nx nastran 5 release guide

166
NX Nastran 5 Release Guide

Upload: msc-nastran-beginner

Post on 03-Jan-2016

287 views

Category:

Documents


5 download

TRANSCRIPT

Page 1: NX Nastran 5 Release Guide

NX Nastran 5 Release Guide

Page 2: NX Nastran 5 Release Guide

Proprietary & Restricted Rights Notice

© 2007 UGS Corp. All Rights Reserved. This software and related documentation are proprietaryto UGS Corp.

NASTRAN is a registered trademark of the National Aeronautics and Space Administration. NXNastran is an enhanced proprietary version developed and maintained by UGS Corp.

MSC is a registered trademark of MSC.Software Corporation. MSC.Nastran and MSC.Patranare trademarks of MSC.Software Corporation.

All other trademarks are the property of their respective owners.

2 NX Nastran 5.0 Release Guide

Page 3: NX Nastran 5 Release Guide

Contents

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1

Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 2Numerical Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 2Multi-body Dynamic Software Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 3New Optimization Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 3Advanced Nonlinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 3Linear Contact Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 4Strength Ratio Output for Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 5Miscellaneous Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1- 5

Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1

SPC/SPCD Enforced Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2- 2Mode Acceleration Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16Random Response Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-19DDAM Process Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-25Elemental Energy Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-33Modal Frequency Response in Rotor Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-44Modal Contribution Computation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-50Direct Structural Damping Matrix Input . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-53

Numerical Capabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1

Sparse Data Recovery for Modal Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3- 2Distributed Memory Parallel Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3- 2Sparse Cholesky Technique for Linear Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3- 5Eigenvalue Analysis Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3- 6

Multi-body Dynamic Software Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1

RecurDyn Flex Input File Creation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4- 2ADAMS Stress Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-10Component Mode Reduction of the Residual Structure Procedure for Flexbody Solutions . . 4-23

New Optimization Option . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1

Input Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5- 2

Advanced Nonlinear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1

Contact Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6- 2Improved Convergence for Contact with Friction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6- 3Surface-to-Surface Glue Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-22Element Birth and Death . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-24Bolt Preload Capability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-26Shell Thickness Result Recovery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-29Iterative Solution Improvement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-30

NX Nastran 5.0 Release Guide 3

Page 4: NX Nastran 5 Release Guide

Contents

Linear Contact Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1

New Options to Improve Contact Solution Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7- 2Support of Contact in SOLs 103, 111, and 112 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7- 6Superelements with Contact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7- 6Shell Element Z-Offset with Contact . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7- 6

Strength Ratio Output for Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-1

Laminate Strength Ratio Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8- 2

Miscellaneous Enhancements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-1

Thermal Expansion of Rigid Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9- 2Bolt Preload Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9- 3Improvements to Surface Glue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9- 6Parameter Specification Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9- 9Punch Output Corrections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-10PARAM K6ROT Update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-12AUTHQUEUE Update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-12Documentation Improvements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9-12

Upward Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-1

Updated and New DMAP Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10- 1Updated and New Datablocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10- 2Updated and New Subdmaps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10- 2

System Description Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-1

4 NX Nastran 5.0 Release Guide

Page 5: NX Nastran 5 Release Guide

Chapter

1 Introduction

This section presents an overview of the features introduced in NX Nastran 5.

NX Nastran 5.0 Release Guide 1-1

Page 6: NX Nastran 5 Release Guide

Chapter 1 Introduction

1.1 Dynamics• The constraint mode method of enforced motion, which was referred to as the relative

method in NX Nastran 4.1, uses a formulation in which the response output is calculatedwith both the normal mode shapes, and the constraint mode shapes. A change in the defaulthas occurred in NX Nastran 5. The absolute displacement formulation is used as the defaultSPC/SPCD method.

• A new mode acceleration formulation with improved performance has been added in thisrelease to supplement the current algorithm. The improved performance makes modeacceleration a more attractive option when increased accuracy is desired.

• Up until the release of NX Nastran version 4.0, only auto spectrum PSD functions ofresponse could be output for random results. It was required to use DMAP alters to obtaincross spectrum results. In NX Nastran 4.1, the option to output cross-power spectraldensity functions became available. Now with NX Nastran 5.0, it is possible to also requestcross-correlation functions.

• The Dynamic Design Analysis Method (DDAM) is a list of procedures to determine modalshock response. To complete DDAM analysis procedures in previous releases of NX Nastran,the use of dmap alter files were required. Now in NX Nastran 5,these DDAM procedureshave been streamlined and automated, thus eliminating the alter requirement.

• The element strain (ESE), kinetic (EKE), and damping (EDE) case control commands havebeen improved to allow the output of element energy information in SORT2 complex format.This format is suitable for plotting complex function data from frequency domain dynamicsolutions. In previous releases, you could only output the real magnitude of element energy,thus no phase information was written. In addition, only SORT1 format could be output,which is not convenient for function plotting.

• Rotor dynamics can now be used in a modal frequency response solution, SOL 111, tocalculate the dynamic response of a rotating system. This new response calculation addsto the existing rotor dynamics capability, direct complex mode solution, SOL 110, whoseoutput is used to create Campbell diagrams.

• When the modal method is used for dynamic response calculations, understanding whichmodes contribute to the response helps understand the dynamic behavior of the simulatedsystem, and can provide insight as to how to improve the dynamic performance. In thisrelease a new capability has been introduced to compute the modal contributions in modalfrequency response.

• A structural damping matrix can now be included with a model in NX Nastran 5 using thenew K42GG case control command. This ability adds to the other direct matrix capabilitiesof stiffness, mass, and damping matrices using DMIG bulk entries.

1.2 Numerical Capabilities• A new sparse data recovery option is available for the modal frequency response analysis

(SOL 111), modal transient response analysis (SOL 112), and optimization (SOL 200) as thenew default. For SOL 111 and 112, this feature reduces the cost of matrix-multiplicationsinside DDRMM modules when large amount of data are requested to recover in the modalanalysis. Similarly, SOL 200 utilize the partitioning of eigenvector matrix in order to reducethe cost of matrix-matrix multiplies.

1-2 NX Nastran 5.0 Release Guide

Page 7: NX Nastran 5 Release Guide

Introduction

• The parallel processing method Distributed Memory Parallel (DMP) has been enhanced inNX Nastran 5 with the following additions:

The normal mode calculation portion of an optimization solution (SOL 200) can now use DMP.

A new linear statics (SOL 101) DMP option, Load Domain Static Analysis (LDSTAT), nowexists to decrease the solution times when large numbers of load cases exist.

• This NX Nastran release includes support for a new sparse Cholesky decomposition.All users of Cholesky decomposition (including the Givens and Householder eigensolvermethods) will utilize sparse decomposition by default, which can be considerably faster thanthe non-sparse Cholesky decomposition method used in earlier NX Nastran versions.

• NX Nastran Version 5.0 introduces the new REDMULT performance option for use whensolving vibration problems with the Lanczos method. This option reduces the cost ofmatrix-vector multiplies inside the READ module when the mass matrix involved isrelatively dense, which can occur when a large number of MFLUID is present in the model.

1.3 Multi-body Dynamic Software Interfaces• This release includes interoperability between NX Nastran and the FunctionBay RFI

(RecurDyn Flex Input) file product. You can now create a RecurDyn Flex Input file (RFI)directly from NX Nastran. The RFI contains the reduced order matrices from the results of aNX Nastran non-restart SOL 103 analysis. The RFI can be imported into RecurDyn andused to represent a flexible component in a multi-body dynamics analysis. This direct RFIexport capability streamlines the process of creating flexible components from FE models,making it possible to obtain more accurate results from multi-body simulations.

• NX Nastran 5 enhances the ADAMS MNF file creation process with a new results recoverycapability. The results from an ADAMS multi-body dynamics analysis, along with anoptional component modal definitions file (OUTPUT2 format), are used in a consecutive NXNastran SOL 103 results recovery solution.

• The set-up of a flex body modal solution in NX Nastran for export to ADAMS MNF orRecurdyn RFI files requires special considerations for the modal solution. This is because flexbodies will be attached to other components in the multi-body dynamic (MBD) simulationand local flexibility effects at the connection locations are thus important. A modalsolution method called Component Mode Reduction of Residual Structure (CMR of RS) isrecommended for flex body solutions because it includes both global and local effects. Thedetails of this method are presented in the multi-body dynamic chapter.

1.4 New Optimization OptionNX Nastran 5 introduces a new optimizer option, UGS-ADS, which is based on public domainADS code.

1.5 Advanced Nonlinear• The CTDISP option has been created on the NXSTRAT bulk entry to prevent contact

conditions from updating.

NX Nastran 5.0 Release Guide 1-3

Page 8: NX Nastran 5 Release Guide

Chapter 1 Introduction

• A new option on the NXSTRAT bulk entry, CTDAMP, has been created to stabilize theportions of the model experiencing rigid body motion, thus helping the solution to continueand converge.

• A new option when INIPENE=3 is available which will include both penetration and nowgaps. This option is particularly beneficial when contact conditions are defined on the facesof concentric cylinders.

• In NX Nastran 4.1, a new contact segment option became available for SOL 601 whichimproved contact results. This improvement matches the order of the contact segmentswith the order of the elements. The restrictions which existed in NX Nastran 4.1 whenCSTYPE=1 have been eliminated in NX Nastran 5.

• The contact algorithm has been improved in NX Nastran 5 to be more robust for contactwith friction.

• The option to “glue” element faces together during a 601 solution is available in NX Nastran5. The glue option connects predefined surfaces together and prevents relative motion inall directions. Predefined regions of element free faces are used to detect where the glueelements are created.

• An option to define element birth and death times for a specific set of elements is available inNX Nastran 5.

• NX Nastran 5 offers an automated method to simplify the multisolution process of boltpreloads with SOL 601.

• A new shell thickness result output option is available when using advanced nonlinear tosolve large strain problems.

• A new 3D-iterative solution option is now available in NX Nastran 5 for SOL 601 toefficiently solve large models containing mainly higher order 3-D solid elements (e.g., 10-nodeCTETRA, 20-node CHEXA, etc.).

1.6 Linear Contact Enhancements• Two new contact solution parameters, INTORD and REFINE, are now available in NX

Nastran 5 on the BCTPARM bulk entry to improve the accuracy of the contact solution.

• Now in NX Nastran 5, a contact definition can be included in a normal mode solution (SOL103), and in an optional modal dynamic response calculation (SOLs 111 and 112).

• In NX Nastran 5, the use of superelements is now permitted in solution sequences whichsupport contact (SOLs 101, 103, 111 and 112). The only requirement is that the contactdefinition must occur in the residual structure.

• Now in NX Nastran 5, the linear contact solution can include the shell element ZOFFS whenevaluating the contact surfaces.

1-4 NX Nastran 5.0 Release Guide

Page 9: NX Nastran 5 Release Guide

Introduction

1.7 Strength Ratio Output for Composites• Strength ratio is now output together with the failure index when using the PCOMP

bulk entry. Strength ratio is a more direct indicator of failure than failure index since itdemonstrates the percentage of applied load to the failure criteria.

1.8 Miscellaneous Enhancements• The option to include rigid elements in thermal expansion calculations is now available.

• When analyzing preloaded bolts, you may be interested in obtaining the stresses due to thepreload condition alone, or due to a combination of the bolt preload and service load. Youcan manually determine the preloaded bolt condition by using equivalent thermal loads,although using this method is approximate and typically requires many solution iterationswhen multiple bolts exist. NX Nastran 5 offers an automated method to simplify thismultisolution process.

• The surface-to-surface glue capability was supported by the global iterative solver inNX Nastran 4.1, but not by the element iterative solver. Now in NX Nastran 5, thesurface-to-surface glue definitions can be included when using the element iterative solver.

• New glue parameters, INTORD and REFINE, are available on the new BGPARM bulk entryto improve the accuracy of the surface-to-surface glue condition.

• Now in NX Nastran 5, the surface-to-surface glue capability can include the shell elementZOFFS when evaluating the glue surfaces.

• NX Nastran 5 now allows the specification of PARAM statements in the nast5rc (UNIX)and nast5.rcf (WINDOWS). Parameters can also be assigned a user defined keyword. Thenew keyword can then be used to specify a value for the parameter on the command lineor in the nastran resource file. The keywords are defined in the “nastran.params” file inthe architecture directory.

• In previous releases of NX Nastran, the algorithm which writes punch files incorrectlyassumed that the first item in an entry was either an integer or real, and that all consecutiveitems were real. These were sometimes invalid assumptions since there are occasions inwhich the first item in a data entry is a character string, and there are also occasions in whichsome consecutive items are integers and not real. This has been corrected in NX Nastran 5.

• The default for parameter K6ROT has been modified from 0 to 100 in NX Nastran 5.Assigning PARAM K6ROT to 100 has shown to be a good, general penalty stiffnessvalue which will not adversely effect results, yet will properly constrain the rotationdegrees-of-freedom.

• When an NX Nastran job fails because of a failed license request, an option to have thejob retried automatically became available in NX Nastran 4.1. NX Nastran will retry afailed license request job every minute up to the value of the AUTHQUEUE keyword. TheAUTHQUEUE keyword default was 20 minutes in NX Nastran 4.1. Based on customerfeedback, the default has been modified to 0 in NX Nastran 5. You must now explicitly definethe AUTHQUEUE keyword in the nast5rc/nast5.rcf files to take advantage of this capability.

• The following documentation improvements have occurred in NX Nastran 5:

NX Nastran 5.0 Release Guide 1-5

Page 10: NX Nastran 5 Release Guide

Chapter 1 Introduction

The NX Nastran documentation is now available in both PDF and HTML format. A singlebookshelf.html provides the links to the PDF and the HTML documentation. HTML searchis available. You can search a single, or all HTML documents by selecting the magnifyingglass icon located at the top left corner of the bookshelf page.

The new NX Nastran Parallel Processing User’s Guide is included with the NX Nastrandocumentation. This guide provides information on NX Nastran parallel methods for Linuxand UNIX systems including parallel processing basics, detailed computational methods, anoverview of execution methods, and examples using various solution sequences.

The NX Nastran Numerical Methods User’s Guide has been rewritten in this release.

1-6 NX Nastran 5.0 Release Guide

Page 11: NX Nastran 5 Release Guide

Chapter

2 Dynamics

NX Nastran 5.0 Release Guide 2-1

Page 12: NX Nastran 5 Release Guide

Chapter 2 Dynamics

2.1 SPC/SPCD Enforced MotionThe SPC/SPCD method of enforced motion allows you to directly specify displacements, velocitiesor accelerations via SPC/SPC1/SPCD data, thus eliminating the need to employ the largemass method or Lagrange Multipliers. The SPC/SPCD method is preferred because it is morestraightforward to use than other methods. Prior to NX Nastran 4.1, the SPC/SPCD method wasbased on the absolute displacement formulation which gives unintended consequences whenused with modal damping. Specifically, external and internal damping forces are created, whengenerally modal damping is intended to simulate only internal damping. The external dampingforces can result in noticeable differences in low frequency response compared to systems withonly internal damping. Another concern with the absolute displacement formulation is thatresidual vectors are generally needed for accurate stress, strain, and force response.

A new constraint mode method was introduced in NX Nastran 4.1 that did not suffer from theseeffects. The new method, referred to as the "relative displacement method" in NX Nastran 4.1,uses a formulation in which the response output is calculated from both normal mode shapes andconstraint mode shapes. In contrast, the original absolute method uses only normal mode shapes.Residual vectors can be used with the constraint mode method but are not critical for accuracy.

The selection of the enforced motion method is enabled by a new system cell 422 (ENFMOTN).The original absolute method is invoked with ENFMOTN=1 and the new constraint modemethod is invoked with ENFMOTN=0. Since many users are dependent on the original absolutemethod as part of their workflow, it is the default for NX Nastran 5.

You may prefer the constraint mode method when you use modal damping. The two methods givealmost identical results when no modal damping is present and residual vectors are used withthe absolute displacement method. However, if modal energy (via the MODALE case controlcommand) or modal contributions (via the MODCON case control command) are requested, youmust use the absolute method. MODALE and MODCON are not supported for the constraintmode method.

Another issue with the constraint mode method is its sensitivity to very high acceleration inputs.For example, an enforced displacement that has a large change in value over a small time step(almost a “discontinuity”) will produce large acceleration inputs. This can produce a “bump” inthe response calculations around the time that this “discontinuity” occurs. It is recommendedthat such large acceleration inputs not be used with the constraint mode method, especiallyoccurring over a single time step. If such a step change is required, it is recommended that a finetime step be used to reduce the high acceleration affect and, thus lesson the “bump” that occursin the area of the “discontinuity”.

The describers ABS/REL were added in version 4.1 to the DISPLACEMENT, VELOCITY, andACCELERATION case control entries. These describers allow you to request the total responseoutput (ABS), or the response output relative to the enforced motion (REL). Having these optionson each output entry allows you the flexibility to request, for example, relative displacementoutput with absolute acceleration output in the same run. The new describers are also supportedon a subcase level to allow for different output requests in different subcases. The new systemcell 422, along with the updated DISPLACEMENT, VELOCITY, and ACCELERATION casecontrol commands are included in the Input Updates section below.

Sparse data recovery has also been introduced for the constraint mode method. This technique isanalogous to the sparse data recovery for modal solutions discussed in section 3.1, and is activatedby the same system cell. The difference is that this particular sparse recovery is applied muchearlier in the analysis procedure to reduce the size of matrices and, thus, computation time. Thisrecovery process keys off the output requests and currently supports the following case controlentries: DISPLACEMENT, VELOCITY, ACCELERATION, OLOAD, STRESS/ELSTRESS,STRAIN/ELSTRAIN, FORCE/ELFORCE, EDE, EKE, ESE, XYPEAK, XYPRINT, XYPLOT,and XYPUNCH. The sparse data recovery capability is currently automatically deactivated if

2-2 NX Nastran 5.0 Release Guide

Page 13: NX Nastran 5 Release Guide

Dynamics

the following output requests are made: SPCFORCE, MPCFORCE, GPSTRESS, GPSTRAIN,GPFORCE, OUTPUT(PLOT), and OUTPUT(POST). Also, if a PARAM,POST request is made,the sparse data recovery is deactivated since most post processors require the full set of results.

If you plan on running a mini restart (scr=mini) in order to recover additional results later froma constraint mode enforced motion analysis, you must deactivate the sparse data recovery (i.e.system cell 421/SPARSEDR=0; see section 3.1 for more information). Otherwise, only resultswhich were requested in the original run will be output; any additional result requests in themini restart run will be zero. Since the sparse data recovery method is so much more efficient, itis advised to not use the mini restart method when sparse data recovery is active. Either requestall necessary outputs in the initial run or use a regular restart.

Enforced Motion Example

The following example demonstrates the improvements of the new constraint mode methodover the absolute method.

Figure 2-1. Input is enforced acceleration:

The geometry is a cantilever which is modeled with shell elements. One end of the cantilever ismade rigid with an RBE2 element. The independent grid (Grid 100) of the RBE2 is restrained byan SPC in all six DOF. In addition, the z DOF (direction normal to the plane of the shells) is anSPCD DOF which identifies it as the location of input motion.

A random PSD analysis was performed with a constant base acceleration input spectrum of100.0 across the frequency range from 1 to 1000 Hz. Results have been computed three differentways: using the old absolute displacement formulation, the new constraint mode formulation,and the large mass method.

Figure 2-2 shows the resulting PSD acceleration at the mid point, grid 32. The accelerationresponse is very similar for all three methods.

Figure 2-3 shows the stress response at a location near the base motion input. The results forthe new relative displacement formulation and the large mass method are very similar andaccurate. However, the absolute formulation result is only accurate above 10 Hz. Below 10 Hzthe predicted response deviates noticeably from the correct response.

The new constraint mode formulation is always more accurate than the absolute formulation,and it will match the response of the large mass method. The absolute formulation method givesgood accuracy for most responses, but it is not as accurate for responses involving relativedisplacement quantities (such as stress, strain, or force), particularly when significant dampingexists in the model.

NX Nastran 5.0 Release Guide 2-3

Page 14: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Vertical Axis = AccelerationHorizontal Axis = Frequency

Absolute Displacement Method

Constraint Mode Method

Large Mass Method

Figure 2-2. Acceleration Response

2-4 NX Nastran 5.0 Release Guide

Page 15: NX Nastran 5 Release Guide

Dynamics

Demonstrates the effect of external damping

Vertical Axis = StressHorizontal Axis = Frequency

Absolute Displacement Method

Constraint Mode Method

Large Mass Method

Figure 2-3. Stress Response

The file used in this example plate_111.dat can be found in the install locationNX_Nastran_Install_Path/nxn5/nast/tpl.

NX Nastran 5.0 Release Guide 2-5

Page 16: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Mathematical Details for Constraint Mode Method

The original SPC/SPCD enforced motion formulation uses an absolute displacement formulationper the equation:

Equation 2-1.

Here the set nomenclature is simplified with f representing the free physical degrees of freedom(DOF), and s representing the SPCD enforced motion DOF, thus

uf = absolute displacements of the f-set DOF

us = applied enforced motion of the s-set DOF

M, B, K, and are the mass, viscous damping, stiffness, and structural damping matrices. Thevector P is the applied force load. The structural damping matrix, , consists of the uniform andelemental structural damping components so that

Equation 2-2.

where G is the uniform structural damping coefficient, GE are elemental structural coefficients,and KE are elemental stiffness matrices.

The formulation for the constraint mode method is obtained by separating the absolute

displacement into a static enforced motion component, , and a dynamic relativedisplacement component, vf, as

Equation 2-3.

Substituting into the equation of motion, a new equation of motion in terms of the dynamicdisplacement component is obtained

Equation 2-4.

The above equation is solved for vf and then the absolute displacement is obtained by adding thestatic enforced motion displacement.

Since the constraint mode method uses static mode shapes, it is not as critical that residualvectors be computed as for the absolute displacement method. But it is still possible to useresidual vectors with the constraint mode method. In that case, residual vectors are computedfor inertial, damping, and structural damping loads. Thus the residual vectors are computedfrom the following terms

2-6 NX Nastran 5.0 Release Guide

Page 17: NX Nastran 5 Release Guide

Dynamics

Equation 2-5.

Input Updates for New Constraint Mode Method

New system cell 422–ENFMOTN:

422 ENFMOTN Controls which formulation is used for enforced motionresponse analysis (and mode acceleration, if requested).

0: Constraint mode method of enforced motion formulation(and new mode acceleration method)

1: Absolute displacement enforced motion formulation (andold mode acceleration method) (Default). (ABS/REL optionon DISPLACEMENT, VELOCITY, and ACCELERATIONcase control entries is disabled).

Updated DISPLACEMENT, VELOCITY, and ACCELERATION case control commands:

DISPLACEMENT

Displacement Output Request

Requests the form and type of displacement or pressure vector output. Note:PRESSURE and VECTOR are equivalent commands.

FORMAT:

EXAMPLES:DISPLACEMENT=5DISPLACEMENT(REAL)=ALLDISPLACEMENT(SORT2, PUNCH, REAL)=ALL

DESCRIBERS:

Describer Meaning

SORT1 Output will be presented as a tabular listing of grid pointsfor each load, frequency, eigenvalue, or time, depending onthe solution sequence.

NX Nastran 5.0 Release Guide 2-7

Page 18: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Describer Meaning

SORT2 Output will be presented as a tabular listing of load,frequency or time for each grid point.

PRINT The printer will be the output medium.

PUNCH The punch file will be the output medium.

PLOT Generates, but does not print, displacement data.

REAL or IMAG Requests rectangular format (real and imaginary) ofcomplex output. Use of either REAL or IMAG yields thesame output.

PHASE Requests polar format (magnitude and phase) of complexoutput. Phase output is in degrees.

