formulation, the boundary matrices from the creation run are invariant. Therefore, only
the assembly run for Case B is required. Comparing the time spent on this assembly run with the single
run as shown in Case B in the bottom half of the table, the speed up is 16X.
Initial Final Init Max Final Max # Design Clock Time

Case A OBJ OBJ Const. Const. Cycle (Minutes)
Full Model Run 0.1534 -0.0639 0.1329 -0.2102 9 76
AESO Creation N/A N/A N/A N/A N/A 10
Run
AESO Assembly 0.1534 -0.0639 0.1329 -0.2102 9 7
Run
Total time of two 16
AESO runs
Performance 5
Ratio

Case B OBJ OBJ Const Const Cycle (Minutes)
Full Model 0.0713 0.0586 0.2855 0.0081 14 110
Run
AESO N/A N/A N/A N/A N/A 0
Creation
Run
AESO 0.0713 0.0584 0.2855 -0.0108 14 7
assembly
Run
Road Response Optimization with CAMARO Model with Analysis=DFREQ

In general, direct frequency approach is more expensive than the modal approach. However, when the
size of the residual model is small, the Direct Frequency approach can also be performed efficiently as
shown here. Case A of the same road response optimization problem are solved with both a regular direct
frequency response optimization and the direct frequency AESO. The table below presents the results.
Again, the same final designs are achieved. However, the speed up obtained by the AESO over the
regular job is 27X.

Case A OBJ OBJ Const Const Cycle (Minute)
Full Model Run 0.1535 0.1327 -0.0631 -0.2073 9 388
AESO Creation Run N/A N/A N/A N/A N/A 10
AESO Assembly Run 0.1534 0.1327 -0.0636 -0.2062 9 5
Total time of two 15
AESO runs
Performance Ratio 27
The following figure shows the plot of the total clock time spent by two runs vs. design cycle and it
clearly shows that the AESO is much more efficient than the regular run because the cost per design cycle
for AESO (or the slope of the curve for the AESO run) is much smaller than that for a regular run (the
slope of the curve for the regular run). Although for this example, the one time cost for the AESO due to
the creation run is smaller than that for a regular run, in general, this one time cost may be larger because
the time spent by CMS procedure on mass matrix reductions could be expensive.
450
400
Clock Time (Minutes)
350
300
250
200
150
100
50
0
0 2 4 6 8 10
Design Cycle
Total Time (Regular Run) Total Time (AESO)

CHAPTER 6 91
Optimization
Miscellaneous
Prior to this release, SOL200 supports a limited Part optimization capability while it fully supports the
traditional Superelement optimization. Notice that the traditional SE is defined with SESET Bulk Data
entry while the Part is defined with Begin Super Bulk Data entry. In addition, the Part may be defined
either as External or Internal.
The enhancement added in this release provides a robust Part optimization capability as long as the
design model is in the residual (or an upstream Part can not be designed or constrained). One example
of its application would be: an airplane wing is modeled as an internal Part while the engine is modeled
as an external Part. Then these two are attached to the fuselage, a residual that will be designed.
The Table below summarizes the features of Superelement Optimization and Part Optimization in
SOL200 in this new release.
Design Model in Upstream