ABS For enforced motion dynamic analysis, displacement resultswill be output as absolute displacement.

REL For enforced motion dynamic analysis, displacement resultswill be output relative to the enforced motion input.

PSDF Requests the power spectral density function be calculatedfor random analysis post-processing. The request mustbe made above the subcase level and RANDOM must beselected in the Case Control. See remark 8.

ATOC Requests the autocorrelation function be calculated forrandom analysis post-processing. The request must bemade above the subcase level and RANDOM must beselected in the Case Control. See remark 8.

CRMS Requests the cumulative root mean square function becalculated for random analysis post-processing. Requestmust be made above the subcase level and RANDOM mustbe made in the Case Control. See remark 8.

RALL Requests all of PSDF, ATOC, and CRMS be calculatedfor random analysis post-processing. The request mustbe made above the subcase level and RANDOM must beselected in the Case Control. See remark 8.

RPRINT Writes random analysis results to the print file. (Default)See remark 8.

NORPRINT Disables the writing of random analysis results to the printfile. See remark 8.

RPUNCH Writes random analysis results to the punch file. Seeremark 8.

2-8 NX Nastran 5.0 Release Guide

Page 19: NX Nastran 5 Release Guide

Dynamics

Describer Meaning

ALL Displacements for all points will be output.

NONE Displacement for no points will be output.

n Set identification of a previously appearing SET command.Only displacements of points with identification numbersthat appear on this SET command will be output.(Integer>0)

REMARKS:1. Both PRINT and PUNCH may be requested.

2. The defaults for SORT1 and SORT2 depend on the type of analysis:

• SORT1 is the default in static analysis, frequency response, steady stateheat transfer analysis, real and complex eigenvalue analysis, flutteranalysis, and buckling analysis.

• SORT2 is the default in transient response analysis (structural and heattransfer). SORT2 is not available for real eigenvalue (including buckling),complex eigenvalue, or flutter analysis. If SORT1 is selected in a transientsolution for one or more of the commands ACCE, DISP, ENTH, FORC,HDOT, MPCF, OLOA, SPCF, STRA, STRE, and VELO then the remainingcommands will also be output in SORT1 format. If SORT2 is selected in afrequency response solution for one or more of the commands ACCE, DISP,FORC, MPCF, OLOA, SPCF, STRA, STRE, and VELO then the remainingcommands will also be output in SORT2 format.

• XY plot requests will force SORT2 format thus overriding SORT1 formatrequests.

3. VECTOR and PRESSURE are alternate forms and are entirely equivalentto DISPLACEMENT.

4. DISPLACEMENT = NONE overrides an overall output request.

5. The PLOT option is used when curve plots are desired in the magnitude/phaserepresentation and no printer output request is present for magnitude/phaserepresentation.

6. The units of translation are the same as the units of length of the model.Rotations are in units of radians.

7. Displacement results are output in the global coordinate system (see fieldCD on the GRID Bulk Data entry).

8. The option of PSDF, ATOC, CRMS, and RALL, or any combination of themcan be selected for random analysis. The results can be either printed to the.f06 file, punched to the punch file, or output to both files using RPRINT andRPUNCH.

NX Nastran 5.0 Release Guide 2-9

Page 20: NX Nastran 5 Release Guide

Chapter 2 Dynamics

9. When doing enforced motion dynamic analysis and relative output is requested(using the REL describer), the output will be relative to the input as describedby the equation:

where uf = absolute displacement

yf = relative displacement

us = enforced motion.

REMARKSRELATED TO

SOLS 601 AND701:

1. Output is restricted to REAL format. IMAG, PHASE, PSDF, ATOC and RALLare ignored.

2. Displacements, velocities and accelerations must be output for the same set ofgrid points if requested. Output requested for set n in this command will becombined with the sets requested in the VELOCITY and ACCELERATIONcommands, and displacements will be output at the grid points of the combinedset.

VELOCITY

Velocity Output Request

Requests the form and type of velocity vector output.

FORMAT:

EXAMPLES:VELOCITY=5VELOCITY(SORT2,PHASE,PUNCH)=ALL

DESCRIBERS:

Describer Meaning

SORT1 Output will be presented as a tabular listing of grid pointsfor each load, frequency, eigenvalue, or time, depending onthe solution sequence.

SORT2 Output will be presented as a tabular listing of frequencyor time for each grid point.

2-10 NX Nastran 5.0 Release Guide

Page 21: NX Nastran 5 Release Guide

Dynamics

Describer Meaning

PRINT The printer will be the output medium.

PUNCH The punch file will be the output medium.

PLOT Generates, but does not print, velocities.

REAL or IMAG Requests rectangular format (real and imaginary) ofcomplex output. Use of either REAL or IMAG yields thesame output.

PHASE Requests polar format (magnitude and phase) of complexoutput. Phase output is in degrees.

ABS For enforced motion dynamic analysis, velocity results willbe output as absolute velocity.

REL For enforced motion dynamic analysis, velocity results willbe output relative to the enforced motion input.

PSDF Requests the power spectral density function be calculatedfor random analysis post-processing. The request mustbe made above the subcase level and RANDOM must beselected in the Case Control. See remark 7.

ATOC Requests the autocorrelation function be calculated forrandom analysis post-processing. The request must bemade above the subcase level and RANDOM must beselected in the Case Control. See remark 7.

CRMS Requests the cumulative root mean square function becalculated for random analysis post-processing. Requestmust be made above the subcase level and RANDOM mustbe made in the Case Control. See remark 7.

RALL Requests all of PSDF, ATOC, and CRMS be calculatedfor random analysis post-processing. The request mustbe made above the subcase level and RANDOM must beselected in the Case Control. See remark 7.

RPRINT Writes random analysis results to the print file. (Default)See remark 7.

NORPRINT Disables the writing of random analysis results to the printfile. See remark 7.

RPUNCH Writes random analysis results to the punch file. Seeremark 7.

ALL Velocity for all solution points will be output.

NONE Velocity for no solution points will be output.

NX Nastran 5.0 Release Guide 2-11

Page 22: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Describer Meaning

n Set identification of a previously appearing SET command.Only velocities of points with identification numbers thatappear on this SET command will be output. (Integer>0)

REMARKS:1. Both PRINT and PUNCH may be requested.

2. Velocity output is only available for transient and frequency response problems.

3. See DISPLACEMENT case control card for a discussion of SORT1 and SORT2.

4. VELOCITY=NONE overrides an overall output request.

5. The PLOT option is used when curve plots are desired in the magnitude/phaserepresentation and no printer request is present for magnitude/phaserepresentation.

6. Velocity results are output in the global coordinate system (see field CD onthe GRID Bulk Data entry).

7. The option of PSDF, ATOC, CRMS, and RALL, or any combination of themcan be selected for random analysis. The results can be either printed to the.f06 file, punched to the punch file, or output to both files using RPRINT andRPUNCH.

8. When doing enforced motion dynamic analysis and relative output is requested(using the REL describer), the output will be relative to the input as describedby the equation:

where uf = absolute displacement

yf = relative displacement

us = enforced motion.

REMARKSRELATED TO

SOLS 601 AND701:

1. Output is restricted to REAL format. IMAG, PHASE, PSDF, ATOC and RALLare ignored.

2. Displacements, velocities and accelerations must be output for the same setof grid points if requested. Output requested for set n in this commandwill be combined with the sets requested in the DISPLACEMENT andACCELERATION commands, and velocities will be output at the grid points ofthe combined set.

2-12 NX Nastran 5.0 Release Guide

Page 23: NX Nastran 5 Release Guide

Dynamics

ACCELERATION

Acceleration Output Request

Requests form and type of acceleration vector output.

FORMAT:

EXAMPLES:ACCELERATION=5ACCELERATION(SORT2, PHASE)=ALLACCELERATION(SORT1, PRINT, PUNCH, PHASE)=17

DESCRIBERS:

Describer Meaning

SORT1 Output will be presented as a tabular listing of grid pointsfor each load, frequency, eigenvalue, or time, depending onthe solution sequence.

SORT2 Output will be presented as a tabular listing of frequencyor time for each grid point.

PRINT The printer will be the output medium.

PUNCH The punch file will be the output medium.

PLOT Computes, but does not print or punch, acceleration output.

REAL or IMAG Requests rectangular format (real and imaginary) of complexoutput. Use of either REAL or IMAG yields the same output.

PHASE Requests polar format (magnitude and phase) of complexoutput. Phase output is in degrees.

ABS For enforced motion dynamic analysis, acceleration resultswill be output as absolute acceleration.

REL For enforced motion dynamic analysis, acceleration resultswill be output relative to the enforced motion input.

PSDF Requests the power spectral density function be calculatedfor random analysis post-processing. The request mustbe made above the subcase level and RANDOM must beselected in the Case Control. See remark 7.

NX Nastran 5.0 Release Guide 2-13

Page 24: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Describer Meaning

ATOC Requests the autocorrelation function be calculated forrandom analysis post-processing. The request must be madeabove the subcase level and RANDOM must be selected inthe Case Control. See remark 7.

CRMS Requests the cumulative root mean square function becalculated for random analysis post-processing. Requestmust be made above the subcase level and RANDOM mustbe made in the Case Control. See remark 7.

RALL Requests all of PSDF, ATOC, and CRMS be calculated forrandom analysis post-processing. The request must be madeabove the subcase level and RANDOM must be selected inthe Case Control. See remark 7.

RPRINT Writes random analysis results to the print file. (Default)See remark 7.

NORPRINT Disables the writing of random analysis results to the printfile. See remark 7.

RPUNCH Writes random analysis results to the punch file. See remark7.

ALL Accelerations at all points will be output.

NONE Accelerations at no points will be output.

n Set identification of a previously appearing SET command.Only accelerations of points with identification numbers thatappear on this SET command will be output. (Integer>0)

REMARKS:1. Both PRINT and PUNCH may be requested.

2. Acceleration output is only available for transient and frequency responseproblems.

3. See Remark 2 under “DISPLACEMENT” case control command for adiscussion of SORT1 and SORT2.

4. ACCELERATION = NONE allows overriding an overall output request.

5. The PLOT option is used when curve plots are desired in the magnitude/phaserepresentation and no printer request is present for magnitude/phaserepresentation.

6. Acceleration results are output in the global coordinate system (see field CD onthe GRID Bulk Data entry).

7. The option of PSDF, ATOC, CRMS, and RALL, or any combination of themcan be selected for random analysis. The results can be either printed to the

2-14 NX Nastran 5.0 Release Guide

Page 25: NX Nastran 5 Release Guide

Dynamics

.f06 file, punched to the punch file, or output to both files using RPRINT andRPUNCH.

8. When doing enforced motion dynamic analysis and relative output is requested(using the REL describer), the output will be relative to the input as describedby the equation:

where uf = absolute displacement

yf = relative displacement

us = enforced motion.

REMARKSRELATED TO

SOLS 601 AND701:

1. Output is restricted to REAL format. IMAG, PHASE, PSDF, ATOC and RALLare ignored.

2. Displacements, velocities and accelerations must be output for the same set ofgrid points if requested. Output requested for set n in this command will becombined with the sets requested in the VELOCITY and DISPLACEMENTcommands, and accelerations will be output at the grid points of the combinedset.

NX Nastran 5.0 Release Guide 2-15

Page 26: NX Nastran 5 Release Guide

Chapter 2 Dynamics

2.2 Mode Acceleration MethodDynamic analysis using modal methods is typically performed in the following sequence:

• A set of normal modes are calculated and used to define the dynamic model in modalcoordinates.

• The dynamic response is calculated for each normal mode in modal coordinates.

• The physical response is recovered from the dynamic response.

The accuracy of the recovered response depends on the number of normal modes computed. Ageneral rule is to compute modes to frequencies twice the range of interest. Still in some cases,the effect of the modes not computed (the truncated modes) can be significant. This is especiallytrue for element forces, stresses and strains. It may not be practical to solve for more modes andthe mode acceleration offers an alternative approach for improving accuracy.

Mode acceleration accounts for both the dynamic effects of the computed normal modes plusthe static response of the truncated modes. There is an added computational expense forperforming the mode acceleration computations but a new mode acceleration formulation withimproved performance has been added in this release as an alternative to the older algorithm.The improved performance makes mode acceleration a more attractive option when increasedaccuracy is desired.

The MODACC parameter will continue to be used to turn on this capability, and the defaultremains as -1 for off. The current requirement that mode acceleration must always usePARAM,DDRMM,-1 still exists, although a change in the software has occurred where theDDRMM will always be assigned to -1 when either the old or new mode acceleration method isused, regardless of the value assigned.

The new mode acceleration will be used for both applied loads and enforced motion terms and issupported in SOLs 111, 112, 146, and 200 (modal transient and modal frequency response only).

Also, the ability to restart from the same solution sequence as well as SOL 103 is supported.You will be able to choose between the old and new mode acceleration methods with the systemcell 422 (ENFMOTN), which is the same system cell used to select the enforced motion method.This is because the old mode acceleration method is based on the absolute displacement methodfor the enforced motion component, whereas the new mode acceleration method is based on thenew constraint mode method for the enforced motion component, which was introduced in NXNastran 4.1. Setting system cell 422 to 0 (with PARAM, MODACC, 1) will choose the new modeacceleration method; setting it to 1 (default) will choose the old mode acceleration method.However, if modal energy (via the MODALE case control command) or modal contributions(via the MODCON case control command) are requested, the absolute method must be used.MODALE and MODCON are not supported for the constraint mode method.

The new MODACC parameter, and the system cell 422 (ENFMOTN) are included below underInput Updates for Mode Acceleration.

Mathematical Details for New Mode Acceleration Method:

The dynamic equation of motion with the constraint mode method is

Equation 2-6.

2-16 NX Nastran 5.0 Release Guide

Page 27: NX Nastran 5 Release Guide

Dynamics

Where the total response, uf, is computed as the sum of a static enforced motion component,

, and a dynamic relative displacement component, vf , as

Equation 2-7.

The other terms in the above equations have been previously defined in section 2–1.

The solution to equation 2–6 can be obtained efficiently in modal coordinates, , by using themodal transformation

Equation 2-8.

Here is the mode shape matrix of normal modes.

Substituting into equation 2–6 gives

Equation 2-9.

where

Equation 2-10.

The matrices m and k are modal diagonal matrices. The modal viscous damping matrix, b, and

modal structural damping matrix, , are generally not diagonal.

Standard modal analysis involves solving equation 2–9 for , and substituting back intoequations 2–8 and 2–9 to get the physical displacement uf.

In the mode acceleration method, the same approach is used again to compute the modaldisplacement. But a more accurate physical displacement response is computed by solving forthe displacement vf from a direct solution of equation 2–6 with the inertia and damping termstaken to the right hand side and computed in modal terms. This gives the response computed as

Equation 2-11.

Transforming back into total displacement coordinates, uf , and using properties of the modalmatrices, the absolute displacements can be computed as

NX Nastran 5.0 Release Guide 2-17

Page 28: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Equation 2-12.

where Ω is a square modal size matrix with the modal frequencies on the diagonal. The termson the right hand side of equation 2–12 are essentially different types of mode shapes. Thenames for shapes are

= real normal modes

= static constraint modes

= applied force attachment modes

= inertia attachment modes

= viscous damping attachment modes

= structural damping attachment modes

Input Updates for Mode Acceleration

New parameter MODACC:

MODACC Default = -1MODACC = 0 selects the mode acceleration method for data recovery indynamic analysis. See “Formulation of Dynamic Equations in SubDMAP GMA”in the NX Nastran User’s Guide for further discussion.

MODACC = 1 is the same as MODACC = 0 except if SUPORTi entries arepresent then the displacements are corrected for rigid body motion and thelocal flexibility at the SUPORTi degrees-of-freedom.

MODACC ≥ 0 is not recommended for use in hydroelastic problems.

New system cell 422–ENFMOTN (repeated here for convenience):

422 ENFMOTN Controls which formulation is used for enforced motionresponse analysis (and mode acceleration, if requested).

0: Constraint mode method of enforced motion formulation(and new mode acceleration method)

1: Absolute displacement enforced motion formulation (andold mode acceleration method) (Default). (ABS/REL optionon DISPLACEMENT, VELOCITY, and ACCELERATIONcase control entries is disabled).

2-18 NX Nastran 5.0 Release Guide

Page 29: NX Nastran 5 Release Guide

Dynamics

2.3 Random Response OutputIn many applications, the loading on a structure is not known in a definite (or deterministic)sense, and is instead quantified in a statistical sense with properties such as the mean load andstandard deviation of load. Such loadings are known as random or stochastic loads. Examples ofrandom loads include: rough road surface loads on automobiles, ocean wave loads on offshoreplatforms, wind loads on buildings, vibration loads from rocket engines, and earthquake groundmotions.

You should expect the response of the structure due to random loads to also be random in nature,and analysis methods that determine the statistical properties of response are appropriate. Themost commonly used approach for random analysis uses Power Spectral Density (PSD) functionsto define loading and the calculated response. A PSD function describes how the variance of atime series of the product of two variables is distributed with frequency. If the two variablesare the same, the PSD function is known as an auto spectrum; if the variables are different,the PSD function is called a cross spectrum.

NX Nastran performs random analysis in a two-step procedure. In the first step, a frequencyresponse analysis is performed to calculate transfer functions which are the ratio of output to aunit input. In the second step, the PSD loading function is used to scale the transfer functionsto compute the random PSD response. The input PSD functions represent the statistics of theapplied loads. If you have several applied loads that are independent of each other, you onlyneed to apply the auto spectra of the loads. If, on the other hand, the loads are not independentof each other, you must also input the cross spectrum of the loads to define the correlation ofthe loadings. Note: Although the process is described as two steps, it can be performed withNX Nastran as only one solution submittal.

Random response output consists of the response PSD, autocorrelation functions, number of zerocrossings with positive slope per unit time, and RMS (root-meansquare) values of response. Priorto NX Nastran version 4.0, only auto spectrum PSD functions of response could be output forrandom results. You were required to use a DMAP alter if you wanted cross spectrum results.

In NX Nastran 4.1, the option to output cross-power spectral density functions became available.Now with NX Nastran 5.0, you can also request cross-correlation functions. The Case Controlcommand, RCROSS and Bulk Data entry, RCROSS were created in version 4.1 for this purpose.The new formats of these entries are included at the end of this section. The followingexamples have been provided in the install location NX_Nastran_Install_Path/nxn5/nast/tpl todemonstrate capabilities and to show the correct syntax for using the new RCROSS commands:rcross01.dat, rcross02.dat, rcross03.dat, rcross04.dat.

Theory for Cross-power Spectral Density and Cross-correlation FunctionsAssuming that a random process is ergodic, the cross correlation function between a pair of timerecords ua(t) and ub(t) is defined as:

Equation 2-13.

where ua(t) and ub(t) can be displacement, velocity, acceleration or single-point constraintforce responses at the same or different grid points; or stress, strain, and force components inthe same or different elements. The assumption of ergodicty means that as the length of therecord is increased the average approaches the corresponding ensemble averages. The operationof integration over a time record of length T and dividing that quantity by T represents anaveraging process over time.

NX Nastran 5.0 Release Guide 2-19

Page 30: NX Nastran 5 Release Guide

Chapter 2 Dynamics

If we denote the Fourier transform of a time history as:

Equation 2-14.

we can define the cross-power spectral density (cross-PSD) as:

Equation 2-15.

where ω is angular frequency in rad/sec and the symbol (*) denotes complex conjugate. Simplystated, the cross-PSD is the product of the Fourier transform of two records.

It can be shown using equations 2–6 and 2–7 that the cross correlation function and crossspectrum are related as Fourier transform pairs given as:

Equation 2-16.

Equation 2-17.

It is noted that the cross correlation function is always real valued and the cross spectrumfunction is in general complex valued. In addition, it can be shown to possess the followingproperties:

Equation 2-18.

Equation 2-19.

It should also be noted that we have been referring to the general case where ua(t) and ub(t) aredifferent responses. In many situations, only a single response, rather than the joint response isof interest. In that case, auto-correlation functions, Raa(τ),and auto-spectrum functions, Saa(ω),are computed. The auto-spectrum by definition is completely real.

Returning to the more general case, we can rewrite in terms of its real and imaginary parts as:

Equation 2-20.

From the properties in equation 2–10 it can be determined that Srab(ω) is an even function of ωand SIab(ω) is an odd function of ω.

2-20 NX Nastran 5.0 Release Guide

Page 31: NX Nastran 5 Release Guide

Dynamics

Substituting equation 2–13 into the equation 2–12 and recalling that eiωt = cos(ωt)+isin(wt), itcan be shown that:

Equation 2-21.

A particular case of interest is the value of the cross-correlation function for τ=0. As evidenced byequation 2–6 it represents the mean cross product. Using equation 2–14 it can be computed as:

Equation 2-22.

where (—) denotes mean average.

The mean cross product is an important statistical quantity. For cross-correlation functions, is itknown as the covariance. For auto-correlation functions the mean cross-product is known as thevariance and its square root is the standard deviation. By using the standard assumption thatloadings and response are Gaussian (normally distributed), knowing the covariance allows you todefine the probability distribution and thus determine likelihood of occurrence of response levels.For example, if the standard deviation of response ua(t) is determined to be σa, the likelihoodthat at any time the response is less than 3σa is 99.9%.

If a system is subject to multiple sources of excitation, the cross-PSD of responses ua(t) and ub(t)is related to the cross-PSDs of the excitation sources and the frequency response functionsby the equation:

Equation 2-23.

where Sjk(ω) is the cross-PSD of loadings Pj(t) and Pk(t). Haj(ω) is the frequency responsefunction of ua(t) corresponding to the excitation source Pj(t) and similarly Hbk(ω)is the complexconjugate of the frequency response function of ub(t) corresponding to the excitation source Pk(t).If the two excitation sources are not correlated, we have Sjk(ω).= 0.

New RCROSS Case Control Command

RCROSS

Cross-Power Spectral Density and Cross-correlation Function Output Requests

Requests computation and output of cross-power spectral density andcross-correlation functions in random analysis.

FORMAT:

NX Nastran 5.0 Release Guide 2-21

Page 32: NX Nastran 5 Release Guide

Chapter 2 Dynamics

EXAMPLES:RCROSS(PHASE)=10RCROSS(PSDF,NOPRINT,PUNCH)=20

DESCRIBERS:

Describer Meaning

REAL orIMAG

Requests rectangular format (real and imaginary) of complexoutput (for cross-power spectral density function). Use of eitherREAL or IMAG yields the same output. (Default)

PHASE Requests polar format (magnitude and phase) of complexoutput (for cross-power spectral density function). Phaseoutput is in degrees.

PRINT Writes output to print file. (Default)

NOPRINT Does not write output to the print file.

PUNCH Writes output to punch file.

PSDF Requests the cross-power spectral density function becalculated and output for random analysis post-processing.(Default)

CORF Requests the cross-correlation function be calculated andoutput for random analysis post-processing.

RALL Requests both the cross-power spectral density andcross-correlation function be calculated and output for randomanalysis post-processing.

n Identification number of RCROSS bulk data entry to be used inrandom analysis. (Integer > 0)

REMARKS:1. The case control RCROSS entry must be used in conjuction with the case

control RANDOM entry. See remarks under the RANDOM case control entry.

2. Response quantities, such as DISPLACEMENT, STRESS, and FORCE,must be requested by corresponding case control entries in order to computecross-power spectral density and cross-correlation functions between thetwo response quantities specified by the RCROSS bulk data entry. It isrecommended that the DISPLACEMENT, STRESS, and FORCE requests beput above the subcase level to ensure that these response quantities existwhen the random analysis post-processing occurs.

3. The response quantities must belong to the same superelement. Thecross-power spectral density and cross-correlation functions between tworesponses that belong to different superelements are not supported.