Item Feature Support in SOL 200 SE or Part
1 Traditional SE Yes Yes
2 Ext. Part Yes No
3 Int. Part Yes No
4 2+3 Yes No
1+2,1+3,1+2+3 No No
To illustrates how to create a Part optimization job, the sample model as shown in Fig 7-6 is used here.
The top left squared region is modeled as Part 1 (elements 18 to 42) while the top right squared region
is modeled as Part 2 (elements 43 to 67). The rest is modeled as a residual (elements 1 to 17). Grid points
1 and 2 are fully constrained. The design task is to maximize the first natural frequency by varying the
thickness of elements 1 to 17 within a given range.
The following listing is a condensed version of dseopt18.dat that shows three major Bulk Data sections.
Its full version can be obtained from the Test Problem Library. The unique feature of a Part optimization
job is the multiple Bulk Data sections and how the design model definition is placed. Three sections in
this job are:
1. Main section (between Begin Bulk and Begin Super = 1) that defines the residual model. Notice
that all the designed entries such as DESVAR, DVPREL1, DRESP1 entries are defined in this
section because the design model must be within the residual.
2. Section for Part 1 (between Begin Super = 1 and Begin Super =2). This section defines Part 1
model.
3. Section for Part 2 (between Begin Super = 2 and Enddata). It defines Part 2 model.
Notice that the design model must be placed in the main Bulk Data section because the design model is
required to be in the residual. The detailed steps to define Parts and residual model are not skipped here.
SOL 200
diag 8,15,56
CEND
TITLE = Test for Internal Part Optimization
echo = sort
desobj(max) = 1
analysis = modes
subcase 10
spc = 10
method = 1
$
$ residual structure model
$
BEGIN BULK
eigrl,1,,,9
param,post,0
PARAM,GRDPNT,0
PARAM,WTMASS,.00259
CQUAD4 5 1 13 14 24 23
$
$GRDSET 6
GRID 13 -.4 3.6 0.
GRID 14 .4 3.6 0.
GRID 23 -.4 4.4 0.
GRID 24 .4 4.4 0. Main Bulk
$ Data Section
MAT1,1,30.+6,,.3,.283
PSHELL,1,1,.05,1,,1
$
CQUAD4 14 1 19 20 30 29
CQUAD4 15 1 29 30 36 35
$$
spc1,10,123456,1,2
$
$ design model definition must be in the main BULK Data section
$ 2 3 4 5
desvar 1 X1 0.05 0.01 2.0
dvprel1 1 pshell 1 T
1 1.0
DRESP1 1 F1 FREQ 1
CHAPTER 6 93
Optimization
$
$ Part 1
$
begin super = 1
$
$ define modal coordinates for CMS
$
$ define q-set for component modes and residual vectors
SPOINT 11001 THRU 11020
QSET1 0 11001 THRU 11020
$
$ define which dofs will be retained (i.e. which dofs
$ will form the attachment to the system model when we
$ create SE10 in se10.dat)
$
ASET1 123456 35 36
$
$ part1.dat
$
CQUAD4 18 9 33 34 46 45
CQUAD4 42 9 85 86 98 97
$
param,k6rot,100.
$ Section for
$ boundary grids Part 1 Model
$
GRID 35 -3.6 6. 0.
GRID 36 -2.8 6. 0.
$
GRID 97 -2. 10. 0.
GRID 98 -1.2 10. 0.
$
MAT1,1,30.+6,,.3,.283
PARAM,WTMASS,.00259
PARAM,AUTOSPC,YES
PSHELL,9,1,.05,1,,1
$
$ plotels to outline component in assembly run
$
plotel,101,33,35
plotel,102,33,93
plotel,103,93,98
plotel,104,98,38
plotel,105,38,36
plotel,106,35,36
eigrl,1,,,9
$ Part 2
$
begin super = 2
$
$ define modal coordinates for CMS
$
$ define q-set for component modes and residual vectors
SPOINT 21001 THRU 21020
QSET1 0 21001 THRU 21020
$
$ define which dofs will be retained (i.e. which dofs
$ will form the attachment to the system model when we
$ create SE10 in se10.dat)
$
ASET1 123456 4142
$
$
$ part2.dat
$
CQUAD4 43 2 39 40 52 51
CQUAD4 66 2 90 91 103 102

CQUAD4 67 2 91 92 104 103
$
param,k6rot,100. Section for
$ Part 2 Model
$ boundary grids
$
GRID 41 2.8 6. 0.
GRID 42 3.6 6. 0.
$
GRID 103 4.4 10. 0.

GRID 104 5.2 10. 0.
$
$
$
MAT1,1,30.+6,,.3,.283
PARAM,WTMASS,.00259
PARAM,AUTOSPC,YES
PSHELL,2,1,.05,1,,1
$
$ plotels to outline component in assembly run
$
plotel,201,39,41
plotel,202,39,99
plotel,203,99,104
plotel,204,104,44
plotel,205,44,42
plotel,206,41,42
$
eigrl,1,,,9
enddata
CHAPTER 6 95
Optimization
Randomization of a User’s Input Data File (Pre-

release)
Introduction
The stochastic capability in MSC Nastran is the first step towards a complete and automatic self-
randomization of a Finite Element model. The capability currently offers the user the possibility to
automatically distribute tolerances and uncertainties with minimum effort. This reduces dramatically the
burden on a user wishing to perform large-scale stochastic simulations. In fact, once the stochastic option
is triggered, the entire Bulk Data Deck is randomized automatically and without further user
intervention. The resulting model, which needs to be incorporated in a Monte Carlo Simulation loop -
there are numerous off-the-shelf products which support this capability - possesses unprecedented levels
of realism.
In order to make full use of this new development, it is necessary to resort to a multi-run environment,
which can spawn a certain number of independent MSC Nastran executions, collect the results, allow the
user to perform statistical post-processing. With the self-randomization capability in MSC Nastran, all
the user needs to define are the outputs he wishes to monitor, such as stresses, eigen-frequencies,
temperatures, displacements, etc. There is no need to define inputs, as these are defined automatically by
MSC Nastran. The process is, essentially, error-free.
Benefits
A basic assumption of MSC Nastran is that the inputs to the analysis are known exactly so that the
computed responses are also known exactly. This is, of course, an invalid assumption in that there will
always be some uncertainty in the inputs with a corresponding variations in the outputs. The MSC
Nastran 2007 r1 release of Nastran provides a way of introducing this uncertainty into the analysis
process by automatically randomizing user input real numbers based on the input values and statistical
quantities that characterize the variation.
Input
The randomization capability is driven by a new STOCHASTICS case control command as shown in the
Quick Release Guide. If STOCHASTICS=ALL is used, all real quantities on connectivity (those starting
with C), Material and Property entries as well as any loads and SPCD quantities are modified based on
a covariance factor of 0.05. A Gaussian distribution is used to randomly select the perturbed quantity
with the restriction that the value can be no more that a specified number of standards deviations from
the user input mean value. The default number of maximum standard deviations is 3.
Alternatively, the STOCHASTICS command can point to a STOCHAS bulk data entry that provides the
ability to selectively randomize different types of input quantities using user specified covariance values
and number of allowed standard deviations. In this case, only the types of input specified are randomized
so that, for example, it is possible to randomize the loads input while leaving the property values
unchanged.
Output
There is no new output produced by this capability at present.

The randomization algorithm involves using a random number generator, a Gaussian distribution, the
prescribed covariance and a mean value based on the user input to come up with a randomized value that
is to be used in the analysis. In order to avoid physically meaningless properties, the random value is
prescribed to be within m standard deviations of the input value, where m is a user input value with a
default value of 3.0.
The product of m * COV should not be greater than 1.0 to eliminate the possibility of the property
changing sign.
The full benefit of this beta capability requires submitting multiple runs with the same randomization
parameters. Each would produce a unique randomization and it is possible to collect the results of each
of these analyses and produce statistical information on the variability of the responses. At present, MSC
does not have software that performs these functions and it is unlikely that users would carry out this type
of multiple run analysis on their own.
If the user input property value is 0.0, no randomization occurs. If it is recommended that any property
values that are 0.0 (say orientation angle on a PCOMP entry) be set to some non-zero value that is not
negligible.
CHAPTER 6 97
Optimization
Random Elimination of Element Types (Pre-release)
Introduction
There has been a long time capability in Nastran that allows the user to specify the random elimination
of a specified percentage of the CWELD elements that are contained in a bulk data file. This was done
using Nastran PARAM CWRANDEL, with an additional CWDIAGP PARAM providing the option of
printing the ID’s of the deleted elements. This capability has been extended to the CELASi, CFAST,
CSEAM, and 1-D mass (CMASSi, CONM1 and CONM2) elements. Further, the user interface has been
changed from the NASTRAN PARAM input to the MDLPRM entry.
Benefits
The ability to randomly delete various 1-D elements provides the user with some assessment of the
integrity of the design. For instance, if randomly deleting 20%, say, of the CWELD’s from a model
caused a negligible change in the first ten natural frequencies, this was taken as an indication of the
robustness of the structure. Extending this to other element types provides that many more options in this
type of analysis. Placing the input on the MDLPRM entry consolidates that input so that the user does
not have to deal with the PARAM entry.
Input
The MDLPRM entry has 10 new PARAMi names that support this capability. Five of them (e.g,
DELELAS) select the element type to which this random elimination applies and the ratio to be deleted
while an additional five (e.g., PRTELAS) provide control as to whether the ID’s of the deleted elements
are to be printed. The default is that the ID’s will not be printed.
Output
There is no new output produced by this capability.

The deletion ratio is input as a real number between 0.0 and 1.0 with 0.0 indicating no deletion is to take
place while 1.0 would eliminate all elements of the specified type.
It is possible that the elimination of a series of elements will introduce mechanisms in the structure that
will cause the analysis to fail. It is the user’s responsibility to determine whether this failure has occurred.
A likely scenario for the use of this capability would be to submit the same deck multiple times and
determine in the variation in the results. MSC does not offer an automated way of doing this at this time.
Ch. 7: Rotor Dynamics & Aeroelasticity MSC Nastran 2007 Release Guide
7 Rotor Dynamics & Aeroelasticity

Changes to Rotordynamics for MSC Nastran 2007

Updating/Summing of Monitor Points
Stripwise Aerodynamic Results

Input of an Aerodynamic Mesh

Rigid Body Spline
Wendland Spline Functions for the Spline4/5

Spline Blending

Export of the Spline Matrix
Changes to Rotordynamics for MSC Nastran 2007
Unbalance Entry for Frequency Response

The UNBALNC loading has been available for rotordynamic transient response for several versions.
Unfortunately, unbalance loading for frequency response had to be input using RLOADi entries. This
manual input was uncommonly prone to error. And backward (opposite the rotor spin direction) loading
was easily input. To prevent this problem and provide a more natural input, the UNBALNC Bulk Data
entry can now be used for frequency response with the rotordynamics option. The entry is selected using
the DLOAD Case Control command (note that the transient response solutions must still specify the
transient unbalance using the RGYRO Case Control command).
The format for the UNBALNC in frequency response is the same for transient response, except the
continuations for force output via EPOINTs are ignored. For further information on the UNBALNC Bulk
Data entry see the MSC Nastran Quick Reference Guide.
New Parameter and Hybrid Damping Specifications