2-22 NX Nastran 5.0 Release Guide

Page 33: NX Nastran 5 Release Guide

Dynamics

New RCROSS Bulk Entry

RCROSS

Cross-Power Spectral Density and Cross-Correlation Function Output

Defines a pair of response quantities for computing the cross-power spectraldensity and cross-correlation functions in random analysis.

FORMAT:

1 2 3 4 5 6 7 8 9 10RCROSS SID RTYPE1 ID1 COMP1 RTYPE2 ID2 COMP2 CURID

EXAMPLE:

RCROSS 20 DISP 50 2 STRESS 150 8 4

FIELDS:

Field Contents

SID Case control RCROSS identification number for cross-powerspectral density and cross-correlation functions. (Integer>0)

RTYPEi Type of response quantity. At least one field must be selected. Seeremark 2. (Character or blank)

IDi Element, grid, or scalar point identification number. (Integer > 0)

COMPi Component code (item) identification number. See remark 3.(Integer > 0)

CURID Curve identification number. See remark 4. (Integer > 0 or blank)

REMARKS:1. This entry is required for computing the cross-power spectral density and

cross-correlation functions. SID must be selected with the case controlcommand (RCROSS=SID). Fields RTYPE1, ID1, and COMP1 represent thefirst response quantity; fields RTYPE2, ID2, and COMP2 represent the secondresponse quantity.

2. The keywords for field RTYPEi are listed as follows:

Keyword Meaning

DISP Displacement Vector

VELO Velocity Vector

ACCEL Acceleration Vector

OLOAD Applied Load Vector

SPCF Single-point Constraint Force Vector

NX Nastran 5.0 Release Guide 2-23

Page 34: NX Nastran 5 Release Guide

Chapter 2 Dynamics

MPCF Multi-point Constraint Force Vector

STRESS Element Stress

STRAIN Element Strain

FORCE Element Force

If either RTYPE1 or RTYPE2 is blank, then the blank field takes the defaultfrom the defined field.

3. For elements, the item code COMPi represents a component of the elementstress, strain or force and is described in Tables “Element Stress-Strain ItemCodes” and “Element Force Item Codes”. For an item having both a real andimaginary part, the code of the real part must be selected. This is required forcomputing both the cross-power spectral density function and cross-correlationfunction.

For grid points, the item code is one of 1,2,3,4,5, and 6, which represent themnemonics T1, T2, T3, R1, R2, and R3, respectively. For scalar points, alwaysuse 1.

4. Field CURID is optional. It is for your convenience to identify the output byusing a single index.

2-24 NX Nastran 5.0 Release Guide

Page 35: NX Nastran 5 Release Guide

Dynamics

2.4 DDAM Process AutomationThe Dynamic Design Analysis Method (DDAM) is a list of procedures to determine the modalshock response of on-board ship equipment due to underwater explosions. To complete DDAManalysis procedures in previous releases of NX Nastran, you were required to use dmap alterfiles. In NX Nastran 5, these DDAM procedures have been streamlined and automated, thuseliminating the alter requirement.

The NX Nastran 5 DDAM analysis has been implemented as a new solution sequence, SOL 187.This solution sequence has 3 separate phases which can all occur automatically from a single NXNastran job submittal, or can optionally be completed interactively:

• Phase 1: A modal analysis (SOL 187) runs to calculate the natural frequencies. Then theparticipation factors and modal effective weights are calculated for each mode. The modes,participation factors, and modal effective weights are written to an ASCII OUTPUT4 file.

• Phase 2: The Naval Shock Analysis (NAVSHOCK) FORTRAN program is automaticallyinvoked to compute the modal shock responses. NAVSHOCK uses the following files as input:

The OUTPUT4 file created in phase 1 by NX Nastran.

A required, user-created DDAM Control file storing various runtime options.

An optional, user-created DDAM Coefficient file containing the weighting factors used forthe response calculations, the directional scaling factors, as well as the modal mass cutoffvalue. This file must be listed in the DDAM Control file for NAVSHOCK to use it.

An optional, user-created Shock Spectra file which defines the input shock spectrum asdata pairs of frequency and displacement, velocity, acceleration. This file must also belisted in the DDAM Control file for NAVSHOCK to use it.

• Phase 3: The modal shock responses created in phase 2 are read by NX Nastran, and resultsare recovered and output for post-processing.

DDAM Input Requirements

The following are the input requirements needed to run a DDAM analysis. When you run DDAMfrom a single NX Nastran job submittal, the NX Nastran input file creates a wrapper around theentire DDAM analysis process. The details of this NX Nastran input file, plus all files used bythe NAVSHOK program are discussed below.

NX Nastran Input File Requirements

File Management and Executive Control

You must select solution sequence SOL 187 in the executive control section.

Note: The analysis inputs and options for SOL 187 are the same as for SOL 103.

The DDAM analysis requires 3 ASSIGN statements in the NX Nastran input file. Thesestatements link physical files required by the FORTRAN NAVSHOK program to the NX NastranOUTPUT4 and INPUT4 files. DDAM analysis in NX Nastran uses the units 11, 13 and 22as described below.

NX Nastran 5.0 Release Guide 2-25

Page 36: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Unit 11 is assigned to an OUTPUT4 file and is where NX Nastran writes the matrices whichare used as input to the FORTRAN program. The following ASSIGN statement assigns thefile ddam.f11 to be used as input to the FORTRAN program. The “DELETE” qualifier tellsNX Nastran to delete any existing version of the file and replace it with the one generatedin the current submittal.

ASSIGN OUTPUT4=‘ddam.f11’, UNIT=11, FORM=FORMATTED, DELETE

Unit 13 is assigned to an INPUT4 file that stores the output from the FORTRAN program.Once the FORTRAN program has completed, NX Nastran reads the results from this file. Thefollowing ASSIGN statement assigns the file ddam.f13 to store the output of the FORTRANprogram. The new DEFER keyword tells NX Nastran not to check for the existence of the file atthe start of the NX Nastran job. This keyword is needed because the FORTRAN output file isnot created until the FORTRAN program runs later in the job.

ASSIGN INPUTT4=‘ddam.f13’, UNIT=13, FORM=FORMATTED, DEFER

Unit 22 is assigned to an INPUT4 file that stores the input control options for the FORTRANprogram. The FORTRAN program reads from this file the keyboard responses which wouldnormally be provided interactively. The following ASSIGN statement assigns unit 22 to theddam.inp file:

ASSIGN INPUTT4=‘ddam1.inp’, UNIT=21, FORM=FORMATTED

Case Control

The Case Control section for SOL 187 is similar to SOL 103, but there are a few exceptions:

• To obtain the needed modal mass for final response post-processing, it is important that youdefine a large enough frequency range for the natural frequency solution.

• Superelement reduction procedures are supported, although NRL-summed results recoveryonly occurs for the residual.

• The number of supported case control output requests is limited as compared to SOL 103.The available output requests for SOL 187 are STRESS, FORCE, DISPLACEMENT,VELOCITY and ACCELERATION. Mode shapes are written by default.

Parameter DPREONLY

The new DPREONLY parameter is used to signal the SOL 187 analysis that it is to stopexecution immediately before the FORTRAN NAVSHOK program is run. This option is usedwhen you want to run the FORTRAN program outside the NX Nastran job. Often for securityreasons, you may be required to run the FORTRAN program on a separate, secure computer,or you may be required to manually enter the shock coefficients from the keyboard. Settingthis parameter to “YES” will stop the analysis after the OUTPUT4 files are written, but justbefore the FORTRAN program is executed.

Once the FORTRAN program has been run manually, you use a NX Nastran restart solution tocontinue the analysis. Note, that the restart input file must use the /, directive to remove theDPREONLY parameter from the restart database. This is done by looking at the sorted BulkData from the original run and finding the sequence number (n) of the DPREONLY parameterstatement, then including a "/,n" directive after the BEGIN BULK in the restart input file. Thesorted Bulk Data is output by default. If the ECHO Case Control command is specified in theoriginal run, it should be set to output the sorted Bulk Data (for example, ECHO=SORT). See theRestarts chapter in the NX Nastran User’s Guide for more details.

2-26 NX Nastran 5.0 Release Guide

Page 37: NX Nastran 5 Release Guide

Dynamics

Bulk Data

In addition to all the information needed for a typical SOL 103 analysis, the bulk data sectionrequires the following:

• A SUPORT entry is required to define the single fixed base location for the 3 translationaldegrees-of-freedom. Rigid elements or MPCs can be used to connect this single locationacross a distributed region of the model. If superelements are defined, the fixed base pointmust be part of the residual, and in the Aset. The directions of the SUPORT location shouldreference a rectangular coordinate system. Although the shock response is only computed intranslational DOF, it is still important to SPC any of the 6 DOF directions at the fixed pointwhich are not included on the SUPORT entry. If a massless DOF is included on the SUPORTentry, the following error occurs: “MR MATRIX has NULL column”.

• The eigenvectors must be normalized to “MASS.” Failure to do so will result in incorrectmodal masses and participation factors.

• The NX Nastran model must be defined in the units of length=inches and mass=lb-sec2/insince this is a requirement of the NAVSHOK program.

Required DDAM Control File

The NAVSHOK program requires the ASCII control file which supplies runtime options. Asdescribed above in the file management descriptions, this is the file assigned to unit 21. Theformat of this control file will be one of three, depending on which options you request.

Default user options:

F Fnsurf nstruc nplastpreff/a_axis vert_axisf11 filenamef13 filename.ver filename

User coefficient option:

T Fcoef.dat filenamensurf nstruc nplastpreff/a_axis vert_axis.f11 filename.f13 filename.ver filename

User spectrum Option:

F Tspec.dat filename

NX Nastran 5.0 Release Guide 2-27

Page 38: NX Nastran 5 Release Guide

Chapter 2 Dynamics

preff/a_axis vert_axis.f11 filename.f13 filename.ver filename

Specific file formats:

First Line - spectrum control - format a1,1x,a1First Item - Coefficients from File or form compiled sourceT=coefficients from external fileF=use built-in coefficientsSecond item - DDAM or general spectrum run flagT=General non-DDAM spectrum run

F=DDAM

Second Line - file name (if needed) -format a80If 1st item on line 1 is T, Name of coefficient fileIf 2nd item on line 1 is T, Name spectrum fileIf neither are T, line is not needed

Third Line - location flags - format i1,1x,i1,1x,i1First Item - Surface or Submarine1=Surface2=SubmarineSecond Item - equipment location1=Deck2=Hull3=ShellThird Item - coefficient type1=Elastic2=Elastic/Plastic

4th Line - Weight cutoff percentage - format F8.3 (0. To 100.)5th Line - Axis Orientation - format a1,1x,a1

First Item - F/A axis,X,Y, or ZSecond Item - Vertical Axis X,Y, or Z

6th Line - Input file name -format a807th Line - Output file name - format a808th Line - Verification file name -format a80

2-28 NX Nastran 5.0 Release Guide

Page 39: NX Nastran 5 Release Guide

Dynamics

Note: The capitalization and spacing of the first line and the axis definition line areimportant. Below is an example to demonstrate proper formatting:

F F1 1 180.X Znavs1.f11navs2.f11navs1.ver

Optional DDAM Coefficient File

The DDAM coefficient file, which is formatted similar to an NX Nastran input file, contains theweighting factors used for the response calculations, the directional scaling factors, and themodal mass cutoff value. The equations which describe these factors are:

(M is the modal weight in kips.)

For surface ship, hull and shell mount, the equation is:

A complete set of AA, AB, AC, and (when needed) AD weighting factors for each of the possibleanalysis configurations (surface ship and submerged ship, deck mount, shell mount and hullmount) must be included. In addition, there is a set of three AF (acceleration factors) and VF(velocity factors) for each of these, one for each shock direction. Additional factors are alsorequired for Elastic-Plastic conditions.

The coefficients and factors are defined on a COEF entry:

COEF nsurf nstruc nplastVF(1) VF(2) VF(3) AF(1) AF(2) AF(3)VA VB VC AA AB AC AD

nsurf Ship type. Allowable values are SUB (submerged) and SURF (surfaceship)nstruc Mounting location. Allowable values are DECK, HULL, and SHELLnplast Elastic or elastic-plastic factors. Allowable values are ELASTIC and ELPL

The (1), (2), and (3) after VF and AF refer to the directions: (1)=fore/aft, (2)=athwartship, and(3)=vertical.

To use the default in the NAVSHOK program, enter a blank or an asterisk *. In addition to theCOEF entry, there is a CUTOFF entry which defines the modal mass cutoff percentage:

1 2 3 4 5 6 7 8 9 10CUTOFF pref

NX Nastran 5.0 Release Guide 2-29

Page 40: NX Nastran 5 Release Guide

Chapter 2 Dynamics

pref is the cutoff weight percentage for the modal mass calculation. Enter as a percentage, not adecimal fraction (that is 85. instead of .85).

# Example factors# - deck coefficientsCOEF SURF DECK ELPL

.30 .40 1.0 .27 .60 1.010. 20. 45. 13. 31.6 6.

# hull coefficientsCOEF SURF HULL ELPL

.4 .5 1.0 .25 .30 1.04. 8. 38. 38. 53.4 6.5 15.

# shell coefficientsCOEF SURF SHELL ELPL

.25 .50 1.0 .25 .50 1.0* * * 10. 44.6 7.2 17

All fields must be 8 characters long, and can be achieved by adding blanks to the ends oflines.

Optionally defined Shock Spectra

The Shock Spectra file allows you to define the spectrum as data pairs of frequency versusdisplacement, velocity or acceleration. To span several orders of magnitude, you can use eitherthe logarithmic or linear scale for the data.

The format of the ASCII file is as follows:

# = comments, anywhere in fileDATATYP type data1 dir freq interpBEGIN DATAf1, data1f2, data2fn, datan[BEGIN DATA][f1, data1][f2, data2][fn, datan][BEGIN DATA][f1, data1][f2, data2][fn, datan]END FILE

The data on the DATATYP entry are as follows:

type One of the following: DISP, VELO, or ACCE

data

Units of displacement, velocity or acceleration data as described:

G (acceleration data in Gs), F (Displacement, Velocity or Acceleration data in feet,ft/sec, or ft/sec2), I (displacement, velocity or acceleration in inches, in/sec or in/sec2),or M (displacement, velocity or acceleration in meters, m/sec or m/sec2)

2-30 NX Nastran 5.0 Release Guide

Page 41: NX Nastran 5 Release Guide

Dynamics

dirNumber of spectra in the file. A value of 1 indicates a single spectrum will be usedfor all three shock directions. A value of 3 indicates there are three spectra, one foreach direction.

freq Units of frequency terms. RAD (radians) or HERTZ (frequency in hertz).

interp

Flags the interpolation scheme for the table values. Options are: LOGLOG (bothaxes are logarithmic), LINLIN (neither axis is logarithmic), LOGLIN (the frequencyaxis is logarithmic, the other is not), or LINLOG (the frequency range is linear, theother is logarithmic).

A BEGIN DATA entry should precede each frequency/motion data section. If dir=1, there willbe one BEGIN DATA entry, if dir=3, there will be three. The data section is completed withthe END FILE entry.

Example:

# sample spectrum fileDATATYP ACCE I 1 HERTZ LOGLIN# acceleration vs frequency file - acceleration in in/sec**2BEGIN DATA1. 1.10. 10.100. 1000.# point added to define range500. 800.1000. 500.END FILE

Running ddam.exe interactively

You may choose to end the NX Nastran job after the natural frequency solution has completed,then run the NAVSHOK program interactively. When running interactively, NAVSHOK willprompt you for all needed inputs.

The NAVSHOK program is included in the NX Nastran installation at /nxn5/arch/ddam.exe.

Selection of Specific Modes

DMI (Direct Matrix Input) is a way to choose which modes go into the NRL summation. You candefine a partitioning vector (PARTNVEC) on the DMI entries which will break the eigenvectormatrix and UHV matrix into specific pieces. PARTNVEC is a multi-row, 3-column matrix wherethe rows are mode numbers, and the columns are shock direction. If the matrix has a “1”, thatmode is retained for a shock direction. The example below demonstrates a PARTNVEC definition:

1 2 3 4 5 6 7 8 9$ define the matrix as a 12 row by 3 column matrixDMI PARTNVEC 0 2 1 12 3$ define col 1 (keep all modes – 1 -12) (f/a shock)DMI PARTNVEC 1 1 1. THRU 12$ define col 2 (keep modes 1, 2 and 3) (athw shock)DMI PARTNVEC 2 2 1. 2 1. 3 1.$ define col 3 (keep only modes 5) (vert. shock)DMI PARTNVEC 3 5 1. 1 1 1 1

NX Nastran 5.0 Release Guide 2-31

Page 42: NX Nastran 5 Release Guide

Chapter 2 Dynamics

You will need to request all the modal mass (“100.”) if you use the PARTNVEC option, since itwill perform its own partitioning on the output from the DDAM program.

OUTPUT4 Data Written in Phase 1

Phase I of the SOL 187 is a modal analysis to calculate the natural frequencies of the system.Then the participation factors and modal effective weights are calculated for each mode and arewritten to Output 4 files in ASCII format. Phase I computes and outputs the following:

An Output4 file containing the following matrices:OMEX Natural frequenciesPAB Participation factorsMTOT Total mass in each directionMFRACT Percent of total mass from the modes usedThe program also prints the following for user verification:

The 4 matrices written to the above Output4 fileMTOTC Total rigid body mass matrixMEFC Total modal effective mass matrixMEFF Diagonal Terms of MEFFCPHBASE Rigid body vector set referenced to R set

2-32 NX Nastran 5.0 Release Guide

Page 43: NX Nastran 5 Release Guide

Dynamics

2.5 Elemental Energy OutputThe element strain (ESE), kinetic (EKE), and damping (EDE) case control commands havebeen improved to allow the output of element energy information in SORT2 complex format.This format is suitable for plotting function data from frequency domain dynamic solutions.In previous releases, you could only output the real magnitude of element energy, thus nophase information was written. In addition, only SORT1 format could be output, which is notconvenient for function plotting.

To control the output format, new describer sets [RMAG, REAL or IMAG, PHASE] and [SORT1,SORT2] have been added to the ESE, EKE, and EDE case control cards. These new sets areconsistent with the other NX Nastran case control commands, but the following exception:

• The RMAG option allows you to output the energy magnitude as a real number only, which isneeded to retain the same data output as used in previous versions.

The updated case control formats for EDE, EKE, and ESE are included at the end of this section.

The table below summarizes the output result with different describer selections:

RMAG REAL or IMAG PHASE

AVERAGE Eavg (Eavg , 0.) (Eavg , 0.)

AMPLITUDE Eamp_mag (Eamp_real , Eamp_imag ) (Eamp_mag , θamp )

PEAK Epeak - -

AVGAMP -(Eavg , 0.)

(Eamp_real , Eamp_imag )

(Eavg , 0.)

(Eamp_mag , θamp )

Table of Energy Output Results

Example Element Energy Output Formats

Example 1:

ESE(PUNCH, AVGAMP, REAL, SORT2 )=ALL

PUNCH Output$TITLE = MODAL FREQUENCY RESPONSE 1$SUBTITLE= 2$LABEL = 3$AVERAGE ELEMENT STRAIN ENERGY 4

NX Nastran 5.0 Release Guide 2-33

Page 44: NX Nastran 5 Release Guide

Chapter 2 Dynamics

$REAL-IMAGINARY OUTPUT 5$SUBCASE ID = 1 6$ELEMENT NUMBER = 201 7

1.000000E-01 4.918013E+00 0.000000E+00 81.100000E-01 5.975217E+00 0.000000E+00 91.200000E-01 7.143042E+00 0.000000E+00 101.300000E-01 8.424302E+00 0.000000E+00 111.400000E-01 9.822133E+00 0.000000E+00 121.500000E-01 1.133997E+01 0.000000E+00 13

.

.

.$TITLE = MODAL FREQUENCY RESPONSE 7993$SUBTITLE= 7994$LABEL = 7995$AMPLITUDE ELEMENT STRAIN ENERGY 7996$REAL-IMAGINARY OUTPUT 7997$SUBCASE ID = 1 7998$ELEMENT NUMBER = 201 7999

1.000000E-01 4.914062E+00 1.970973E-01 80001.100000E-01 5.969384E+00 2.639417E-01 80011.200000E-01 7.134707E+00 3.449702E-01 80021.300000E-01 8.412709E+00 4.418095E-01 80031.400000E-01 9.806374E+00 5.561835E-01 80041.500000E-01 1.131896E+01 6.899223E-01 80051.600000E-01 1.295405E+01 8.449761E-01 8006

Example 2: ESE(PUNCH, AVGAMP, REAL, SORT1 )=ALL

PUNCH Output

$TITLE = MODAL FREQUENCY RESPONSE 1$SUBTITLE= 2$LABEL = 3$AVERAGE ELEMENTAL STRAIN ENERGY 4$REAL-IMAGINARY OUTPUT 5$SUBCASE ID = 1 6$FREQUENCY = 0.1000000E+00 7

201 4.918013E+00 0.000000E+00 8202 1.939230E-05 0.000000E+00 9203 2.092620E-03 0.000000E+00 10204 4.920125E+00 0.000000E+00 11

$TITLE = MODAL FREQUENCY RESPONSE 12

.

.

.$TITLE = MODAL FREQUENCY RESPONSE 21924$SUBTITLE= 21925$LABEL = 21926$AMPLITUDE ELEMENTAL STRAIN ENERGY 21927$REAL-IMAGINARY OUTPUT 21928$SUBCASE ID = 1 21929$FREQUENCY = 0.1000000E+00 21930

201 7.134707E+00 3.449702E-01 21931202 2.792866E-05 2.065627E-08 21932203 3.013540E-03 9.193350E-07 21933204 7.137749E+00 3.449712E-01 21934

$TITLE = MODAL FREQUENCY RESPONSE 21935

2-34 NX Nastran 5.0 Release Guide

Page 45: NX Nastran 5 Release Guide

Dynamics

Updated Case Control Commands EDE, EKE, and ESE

EDE

Element Energy Loss Per Cycle Output Request

Requests the output of the energy loss per cycle in selected elements.

FORMAT:

EXAMPLES:EDE=ALLEDE(PUNCH, THRESH=.0001)=19

DESCRIBERS:

Describer Meaning

PRINT Write energies to the print file (default).

PUNCH Write energies to the punch file.

PLOT Do not write energies to either the punch file or the print file.

AVERAGE Requests average energy in frequency response analysis only.

AMPLITUDE Requests amplitude of energy in frequency response analysisonly.

PEAK Requests peak energy for frequency response analysis only.PEAK is the sum of AVERAGE and AMPLITUDE.

AVGAMP Requests both the average and amplitude energy in frequencyresponse analysis only.

RMAG Outputs the energy magnitude in real data format.

REAL orIMAG

Requests rectangular format (real and imaginary) of complexoutput. Use of either REAL or IMAG yields the same output.

PHASE Requests polar format (magnitude and phase) of complexoutput. Phase output is in degrees.

NX Nastran 5.0 Release Guide 2-35

Page 46: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Describer Meaning

SORT1 Output will be presented as a tabular listing of grid pointsfor each load, frequency, eigenvalue, or time, depending onthe solution sequence.

SORT2 Output will be presented as a tabular listing of frequency ortime for each grid point.

THRESH Suppresses energies for elements having an energy value ofless than p% in all output files. THRESH overrides the valueof TINY described in Remark 1. (Default=0.001)

ALL Computes energy for all elements.

n Set identification number. Energy for all elements specified onthe SET n command will be computed. The SET n commandmust be specified in the same subcase as the EDE commandor above all subcases. (Integer>0)

NONE Element energy will not be output.