The specification of damping using parameter input (PARAM,G for structural damping and PARAMs
ALPHA1 and ALPHA2 for Rayleigh damping) often resulted in users not knowing what damping, if any,
was actually being specified in the model. This is because parameters can be entered anywhere in the
Case Control or Bulk Data Sections. For large models with many ‘include’ files, a rigorous search would
be required. To make the damping specifications easier and also allow new damping formulations, such
as hybrid damping, a new damping entry has been developed that is selected in the case control. The new
entry allows selection of parameter damping, such as structural damping (PARAM, G) and Rayleigh
damping (PARAMs ALPHA1 and ALPHA2), and a new hybrid damping. This entry also allows the
scaling of material damping (GE on material entries, MATi). The new Bulk Data entry is named
DAMPING and is selected by new Case Control commands:
For specifying damping of superelements, use
SEDAMP= n
For specifying damping of the residual structure, use
RSDAMP(STRUCTURE, FLUID, or BOTH)= n
Default= STRUCTURE
Where n specifies the DAMPING Bulk Data entry.
For more information on the DAMPING Bulk Data entry see the MSC Nastran Quick Reference Guide.
CHAPTER 7 101
Updating/Summing of Monitor Points

Introduction
MSC Nastran 2005 r3 greatly expanded the concept of monitor points by modifying the MONPNT1 and
introducing the MONPNT2, MONPNT3 and MONDSP1. The MSC Nastran 2007 r1 release of MSC
Nastran further expands the monitor points by providing two additional capabilities
1. Allowing for the modification of existing component results for MONDSP1, MONPNT1 and
MONPNT3 by a scalar multiple to allow, for example, a change in sign or units.
2. Enabling the weighted summation or two or more MONDSP1, MONPNT1 or MONPNT3’s that
are of the same type.
Benefits
The ability to update an existing monitor point result is done by applying factors to specified
components. An example of the utility of this feature is if the model has been constructed in one set of
units (e.g., centimeters) and it is desired to see the results in a different set (e.g, millimeters).
The ability to sum existing monitor point results could be used; e.g., to better present running results
along a wing or fuselage.
Input
The new MONSUM Bulk Data entry is used to implement both updating and summing of monitor point
results. The description of this entry in the MSC Nastran Quick Reference Guide includes examples and
numerous remarks that help in understanding the capabilities of this entry.
Output
The output of the monitor point results from the MONSUM Bulk Data entry is identical to that of the
underlying monitor point. In the updating scenario, if the NAME field on the entry, is the same as the
NAMEij fields, only the final updated result is given. If a new NAME is used, both results are given.
Examples (monsum.dat and monsum3.dat)

Two small test files demonstrate the use of the MONSUM Bulk Data entry. They are both variations of
the familiar forward swept wing and are:
monusm.dat- contains some MONSUM Bulk Data entries that are simple enough that the results can be
checked by hand
monsum3.dat – sums MONPNT3 results. An interesting feature of this example is that it demonstrates
the DMAP enhancements that were required to sum MONPNT3 results produced by using what is
referred to as the “mini-ema” method with results using the grid point force recovery method.

The remarks for the MONSUM Bulk Data entry in the MSC Nastran Quick Release Guide provide a
number of guidelines that should be reviewed before applying this new entry. Because it performs both
an “update-in-place” and a combining function, it represents a powerful, compact capability.
The two examples shown in the MSC Nastran Quick Reference Guide provide a further explanation of
this entry. In the first, a new monitor point is being synthesized from three existing monitor point results
to provide the user with a blended result that has special meaning to the designer. In this case, results at
three stations along the root chord are being summed with different weights for the desired shear and
bending moment resultants. Since the COEFi fields are left blank for the shear components, the default
of 1.0 is applied. In the second example, the MONSUM is used to convert the units of the monitor point
output from English to metric units. Note that in the first case, monitor point results will be presented for
the summed quantity as well as for each of the original MONPNT1 quantities that contribute to the sum.
In the second case, only the final result is presented since its NAME is shared with the NAMEij attributes.
If output in both units systems is desired, one needs to simply provide a unique NAME in order to get
both results.
CHAPTER 7 103
Stripwise Aerodynamic Results