REMARKS:1. If THRESH = p is not specified, then p defaults to the values specified by

user parameter TINY.

2. The energy calculations include the contribution of initial thermal strain.

3. Energy density (element energy divided by element volume) is also computedin some solution sequences. It can be suppressed by use of PARAM,EST,-1.

4. The equations used to calculate elemental damping energy components aregiven below.

Average Damping Energy:

where:

E = elemental energy

ur = displacement - real part

ui = displacement - imaginary part

[Be] = elemental damping

Real Part of Damping Energy Amplitude:

Imaginary Part of Damping Energy Amplitude:

2-36 NX Nastran 5.0 Release Guide

Page 47: NX Nastran 5 Release Guide

Dynamics

Magnitude of Damping Energy Amplitude:

Phase of Damping Energy Amplitude:

Peak Damping Energy:

5. In SOL 111, EDE can only be requested if PARAM,DDRMM,-1 is used.

6. Only damping from the viscous dampers (e.g., CVISC, CDAMPi, etc.) areincluded. Structural damping is not included in the calculation.

EKE

Element Kinetic Energy Output Request

Requests the output of the kinetic energy in selected elements.FORMAT:

EXAMPLES:EKE=ALLEKE(PUNCH, THRESH=.0001)=19

DESCRIBERS:

Describer Meaning

PRINT Write energies to the print file (default).

PUNCH Write energies to the punch file.

PLOT Do not write energies to either the punch file or the print file.

AVERAGE Requests average energy in frequency response analysis only.

NX Nastran 5.0 Release Guide 2-37

Page 48: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Describer Meaning

AMPLITUDE Requests amplitude of energy in frequency response analysisonly.

PEAK Requests peak energy for frequency response analysis only.PEAK is the sum of AVERAGE and AMPLITUDE.

AVGAMP Requests both the average and amplitude energy in frequencyresponse analysis only.

RMAG Outputs the energy magnitude in real data format.

REAL orIMAG

Requests rectangular format (real and imaginary) of complexoutput. Use of either REAL or IMAG yields the same output.

PHASE Requests polar format (magnitude and phase) of complexoutput. Phase output is in degrees.

SORT1 Output will be presented as a tabular listing of grid pointsfor each load, frequency, eigenvalue, or time, depending onthe solution sequence.

SORT2 Output will be presented as a tabular listing of frequency ortime for each grid point.

THRESH Suppresses energies for elements having an energy value ofless than p% in all output files. THRESH overrides the valueof TINY described in Remark 1. (Default=0.001)

ALL Computes energy for all elements.

n Set identification number. Energy for all elements specified onthe SET n command will be computed. The SET n commandmust be specified in the same subcase as the EKE commandor above all subcases. (Integer>0)

NONE Element energy will not be output.

REMARKS:1. If THRESH = p is not specified, then p defaults to the values specified by

user parameter TINY.

2. The energy calculations include the contribution of initial thermal strain.

3. Energy density (element energy divided by element volume) is also computedin some solution sequences. It can be suppressed by use of PARAM,EST,-1.

4. The equations used to calculate elemental kinetic energy components aregiven below.

Average Kinetic Energy:

2-38 NX Nastran 5.0 Release Guide

Page 49: NX Nastran 5 Release Guide

Dynamics

where:

E = elemental energy

vr = velocity - real part

vi = velocity - imaginary part

[Me] = elemental mass

Real Part of Kinetic Energy Amplitude:

Imaginary Part of Kinetic Energy Amplitude:

Magnitude of Kinetic Energy Amplitude:

Phase of Kinetic Energy Amplitude:

Peak Kinetic Energy:

5. In SOL 111, EKE can only be requested if PARAM,DDRMM,-1 is used.

ESE

Element Strain Energy Output Request

Requests the output of the strain energy in selected elements.

NX Nastran 5.0 Release Guide 2-39

Page 50: NX Nastran 5 Release Guide

Chapter 2 Dynamics

FORMAT:

EXAMPLES:ESE=ALLESE (PUNCH, THRESH=.0001)=19

DESCRIBERS:

Describer Meaning

PRINT Write energies to the print file (default).

PUNCH Write energies to the punch file.

PLOT Do not write energies to either the punch file or the print file.

AVERAGE Requests average energy in frequency response analysis only.

AMPLITUDE Requests amplitude of energy in frequency response analysisonly.

PEAK Requests peak energy for frequency response analysis only.PEAK is the sum of AVERAGE and AMPLITUDE.

AVGAMP Requests both the average and amplitude energy in frequencyresponse analysis only.

RMAG Outputs the energy magnitude in real data format.

REAL orIMAG

Requests rectangular format (real and imaginary) of complexoutput. Use of either REAL or IMAG yields the same output.

PHASE Requests polar format (magnitude and phase) of complexoutput. Phase output is in degrees.

SORT1 Output will be presented as a tabular listing of grid pointsfor each load, frequency, eigenvalue, or time, depending onthe solution sequence.

SORT2 Output will be presented as a tabular listing of frequency ortime for each grid point.

THRESH Suppresses energies for elements having an energy value ofless than p% in all output files. THRESH overrides the valueof TINY described in Remark 1. (Default=0.001)

2-40 NX Nastran 5.0 Release Guide

Page 51: NX Nastran 5 Release Guide

Dynamics

Describer Meaning

ALL Computes energy for all elements.

n Set identification number. Energy for all elements specified onthe SET n command will be computed. The SET n commandmust be specified in the same subcase as the ESE command orabove all subcases. (Integer>0)

NONE Element strain energy will not be output.

REMARKS:1. If THRESH = p is not specified then p defaults to the values specified by user

parameter TINY.

2. The energy calculations include the contribution of initial thermal strain.

3. Energy density (element strain energy divided by element volume) is alsocomputed in some solution sequences. It can be suppressed by use ofPARAM,EST,-1.

4. The equations used to calculate elemental strain energy components aregiven below.

Average Strain Energy:

where:

E = elemental energy

ur =

displacement - real part

ui = displacement - imaginary part

[Ke] = elemental stiffness

Real Part of Strain Energy Amplitude:

Imaginary Part of Strain Energy Amplitude:

Magnitude of Strain Energy Amplitude:

NX Nastran 5.0 Release Guide 2-41

Page 52: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Phase of Strain Energy Amplitude:

Peak Strain Energy:

5. In SOL 111, ESE can only be requested if PARAM,DDRMM,-1 is used.

6. Element data recovery for thermal loads is not currently implemented indynamics.

7. Element strain energy is available for nonlinear static analysis (SOL 106).All other nonlinear solution sequences do not support element strain energyoutput.

8. The strain energy for nonlinear elements is calculated by integrating thespecific energy rate, which is the inner product of strain rate and stress, overelement volume and time.

Equation 2-24.

where:

σ = stress tensor

= tensor of the strain rate

V = element volume

t = actual time in the load history

Loads from temperature changes are included in Equation 2-24. If weassume a linear variation of temperatures from subcase to subcase, then thestrain energy in Equation 2-24—for the special case of linear material andgeometry—becomes

Equation 2-25.

where Pet is the element load vector for temperature differences.

For linear elements, the default definition of element strain energy is

2-42 NX Nastran 5.0 Release Guide

Page 53: NX Nastran 5 Release Guide

Dynamics

Equation 2-26.

where Pet is the element load vector for temperature differences and elementdeformation.

In Equation 2-26, the temperatures are assumed to be constant within asubcase. The default definition of the strain energy for linear elements differsfrom the definition for nonlinear elements by a factor of 1/2 in the temperatureloads. To request the strain energy for linear elements using Equation 2-25,set the parameter XFLAG to 2; the default value for XFLAG is 0, which usesEquation 2-26 for the strain energy of linear elements.

NX Nastran 5.0 Release Guide 2-43

Page 54: NX Nastran 5 Release Guide

Chapter 2 Dynamics

2.6 Modal Frequency Response in Rotor DynamicsRotor dynamics can now be used in a modal frequency response solution, SOL 111, to calculatethe dynamic response of a rotating system. This new response calculation adds to the existingrotor dynamics capability, direct complex mode solution, SOL 110, whose output is used tocreate Campbell diagrams.

The dynamic response calculations can be done using an asynchronous solution in which therotor speed is defined, but is independent of the applied excitation, or using a synchronoussolution in which the external excitation is synchronous with the rotor speed. A new option onthe ROTORD bulk entry, SYNC, has been created to select the calculation method.

Asynchronous solution inputs:

The asynchronous solution runs when SYNC is set to “0”. The rotor speeds are defined by theRSTART entry for each rotor definition. The RSTEP and NUMSTEP entries in this case areignored. The frequency definition comes from the standard entries for frequency responseanalysis such as FREQ, DLOAD, RLOAD, which also define the dynamic loading scenario.

Synchronous solution inputs:

The synchronous analysis runs when SYNC is set to “1”. The rotor speeds are defined byRSTART, RSTEP, and NUMSTEP on the ROTORD card. The frequency corresponding to theserotation speeds is computed from the rotation speed, which also defines the dynamic loading.

See the Rotor Dynamics Input Updates section below for the version 5 updates.

Mathematical Details

The general equation of motion in the frequency domain for a rotating system with appliedloads is:

In the above, ω is the frequency of applied loading, Ω is the rotor rotation speed, u(ω) is thedisplacement, and F(ω) is the applied load. The matrices m, b, k are the modal mass, dampingand stiffness. The c is the modal Coriolis matrix and the kb is the internal structural damping.

There are two forms of the general equation which depend on whether the applied loadingis speed dependent or not (synchronous or asynchronous). If the loading is speed dependent(synchronous), then ω = Ω, and the general equation has the form:

The loading in the synchronous case comes from any mass unbalance on the rotating rotors(not from an external excitation).

When the loading is not speed dependent (asynchronous), then ω≠ Ω, and the general equationhas the form:

The loading in the asynchronous case comes from an external excitation other than the rotationof the rotor. Each rotor in the asynchronous case has a defined speed Ω, which remains constantfor the response calculation.

2-44 NX Nastran 5.0 Release Guide

Page 55: NX Nastran 5 Release Guide

Dynamics

Rotor Dynamics Input Updates

Case Control

The following new case control card, MODSEL=n, has been created to select or deselect the modenumbers in dynamic response and aero-elasticity solutions:

MODSEL

Selects/deselects mode numbers

Used to select or deselect mode numbers in dynamic response and aero-elasticitysolutions.

FORMAT:MODSEL = n

EXAMPLES:MODSEL=3

DESCRIBERS:

Describer Meaning

n Points to a SET card that holds the list of selected and retainedmode numbers. The default is to use all modes. The modenumbers omitted from the list are removed from the modal space.Numbers larger than the number of eigenvalues computed areignored. (Integer)

Bulk Entries

The ROTORD entry has been updated with the new SYNC option, in addition, the RSETi optionwill now reference the RSETID on the new ROTORG entry. The ROTORG is now the requiredmethod of selecting the grids which compose a rotor. Both the updated ROTORD and the newROTORG are shown below:

ROTORD

Defines rotor dynamics solution options (SOLs 110 and 111).

Defines rotor dynamics solution options for SOLs 110 and 111.

FORMAT:

1 2 3 4 5 6 7 8 9 10ROTORD SID RSTART RSTEP NUMSTEP REFSYS CMOUT RUNIT FUNIT ++ ZSTEIN ORBEPS ROTPRT SYNC ETYPE EORDER+ RID1 RSET1 RSPEED1 RCORD1 W3_1 W4_1 RFORCE1+ RID2 RSET2 RSPEED2 RCORD2 W3_2 W4_2 RFORCE2....

+ RID10 RSET10 RSPEED10 RCORD10 W3_10 W4_10 RFORCE10

NX Nastran 5.0 Release Guide 2-45

Page 56: NX Nastran 5 Release Guide

Chapter 2 Dynamics

EXAMPLE:

ROTORD 998 0.0 5000.0 58 fix -1.0 cps +r0

+r0 no +r1

+r1 1 11 1 0.0 0.0 1 +r2

+r2 2 12 1 0.0 0.0 +r3

+r3 3 13 1 0.0 0.0 +r4

+r4 4 14 1 0.0 0.0 +r5

+r5 5 15 1 0.0 0.0 +r6

+r6 6 16 1 0.0 0.0 +r7

+r7 7 17 1 0.0 0.0 +r8

+r8 8 18 1 0.0 0.0 +r9

+r9 9 19 1 0.0 0.0 +r10

+r10 10 20 1 0.0 0.0 10

FIELDS:

Field Contents

SID Set identifier for all rotors. Must be selected in the case controldeck by RMETHOD=SID.

RSTART Starting value of rotor speed. (Real ≥ 0.0)

RSTEP Step-size of rotor speed. (Real > 0.0)

NUMSTEP Number of steps for rotor speed including RSTART.(Integer > 0)

REFSYS Reference system. (default=‘ROT’)

= ‘FIX’ analysis is performed in the fixed reference system.

= ‘ROT’ analysis is performed in the rotational reference system.

CMOUT Rotor speed for complex mode output request. (Integer ≥ 0)

= -1.0 output complex modes and whirl direction for all RPM.

= 0.0 no complex modes and no whirl direction are output.(default)

> 0.0 rotor speed value for which complex modes will becalculated and written to F06 or OP2.

RUNIT Units used for rotor speed input (RSTART and RSTEP) andoutput (units for output list and Campbell’s diagram output).

= ‘RPM’ revolutions per minute. (default)

= ‘CPS’ cycles per second.

= ‘HZ’ cycles per second.

= ‘RAD’ radians per second.

2-46 NX Nastran 5.0 Release Guide

Page 57: NX Nastran 5 Release Guide

Dynamics

Field Contents

FUNIT Units used for frequency output.

= ‘RPM’ revolutions per minute. (default)

= ‘CPS’ cycles per second.

= ‘HZ’ cycles per second.

= ‘RAD’ radians per second.

ZSTEIN Option to incorporate Steiner’s inertia terms. (default=NO)

ORBEPS Threshold value for detection of whirl direction. (default=1E-6)

ROTPRT Controls .f06 output options.

= 0 no print. (default)

= 1 print generalized matrices.

= 2 print eigenvalue summary and eigenvectors at each RPM.

= 3 combination of 1 & 2.

SYNC Specifies if a rotor dynamics response calculation (SOL 111) issynchronous or asynchronous. (Integer)

=0 synchronous (default)

=1 asynchronous

ETYPE Excitation type.

=1 Mass unbalanced (default). Specify mass unbalance = m x ron DLOAD bulk entry and the program will multiply by Ω2.

=0 Force excitation. Specify force = m x r x Ω2 on DLOAD bulkentry.

EORDER Excitation order

= 1.0 (default)

= 0.0 Forward whirl (modes crossing with 0P line)

= 2.0 Backward whirl (modes crossing with 2P line)

This value is always 1.0 for fixed reference system.

RIDi Rotor ID for output identification (default=i); must be uniquewith respect to all other RIDi values.

RSETi Refers to a RSETID value on a ROTORG bulk entry. EachROTORG bulk entry specifies the grids for a single rotor.(Integer)

NX Nastran 5.0 Release Guide 2-47

Page 58: NX Nastran 5 Release Guide

Chapter 2 Dynamics

Field Contents

RSPEEDi(placeholder)

Once supported, RSPEED will adjust the rotor speed for rotor iin a rotor dynamics response calculation. RSPEED is currentlyhard coded to equal 1.

RCORDi Points to CORD** entry specifying rotation axis or rotor i as z.(default=0 for global z axis)

W3_i Reference frequency for structural damping defined by PARAMG for rotor i. (default=0.0)

W4_i Reference frequency for structural damping defined on materialcard for rotor i. (default=0.0)

RFORCEi Points to RFORCE bulk data entry for rotor i. (Default=0 forno rotational force applied; a rotational force is required fordifferential stiffness to be calculated.)

REMARKS:1. Any entries where defaults exist are optional. Thus, if the defaults are

acceptable for a particular model, then a continuation card would not beneeded.

2. There is a maximum limit of 10 rotors (i.e. 11 continuation cards).

3. The Steiner’s term option (ZSTEIN) should only be used for analyzing the“stiff” part of the rotor. In the rotating system analysis, there is a stiffeningeffect of the centrifugal forces. The gyroscopic matrix is calculated in the fixedsystem.

4. The W3 parameter defines the reference frequency for the structural dampingdefined by PARAM G.

5. The W4 parameter defines the reference frequency for the structural dampingdefined on the material cards.

6. The static centrifugal force is calculated for unit speed measured in rad/sec. Onthe RFORCE card, the unit of Hz is used, thus the conversion 1/2(π)=0.159155must be used.

7. For calculating the modal frequency response using synchronous analysis, therotation speeds are defined by the RSTART, RSTEP and NUMSTEP fields onthe ROTORD bulk entry. The frequencies corresponding to these rotationspeeds are computed and the dynamic loads are calculated accordingly.

8. For calculating the modal frequency response using asynchronous analysis,the unique rotation speed is defined by the RSTART field on the ROTORDbulk entry. The RSTEP and NUMSTEP fields in this case will be ignored. Thefrequency and dynamic load definitions are taken from the standard FREQ,DLOAD, and RLOAD bulk entries.

2-48 NX Nastran 5.0 Release Guide

Page 59: NX Nastran 5 Release Guide

Dynamics

ROTORG

Rotor Grids Selection

Specifies the grids that compose each rotor.

FORMAT:

1 2 3 4 5 6 7 8 9 10ROTORG RSETID GRID1 GRID2 ... GRIDn

or

ROTORG RSETID GRID1 THRU GRID2 BY INC

EXAMPLE:

ROTORG 14 6 10 15 18

FIELDS:

Field Contents

RSETID Identification number of rotor. References a RSETi on theROTORD bulk entry. (Integer>0)

GRIDi Grids defining a rotor. (Integer>0)

THRU Specifies a range of grid ID’s (Optional)

BY Specifies an increment when using THRU option.

INC Increment used with THRU option. (Integer>0)

NX Nastran 5.0 Release Guide 2-49

Page 60: NX Nastran 5 Release Guide

Chapter 2 Dynamics

2.7 Modal Contribution ComputationWhen you use the modal method for dynamic response calculations, understanding which modescontribute to the response helps understand the dynamic behavior of the simulated system, andcan provide insight as to how to improve the dynamic performance.

In this release, a new capability has been introduced to compute the modal contributions inmodal frequency response.

The new command has the form:

MODCON = YES/NO

with a default of no, meaning the modal contribution is not output. The option to PRINT, PLOT,or PUNCH the modal contribution data can be selected using the existing DISP, VELO, ACCEand the SDISP, SVELO and SACCELERATION case control methods.

There are no new bulk data entries required, however, you can use the USET card with U3 todefine the l=1,....n degrees-of-freedom which modal contributions are computed.

The new MODCON case control command is included below:

New MODCON Case Control Command

MODCON

Modal Contribution Request

Requests the output of modal contribution data for SOL 111.FORMAT:

MODCON=YES/NOEXAMPLES:

MODCON=YES

DESCRIBERS:

Describer Meaning

NO Do not output modal contribution data. (Default)

YES Modal contribution data will be output.

REMARKS:1. The option to PRINT, PLOT, or PUNCH the modal contribution data

can be selected by the DISP, VELO, ACCE and the SDISP, SVELO andSACCELERATION case control methods.

2. Output of modal contribution can only occur on grids. Element output is notsupported at this time.

3. No new bulk data entries are required, however, the USET card with U3 canbe used to define the l=1,....n degrees-of-freedom which modal contributionsare computed.

4. The complex and actual modal contributions are assigned with dummy subcaseIDs for each DOF in the f06 and punch files to distinguish each set of results

2-50 NX Nastran 5.0 Release Guide

Page 61: NX Nastran 5 Release Guide

Dynamics

while post processing. It is recommended to use a large increment whenassigning the SUBCASE IDs in the case control section to avoid a dummy IDduplicating a case control ID.

5. This capability is NOT supported for the constraint mode method of enforcedmotion.

6. This capability is not supported for models having rigid body or fluid modes.

New Output DatablocksThe new datablocks, MODCONPS and MODALCBN, have been created to output the absoluteand actual complex modal contributions. The computed results are printed under the followingheaders:

*****************************************************************************************THE TABLE(S) STARTING ON THE FOLLOWING PAGE CONTAIN(S) THE COMPLEX MODAL CONTRIBUTIONS*****************************************************************************************

*****************************************************************************************THE TABLE(S) STARTING ON THE FOLLOWING PAGE CONTAIN(S) THE ACTUAL MODAL CONTRIBUTIONS*****************************************************************************************

Mathematical DescriptionThe general equation for the calculation of the physical response using the modal method is:

In the above equation,

represents some type of physical response from a modal dynamic simulation (displacement,velocity, acceleration, stress, strain ….). The subscript l represents a specific component of thephysical response (e.g. x-direction of node 22). The variable:

,

is the modal displacement response of mode number m. There are a total of M modes. Thevariables:

are elements of the mode shape matrix. Finally,

are the modal contributions of the physical response.

It is the terms:

NX Nastran 5.0 Release Guide 2-51

Page 62: NX Nastran 5 Release Guide

Chapter 2 Dynamics

that are to be computed for the new modal contribution output. From the above equation, themodal contributions are calculated as:

2-52 NX Nastran 5.0 Release Guide

Page 63: NX Nastran 5 Release Guide

Dynamics

2.8 Direct Structural Damping Matrix InputA structural damping matrix can now be included with a model in NX Nastran 5 using the newK42GG case control command. K42GG is repeated below for convenience. This ability adds tothe other direct matrix capabilities of stiffness, mass, and damping matrices using DMIG bulkentries. For more information on direct matrix inputs, see “Matrix Assembly Operations inSubDMAP SEMG” in the NX Nastran User’s Guide.

New K42GG Case Control Command

K42GG

Direct Input Structural Damping Matrix Selection

Selects direct input structural damping matrices.

FORMAT:K42GG=name

EXAMPLES:K42GG=K4DMIGK42GG=K1, K2, K3SET 100=K4DMIG, K1, K8K42GG=100

DESCRIBERS:

Describer Meaning

name Name of a [k4,2gg] matrix that is input on the DMIG Bulk

Data entry. (Character) See “Matrix Assembly Operations inSubDMAP SEMG” in the NX Nastran User’s Guide.

REMARKS:1. DMIG matrices will not be used unless selected.

2. Terms are added to the structural damping matrix before any constraintsare applied.

3. The matrix must be symmetric and field 4 on the DMIG,name entry mustcontain the integer 6.

4. A scale factor may be applied to this input via the PARAM, CK42 entry.

5. The matrices are additive if multiple matrices are referenced on the K42GGcommand.

NX Nastran 5.0 Release Guide 2-53

Page 64: NX Nastran 5 Release Guide
Page 65: NX Nastran 5 Release Guide

Chapter

3 Numerical Capabilities

NX Nastran 5.0 Release Guide 3-1

Page 66: NX Nastran 5 Release Guide

Chapter 3 Numerical Capabilities

3.1 Sparse Data Recovery for Modal SolutionsA new sparse data recovery option is available for the modal frequency response analysis (SOL111), modal transient response analysis (SOL 112), and optimization (SOL 200) as the newdefault. For SOL 111 and 112, this feature reduces the cost of matrix-multiplications insideDDRMM modules when you want to recover large amounts of data in the modal analysis.Similarly, SOL 200 utilizes the partitioning of the eigenvector matrix in order to reduce the costof matrix-matrix multiplies.

Inputs

The option is on by default, but can be deactivated with system cell 421:

NASTRAN SYSTEM(421)=0

or by using the nastran keyword SPARSEDR:

NASTRAN SPARSEDR =0.

Performance Data

For a SOL 111 model, the SPARSEDR option reduces the time spent by more than 80%:

Version 4 Version 5 UnitsDDRMM time 103:09 20:50 min:secMemory HiW 61.5 65.5 MwordsDisk HiWater 128 127 Gbytes

In a SOL 200 model, the SPARSEDR option reduces the time spent by more than 23 %:

With sparsedr:

1:48:50 490:20 2042.1G .0 13008.4 .0 EXITOPT 73 EXIT BEGN

Without sparsedr:

2:51:18 636:22 2673.8G .0 16838.7 .0 EXITOPT 73 EXIT BEGN

3.2 Distributed Memory Parallel ImprovementsDue to the increased size and complexity of models today, reducing solution times is becomingmore important. The NX Nastran parallel processing methods allow simultaneous use ofmultiple processors on one or more machines to decrease solution times.

The parallel processing method Distributed Memory Parallel (DMP) has been enhanced inNX Nastran 5 with the following additions:

• The normal mode calculation portion of an optimization solution (SOL 200) can now use DMP

• A new linear statics (SOL 101) DMP option, Load Domain Static Analysis (LDSTAT), nowexists to decrease the solution times when large numbers of load cases exist.

The details of these improvements are as follows. See the NX Nastran Parallel Processing User’sGuide for complete details on all parallel processing methods.

3-2 NX Nastran 5.0 Release Guide

Page 67: NX Nastran 5 Release Guide

Numerical Capabilities

Optimization, SOL 200 with Normal Modes

When an optimization solution (SOL 200) requires normal mode frequencies as part of theoptimization problem, the normal mode computation portion of the solution can now takeadvantage of DMP. When DMP is activated in this situation, SOL 200 will run in a serialenvironment on all processes, except the normal mode analysis.

Input Requirements

In a SOL 200 solution, the DMP option can be activated using nastran keywords similar toa SOL 103 DMP run. The DMP methods supported are Frequency Domain Normal ModesAnalysis (FDMODES), Geometric Domain Normal Modes Analysis (GDMODES), and HierarchicDomain Normal Modes Analysis (HDMODES).

With GDMODES and HDMODES, the keyword gpart=1 is required.

Depending on the desired DMP method, one of the following inputs should be included in theSOL 200 submittal:

GDMODES: NASTRAN gpart=1 dmp=p

HDMODES: NASTRAN gpart =1 dmp =p nclust=c

FDMODES: NASTRAN dmp=p

where p is the number of processors and c is the number of clusters.

Linear Statics Improvement

The new computational technique for DMP linear static analysis (SOL 101), and Load DomainStatic Analysis (LDSTAT) is now available. LDSTAT will split the analysis tasks such thateach processor will solve the problem for a single load case. Solving multiple load casessimultaneously across multiple processors can shorten the solution time dramatically when largenumbers of load cases are included in the run.

Input Requirements

The option is activated with the nastran keyword gpart and dmp:

LDSTAT: NASTRAN gpart=1 dmp=p

where p is the number of processors.

Example

A model with 70 load cases was solved on a single processor machine (SERIAL case), then wassolved across 4 processors using the new DMP option LDSTAT. The solve time was reducedby more than 74 %:

SERIAL

17:01:18 25:13 53849.0 0.0 1323.4 0.0 STATRS 96 SSG3 BEGN17:03:06 27:01 59732.0 5883.0 1430.9 107.5 STATRS 96 SSG3 END

LDSTAT ( p=4 )

NX Nastran 5.0 Release Guide 3-3

Page 68: NX Nastran 5 Release Guide

Chapter 3 Numerical Capabilities

13:16:42 17:46 38790.0 0.0 717.8 0.0 STATRS 102 SSG3 BEGN13:17:10 18:14 42896.0 4106.0 745.9 28.1 STATRS 102 SSG3 END

3-4 NX Nastran 5.0 Release Guide

Page 69: NX Nastran 5 Release Guide

Numerical Capabilities

3.3 Sparse Cholesky Technique for Linear SolutionsThis NX Nastran release includes support for a new sparse Cholesky decomposition. All users ofCholesky decomposition (including the Givens and Householder eigensolver methods) will utilizesparse decomposition by default, which can be considerably faster than the non-sparse Choleskydecomposition method used in earlier NX Nastran versions. For example, the decomposition costof a typical structural stiffness matrix of order 8000 compares as follows:

Decomposition type Wall time (min:sec)Non-sparse Cholesky 1:40Sparse Cholesky 0:16

With the sparse method, the DCMP module produces a sparse Cholesky factor, identified by form14 in its trailer. To use the non-sparse Cholesky method, set system cell 424 to :

NASTRAN SYSTEM(424)=0

or use the SPCHOL keyword:

NASTRAN SPCHOL=0.

Not all FBS options are available with sparse Cholesky; for details, see the Numerical User’sGuide.

NX Nastran 5.0 Release Guide 3-5

Page 70: NX Nastran 5 Release Guide

Chapter 3 Numerical Capabilities

3.4 Eigenvalue Analysis ImprovementNX Nastran Version 5.0 introduces the new REDMULT performance option for use when solvingvibration problems with the Lanczos method. This option reduces the cost of matrix-vectormultiplies inside the READ module when the mass matrix involved is relatively dense, whichcan occur when a large number of MFLUID is present in the model. In such cases, matrix-vectormultiplies can be a considerable portion of the overall Lanczos time, as indicated by the followingsummary:

*** USER INFORMATION MESSAGE 5403 (LNNRIGL)BREAKDOWN OF CPU TIME (SEC) DURING LANCZOS ITERATIONS:OPERATION REPETITIONS AVERAGE TOTALFBS (BLOCK SIZE= 7) 75 10.97 822.60MATRIX-VECTOR MULTIPLY 173 10.15 1756.69SHIFT AND FACTOR 5 1699.75 8498.73LANCZOS RUN 3 9.40 28.21

Using the REDMULT option reduces the time spent in the READ module by more than 15%:

*** USER INFORMATION MESSAGE 5403 (LNNRIGL)BREAKDOWN OF CPU TIME (SEC) DURING LANCZOS ITERATIONS:OPERATION REPETITIONS AVERAGE TOTALFBS (BLOCK SIZE= 7) 81 12.34 999.39MATRIX-VECTOR MULTIPLY 187 .00 .69SHIFT AND FACTOR 5 1687.57 8437.84LANCZOS RUN 3 10.41 31.24

Inputs

The option is activated by setting system cell 426:

NASTRAN SYSTEM(426)=1

or by using the REDMULT keyword:

NASTRAN REDMULT=1.

The REDMULT option may be used independently of or in conjunction with the REDORTHoption, introduced in Version 4.0.

3-6 NX Nastran 5.0 Release Guide

Page 71: NX Nastran 5 Release Guide

Chapter

4 Multi-body DynamicSoftware Interfaces

NX Nastran 5.0 Release Guide 4-1

Page 72: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

4.1 RecurDyn Flex Input File CreationThis release includes interoperability between NX Nastran and the FunctionBay RFI (RecurDynFlex Input) file product. You can now create a RecurDyn Flex Input (RFI) file directly fromNX Nastran. The RFI contains the reduced order matrices from the results of a NX Nastrannon-restart SOL 103 analysis. The RFI can be imported into RecurDyn and used to representa flexible component in a multi-body dynamics analysis. This direct RFI export capabilitystreamlines the process of creating flexible components from FE models, making it possible toobtain more accurate results from multi-body simulations.

The component modes determined by NX Nastran during the SOL 103 analysis represent a setof constraint modes for boundary coordinates plus a truncated set of fixed-interface normalmodes. This combination of mode types is referred to as the Craig-Bampton modes. The RFI filecontains data of this representation that FunctionBay can use to perform flexible componentdynamics analyses.

The RFI creation is initiated by the NX Nastran case control command RECURDYNRFIFLEXBODY=YES, along with the inclusion of the Bulk Data entry DTI,UNITS.

You can define the flexible body attachment points in NX Nastran by either defining thecomponent as a superelement or part superelement, in which case the physical external (a-set)grids become the attachment points. For a residual-only type model, standard NX Nastran ASETBulk Data entries are used to define the attachment points.

The new RECURDYNRFI is included below for convenience. The documented descriptions andremarks on this command contain a comprehensive list of use requirements for RFI export.These include the equations representing the nine mass variants, and the DTI,UNITS bulkentry requirement for units control of the RFI.

New Case Control Command

RECURDYNRFI

Generates RecurDyn Flex Input File

Generates a RecurDyn Flex Input (RFI) file during SOL 103.

FORMAT:

4-2 NX Nastran 5.0 Release Guide

Page 73: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

EXAMPLES:RECURDYNRFI FLEXBODY=YES

DESCRIBERS:

Describer Meaning

FLEXBODY Requests the generation of RFI (required).

NO Standard NX Nastran solution without RFI creation (default).

YES RFI generation requested.

FLEXONLY Determines if standard DMAP solution and data recoveryruns or not after RFI creation is complete.

YES Only RFI creation occurs (default).

NO RFI file creation occurs along with standard DMAP solutionand data recovery.

OUTGSTRS Determines if grid point stress is written to RFI.

NO Do not write grid point stress to RFI (default).

YES Write grid point stress to RFI.

OUTGSTRN Determines if grid point strain is written to RFI.

NO Do not write grid point strain to RFI (default).

YES Write grid point strain to RFI.

MINVAR Determines how mass invariants are computed.

PARTIAL Mass invariants 6 and 8 are not computed.

CONSTANT Mass invariants 1,2,3 and 9 are computed.

FULL All nine mass invariants are computed.

NONE No mass invariants are computed.

PSETID Selects a set of elements defined in the OUTPUT(PLOT)(including PLOTEL) whose connectivity is exported to facegeometry into the RFI. (See Remark 16)

NONE No specific sets are selected, thus all grids, geometry andassociated modal data are written to RFI (default).

setid The connectivity of a specific element set is used to exportface geometry.

NX Nastran 5.0 Release Guide 4-3

Page 74: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

Describer Meaning

ALL The connectivity of all element sets are used to export facegeometry.

REMARKS:1. The creation of the RecurDyn Flex Input file is applicable in a non-restart

SOL 103 analysis only.

2. The creation of the RecurDyn Flex Input file is initiated by RECURDYNRFIFLEXBODY=YES (other describers are optional) along with the inclusionof the bulk data entry DTI,UNITS.

3. The Data Table Input Bulk Data entry DTI,UNITS, which is required for anRECURDYNRFI FLEXBODY=YES run, is used to specify the system of unitsfor the data stored in the RFI (unlike NX Nastran, RecurDyn is not a unitlesscode). Once identified, the units will apply to all superelements in the model.The complete format is:

DTI UNITS 1 MASS FORCE LENGTH TIME

All entries are required. Acceptable character strings are listed below.

Mass:

KG - kilogram

LBM – pound-mass

SLUG – slug

GRAM – gram

OZM – ounce-mass

KLBM – kilo pound-mass (1000.lbm)

MGG – megagram

Force:

N – Newton

LBF – pound-force

KGF – kilograms-force

OZF – ounce-force

DYNE – dyne

KN – kilonewton

KLBF – kilo pound-force (1000.lbf)

MN - millinewton

Length:

KM – kilometer

4-4 NX Nastran 5.0 Release Guide

Page 75: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

M – meter

CM – centimeter

MM – millimeter

MI – mile

FT – foot

IN – inch

Time:

H – hour

MIN-minute

S – second

MS – millisecond

4. Since DTI,UNITS determines all units for the RFI, the units defined inWTMASS, which are important for units consistency in NX Nastran, areignored in the output to the RFI. For example, if the model mass is kilograms,force in Newtons, length in meters, and time in seconds, then WTMASS wouldequal 1 ensuring that NX Nastran works with the consistent set of kg, N, andm. The units written to the RFI would be: “DTI,UNITS,1,KG,N,M,S”.

5. You can create flexible body attachment points by defining the component asa superelement or part superelement, in which case the physical external(a-set) grids become the attachment points; or for a residual-only type model,you can use standard NX Nastran ASET Bulk Data entries to define theattachment points.

6. The eight mass variants are:

NX Nastran 5.0 Release Guide 4-5

Page 76: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

sp = [xyz]T are the coordinates of grid point p in the basic coordinate system.

φp=partitioned orthogonal modal matrix that corresponds to the translationaldegrees of freedom of grid p.

Ip=inertia tensor p.

φp*=partitioned orthogonal modal matrix that corresponds to the rotational

degrees of freedom of grid p.

=skew-symmetric matrix formed for each grid translational degree offreedom for each mode.

M=number of modes.

N=number of grids.

7. To accurately capture the mode shapes when supplying SPOINT/QSETcombinations, the number of SPOINTS (ns) should be at least ns=n+(6+p),assuming that residual flexibility is on. In the above equation for ns, thenumber of modes (n) is specified on the EIGR or EIGRL Bulk Data entries; thenumber of load cases is p. In general, you can’t have too many SPOINTs, asexcess ones will simply be truncated with no performance penalty.

8. For FLEXBODY=YES runs, residual vectors for the component shouldalways be calculated as they result in a more accurate representation of thecomponent shapes at little additional cost.

9. OMIT or OMIT1 Bulk Data entries are not supported.

10. Lumped mass formulation (default) is required. Either leavePARAM,COUPMASS out of the input file or supply PARAM,COUPMASS,-1(default) to ensure lumped mass.

11. P-elements are not allowed because they are always use a coupled massformulation.

12. PARAM,WTMASS,value with a value other than 1.0 may be used with an NXNastran run generating an RFI. It must have consistent units with regard tothe DTI,UNITS Bulk Data entry. Before generating the RFI, NX Nastran willappropriately scale the WTMASS from the physical mass matrix and modeshapes.

4-6 NX Nastran 5.0 Release Guide

Page 77: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

13. There is a distinction between how an RECURDYNRFI FLEXBODY=YES runhandles element-specific loads (such as a PLOAD4 entry) versus those thatare grid-specific (such as a FORCE entry), especially when superelementsare used. The superelement sees the total element-specific applied load.For grid-specific loads, the loads attached to an external grid will movedownstream with the grid. That is to say, it is part of the boundary and notpart of the superelement. This distinction applies to a superelement run andnot to a residual-only or parts superelement run.

14. The loads specified in NX Nastran generally fall into two categories:non-follower or fixed direction loads (non-circulatory) and follower loads(circulatory). The follower loads are nonconservative in nature. Examplesof fixed direction loads are the FORCE entry or a PLOAD4 entry when itsdirection is specified via direction cosines. Examples of follower loads arethe FORCE1 entry or the PLOAD4 entry when used to apply a normalpressure. By default in NX Nastran, the follower loads are always active inSOL 103 and will result in follower stiffness being added to the differentialstiffness and elastic stiffness of the structure. In a run with RECURDYNRFIFLEXBODY=YES and superelements, if the follower force is associatedwith a grid description (such as a FORCE1) and the grid is external to thesuperelement, the follower load will move downstream with the grid. Thus,the downstream follower contribution to the component’s stiffness will be lost,which could yield poor results. This caution only applies to a superelement runand not to a residual-only or a part superelement run.

15. OUTGSTRS and OUTGSTRN entries require the use of standard NX NastranSTRESS= or STRAIN= used in conjunction with GPSTRESS= or GPSTRAIN=commands to produce grid point stress or strain. GPSTRESS(PLOT)= orGPSTRAIN(PLOT)= will suppress grid stress or strain print to the NXNastran .f06 file.

16. To reduce the FE mesh detail for dynamic simulations, PSETID (on theRECURDYNRFI Case Control command) defined with a SET entry is used todefine a set of PLOTELs or other elements used to select grids to display thecomponents in RecurDyn. This option can significantly reduce the size of theRFI without compromising accuracy in the FunctionBay simulation providingthat the mass invariant computation is requested. With superelementanalysis, for any of these elements that lie entirely on the superelementboundary (all of the elements’ grids attached only to a-set or exterior grids), aSEELT Bulk Data entry must be specified to keep that display element withthe superelement component. This can also be accomplished using PARAM,AUTOSEEL,YES. The SEELT entry is not required with parts superelements,as boundary elements stay with their component.

If the SET entry points to an existing set from the OUTPUT(PLOT) section,this single set is used explicitly to define elements used to select grids todisplay the component in RecurDyn. If PSETID does not find the set ID inOUTPUT(PLOT), it will search sets in the case control for a matching set ID.This matching set ID list then represents a list of OUTPUT(PLOT) definedelements’ sets, the union of which will be used to define a set of PLOTELs orother elements used to select grids to display the component in RecurDyn. Ifthe user wishes to select all of the sets in the OUTPUT(PLOT) section, thenuse PSETID=ALL.

NX Nastran 5.0 Release Guide 4-7

Page 78: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

The following element types are not supported for writing to an RFI, norare they supported as a ‘type’ entry in a set definition in OUTPUT(PLOT):CAABSF, CAEROi, CDUMi, CHACAB, CHACBR, CHBDYx, CDAMP3,CDAMP4, CELAS3, CELAS4, CFLUIDi, CMASS3, CMASS4, CRAC2D,CRAC3D, CTRMEM, CTWIST, CWEDGE, CWELD, and GENEL.

17. Typical NX Nastran data entry requirements are described below.

Typical Parameters:

• PARAM,RESVEC,character_value – controls calculation of residualflexibility (including inertia relief) modes. In SOL 103, residual flexibilityis on by default for only component modes (o-set).

• PARAM,GRDPNT, value - mass invariants 1I, 2I, and 3I will be computedusing results of NX Nastran grid point weight generator execution in thebasic coordinate system.

Typical Case Control:

• RECURDYNRFI FLEXBODY=YES is required for RFI generation.

• METHOD=n is required before or in the first subcase for modal solutions.

• SUPORT1=seid is necessary to select a static support set for a residualonly linear preload run.

• SUPER=n,SEALL=n is useful with multiple superelement models to selectan individual superelement as a flexible body. Cannot be used with alinear STATSUB(PRELOAD) run.

• OUTPUT(PLOT) is necessary to define elements used to select grids todisplay the component in RecurDyn when PSETID=ALL or setid.

SET n=list of elements (including PLOTELs) is used to select grids todisplay the component.

• OUTPUT(POST) is necessary to define volume and surface for grid stressor strain shapes.

SET n=list is a list of elements for surface definition for grid stress orstrain shapes.

Stress and strain data in the RFI is limited to the six components (i.e. 3normal and 3 shear) for a grid point for a given mode.

SURFACE n SET n NORMAL z3 is used to define a surface for writingstress and strain data. Only one FIBER selection is allowed for eachSURFACE, thus the use of the FIBRE ALL keyword on the SURFACE casecontrol command will write stresses to the RFI at the Z1 fiber location only.

Since the FIBRE keyword only applies to stresses, strain data will alwaysbe written to the RFI at the MID location.

Stress and strain data at grid points can only be written to the RFI forsurface and volume type elements (e.g. CQUAD and CHEXA).

VOLUME n SET n is a volume definition.

4-8 NX Nastran 5.0 Release Guide

Page 79: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

The default SYSTEM BASIC is required with SURFACE or VOLUME.

• STRESS(PLOT) is necessary for stress shapes.

• STRAIN(PLOT) is necessary for strain shapes.

• GPSTRESS(PLOT) is necessary for grid point stress shapes to be includedin the RFI.

• GPSTRAIN(PLOT) is necessary for grid point strain shapes to be includedin the RFI.

Typical Bulk Data:

• DTI,UNITS,1,MASS,FORCE,LENGTH,TIME is required for RFIgeneration. For input files containing superelements, this command mustreside in the main bulk data section.

• SPOINT,id_list defines and displays modal amplitude.

• SESET,SEID,grid_list defines a superelement (see GRID and BEGINBULK SUPER=). The exterior grids will represent the attachment pointsalong with the q-set.

• SEELT,SEID,element_list reassigns superelement boundary elements toan upstream superelement.

• RELEASE,SEID,C,Gi is an optional entry that removes DOFs from anattachment grid for which no constraint mode is desired. For example, thisallows the removal of rotational degrees of freedom from an analysis whereonly translational degrees of freedom are required.

• SEQSET,SEID,spoint_list defines modal amplitudes of a superelement(see SEQSET1).

• SENQSET,SEID,N defines modal amplitudes of a part superelement. Itmust reside in the main Bulk Data Section.

• ASET,IDi,Ci defines attachment points for a residual-only run (see ASET1).

• QSET1,C,IDi defines modal amplitudes for the residual structure or modalamplitudes for a part superelement (see QSET).

• SUPORT1,SID,IDi,Ci defines the static support for a preload conditionwith a residual-only run. This entry is case control selectable. Do notuse SUPORT.

• PLOTEL,EID,Gi can be used, along with existing model elements, to defineelements used to select grids to display the components in RecurDyn.

• EIGR,SID,METHOD,… obtains real eigenvalue extraction (see EIGRL).

18. RECURDYNRFI and ADAMSMNF case control entries cannot be used in thesame analysis run. In other words, a RecurDyn RFI file or an ADAMS MNFfile can be generated during a particular NX Nastran execution, but not both

NX Nastran 5.0 Release Guide 4-9

Page 80: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

files at the same time. Attempting to generate both files in the same analysiswill cause an error to be issued and the execution to be terminated.

4.2 ADAMS Stress RecoveryNX Nastran 3 introduced the ability to write an ADAMS modal neutral file (MNF) directly froman NX Nastran SOL 103 analysis using the ADAMSMNF case control command. The resultingMNF is imported into ADAMS and used to represent a flexible component in a multi-bodydynamics analysis.

NX Nastran 5 enhances this process with a new results recovery capability. The results from anADAMS multi-body dynamics analysis, along with an optional component modal definitions file(OUTPUT2 format), are used in a consecutive NX Nastran SOL 103 results recovery solution.The component modal definitions file can be created by NX Nastran during the original solutionin which the MNF is created (most efficient method), or the component modal definitions canbe recalculated during the final recovery solution. The final solution recovers results such asstresses, strains, and element forces. These can be written to an OUTPUT2 file and importedinto an NX Nastran compatible post-processor.

Procedure for Results Recovery

The following steps describe the new procedure of results recovery, and introduce the new inputoptions.

1. Initial NX Nastran Sol 103, MNF creation

In this step, the ADAMS MNF creation occurs, which is used in step 2. The optionalcomponent modal definitions OUTPUT2 file is also created for use in the results recoverysolution in step 3. The input file used in step 1 is similar to a SOL 103, but with the additionof the ADAMSMNF case control command, and the DTI,UNITS bulk entry which is used tospecify the system of units for the data stored in the MNF.

If you choose to create the optional component modal definitions file in this step, you needthe new ADMOUT=YES option on the ADAMSMNF case control command, along with thefollowing ASSIGN statement in the file management section:

ASSIGN OUTPUT2=’name.out’ STATUS=UNKNOWN UNIT=20 FORM=UNFORM

The updates to the ADAMSMNF case control command include the new ADMOUT option,plus the new remarks 17 and 18. The updated ADAMSMNF case control command is givenbelow.

2. Flexbody Analysis in ADAMS

The MNF created in step 1 is imported into ADAMS, along with any other MNFs whichrepresent the components of a multi-body assembly. The ADAMS solution stores thedynamic responses into a modal deformations file (MDF) in binary format (OUTPUT2), orin ASCII format (Punch).

3. Final NX Nastran SOL 103, results recovery

The MDF created by ADAMS in step 2 is used in another NX Nastran SOL 103 solution torecover results, which can be used to evaluate the stress condition on specific components ofthe assembly. The new case control command ADMRECVR has been created to enter theneeded recovery options. The ADMRECVR command is repeated below for convenience.