Introduction
An additional monitor point, MONCNCM, has been provided that streamlines the task of providing
stripwise lift and pitching moment results for a doublet-lattice type of aerodynamic model.
Benefits
Stripwise aerodynamics; i.e., aerodynamic lift and pitching moments, can be quite useful in visualizing
the aerodynamic results on lifting surfaces. In particular, these can be compared with wind tunnel results
with the possibility of weighting the computed aerodynamics to match the test results. It is possible to
generate these results using an aerodynamic MONPNT1, but the preparation of the input data is a
tedious, error-prone process. The new entry automates this process so that it is possible to obtain the
results for each aerodynamic strip in the model with a single MONCNCM entry.
Input
The new MONCNCM Bulk Data entry is used to provide the stripwise aerodynamic results. The remarks
of this entry in the MSC Nastran Quick Release Guide provide guidance in its use. In particular, use of
CAERID1=ALL produces results at each aerodynamic strip. The code internally determines if the flat
plate panels are abutting and, if they are, considers this a single strip from the leading edge to the trailing
edge of the surface made up of the multiple CAERO1 entries.
Output
A new table in the .f06 file that is included as part of the aerodynamic monitor point output. A sample
output from the moncncm test case discussed below is shown here:
AERODYNAMIC MONITOR POINT INTEGRATED LOADS
CONFIGURATION = AEROSG2D XY-SYMMETRY = ASYMMETRIC XZ-SYMMETRY = SYMMETRIC
MACH = 9.000000E-01 Q = 4.000000E+01
MONCNCM NAME = ALL INT GROUP ID = 1 CLASS = STRIP
LABEL = ALL STRIPS FOR THE AIRPLANE
CP = 1 AERODYNAMIC COORDINATE SYSTEM
FORCE MOMENT
STRP YS ZS XREF REFC REFS RIGID ELASTIC RIGID ELASTIC
____ __________ __________ __________ ___________ ___________ ___________ ___________ ___________ ___________
1 1.250E+00 0.000E+00 0.000E+00 1.0000E+01 2.5000E+01 1.5715E+00 1.5735E+00 1.4352E-01 1.4300E-01
2 1.250E+00 0.000E+00 1.428E+01 1.0000E+01 2.5000E+01 -3.1044E-01 -2.9946E-01 -6.2361E-02 -6.1159E-02
3 3.750E+00 0.000E+00 0.000E+00 1.0000E+01 2.5000E+01 1.2319E+00 1.2336E+00 1.1194E-01 1.1147E-01
4 3.750E+00 0.000E+00 1.283E+01 1.0000E+01 2.5000E+01 4.9708E-03 1.6573E-02 -9.2159E-03 -8.0990E-03
5 6.250E+00 0.000E+00 1.139E+01 1.0000E+01 2.5000E+01 1.1896E+00 1.2036E+00 1.7237E-01 1.7369E-01

6 8.750E+00 0.000E+00 9.948E+00 1.0000E+01 2.5000E+01 1.1173E+00 1.1319E+00 9.4866E-02 9.5348E-02
7 1.125E+01 0.000E+00 8.505E+00 1.0000E+01 2.5000E+01 1.0168E+00 1.0308E+00 3.8398E-02 3.8538E-02
8 1.375E+01 0.000E+00 7.061E+00 1.0000E+01 2.5000E+01 8.8463E-01 8.9719E-01 3.4616E-03 3.4251E-03
9 1.625E+01 0.000E+00 5.618E+00 1.0000E+01 2.5000E+01 7.1419E-01 7.2468E-01 -1.3503E-02 -1.3636E-02
10 1.875E+01 0.000E+00 4.175E+00 1.0000E+01 2.5000E+01 4.8032E-01 4.8760E-01 -1.6023E-02 -1.6161E-02
The strips are arranged based on increasing y-coordinate of the strip in the aerodynamic coordinate
system. The z coordinate and the x location of the point at which the pitching moment is computed (based
on the MREF field of the MONCNCM entry) is given as are the local chord length and strip area. This
is followed by the lift and pitching moment at the trim state for the rigid vehicle and for the elastically
deforming vehicle.
Examples (moncncm and monbodi)

Two small test files demonstrate the use of the MONCNCM Bulk Data entry. They are both variations of
the familiar forward swept wing and are:
moncncm provides tests of various ways of merging data across strips that may not be practical from an
engineering standpoint but that exercise different code paths.
monbodi - demonstrates the application of MONCNCM when there are multiple interference groups and
aerodynamic bodies. It is seen that different interference groups are not merged across aerodynamic
panels and that there are no results given for CAERO2 entries.

The MONCNCM entry only supports Doublet Lattice like models and only for lifting surfaces.
Aerodynamic bodies are not supported. The output is in the aerodynamic panel coordinate systems so
that the sign of the forces and moments are a function of the numbering of the corners of the panel (see
the description of the CAERO1 entry in the MSC Nastran Quick Reference Guide). The results are given
at the aeroelastic trim state.
CHAPTER 7 105
Input of an Aerodynamic Mesh

Introduction
Four new bulk data entries have been provided that allow the user to define an aerodynamic mesh that is
distinct from the traditional aerodynamic model that is input using a combination of CAEROi/PAEROi
entries.
Benefits
The ability to input an aerodynamic mesh into MSC Nastran opens up various capabilities that are
considered embryonic at this point. Among these are:
It enables the viewing of an external aero mesh on an MSC Nastran structural model using a tool such
as Patran.
It is possible to spline results from the structural mesh to the new aerodynamic mesh and vice versa. The
spline export feature described in section 1.6 below provides this spline for manipulation outside of MSC
Nastran
The aerodynamic grid points can be loaded using the AEFORCE entry, thereby enabling the import of
rigid aerodynamic loads from an external aerodynamics program.
The AESCALE entry described below opens up some interesting possibility for morphing models that
have yet to be explored.
Input
The aerodynamic mesh can be input using four new bulk data entries
• The AEGRID provides coordinates of the mesh.
• The AEQUAD4 entry connects the mesh using a quad element
• The AETRIA3 connects the mesh using a triangular element
• The AESCALE entry allows for a scaling of the AEGRID data on a grid-by-grid basis
Note that the element data (AEQUAD4 and AETRIA3) are strictly for display and do not participate in
the creation of any aerodynamic matrices.
Output
There is no new output.
Example (aegridf)
This simple example is a variation on the 15 degree swept wing model named ha144c.dat in the
MSC.Nastran User’s Guide for Aeroelasticity. Relative to the example in the guide, the aegridf.dat
example:
• Removes the doublet lattice model

• Adds AEGRID quantities that coincide with all the structural grids in the model plus four more
that represent an aerodynamic stabilizer
• Uses a SPLINE4 entry to spline the wing forces at the AEGRID’s to the structure
• Uses a SPLINRB (see ?.4) to spline the AEGRID’s on the stabilizer to the support point
• Removes the FORCE/GRAV artifice used to simulate constrained wind tunnel condition. Instead
a 1g cruise condition is specified with the aircraft free to pitch and plunge.
• AEPARM entries specify ALPHA and STAB states.
• AEFORCE entries are used to set up load vectors for the intercept, unit angle of attack and unit
stabilizer setting.
The stability derivative print from running the example includes:
NON-DIMENSIONAL STABILITY AND CONTROL DERIVATIVE COEFFICIENTS
MACH = 4.5000E-01 Q = 2.0000E+00
CHORD = 2.0705E+00 SPAN = 1.1050E+01 AREA = 1.1440E+01
CONTROLLER STATE: INTERCEPT ONLY, ALL CONTROLLERS ARE ZERO
TRIM VARIABLE COEFFICIENT RIGID ELASTIC INERTIAL

UNSPLINED SPLINED RESTRAINED UNRESTRAINED RESTRAINED UNRESTRAINED
REF. COEFF. CX 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CY 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CZ 3.496504E-02 3.496504E-02 3.496504E-02 3.496504E-02 0.000000E+00 3.496504E-02
CMX 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CMY -1.249997E-02 -1.249997E-02 -1.249997E-02 -1.249997E-02 0.000000E+00 -1.249997E-02
CMZ 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
ALPHA CX 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CY 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CZ 2.596154E+00 2.596154E+00 2.596154E+00 2.596154E+00 0.000000E+00 2.596154E+00

CMX 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CMY -1.296420E+00 -1.296420E+00 -1.296420E+00 -1.296420E+00 0.000000E+00 -1.296420E+00
CMZ 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
STAB CX 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00

CHAPTER 7 107
CY 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CZ 7.517483E-01 7.517483E-01 7.517483E-01 7.517483E-01 0.000000E+00 7.517483E-01
CMX 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00

CMY -2.760425E+00 -2.760425E+00 -2.760425E+00 -2.760425E+00 0.000000E+00 -2.760425E+00
CMZ 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
URDD3 CX 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CY 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CZ 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 1.793749E-03 0.000000E+00
CMX 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
CMY 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 -6.412643E-04 0.000000E+00
CMZ 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00 0.000000E+00
It is seen that the splined and unsplined rigid aerodynamic terms are identical, which means the splines
are functioning well, and that the rigid and elastic results are the same, which reflects the fact that there
is no aeroelastic feedback in this case.
Trim results from the run include:
AEROSTATIC DATA RECOVERY OUTPUT TABLES

MACH = 4.500000E-01 Q = 2.000000E+00
CHORD = 2.0705E+00 SPAN = 1.1050E+01 AREA = 1.1440E+01
TRIM ALGORITHM USED: LINEAR TRIM SOLUTION WITHOUT REDUNDANT CONTROL SURFACES.
AEROELASTIC TRIM VARIABLES
ID LABEL TYPE TRIM STATUS VALUE OF UX
INTERCEPT RIGID BODY FIXED 1.000000E+00
501 ALPHA GENERAL CONTROL FREE -1.334841E-02 RADIANS
502 STAB GENERAL CONTROL FREE 1.973043E-03 RADIANS
503 URDD3 RIGID BODY FIXED 1.000000E+00 LOAD FACTOR
505 URDD5 RIGID BODY FIXED 0.000000E+00 RAD/S/S PER G
It is seen that the very lightweight structure is in trim based on intercept aerodynamics, a negative angle
of attack and a slightly positive stabilizer.