4-10 NX Nastran 5.0 Release Guide

Page 81: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

If the component modal definitions were computed in step 1, or created using the originaldmap alter mnfx.alt, the resulting OUTPUT2 file can be used in step 3. The following linein the file management section of the NX Nastran input file is required to reference theexisting OUTPUT2 file:

ASSIGN INPUTT2=’<OUTPUT2_filename>’ UNIT=20

If no component modal definitions exist, there is an option to recompute them in the resultsrecovery solution. It is more efficient to calculate them in step 1 and have the OUTPUT2 fileready for the results recovery step. See the MSRMODE option on the new ADMRECVR casecontrol command below for more information on these options.

Updated/New Case Control Commands

ADAMSMNF

Generates ADAMS Interface Modal Neutral File

Generates ADAMS Interface Modal Neutral File (MNF) during SOL 103.

FORMAT:

EXAMPLES:ADAMSMNF FLEXBODY=YES

DESCRIBERS:

Describer Meaning

FLEXBODY Requests the generation of MNF.

NO Standard NX Nastran solution without MNF creation(default).

YES MNF generation requested.

NX Nastran 5.0 Release Guide 4-11

Page 82: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

Describer Meaning

FLEXONLY Determines if standard DMAP solution runs or not after MNFcreation is complete.

YES Only MNF creation occurs (default).

NO MNF file creation occurs along with standard DMAP solution.

OUTGSTRS Determines if grid point stress is written to MNF.

NO Do not write grid point stress to MNF (default).

YES Write grid point stress to MNF.

OUTGSTRN Determines if grid point strain is written to MNF.

NO Do not write grid point strain to MNF (default).

YES Write grid point strain to MNF.

MINVAR Determines how mass invariants are computed.

PARTIAL Mass invariants 5 and 9 are not computed.

CONSTANT Mass invariants 1,2,6 and 7 are computed.

FULL All nine mass invariants are computed.

NONE No mass invariants are computed.

PSETID Selects a set of elements defined in the OUTPUT(PLOT)section (including PLOTEL) or on a sketch file whoseconnectivity is exported to face geometry to be used inADAMS. (See Remark 15)

NONE All grids, geometry and associated modal data is written toMNF (default).

setid The connectivity of a specific element set is used to exportface geometry.

ALL The connectivity of all element sets are used to export facegeometry.

sktunit The connectivity of element faces defined on a sketch file isused to export face geometry. Note that the value must be anegative number to distinguish it from a setid value.

ADMOUT Requests that the FLEXBODY run output an NX NastranOP2 file for use in post processing of ADAMS/Flex results.

NO OP2 file will not be generated (default).

4-12 NX Nastran 5.0 Release Guide

Page 83: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

Describer Meaning

YES OP2 file will be generated.

CHECK Requests debug output be written to the f06 file whenADMOUT=YES (See Remark 19).

NO No debug output will be written (default).

YES Debug output will be written.

REMARKS:1. The creation of the Adams MNF, which is applicable in a non-restart SOL 103

analysis only, is initiated by ADAMSMNF FLEXBODY=YES (other describersare optional) along with the inclusion of the bulk data entry DTI,UNITS.

2. The Data Table Input Bulk Data entry DTI,UNITS, which is required for anADAMSMNF FLEXBODY=YES run, is used to specify the system of units forthe data stored in the MNF (unlike NX Nastran, ADAMS is not a unitlesscode). Once identified, the units will apply to all superelements in the model.The complete format is:

DTI UNITS 1 MASS FORCE LENGTH TIME

All entries are required. Acceptable character strings are listed below.

Mass:

KG - kilogram

LBM – pound-mass (0.45359237 kg)

SLUG – slug (14.5939029372 kg)

GRAM – gram (1E-3 kg)

OZM – ounce-mass (0.02834952 kg)

KLBM – kilo pound-mass (1000 lbm) (453.59237 kg)

MGG – megagram (1E3 kg)

MG – milligram (1E-6 kg)

MCG – microgram (1E-9 kg)

NG – nanogram (1E-12 kg)

UTON – U.S. ton (907.18474 kg)

SLI – slinch (175.1271524 kg)

Force:

N – Newton

LBF – pound-force (4.44822161526 N)

KGF – kilograms-force (9.80665 N)

OZF – ounce-force (0.2780139 N)

DYNE – dyne (1E-5 N)

NX Nastran 5.0 Release Guide 4-13

Page 84: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

KN – kilonewton (1E3 N)

KLBF – kilo pound-force (1000 lbf) (4448.22161526 N)

MN – millinewton (1E-3 N)

MCN – micronewton (1E-6 N)

NN – nanonewton (1E-9 N)

Length:

M – meter

KM – kilometer (1E3 m)

CM – centimeter (1E-2 m)

MM – millimeter (1E-3 m)

MI – mile (1609.344 m)

FT – foot (0.3048 m)

IN – inch (25.4E-3 m)

MCM – micrometer (1E-6 m)

NM – nanometer (1E-9 m)

A – Angstrom (1E-10 m)

YD – yard (0.9144 m)

ML – mil (25.4E-6 m)

MCI – microinch (25.4E-9 m)

Time:

S – second

H – hour (3600.0 sec)

MIN-minute (60.0 sec)

MS – millisecond (1E-3 sec)

MCS – microsecond (1E-6 sec)

NS – nanosecond (1E-9 sec)

D – day (86.4E3 sec)

3. Since DTI,UNITS determines all units for the MNF, the units defined inWTMASS, which are important for units consistency in NX Nastran, areignored in the output to the MNF. For example, if the model mass is kilograms,force in Newtons, length in meters, and time in seconds, then WTMASS wouldequal 1 ensuring that NX Nastran works with the consistent set of kg, N, andm. The units written to the MNF would be: “DTI,UNITS,1,KG,N,M,S”.

4. You can create flexible body attachment points by defining the component asa superelement or part superelement, in which case the physical external(a-set) grids become the attachment points; or for a residual-only type model,you can use standard NX Nastran ASET Bulk Data entries to define theattachment points.

4-14 NX Nastran 5.0 Release Guide

Page 85: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

5. The nine mass variants are:

sp = [xyz]T are the coordinates of grid point p in the basic coordinate system.

NX Nastran 5.0 Release Guide 4-15

Page 86: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

φp=partitioned orthogonal modal matrix that corresponds to the translationaldegrees of freedom of grid p.

Ip=inertia tensor p.

φp*=partitioned orthogonal modal matrix that corresponds to the rotationaldegrees of freedom of grid p.

=skew-symmetric matrix formed for each grid translational degree offreedom for each mode.

M=number of modes.

N=number of grids.

6. To accurately capture the mode shapes when supplying SPOINT/QSETcombinations, the number of SPOINTS (ns) should be at least ns=n+(6+p),assuming that residual flexibility is on. In the above equation for ns, thenumber of modes (n) is specified on the EIGR or EIGRL Bulk Data entries; thenumber of load cases is p. In general, you can’t have too many SPOINTs, asexcess ones will simply be truncated with no performance penalty.

7. For FLEXBODY=YES runs, residual vectors for the component shouldalways be calculated as they result in a more accurate representation of thecomponent shapes at little additional cost.

8. OMIT or OMIT1 Bulk Data entries are not supported.

9. Lumped mass formulation (default) is required. Either leavePARAM,COUPMASS out of the input file or supply PARAM,COUPMASS,-1(default) to ensure lumped mass.

10. P-elements are not allowed because they are always use a coupled massformulation.

11. PARAM,WTMASS,value with a value other than 1.0 may be used with an NXNastran run generating an MNF. It must have consistent units with regardto the DTI,UNITS Bulk Data entry. Before generating the MNF, NX Nastranwill appropriately scale the WTMASS from the physical mass matrix andmode shapes.

12. There is a distinction between how an ADAMSMNF FLEXBODY=YES runhandles element-specific loads (such as a PLOAD4 entry) versus those thatare grid-specific (such as a FORCE entry), especially when superelementsare used. The superelement sees the total element-specific applied load.For grid-specific loads, the loads attached to an external grid will movedownstream with the grid. That is to say, it is part of the boundary and notpart of the superelement. This distinction applies to a superelement run andnot to a residual-only or parts superelement run.

13. The loads specified in NX Nastran generally fall into two categories:non-follower or fixed direction loads (non-circulatory) and follower loads(circulatory). The follower loads are nonconservative in nature. Examplesof fixed direction loads are the FORCE entry or a PLOAD4 entry when itsdirection is specified via direction cosines. Examples of follower loads are theFORCE1 entry or the PLOAD4 entry when used to apply a normal pressure.

4-16 NX Nastran 5.0 Release Guide

Page 87: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

By default in NX Nastran, the follower loads are always active in SOL 103 andwill result in follower stiffness being added to the differential stiffness andelastic stiffness of the structure. In a run with ADAMSMNF FLEXBODY=YESand superelements, if the follower force is associated with a grid description(such as a FORCE1) and the grid is external to the superelement, the followerload will move downstream with the grid. Thus, the downstream followercontribution to the component’s stiffness will be lost, which could yieldpoor results. This caution only applies to a superelement run and not to aresidual-only or a part superelement run.

14. OUTGSTRS and OUTGSTRN entries require the use of standard NX NastranSTRESS= or STRAIN= used in conjunction with GPSTRESS= or GPSTRAIN=commands to produce grid point stress or strain. GPSTRESS(PLOT)= orGPSTRAIN(PLOT)= will suppress grid stress or strain print to the NXNastran .f06 file.

15. To reduce the FE mesh detail for dynamic simulations, PSETID (on theADAMSMNF case control command) defined with a SET entry (i.e. setid) isused to define a set of PLOTELs or other elements used to select grids to displaythe components in ADAMS. This option can significantly reduce the size of theMNF without compromising accuracy in the ADAMS simulation providing thatthe mass invariant computation is requested. With superelement analysis, forany of these elements that lie entirely on the superelement boundary (all of theelements’ grids attached only to a-set or exterior grids), a SEELT Bulk Dataentry must be specified to keep that display element with the superelementcomponent. This can also be accomplished using PARAM, AUTOSEEL,YES.The SEELT entry is not required with parts superelements, as boundaryelements stay with their component.

If the SET entry points to an existing set from the OUTPUT(PLOT) section,this single set is used explicitly to define elements used to select grids todisplay the component in ADAMS. If PSETID does not find the set ID inOUTPUT(PLOT), it will search sets in the case control for a matching set ID.This matching set ID list then represents a list of OUTPUT(PLOT) definedelements’ sets, the union of which will be used to define a set of PLOTELs orother elements used to select grids to display the component in ADAMS. Ifthe user wishes to select all of the sets in the OUTPUT(PLOT) section, thenuse PSETID=ALL.

The following element types are not supported for writing to an MNF, norare they supported as a ‘type’ entry in a set definition in OUTPUT(PLOT):CAABSF, CAEROi, CDUMi, CHACAB, CHACBR, CHBDYx, CDAMP3,CDAMP4, CELAS3, CELAS4, CFLUIDi, CMASS3, CMASS4, CRAC2D,CRAC3D, CTRMEM, CTWIST, CWEDGE, CWELD, and GENEL.

PSETID can also point to a sketch file using PSETID=-sktunit, where sktunitreferences an ASSIGN statement of the form

ASSIGN SKT=‘sketch_file.dat’,UNIT=sktunit.

The grids defined for the elements’ faces in the sketch file, along with allexternal (i.e. boundary) grids for the superelements, will be the only grids (andtheir associated data) written to the MNF.

The format of the sketch file, which describes the mesh as a collection of faces,must be as follows:

face_count

NX Nastran 5.0 Release Guide 4-17

Page 88: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

face_1_node_count face_1_nodeid_1 face_1_nodeid_2 ...face_2_node_count face_2_nodeid_1 face_2_nodeid_2 ...<etc>

Faces must have a node count of at least two. For example, a mesh comprisedof a single brick element might be described as follows:

64 1000 1001 1002 10034 1007 1006 1005 10044 1000 1004 1005 10014 1001 1005 1006 10024 1002 1006 1007 10034 1003 1007 1004 1000

Alternatively, the mesh might be described as a stick figure using a collectionof lines (two node faces), as shown below:

82 101 1022 102 1032 103 1042 104 1052 105 1062 106 1072 107 1082 108 109

16. Typical NX Nastran data entry requirements are described below.

Typical Parameters:

• PARAM,RESVEC,character_value – controls calculation of residualflexibility (including inertia relief) modes. In SOL 103, residual flexibilityis on by default for only component modes (o-set).

• PARAM,GRDPNT, value - mass invariants 1I, 2I, and 7I will be computedusing results of NX Nastran grid point weight generator execution in thebasic coordinate system.

Typical case control:

• ADAMSMNF FLEXBODY=YES is required for MNF generation.

• METHOD=n is required before or in the first subcase for modal solutions.

• SUPORT1=seid is necessary to select a static support set for a residualonly linear preload run.

• SUPER=n,SEALL=n is useful with multiple superelement models to selectan individual superelement as a flexible body. Cannot be used with alinear STATSUB(PRELOAD) run.

• OUTPUT(PLOT) is necessary to define elements used to select grids todisplay the component in ADAMS when PSETID=ALL or setid.

SET n=list of elements (including PLOTELs) is used to select grids todisplay the component.

• OUTPUT(POST) is necessary to define volume and surface for grid stressor strain shapes.

4-18 NX Nastran 5.0 Release Guide

Page 89: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

SET n=list is a list of elements for surface definition for grid stress orstrain shapes.

Stress and strain data in the MNF is limited to the six components (i.e. 3normal and 3 shear) for a grid point for a given mode.

SURFACE n SET n NORMAL z3 is used to define a surface for writingstress and strain data. Only one FIBER selection is allowed for eachSURFACE, thus the use of the FIBRE ALL keyword on the SURFACE casecontrol command will write stresses to the MNF at the Z1 fiber locationonly.

Since the FIBRE keyword only applies to stresses, strain data will alwaysbe written to the MNF at the MID location.

Stress and strain data at grid points can only be written to the MNF forsurface and volume type elements (e.g. CQUAD and CHEXA).

VOLUME n SET n is a volume definition.

The default SYSTEM BASIC is required with SURFACE or VOLUME.

• STRESS(PLOT) is necessary for stress shapes.

• STRAIN(PLOT) is necessary for strain shapes.

• GPSTRESS(PLOT) is necessary for grid point stress shapes to be includedin the MNF.

• GPSTRAIN(PLOT) is necessary for grid point strain shapes to be includedin the MNF.

Typical Bulk Data:

• DTI,UNITS,1,MASS,FORCE,LENGTH,TIME is required for MNFgeneration. For input files containing superelements, this command mustreside in the main bulk data section.

• SPOINT,id_list defines and displays modalamplitude.SESET,SEID,grid_list defines a superelement (seeGRID and BEGIN BULK SUPER=). The exterior grids will represent theattachment points along with the q-set.

• SEELT,SEID,element_list reassigns superelement boundary elements toan upstream superelement.

• RELEASE,SEID,C,Gi is an optional entry that removes DOFs from anattachment grid for which no constraint mode is desired. For example, thisallows the removal of rotational degrees of freedom from an analysis whereonly translational degrees of freedom are required.

• SEQSET,SEID,spoint_list defines modal amplitudes of a superelement(see SEQSET1).

• SENQSET,SEID,N defines modal amplitudes of a part superelement. Itmust reside in the main Bulk Data Section.

• ASET,IDi,Ci defines attachment points for a residual-only run (see ASET1).

NX Nastran 5.0 Release Guide 4-19

Page 90: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

• QSET1,C,IDi defines modal amplitudes for the residual structure or modalamplitudes for a part superelement (see QSET).

• SUPORT1,SID,IDi,Ci defines the static support for a preload conditionwith a residual-only run. This entry is case control selectable. Do notuse SUPORT.

• PLOTEL,EID,Gi can be used, along with existing model elements, to defineelements used to select grids to display the components in ADAMS.

• EIGR,SID,METHOD,… obtains real eigenvalue extraction (see EIGRL).

17. ADAMSMNF and RECURDYNRFI case control entries cannot be used in thesame analysis run. In other words, an ADAMS MNF file or a RecurDyn RFIfile can be generated during a particular NX Nastran execution, but not bothfiles at the same time. Attempting to generate both files in the same analysiswill cause an error to be issued and the execution to be terminated.

18. The ADMOUT=YES option is used when you would like results recovery(using the ADMRECVR case control entry) from an ADAMS/Flex analysis.This option requires the following assignment command:

ASSIGN OUTPUT2=’name.out’ STATUS=UNKNOWN UNIT=20FORM=UNFORM

inserted into the file management section of the NX Nastran input file. It willcause an OP2 file with a .out extension to be generated, which then can beused as input into an NX Nastran SOL 103 run using the ADMRECVR casecontrol capability to perform results recovery from an ADAMS/Flex analysis.FLEXBODY=YES is required with its use.

The data blocks output are:

MGGEW - physical mass external sort with weight mass removedMAAEW - modal massKAAE - modal stiffnessCMODEXT - component modes.

This capability is limited to one superelement per NX Nastran model.Residual-only analyses are not supported.

19. Setting CHECK=YES (which is only available when ADMOUT=YES) is notrecommended for models of realistic size due to the amount of data that willbe written to the f06.

20. The ADAMSMNF data routines use the environment variable TMPDIR fortemporary storage during the processing of mode shape data. As a result,TMPDIR must be defined when using ADAMSMNF. TMPDIR should equateto a directory string for temporary disk storage, preferably one with a largeamount of free space.

ADMRECVR

ADAMS stress recovery.

Recovers stress results from an ADAMS/Flex analysis.

4-20 NX Nastran 5.0 Release Guide

Page 91: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

FORMAT:

EXAMPLES:ASSIGN INPUTT2=’adams_results.mdf’ UNIT=13

...

CEND

STRESS(PLOT)=100

ADMRECVR

DESCRIBERS:

Describer Meaning

ADMFORM Specifies the format of the ADAMS/Flex modal deformationsfile (see remarks 1 and 2)

BINARY OUTPUT2 file (default)

ASCII PUNCH file.

MSRMODE Specifies stress recovery type (see remarks 6 and 7).

0 Component modal definitions are stored in an OUTPUT2 file(specifically, a *.out file was created by using ADMOUT=YESon the ADAMSMNF case control in a pre-ADAMS/Flex NXNastran run). The OUTPUT2 files used in this case do notcontain data blocks used for MNF creation (default).

1 Same as option 0, except that the OUTPUT2 file will contain10 additional data blocks used for MNF creation by anADAMS pre-processor (specifically, a *.out file created throughuse of the mnfx.alt DMAP alter capability).

2 No file reference (specifically, component modal definitionswill be recomputed)

RGBODY Requests the addition of rigid body motion with modaldeformations (see remark 5).

NO Do not include rigid body motion (default).

NX Nastran 5.0 Release Guide 4-21

Page 92: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

Describer Meaning

YES Include rigid body motion.

MSGLVL Level of diagnostic output from Lanczos eigensolver whencomponent modal definitions are determined (applies onlywhen MSRMODE=2).

0 No output (default).

1 Warning and fatal messages.

2 Summary output.

3 Detailed output on cost and convergence.

4 Detailed output on orthogonalization.

ADMCHK Requests debug output be written to the f06 file (See Remark9).

NO No debug output will be written (default).

YES Debug output will be written.

REMARKS:1. When modal deformations to be read are in binary

(OUTPUT2) format (specifically, ADMFORM=BINARY), thefollowing statement needs to be specified near the top ofthe NX Nastran input file in the file management section:ASSIGN INPUTT2=’<MDFilename>’ UNIT=13where <MDFilename> is the name of the modal deformationsfile from ADAMS.

2. To input the modal deformations file from ADAMS in ASCII (Punch)format (specifically, ADMFORM=ASCII), the followingstatement needs to be included in the bulk data section:INCLUDE ’<MDFilename>’where ’<MDFilename>’ is the name of the modal deformations file.

3. Dynamic stress/strain output can either be in .f06, PUNCH, and/or OUTPUT2according to standard NX Nastran functionality. However, stress recovery inNX Nastran from ADAMS/Flex results does not support XYPLOT output.

4. If displacements, stresses, and/or strains are to be available for postprocessing, one or more of the following statements must appearin the case control section of the NX Nastran input file:DISP(PLOT) = <set id>STRAIN(FIBER,PLOT) = <set id>STRESS(PLOT) = <set id>

5. Rigid body motions from an ADAMS simulation are included in the modaldeformations file, but they are not applied unless the RGBODY keyword is setto YES and the SORT1 option is included in the DISP(PLOT) command in case

4-22 NX Nastran 5.0 Release Guide

Page 93: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

control. Including rigid body motion affects the display and animation of theflexible component, but it has no effect on dynamic stresses.

6. For MSRMODE=0 or 1, stress recovery references the OUTPUT2 fileobtained from the initial CMS analysis (specifically, ADMOUT=YES on theADAMSMNF case control entry or use of the mnfx.alt DMAP alter capability).No other files are required. The geometric data needs to be included inthe bulk data of the NX Nastran input file because geometry is missingfrom the OUTPUT2 file. This mode of stress recovery is faster than theMSRMODE=2 mode. To reference this OUTPUT2 file the following line needsto be included in the file specification section of the NX Nastran input file:ASSIGN INPUTT2=’<OUTPUT2_filename>’ UNIT=20

7. For MSRMODE=2, no files are referenced for stress recovery. Instead, a fullCMS reanalysis is performed to build the reference data for the stress recoveryanalysis. Obviously, the analysis time is significantly far greater compared tothe MSRMODE=0 or 1 method, but this method frees up disk space. There isalso risk in using this method. If the reanalysis generates slightly differentcomponent eigenvalues or eigenvectors than were generated during thecreation of the ADAMS MNF in the initial NX Nastran run, then the ADAMSresults in the ADAMS MDF (modal deformation file) will be inconsistent andincorrect results will be recovered. Something as simple as a sign change forone eigenvector will cause incorrect results to be recovered. It is, therefore,highly recommended that MSRMODE=0 or 1 always be used.

8. This capability must be performed in SOL 103 and is limited to onesuperelement per NX Nastran model. Residual-only analyses are notsupported.

9. Setting ADMCHK=YES is not recommended for models of realistic size due tothe amount of data that will be written to the f06.

4.3 Component Mode Reduction of the Residual StructureProcedure for Flexbody SolutionsThe set-up of a flex body modal solution in NX Nastran for export to ADAMS MNF or RecurdynRFI files requires special considerations for the modal solution. This is because flex bodieswill be attached to other components in the multi-body dynamic (MBD) simulation and localflexibility effects at the connection locations are thus important.

In a standard normal mode solution, the solved modes give good representation of the globaldynamics of the component, but the local stiffness effects at the connections are typically notcaptured because of modal truncation. A modal solution method called Component ModeReduction of Residual Structure (CMR of RS) is recommended for flex body solutions because itincludes both global and local effects.

The CMR of RS method is a variation to the standard SOL 103 modal solution in NX Nastran. Ithas long been employed by dynamic analysts to include the local stiffness effects in critical areasof a model. The CMR of RS method is a two-step modal solution approach and is equivalent to aCraig-Bampton superelement reduction on the residual structure.

NX Nastran 5.0 Release Guide 4-23

Page 94: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

CMR of RS Theory

The connection degrees-of-freedom (DOF) for the flex body are equivalent to the exterior DOF ofa superelement. In Nastran set terminology, this means the connection DOF are in the t-set.The interior DOF of the flex body component are part of the o-set. The union of the t-set ando-set is the f-set.

The first solution step of the CMR or RS method is to compute normal modes and staticconstraint modes. The normal modes are computed with the specified t-set DOF restrained. Thenormal mode shape matrix, , can be used to transform the physical displacement uf (and itspartitions ut and uo) into the generalized modal DOF, q. The transformation relation is

uu

uqf

t

o

=⎧⎨⎩

⎫⎬⎭

=⎡

⎣⎢

⎦⎥ 0

Φ

Equation 4-1.