The aerodynamic mesh cannot be used in conjunction with the existing CAEROi format for creating an
aero mesh.
The CMPID listed on the aerodynamic elements can be reference on an AECOMP that is used in the
aerodynamic splining. However, the splining is only done based on the AEGRID data that are provided
on an AELIST entry identified with the AECOMP.
Unlike the CAEROi format, there is no central grid on an aerodynamic element. The only aerodynamic
grids are at the element vertices.
There is no integration of user input pressures or downwashes, so the AEFORCE entry is the only means
of loading the aerodynamic mesh and, again, this can only be done at the AEGRID locations.
AIC’s are not supported for this type of mesh.
CHAPTER 7 109
Rigid Body Spline

Introduction
A new spline technique has been introduced that enables the splining of aerodynamic data to exactly six
structural degrees of freedom in a rigid body fashion.
Benefits
The primary application of the rigid body spline is to introduce aerodynamic loads into the structural
analysis process when there is no structure underlying the aero mesh. For example, this enables
introducing loads from the tail component in a wing design when only the wing has structural detail.
Input
The new rigid spline method is enabled via the SPLINRB entry.
Output
There is no new output.
Example (splinrb)
A very simple example is provided that is based on the ha144e test case of Section 7.5 of the
MSC.Nastran User’s Guide for Aeroelastic Analysis. The example in the User’s Guide uses separate 1D
splines (the SPLINE2 entry) to distribute the aerodynamic load on the beam structure. This example
replaces these two splines with SPLINRB entries that spline all the loads to the support point. The effect
of this is that these loads no longer create an aeroelastic affect and the rigid and splined stability
derivatives are identical.

In addition to the loads being transferred to the 6 structural degrees of freedom, it should also be noted
that the motion of the aerodynamic surface corresponds to the rigid body motion defined by these 6 dofs.
Wendland Spline Functions for the Spline4/5

Introduction
Options have been added to the SPLINE4 and 5 bulk data entries to support selection of radial
interpolation functions. Reference 1 contains the mathematical development of these functions while
Reference 2 further develops them in the context of an aeroelastic application.
Benefits
The references indicate that these alternative splining techniques should provide smooth results with
improved performance relative to the existing methods. As indicated in the Guidelines and Limitations
portion of this subsection, MSC does not have enough experience with the new methods to quantify these
benefits.
Input
The METH field for the SPLINE4 now supports RIS in addition to the existing IPS, TPS and FPS
methods. If RIS is selected, the order of the interpolation method is either defaulted to WF2 or can be set
to WF0. RCORE is used to define the radius of support. For the SPLINE5, METH is a new field, with
BEAM (infinite beam spline) the default. If METH=RIS is specified, FTYPE (default=WF2) and
RCORE also require input.
Example (ha144c_ris)
The ha144c example has been modified to change the spline from SPLINE1 to SPLINE4 with
METH=RIS, FTYPE=WF2 and RCORE=5.0. Running the example produces answers that are
equivalent but not identical to ha144c. The differences are comparable to those that results from replacing
the infinite plate spline with a finite plate spline so that it is deemed these results are reasonable.

MSC has had limited experience with the radial interpolation functions. They are being provided in a
“Beta” state so that clients can exercise these new options. This is best done in conjunction with the
Spline Verify option of MSC.Flightloads so that a visual check of the quality of the fit can be made.
Because the radial interpolation functions act over a limited area, specified using RCORE, it is believed
that a single spline can be effectively applied to a large area. For example, with a high aspect ratio wing,
a single spline may be able to suffice.
If the linear SPLINE5 includes the rotational flexibility (DTHX/DTHY > -1.0) from the structural
deformations, the FTYPE =WF2 will be used regardless of the user input value of FTYPE.
CHAPTER 7 111
Spline Blending
Introduction
A capability has been added to blend spline effects across multiple splines.
Benefits
Prior to this MSC Nastran 2007 release, it was an error if the same aerodynamic grid was invoked on
multiple splines. This could result in discontinuities at the panel boundaries so this rule has been relaxed
so that an aero grid can be selected on multiple splines and the results blended in some user specified
fashion. This results in smoother displacement patterns which will be of particularly of benefit as the
aeroelastic analysis includes more CFD aerodynamics.
Input
A parameter has been provided to indicate that blending is allowed to occur. The parameter is
MPTSPLIN and is input using the MDLPRM Bulk Data entry. This parameter has the following
meaning:
MPTSPLIN = 0 – Do not allow an aerodynamic grid to be referenced by more than one spline (Default)
= 1 – Allow an aerodynamic grid to be reference by multiple splines.
This parameter does not apply to aerodynamic grids defined with AEGRID discussed above. These grids
can always be referenced on multiple splines.
Two new Bulk Data entries (SPBLND1 and SPBLND2) have been provided to support spline blending.
If neither of these is used, the splines are averaged. The SPBLND1 performs strip based blending that
can be either averaged (with a user defined weighting), linear or cubic in nature. The SPBLND2
performs a curve based blending based on user input curve.
Example (ha145e_blnd)
The existing ha145e test deck has a single SPLINE1 entry. In this example, this has been replaced by two
splines that overlap the two center rows of boxes. MDLPRM,MPTSPLIN,1 is used to enable this overlap
when using the CAERO1 generated aerodynamic meshes. A SPBLND1 entry is then used to blend the
results in a linear way. If the SPBLND1 entry had not been used, the splines would have been averaged.
For this simple example, the blending has little affect on the flutter results. The intent of the example is
to demonstrated the new features.