The static constraint modes are static deflection shapes computed by applying unit deflectionsat the t-set DOFs. This is the same thing as performing a Guyan reduction on the o-set DOF(see chapter 11 in the NX Nastran Basic Dynamic Analysis User’s Guide for more information).The static constraint modes transform full physical displacement uf into just the physical DOF,ut. The transformation relation is

uu

u

I

K Kuf

t

o oo ott =

⎧⎨⎩

⎫⎬⎭

=−

⎣⎢

⎦⎥ −1

Equation 4-2.

The matrices Koo and Kot are partitions of the Kff stiffness matrix. The normal modetransformation is a good reduction for the global dynamics and the constraint modetransformation is a good reduction for the local stiffness at the connection DOF. Thesetransformations can be combined into a single transformation as

uu

u

I

K K

u

qft

o oo

t =⎧⎨⎩

⎫⎬⎭

=−

⎣⎢

⎦⎥

⎧⎨⎩

⎫⎬⎭

−1

0

Φ

Equation 4-3.

The combined transformation captures both the global dynamics and the local stiffness of thecomponent in the reduction.

In the second solution step, the transformation in equation (3) is used to mathematically reducethe model from the full physical DOF set to a reduced set of generalized modal DOFs and thet-set physical DOFs. The reduced DOF set size is equal to the sum of the number of normalmodes and number of connection DOFs. Since the normal modes and static modes are notorthogonal to each other, the reduced mass and stiffness matrices are not diagonal as they wouldbe with a pure normal mode reduction. However, a second modal solution is performed on thereduced system resulting in a new set of modes.

The final modes are orthogonal to each other and importantly, capture both global dynamicsand local stiffness characterizations of the flex body. It is this second set of modes shapes thatare exported to the flex body file. It should be noted that it is critical that all the modes of

4-24 NX Nastran 5.0 Release Guide

Page 95: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

the reduced system should be solved in the second modal solution so that full global and localeffects embodied in the reduction are retained.

CMR of RS Specification

The CMR of RS method can be set-up in an NX Nastran solution by adding a few additionalsolution control cards to a standard SOL 103 solution. In the case control, an RSMETHOD cardis needed in addition to the standard METHOD card. The purpose of each is:

RSMETHOD Identifies modal solution controls cards (EIGR or EIGRL) to be used for the firstmodal solution. Typically the control cards define a frequency range to computethe modes. Modes outside that range are neglected.

METHOD Identifies modal solution controls cards (EIGR or EIGRL) to be used for the secondmodal solution. This is the modal solution of the reduced system. It is importantto solve for all the modes of the reduced system. This can be done by specifying awide frequency range on the EIGR or EIGRL cards.

The bulk data section of the input file needs definition of the t-set and q-set DOF. These aredefined by these cards:

ASET Defines the connection DOFs (these are in the t-set).SPOINT Creates additional DOFs.QSET Identifies SPOINT DOFs as modal reduction DOFs and places them in the q-set.

For the first modal solution (RSMETHOD), it is recommended that the solved frequency range beat least twice the frequency range of interest in the MBD solution. For example, if the MBDsolution is to be accurate to 1500 Hz, the first modal solution of the components should capturemodes to at least 3000 Hz. The user needs to be sure that the number of q-set DOF created byQSET and SPOINT cards is larger than the actual number of modes computed from RSMETHOD.

As mentioned already, all the modes of the reduced system need to be computed for the secondmodal solution (METHOD). The user can verify that all the modes have been computed byconfirming that the number of modes is equal to the reduced DOF set size (the sum of thenumber of normal modes and number of connection DOFs).

Mixed Boundary Reduction

Another variation of the CMR of RS method is to use a mixed boundary reduction. In thestandard solution, all the t-set DOF are restrained in the first eigenvalue solution. With mixedboundary conditions, the user can designate that some of the t-set DOF will be unrestrained inthis solution. All t-set DOF are treated the same in the constraint mode solution.

The reason for using the mixed boundary approach is to add exterior DOF to the solution thatare more appropriately considered as a free DOF rather than a connection DOF. For example if acomponent will have four nodes used as connection points in the MBD analysis, those pointsshould be restrained exterior points. If there are other points on the component that will not beconnections but may be a marker location to track response, or to apply a force, those nodes maybest be treated as an unrestrained exterior point.

Additional bulk data cards for the mixed boundary approach are:

CSE Defines the exterior DOFs that will be unrestrained in the first eigenvaluesolution.

NX Nastran 5.0 Release Guide 4-25

Page 96: NX Nastran 5 Release Guide

Chapter 4 Multi-body Dynamic Software Interfaces

BSET Defines the exterior DOFs that will be restrained in the first eigenvalue solution.BSET and ASET are treated the same in this case.

Example

Below is a sample input file set-up for doing a CMR of RS solution for a Recurdyne flexible body.

$SOL 103CEND$$ CASE CONTROL$TITLE = Solution 1ECHO = NONERECURDYNRFI,FLEXBODY=yes, FLEXONLY=no $ Creates RFI fileRSMETH=100 $ Points to 1st modal extraction methodMETHOD = 200 $ Points to 2nd modal extraction method$DISPLACEMENT(PLOT) = ALLSTRESS(PLOT,CORNER) = ALL$$ BULK DATA$BEGIN BULKDTI,UNITS,1,KG,N,MM,S $ Units used in MBD solve$$ SOLUTION CARDS$EIGRL, 100, , 400.00 $ Eigenvalue extraction method for 1st modesEIGR, 200, AHOU, , , , 150 $ Eigenvalue extraction method for 2nd modesSPOINT,100001,thru,100101 $ SPOINTS > 1st ModesQSET1,,100001,thru,100101 $ QSET points same as SPOINTSPARAM AUTOSPC YESPARAM GRDPNT 0PARAM POST -2$$ Flex body connection ASET pointsASET1,123456,26ASET1,123456,35ASET1,123456,44ASET1,123456,53ASET1,123456,62ASET1,123456,71ASET1,123456,80ASET1,123456,89$$ Rest of model ….$..

Case Control Section

RECURDYNRFI,FLEXBODY=yes,FLEXONLY=no• Used to request calculation of the flex body.

RSMETH = 100• This card points to the EIGRL card used for the first modal solution

METHOD = 200• This card points to the EIGR card used for the second modal solution

4-26 NX Nastran 5.0 Release Guide

Page 97: NX Nastran 5 Release Guide

Multi-body Dynamic Software Interfaces

Bulk Data Section

DTI,UNITS,1,KG,N,MM,S• This defines the set of units used and is required by the MBD solver

EIGRL, 100, , 400.00• This is the extraction method for the step-1 modal solve. Recall that you want theupper frequency limit to be about twice the maximum frequency of interest in thestructure

EIGR, 200, AHOU, , , , 150• This is the extraction method for the step-2 modes• The Automatic Householder option is recommended since it will assure that you getall the modes of the reduced component• ND (or 150 in this example), is the desired number of eigenvectors, this number hasto be greater that the number of ASET plus the number of step-1 modes• In this example, there are 8 ASET points, each with 6 DOF, and about 74 step-1modes,so we need 122 step-2 modes (8 x 6 + 74, it’s OK to request more)

SPOINT,100001,thru,100101• SPOINTs are scalar DOFs and describe the modal degrees of freedom, you need moreSPOINTs than expected step-1 modes, it doesn’t hurt to have too many• The numbering range has to be outside the range of grid points

QSET1,,100001,thru,100101• The QSET card puts the SPOINTs in the q-set• Use the same range as the SPOINTs

ASET1,123456,grid number• The ASET identifies the connection grids and the DOF at those grids. In this caseall 6 DOF are used as connection DOF, which assumes that all will be connected byfixed joints in the MBD model

NX Nastran 5.0 Release Guide 4-27

Page 98: NX Nastran 5 Release Guide
Page 99: NX Nastran 5 Release Guide

Chapter

5 New Optimization Option

NX Nastran 5.0 Release Guide 5-1

Page 100: NX Nastran 5 Release Guide

Chapter 5 New Optimization Option

NX Nastran 5 introduces a new optimizer option, UGS-ADS. The newly introduced optimizer isbased on the code ADS which was developed by Dr. Garrett N. Vanderplaats around 1985, andincludes modifications to the original code. The original form of the code is described in NASAContractor Report 177985, published in September 1985.

With the addition of UGS-ADS into NX Nastran, there are now two choices for an optimizer:

• DOT, the original optimizer, which remains the default for NX Nastran 5

• UGS-ADS, the new optimizer

5.1 Input RequirementsOnly two points need to be noted to use UGS-ADS as the optimizer of choice with NX Nastran 5:

1. In NX Nastran 5, UGS-ADS supports only the modified feasible directions as the optimizationmethod. Any other user selection with UGS-ADS will automatically revert to the modifiedfeasible directions method. Other optimization methods will be supported in future versions.

2. UGS-ADS can be selected by assigning the value 1 to the NASTRAN system cell 425(ADSOPT). In this case, the NASTRAN statement would look like:

NASTRAN ADSOPT = 1

-or-

NASTRAN SYSTEM(425) = 1

Otherwise, the system cell 425 (ADSOPT) will remain as 0 (default), and the optimizationwill proceed using DOT.

Everything else related to optimization remains the same as in previous versions of NX Nastran.

5-2 NX Nastran 5.0 Release Guide

Page 101: NX Nastran 5 Release Guide

Chapter

6 Advanced Nonlinear

NX Nastran 5.0 Release Guide 6-1

Page 102: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

6.1 Contact Improvements

Contact Small Displacement Option

A typical nonlinear solution can iterate many times as the geometry and material conditionsare updated. The contact conditions also update as the geometry conditions change, which addssolution time. There may be occasions when preventing the contact conditions from updatingbetween nonlinear iterations will not adversely affect the nonlinear results, for example, whendisplacements are small in the region where contact occurs. The CTDISP option has beencreated on the NXSTRAT bulk entry to prevent contact conditions from updating. The CTDISPoptions are as follows:

0 or 2 – Use large displacement formulation (contact conditions are updated) (default)

1 – Use small displacement formulation (contact conditions are not updated)

CTDISP is a global option since it applies to all contact definitions in the model. If you would liketo prevent or allow a specific contact set from updating, use the DISP option on the BCTPARAbulk entry. The DISP options are as follows:

0 – Use the formulation selected by CTDISP in NXSTRAT entry (default)

1 – Use small displacement formulation (contact conditions are not updated)

2 – Use large displacement formulation (contact conditions are updated)

Contact Stability Option with Rigid Body Motion

Rigid body motion in a contact solution can add instability making convergence difficult. A newoption on the NXSTRAT bulk entry, CTDAMP, has been created to stabilize the portions of themodel experiencing rigid body motion, thus helping the solution to continue and converge.

The CTDAMP options on the NXSTRAT bulk entry are as follows:

0 – No stabilization damping is applied (default).

1 – Stabilization damping is applied at the first time step only. The specified damping coefficientsare applied and ramped down to zero by the end of the first time step.

2 – The specified stabilization damping coefficients are applied at all time steps.

Contact Initial Gap and Penetration

The INIPENE parameter on the BCTPARA bulk entry is used when the goal is for a pair ofcontact regions to be initially touching without interference, but due to the faceted nature offinite elements around curved geometry, some of the element faces may have a slight penetration.

A new option is available when INIPENE=3 which will include both penetration and nowgaps. This option is particularly beneficial when contact conditions are defined on the faces ofconcentric cylinders.

The INIPENE options on the BCTPARA bulk entry are as follows:

0 – Initial penetrations are eliminated.

1 – Initial penetrations are eliminated and the list of pentrating nodes is printed.

2 – Initial penetrations are ignored. In successive steps, each contractor node is allowed topenetrate the target up to its initial penetration.

6-2 NX Nastran 5.0 Release Guide

Page 103: NX Nastran 5 Release Guide

Advanced Nonlinear

3 – Initial penetrations or gaps are overridden by GAPVAL. This option is not available forrigid target algorithm (TYPE = 2).

Contact Restrictions Removed

In NX Nastran 4.1, a new contact segment option became available for SOL 601 which improvedcontact results. This improvement matched the order of the contact segments with the orderof the elements. When quadratic contact segments were created on quadratic elements, theresulting contact tractions were significantly better. The new parameter CSTYPE was created onthe bulk data entry NXSTRAT to support the new contact segment creation. CSTYPE=1 wasmore accurate and gave better contact traction results, especially when used with quadraticelements (for example, 10-grid tetrahedral elements).

Restrictions Eliminated in Version 5The following restrictions which existed in NX Nastran 4.1 when CSTYPE=1 was used havebeen eliminated in NX Nastran 5:

• Contact surface offset must be zero (default), that is, OFFSET=0.0 in BCTPARA

• Only frictionless contact (default) or regular Coulomb friction is allowed, that is,FRICMOD=0 or 1 in BCTPARA

• Consistent contact stiffness cannot be used, that is, CSTIFF=0 in BCTPARA

• Continuous normal must be used (default), that is, SEGNORM=0 or 1 on BCTPARA

• Initial penetration cannot be ignored (default), that is, INIPENE=0 or 1 in BCTPARA

Existing RestrictionsThese restrictions remain in NX Nastran 5 when CSTYPE=1 is used:

• Only constraint function algorithm (default) may be used, that is, TYPE=0 in BCTPARA

• Only single-sided contact (default) is allowed, that is, NSIDE=1 in BCTPARA

• No tied contact (default), that is, TIED=0 in BCTPARA

6.2 Improved Convergence for Contact with FrictionContact models with friction are difficult to converge at times. The contact algorithm has beenimproved in NX Nastran 5 to be more robust for contact with friction. The new improvedfriction algorithm is used by default, but the old algorithm can be selected by assigning the newFRICALG parameter on the NXSTRAT bulk entry to “1”.

In addition, a new friction delay option is available by setting the new FRICDLY parameter onthe BCTPARA bulk entry to "1". Delaying the application of friction may also improve theconvergence of the solution. The default for FRICDLY is "0" meaning no delay occurs. Theupdated NXSTRAT and BCTPARA bulk entries are provided below.

NXSTRAT

Strategy Parameters for SOLs 601 and 701

NX Nastran 5.0 Release Guide 6-3

Page 104: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Defines parameters for solution control and strategy in advanced nonlinearstructural analysis.

FORMAT:

1 2 3 4 5 6 7 8 9 10NXSTRAT ID Param1 Value1 Param2 Value2 Param3 Value3

Param4 Value4 Param5 Value5 -etc-

EXAMPLE:

NXSTRAT 1 AUTO 1 MAXITE 30 RTOL 0.005

ATSNEXT 3

FIELDS:

Field Contents

ID Identification number. Currently not used. (Integer ≥ 0)

PARAMi Name of the NXSTRAT parameter. Allowable names are given inTable 8-33. See remark 1 for parameters applicable to SOL 701.(Character)

VALUEi Value of the parameter. See Table 8-33. (Real or integer)

Table 6-1. NXSTRAT Parameters

Name Description

Analysis Control Parameters

SOLVER Selects the solver to use. (Integer; Default = 0)

0 – Direct sparse solver

1 – Multigrid solver

2 – 3D iterative solver. This solver is effective for models with large numbers of higher order 3D solidelements, i.e., CTETRA and CHEXA elements with mid-side nodes.

AUTO Indicates whether automatic incrementation scheme is enabled. (Integer; Default = 0)

0 – No automatic incrementation scheme is used

1 – Automatic time stepping (ATS) scheme is enabled

2 – Automatic load-displacement control (LDC) scheme is enabled

6-4 NX Nastran 5.0 Release Guide

Page 105: NX Nastran 5 Release Guide

Advanced Nonlinear

Table 6-1. NXSTRAT Parameters

Name Description

NPOSIT Indicates whether analysis continues when the system matrix is not positive definite. (Integer; Default= 0)

0 – Analysis may stop

1 – Analysis continues

Notes:

If NPOSIT=0, analysis stops unless the ATS or LDC scheme is enabled; i.e. AUTO=1 or AUTO=2contact analysis is being performed

It is not recommended to set NPOSIT=1 for a linear analysis.

MASSTYP Selects the type of mass matrix to be used in dynamic analysis. (Integer; Default = 1)

0 – Lumped mass is used

1 – Consistent mass is used

Analysis Options

TINT* Integration order for the local t-direction (through thickness) of shell elements with elasto-plasticmaterials. By default, 5-point Newton-Cotes is used for single-layered shell and 3-point Newton-Cotes isused for multi-layered shell. Note that 2-point Gauss integration is always used for all shell elementswith elastic materials. (Integer; Default = 0).

1 ≤ TINT ≤ 6 – Gauss integration method with integration order TINT

-3, -5, -7 – Newton-Cotes integration with order -TINT

ICMODE* Indicates whether incompatible modes are used for 4-node shell elements. (Integer; Default =1 forSOL 601 and 0 for SOL 701)

0 - Incompatible modes are not used

1 - Incompatible modes are used

MSTAB Indicates whether the stiffness matrix stabilization feature is used. (Integer; Default = 0)

0 – Matrix stabilization is not used

1 – Matrix stabilization is used

MSFAC Matrix stabilization factor. (Real; Default = 1.0E-12)

DTDELAY* Element death time delay. (Real; Default = 0.0)

When an element is too deformed and becomes “dead”, its contribution to the overall stiffness of thestructure is removed. By specifying DTDELAY > 0.0, the contribution from the element stiffness isgradually reduced to zero over time DTDELAY instead of being suddenly removed. This may help in theconvergence of the solution.

SDOFANG* Angle used to determine whether a shell mid-surface node is assigned 5 or 6 degrees of freedom.(Real; Default = 5.0)

NX Nastran 5.0 Release Guide 6-5

Page 106: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Table 6-1. NXSTRAT Parameters

Name Description

UPFORM* Indicates whether u/p formulation is used for elements. Note that u/p formulation is always used forhyperelastic elements and always not used for hyperfoam elements and elastic elements with Poisson’sratio less than 0.48. It is also not used for gasket elements (Integer; Default = 0)

0 – u/p formulation is not used

1 – u/p formulation is used instead of displacement-based formulation

ULFORM* Indicates which large strain formulation is used for 4-node shell elements. (Integer; Default = 0)

0 – Updated Lagrangian-Jaumann (ULJ) formulation is used if rigid-target algorithm contact is used orSOL 701 is run. Otherwise, Updated Lagrangian-Hencky (ULH) formulation is used.

1 – Use ULH formulation

2 – Use ULJ formulation

Note: For shell elements which are not 4-noded, the ULJ formulation is always used for large strains.For 3-D solid, plane strain and axisymmetric elements, ULH formulation is always used for largestrains. In restarting from SOL 601 to 701 or vice versa, ULFORM needs to be specified such that bothanalyses use the same formulation.

DISPOPT Indicates whether prescribed displacements are applied to the original configuration or the deformedconfiguration. This option is only applicable for a restart analysis or when a delay (or arrival) time isspecified for the prescribed displacement. (Integer; Default = 0)

0 – Applied to original configuration

1 – Applied to deformed configuration

LOADOPT* Indicates whether prescribed loads (pressure and centrifugal) are deformation-dependent, i.e. thedirection and magnitude of the load may change due to large deformation of the structure. This option isonly applicable for large displacement analysis, i.e. PARAM,LGDISP,1 (Integer; Default = 1)

0 – Load is independent of structural deformation

1 – Load is affected by structural deformation

MAXDISP Specifies a limit for the maximum displacement that is allowed for any node during the analysis. Thisfeature is generally useful for contact analysis where rigid body motion exists in a model. A value of 0.0means there is no limit on displacements. (Real ≥ 0.0; Default = 0.0)

Time Integration

TINTEG Selects the time integration method to be used for nonlinear transient analysis. (Integer; Default = 0)

0 – Use the Newmark method

1 – Use the ADINA composite method

ALPHA Alpha coefficient for the Newmark time integration method. (Real; Default = 0.25)

DELTA Delta coefficient for the Newmark method. (Real; Default = 0.5)

SOL 701 Time Stepping

6-6 NX Nastran 5.0 Release Guide

Page 107: NX Nastran 5 Release Guide

Advanced Nonlinear

Table 6-1. NXSTRAT Parameters

Name Description

XSTEPo Selects time step method used in an explicit time integration analysis. (Integer; Default = 0)

0 - Time step size is calucated by the program based on the critical time step size. The data in theselected TSTEP bulk data entry is used to calculate the total solution time for the analysis.

1 - The number of time steps and the time step size as specified in the selected TSTEP bulk dataentry is used.

XDTCALo Calculation of the critical time step size may be computationally expensive. This parameter specifiesthat the critical time step size be re-calculated every XDTCAL time steps. (Integer > 0, Default = 1)

XDTFACo The critical time step size is calculated based on certain assumptions. It is often necessary, especially fornonlinear analysis, to use a time step size smaller than the calculated critical time step size. The factormultiplied by the calculated critical time step size gives the time step size used in the analysis. (4.0> Real > 0.0, Default = 0.9)

XMSCALEo Specifies the factor to scale the mass (densities) of the entire model (at the beginning of the analysis) toincrease the critical time step size required for stability when the explicit time integration scheme isused. See warning in Remark 2. (Real = 1.0, Default = 1.0)

XDTMIN1o The minimum time step size used to determine if mass scaling will be applied to elements (at thebeginning of the analysis) whose critical time step size is smaller than DTMIN1. The amount of massscaling is calculated for each element so that the critical time step size is equal to DTMIN1. See Remark2 and warning in Remark 3. (Real = 0.0, Default = 0.0)

XDTMIN2o The minimum time step size used to determine whether an element will be removed in an explicit timeintegration analysis. In explicit time integration, the smaller an element size is, the smaller will thecritical time step size be. If the critical time step size for an element is smaller than XDTMIN2, theelement will be removed in the analysis. See Remark 2 and warning in Remark 3. (Real > 0.0, Default= 0.0)

Multigrid Solver

ITEMAX Maximum number of iterations allowed for the multigrid solver to converge. (Integer > 0; Default = 1000)

EPSIA Convergence tolerance EPSIA. (Real; Default = 1.0E-6)

EPSIB Convergence tolerance EPSIB. (Real; Default = 1.0E-4)

EPSII Convergence tolerance EPSII. (Real; Default = 1.0E-8)

Equilibrium Iteration and Convergence

LSEARCH Flag to indicate the use of line searches within the iteration scheme. (Integer; Default = 0)

0– Line search is not used

1 – Line search is used

LSLOWER Lower bound for line search. (0.0 ≤ Real < 1.0; Default = 0.001)

LSUPPER Upper bound for line search. (1.0 ≤ Real; Default = 2.0)

MAXITE Maximum number of iterations within a time step. If the maximum number of iterations is reachedwithout achieving convergence (see CONVCRI parameter), the program will stop unless the automatictime stepping (ATS) or load displacement control scheme is selected (see parameter AUTO). (1 ≤Integer ≤ 999; Default = 15)

NX Nastran 5.0 Release Guide 6-7

Page 108: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Table 6-1. NXSTRAT Parameters

Name Description

CONVCRI Convergence Criteria. (Integer; Default = 0)0 – Convergence based on energy1 – Convergence based onenergy and force2 – Convergence based on energy and displacement3 – Convergence based on force4 –Convergence based on displacement

ETOL Relative energy tolerance. (Real; Default = 0.001)

RTOL Relative force (and moment) tolerance (Real; Default = 0.01)

RNORM Reference force. (Real)

RMNORM Reference moment. (Real)

RCTOL Relative contact force tolerance. (Real; Default = 0.05)

DTOL Relative displacement (translation and rotation) tolerance. (Real; Default = 0.01)

DNORM Reference translation. (Real)

DMNORM Reference rotation. (Real)

STOL Line search convergence tolerance. (Real; Default = 0.5)

RCONSM Reference contact force. (Real; Default = 0.01)

ENLSTH Line search energy threshold. (Real; Default = 0.0)

Automatic Time Stepping (ATS) Scheme

ATSSUBD Number that limits the smallest time step size when the automatic time stepping (ATS) scheme is used.For a time step size of DT, the program will stop if convergence is not achieved and the next subdividedtime step size is less than DT/ATSSUBD. (Integer ≥ 1; Default = 10)

ATSMXDT Factor that limits the maximum time step size when the automatic time stepping (ATS) scheme is used.The ATS scheme may increase the time step size after convergence is achieved. However, for a time stepsize of DT, the program will not use a time step size greater than ATSMXDT * DT. (Real; Default = 3.0)

ATSNEXT Flag controls what time step size to use once convergence is reached after an ATS subdivision. (Integer;Default = 0)

0 – Automatically set by program. For contact analysis, ATSNEXT = 2, otherwise ATSNEXT = 1.