This is another area where MSC has limited experience so there is limited guidance. It can be said that
the blending developments have been made in concert with the alternative aeromeshing of section 1.3
above and is therefore pointed at applications that use external aerodynamics. It is felt that the blending
will be most useful in areas like wing fuselage junctions were the standard splining techniques are likely
to result in gaps in the aerodynamic deformations.
CHAPTER 7 113
Export of the Spline Matrix

Introduction
A capability has been provided to export the spline matrix to either an .op2 file or to the punch file.
Benefits
Users can now easily obtain the spline matrix in a format that can be readily applied in another procedure
or reintroduced into another Nastran run.
Input
The SPLINOUT case control command is used to produce the exported spline matrix. If the .op2 format
is used, the user should also assign a file that will contain the matrix in OUTPUT2 format.
Output
If the punch output is requested, the GPGK spline matrix is written to the .pch file using the DMIG
format. The DMIG print is preceded by a direct table input (DTI) data that provides a map of the matrix
columns to the aerodynamic degrees of freedom. With the default .op2 method, an output2 file is written.
Examples (splinopch and splinoop2)

These two examples show the use of the .pch and .op2 options for exporting a spline mat
References:
Wendland, H., Piecewise Polynomial, Positive Definite and Compactly Supported Radial Functions of
Minimal Degree, Adv. Computational Mathematics, 4(1995), 389-396.
Beckert, A. and Wendland, H,. Multivariate Interpolation for Fluid-Structure-Interaction Problems using
Radial Basis Functions, Aerospace Sci. Technology, 00 (2001), pp 1-11.

Nastran 2007r1 Release Guide

Uploaded by

Copyright:

Available Formats

Nastran 2007r1 Release Guide

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nastran 2007r1 Release Guide

Uploaded by

Copyright:

Available Formats

MSC Nastran 2007 r1

Internet Resources xiii

1 Overview of MSC Nastran 2007 r1

2 Implicit & Explicit Nonlinear Analysis

SUPER ELEMENTS / DMIG 19

3 NVH & Acoustics

Improved Performance with New Sparse Solvers 37

Improved Memory Usage in Lanczos (Pre-release) 40

New MAXRATIO Information Output (Pre-release) 41

Guidelines and Limitations 42

Performance Improvements and Expanded Capabilities for ACMS 45

Relaxed Restrictions for CASI Solver Usage 52

System Dependent Performance Improvements 53

Improved Selection of Reordering Methods 55

5 Elements & Connectors

Nonhomogeneous Multipoint Constraint 66

Modified Type One Responses - Fractional Mass 76

Automatic External Superelement Optimization (AESO) 80

Randomization of a User’s Input Data File (Pre-release) 95

Random Elimination of Element Types (Pre-release) 97

7 Rotor Dynamics & Aeroelasticity

Updating/Summing of Monitor Points 101

Stripwise Aerodynamic Results 103

Input of an Aerodynamic Mesh 105

Example (aegridf) 105

Rigid Body Spline 109

Wendland Spline Functions for the Spline4/5 110

Spline Blending 111

Export of the Spline Matrix 113

mÜçåÉ= råáíÉÇ=pí~íÉë cêáãäÉóI=`~ãÄÉêäÉó

1 Overview of MSC Nastran 2007 r1

Implicit Nonlinear Analysis (SOL 600)

NVH & Acoustics

Elements & Connectors

Rotor Dynamics & Aeroelasticity

List of MSC Nastran Documents Released with MSC Nastran

2 Implicit & Explicit Nonlinear

MSC Nastran Implicit Nonlinear - SOL 600

Bulk Data Parameters

Heat Transfer Examples

Conduction mhqbd1, mhqbd1a, mhqbd1c, mhqbd1s, mhqbd2, mhqbd2c, mhqbd2s,

Other SOL 600 Items

Virtual Crack Closure Technique (VCCT)

Fracture Mechanics J-Integral (LORENZI)

Bilinear model Exponential Model Linear-Exponential Model

Element Type Number of Nodes Characteristic

Element Type Number of Nodes Characteristic

SOL 600 Failure Description – MATF

Primary Format (SOL 600)

Primary Format Example:

Alternate Format (SOL 600 Only)

Example Alternate Format

Element Birth and Death

Composite Element Numerical Analysis

Maximum Number of Improvement in CPU Memory

Large Rotation RBE

The following CWELD parameters are not supported in SOL 600

The following CWELD parameters are supported in SOL 600

If the new CWELD is used, then MSPEEDCW is ignored

The following CFAST parameters are not supported in SOL 600