1 – Use the time step size that gave convergence, i.e., the reduced time step that led to convergence isused again.

2 – Return to the original time step size, i.e., the original time step size before any subdivision tookplace is used.

3 – Use a time step size such that the solution time matches the original solution time specified by theuser.

ATSDFAC Division factor used calculate the sub-increment time step size. If current time step size is DT andconvergence is not achieved, the next time step size will be DT/ATSDFAC. (Real > 1.0; Default = 2.0)

ATSLOWS Flag whether a low-speed dynamics analysis is performed instead of a static analysis. (Integer; Default= 0)0 – Low-speed dynamics option is not activated 1 – Low-speed dynamics is performed

ATSDAMP Damping factor used in low-speed dynamics analysis. (Real ≥ 0.0; Default = 1.0e-4)

Load Displacement Control (LDC) Scheme

6-8 NX Nastran 5.0 Release Guide

Page 109: NX Nastran 5 Release Guide

Advanced Nonlinear

Table 6-1. NXSTRAT Parameters

Name Description

LDCGRID Grid point id at which a displacement is prescribed for the first solution step. (Integer > 0)

LDCDOF Degree of freedom for prescribed displacement at grid point LDCGRID. (1≤ Integer ≤ 6)

1 – X translation

2 – Y translation

3 – Z translation

4 – X rotation

5 – Y rotation

6 – Z rotation

LDCDISP Prescribed displacement at grid point LDCGRID for the first solution step. (Real)

LDCIMAX Displacement convergence factor used to limit the maximum incremental displacement during asolution step. (Real; Default = 3.0)

LDCDMAX Maximum (absolute magnitude) displacement (for the degree of freedom specified by LDCDOF) atthe grid point LDCGRID allowed during the analysis. When the displacement reaches or exceedsLDCDMAX, the program will stop the analysis. See Section 6.2.4 in Advanced Nonlinear Theory andModeling Guide for other criteria that determines when an LDC solution will stop. (Real)

LDCCONT Flag whether the solution is terminated when the first critical point on the equilibrium path is reached.(Integer; Default = 0)

0 – Solution stops

1 – Solution continues

LDCSUBD Maximum number of arc length subdivisions allowed. (Integer ≥ 1; Default = 10)

Contact Control

IMPACT Impact control scheme (Integer; Default = 0)

0 – No special treatment is applied for impact problems

1 – Post impact adjustment of velocities and accelerations is applied

2 – Modified parameters are used in Newmark time integration scheme

NSUPP Number of iterations for pairing contactor node to target segment. If NSUPP > 0, during the firstNSUPP iterations, the pairing target segment is recorded for each contactor node. From iterationNSUPP+1, if a target segment in the recorded list is repeated, it is “frozen” to be the pairing targetsegment for the remaining equilibrium iterations in that time step. Specifying NSUPP > 0 may help inthe convergence for certain problems. (0 ≤ Integer ≤ 99; Default = 0)

RTSUBD Selects the subdivision scheme used in the implicit rigid-target contact algorithm when the tensilecontact force is too large. (Integer; Default = 0)

0 – Subdivision is based on the magnitude of the tensile contact force, i.e., the larger the magnitude,the smaller will be the subdivided time step size.

1 – Subdivision is based on the global automatic time stepping (ATS) subdivsion settings.

NX Nastran 5.0 Release Guide 6-9

Page 110: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Table 6-1. NXSTRAT Parameters

Name Description

CSTYPE Selects the type of contact segment to use. (Integer; Default = 1)

0 – Use linear contact segment

1 – Use element-based contact segment which gives better contact traction results

CTDISP Selects the default displacement formulation used for contact analysis. A different formulation may beselected for each individual contact set via BCTPARA entry. (Integer; Default = 2)

1 – Use small displacement formulation

2 – Use large displacement formulation

RTALG Selects the rigid-target algorithm to use. (Integer; Default = 0)

0 – Use the current algorithm

1 – Use the old (v4) algorithm

FRICALG Selects the friction algorithm to use. (Integer; Default = 0)

0 – Use the current algorithm

1 – Use the old (v4) algorithm

CTDAMP Indicates whether stabilization damping is applied and how it is applied for contact analysis. Thisfeature is generally useful when rigid body motion exists in a model. (Integer; Default = 0)

0 – No stabilization damping is applied

1 – Stabilization damping is applied at the first time step only. The specified damping coefficients areapplied and ramped down to zero by the end of the first time step.

2 – The specified stabilization damping coefficients are applied at all time steps.

CTDAMPN Specifies the normal stabilization damping coefficient. (Real ≥ 0.0, Default = 0.0)

CTDAMPT Specifies the tangential stabilization damping coefficient. (Real ≥ 0.0, Default = 0.0)

Restart Options

MODEX* Indicates the mode of execution. (Integer; Default = 0)

0 – Normal analysis run, i.e. not a restart analysis

1 – Restart analysis

The restart (.res) file from a previous run must exist to do a restart analysis. The filename and locationof the restart file is determined by the “dbs” keyword. By default, dbs points to the current workingdirectory with the prefix of the current job name. Note that keyword scratch=no must be used whenrunning a restart analysis.

TSTART* Solution starting time. If MODEX=1, TSTART must equal a solution time in which data was saved in aprevious run. If TSTART = 0.0, the last time step in the restart file is used. (Real, Default = 0.0)

6-10 NX Nastran 5.0 Release Guide

Page 111: NX Nastran 5 Release Guide

Advanced Nonlinear

Table 6-1. NXSTRAT Parameters

Name Description

IRINT* Frequency of saving the analysis results in the restart file. (Integer; Default = 0)

0 – IRINT is set to 1 when implicit time integration is used and set to the number of steps in the firsttime step block when explicit time integration is used.

> 0 – Restart file is overwritten every IRINT time steps

< 0 – Restart file is appended every IRINT time steps

Other Parameters

NSUBGRP* Number of sub-groups to divide large number of elements with same property ID into. Normally,elements with same type and property ID are placed into a group. If a group contains more than 1000elements and NSUBGRP > 1, the elements are placed into NSUBGRP sub-groups for more efficientprocessing. (Integer > 0; Default = 1)

ELRESCS Indicates the coordinate system used for output of nonlinear 3D element stress/strain results. For linearresults, the coordinate system used for output is specified by CORDM in PSOLID entry. (Integer;Default = 0)

0 – Results are output in element coordinate system

1 – Results are output material coordinate system

Translation Options

ELCV Convert 8-node to 9-node quadrilateral (plane strain, axisymmetric and shell) elements and 20-nodeto 27-node brick elements. Note that this also converts 6-node to 7-node triangular (plane strain andaxisymmetric) elements and 10-node to 11-node tetrahedral elements. (Integer; Default = 0)

0 – No conversion of elements

1 – Convert elements as described above; nodal coincidence is not checked against existing nodes andnew generated nodes are always created.

2 – Convert elements as described above; nodal coincidence is checked against existing nodes and a newnode will not be created at a location if a node already exist at that location.

EQRBAR Indicates how RBAR elements are handled. (Integer; Default = 0)

0 – RBAR is simulated using rigid option in small displacement analysis and using flexible option inlarge displacement analysis.

1 – RBAR is simulated using rigid option (i.e. simulated by rigid link or constraint equations asdetermined by program)

2 – RBAR is simulated using flexible option (i.e. simulated by spring or beam elements as determinedby program)

3 – RBAR is simulated by spring elements

See Section 2.7 of Advanced Nonlinear Theory and Modeling Guide for details on how RBAR elementsare handled.

NX Nastran 5.0 Release Guide 6-11

Page 112: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Table 6-1. NXSTRAT Parameters

Name Description

EQRBE2 Indicates how RBE2 elements are handled. (Integer; Default = 0)

0 – RBE2 is simulated using rigid option in small displacement analysis and using flexible option inlarge displacement analysis.

1 – RBE2 is simulated using rigid option (i.e. simulated by rigid links or constraint equations asdetermined by program)

2 – RBE2 is simulated using flexible option (i.e. simulated by spring or beam elements as determinedby program)

3 – RBE2 is simulated by spring elements

See Section 2.7 of Advanced Nonlinear Theory and Modeling Guide for details on how RBE2 elementsare handled.

SPRINGK Stiffness of spring elements that simulate RBAR or RBE2 elements. (Real, Default = 0.0)

If SPRINGK = 0.0, program automatically sets SPRINGK according to the following calculations.

SPRINGK = EMAX * LMODEL

where EMAX = maximum Young’s Modulus of materials in the model and LMODEL = largest dimensionof the model. If no material is specified in the model, EMAX is set to 1.0E12.

BEAME Young’s Modulus of material assigned to beam elements that simulate RBAR or RBE2 elements.(Real, Default = 0.0)

If BEAME = 0.0, BEAME is set to EMAX * 100.0 where EMAX = maximum Young’s Modulus ofmaterials in the model. If no material is specified in the model, EMAX is set to 1.0E12.

BEAMA Circular cross section area of beam elements that simulate RBAR or RBE2 elements. (Real, Default= 0.0)

If BEAMA = 0.0, program automatically sets BEAMA according to the following calculation:

BEAMA = (LMODEL * .01)2 where LMODEL = largest dimension of the model

RBLCRIT Critical length for determining how RBAR and RBE2 elements are simulated when the rigid or flexibleoption is used to simulate RBAR (see EQRBAR) and RBE2 (see EQRBE2). (Real, Default = 0.0)

If RBLCRIT = 0.0, then

if EQRBAR (or EQRBE2) = 1,

RBLCRIT = LMODEL * 1.0E-6

if EQRBAR (or EQRBE2) = 2,

RBLCRIT = LMODEL * 1.0E-3

REMARKS:1. Parameters applicable to SOL 701 are:

• XSTEP, XDTCAL, XDTFAC, XMSCALE, XDTMIN1 and XDTMIN2 areonly used for SOL 701. These parameters are indicated in the table witha superscript ‘o’.

6-12 NX Nastran 5.0 Release Guide

Page 113: NX Nastran 5 Release Guide

Advanced Nonlinear

• TINT, ICMODE, DTDELAY, SDOFANG, UPFORM, LOADOPT, MODEX,TSTART, IRINT and NSUBGRP parameters can be used for SOL 701.These parameters are indicated in the table with a superscript‘*’.

2. XMSCALE, XDTMIN1 and XDTMIN2 may be used together. XDTMIN1 andDTMIN2 are applied after XMSCALE is applied. If XDTMIN1 and XDTMIN2are both used, XDTMIN1 should be greater than XDTMIN2. If XDTMIN2 =XDTMIN1 is specified, XDTMIN1 will be ignored.

3. WARNING: Specifying XMSCALE > 1.0, XDTMIN1 > 0.0 or XDTMIN2 > 0.0may change the model significantly. Hence, extra care should be exercised inexamining the results when any of these parameters are used.

BCTPARA

Contact Set Parameters (SOLs 601 and 701 only)

Defines parameters for a contact set.

FORMAT:

1 2 3 4 5 6 7 8 9 10

BCTPARA CSID Param1 Value1 Param2 Value2 Param3 Value3

Param4 Value4 Param5 Value5 -etc-

EXAMPLE:

BCTPARA 2 EPSN 1.0E-10 TZPENE 0.1 CSTIFF 1

FIELDS:

Field Contents

CSID Contact set ID. Parameters defined in this command apply tocontact set CSID defined by a BCTSET entry. (Integer > 0)

PARAMi Name of the BCTPARA parameter. Allowable names are given inthe parameter listing below. (Character)

VALUEi Value of the parameter. See Table 6-2 for parameter listing. (Realor Integer)

Table 6-2. BCTPARA Parameters for SOL 601

Name Description

General Parameters

NX Nastran 5.0 Release Guide 6-13

Page 114: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Table 6-2. BCTPARA Parameters for SOL 601

Name Description

TYPE Selects the type of contact algorithm. (Integer; Default=0)

0 – Constraint function algorithm is used.

1 – Segment method algorithm is used.

2 – Rigid target algorithm is used.

NSIDE Flags single or double-sided contact. (Integer; Default=1)

1 – Contact surfaces are single-sided.

2 – Contact surfaces are double-sided.

TBIRTH Birth time for contact set. (Real; Default=0.0)

TDEATH Death time for contact set. (Real; Default=0.0) If TDEATH ≤TBIRTH, it is ignored.

INIPENE Flags how initial penetrations or gaps are handled. (Integer;Default=0)

0 – Initial penetrations are eliminated.

1 – Initial penetrations are eliminated and the list of pentratingnodes is printed.

2 – Initial penetrations are ignored. In successive steps, eachcontractor node is allowed to penetrate the target up to its initialpenetration.

3 – Initial penetrations or gaps are overridden by GAPVAL. Thisoption is not available for rigid target algorithm (TYPE = 2).

GAPVAL Specifies a constant gap distance between the source region(contactor) and the target region when INIPENE = 3. NegativeGAPVAL means initial penetrations which will be eliminated.(Real, Default = 0.0)

PDEPTH Penetration depth for single-sided contact (i.e. NSIDE=1). (Real;Default=0.0)If PDEPTH > 0.0, then penetration is detected.When penetration ≤ PDEPTH, and if penetration > PDEPTH,penetration is deemed not to occur.

SEGNORM Indicates whether a continuous (interpolated) contact segmentnormal is used for the contact surfaces. (Integer; Default=0)

0 – SEGNORM=1 if NSIDE=1, SEGNORM=-1 if NSIDE=2.

1 – Continuous segment normal is used.

-1 – Continuous segment normal is not used.

6-14 NX Nastran 5.0 Release Guide

Page 115: NX Nastran 5 Release Guide

Advanced Nonlinear

Table 6-2. BCTPARA Parameters for SOL 601

Name Description

OFFTYPE Type of offset for contact regions. (Integer; Default=0)

0 – User specified offset value only for single-sided contact (i.e.,NSIDE=1). Use offset of 0.001 for double-sided contact (i.e.,NSIDE=2)

1 – Use specified offset value for either single- or double-sidedcontact.

2 – Half the shell thickness is used for contact regions on shellelements and no offset is used otherwise.

OFFSET Default offset distance for contact regions. (Real; Default=0.0)

Note: For contact algorithm TYPE=0 or 1, individual offsetdistances can be specified for each contact region using theBCRPARA entry to override the default offset distance specifiedhere.

Standard contact algorithm (TYPE=0 or 1)

DISP Selects the displacement formulation used for this contact set.(Integer; Default = 0)

0 – Use the default formulation selected by CTDISP in NXSTRATentry

1 – Use small displacement formulation

2 – Use large displacement formulation

TZPENE Time to eliminate initial penetrations. (Real ≥0.0; Default=0.0) IfTZPENE=0.0 and INIPENE=0 or 1, then the initial penetrationsare eliminated in the first time step. This may cause convergencedifficulties for certain problems. By using TZPENE > 0.0, theinitial penetrations are eliminated gradually over time TZPENE.

CSTIFF Indicates whether consistent contact stiffness is used. (Integer;Default=0)

0 – Consistent contact stiffness is not used

1 – Consistent contact stiffness is used

TIED Indicates whether contact regions in each contact pair are tiedtogether. Currently, tied contact option assumes small rotationsof the contact regions. (Integer; Default=0)

0 – Not tied

1 – Tied

NX Nastran 5.0 Release Guide 6-15

Page 116: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Table 6-2. BCTPARA Parameters for SOL 601

Name Description

TIEDTOL Tolerance used to determine whether contactor nodes are tied tothe target region when TIED=1 is specified. A contactor node istied to its target region if the distance between them is less thanor equal to TIEDTOL. (Real; Default=0.0)

EXTFAC Factor for extending contact surfaces beyond their boundaries.The amount of extension is given by this factor multipliedby the length of the contact segments. (1.0E-6 ≤ Real ≤ 0.1;Default=0.001)

FRICMOD Type of friction model (0 ≤ Integer ≤ 13; Default=0). See Remark 1.

0 – Constant coefficient of friction specified for each contact pair,i.e. FRICi in the BCTSET entry.

1 – Constant coefficient of friction specified by FPARA1.

2 – Friction model 1; uses FPARA1 and FPARA2.

3 – Friction model 2; uses FPARA1, FPARA2, and FPARA3.

4 – Use different static and dynamic friction coefficients; usesFPARA1, FPARA2, and FPARA3.

5 – Friction coefficient varies with sliding velocity; uses FPARA1,FPARA2 and FPARA3.

6 – Anisotropic friction model; uses FPARA1, FPARA2, FPARA3,FPARA4, and FPARA5.

7 – Friction coefficient varies with consistent contact force; usesFPARA1 and FPARA2.

8 – Friction coefficient varies with time; uses FPARA1, FPARA2and FPARA3.

9 – Friction coefficient varies with coordinate values; usesFPARA1, FPARA2, FPARA3, FPARA4, and FPARA5.

12 – Modified friction model 1; uses FPARA1 and FPARA2.

13 – Modified friction model 2; uses FPARA1, FPARA2, andFPARA3.

See Section 4.4 of the Advanced Nonlinear Theory and ModelingGuide for a description of friction models.

FPARA1 Friction parameter A1

FPARA2 Friction parameter A2

FPARA3 Friction parameter A3

6-16 NX Nastran 5.0 Release Guide

Page 117: NX Nastran 5 Release Guide

Advanced Nonlinear

Table 6-2. BCTPARA Parameters for SOL 601

Name Description

FPARA4 Friction parameter A4

FPARA5 Friction parameter A5

FRICDLY Indicates whether the application of friction is delayed, i.e.,friction is applied on a node one time step after the node comesinto contact. Delay of friction application may improve theconvergence of the solution. (Integer; Default = 0)

0 – No delay

1 – Delay

Constraint Function Contact Algorithm (Type=0)

EPSN Parameter for normal constraint function, w. (Real;Default=1.0E-12)

EPST Parameter for frictional constraint function, v. (Real > 0.0;Default=0.001)

CFACTOR1 Compliance factor. (Real; Default=0.0)

Rigid Target Algorithm (TYPE=2)

NCMOD Normal contact modulus. (Real; Default=1.0E11)

Parameters used with current rigid-target algorithm (RTALG=0 in NXSTRAT)

TFORCE The maximum tensile contact force allowed for a convergedsolution. (Real ≥ 0.0, Default = 0.001)

SLIDVEL The maximum sliding velocity used in modeling sticking friction.When the velocity is smaller than SLIDVEL, sticking is assumed;when the velocity is larger than SLIDVEL, sliding is assumed.(Real > 0.0, Default = 1.0E-10)

OCHECK Specifies whether oscillation checking is performed and when it isdone. (Integer >= 0; Default = 5)

If OCHECK=0, no oscillation checking is performed. Otherwise,oscillation checking is performed after equilibrium iterationOCHECK. Oscillation checking consists of two checks:

a) If a contactor node oscillates between two neighboring targetsegments during the equilibrium iterations, oscillation checkingputs the contactor node into contact with the boundary edgebetween the target segments.

NX Nastran 5.0 Release Guide 6-17

Page 118: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Table 6-2. BCTPARA Parameters for SOL 601

Name Description

b) In analysis with friction, if the sliding velocity of a contactornode oscillates during the equilibrium iterations, oscillationchecking puts the contactor node into sticking contact.

GAPBIAS Contact is detected when the distance between the target andcontactor (accounting for any offsets) is less than GAPBIAS.(Real; Default = 0.0)

OFFDET Selects the implementation of offsets. (Integer; Default = 0) 0 –Program chooses the implementation based upon the shape of thetarget surfaces; if a target surface is flat or convex, spheres areused, otherwise, normals are used.

1 – A sphere of radius equal to the offset is placed around eachcontactor node, and contact is detected between the sphere andthe target surface.

2 – Two surfaces are constructed for each contactor surface: anupper surface and a lower surface. These surfaces are constructedusing the offsets and the averaged contactor normals. Contactis then detected between points on the constructed contactorsurfaces and target surface.

Parameters used with old (NX Nastran Version 4) rigid-target algorithm(RTALG=1 in NXSTRAT) which may be removed in a future release.

PENETOL Penetration tolerance which gives the maximum penetrationallowed into a rigid target surface. (Real; Default=1.0E-8)

TCMOD Tangential contact modulus. (Real; Default=0.0)

RFORCE Minimum tensile contact force required to change the state ofa contact node from "node in contact" to "free node", i.e., if thenormal tensile force is greater than RFORCE, a "node in contact"becomes a "free node". (Real; Default=0.001)

LFORCE Limit (maximum) for the sum of all contact forces for nodeschanging from the state of "node in contact" to "free node". If theabsolute value of the sum of the forces is bigger than LFORCE,then the automatic time stepping (ATS) method will be activatedto subdivide the current time step into smaller time increments.(Real; Default=1.0)

6-18 NX Nastran 5.0 Release Guide

Page 119: NX Nastran 5 Release Guide

Advanced Nonlinear

Table 6-2. BCTPARA Parameters for SOL 601

Name Description

RTPCHECK Specifies whether penetration is checked (in addition to checkingthe tensile contact force) against the maximum allowablepenetration when the rigid-target algorithm is used. (Integer;Default=0)

0 – Penetration is not checked. Note that with this setting thereis a possibility that the rigid target surface may penetrate thecontactor surface excessively.

1 – Penetration is checked and subdivision of the time step occursif the penetration exceeds (RTPMAX * maximum model length.)

2 – Penetration is checked and subdivision of time step occurs ifthe penetration exceeds RTPMAX.

RTPMAX Specifies the maximum allowable penetration when the rigidtarget algorithm is used. RTPMAX is either a factor of themodel size or an absolute value depending on the RTPCHECKparameter. (Real > 0.0; Default=0.001)

Table 6-3. BCTPARA Parameters for SOL 701

Name Description

General Parameters

XTYPE Selects the type of contact algorithm. (Integer; Default=0)

0 – Kinematic constraint algorithm is used.

1 – Penalty algorithm is used.

3 – Rigid target algorithm is used.

NSIDE Flags single or double-sided contact. (Integer; Default=1)

1 – Contact surfaces are single-sided.

2 – Contact surfaces are double-sided.

TBIRTH Birth time for contact set. (Real; Default=0.0)

TDEATH Death time for contact set. (Real; Default=0.0) If TDEATH ≤TBIRTH, it is ignored.

NX Nastran 5.0 Release Guide 6-19

Page 120: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

Table 6-3. BCTPARA Parameters for SOL 701

Name Description

INIPENE Flags how initial penetrations are handled. (Integer; Default=0)

0 – Initial penetrations are eliminated.

1 – Initial penetrations are eliminated and the list of penetratingnodes is printed.

2 – Initial penetrations are ignored. In successive steps, eachcontactor node is allowed to penetrate the target up to its initialpenetration.

PDEPTH Penetration depth for single-sided contact (i.e. NSIDE=1). (Real;Default=0.0)If PDEPTH > 0.0, then penetration is detected.When penetration ≤ PDEPTH, and if penetration > PDEPTH,penetration is deemed not to occur.

OFFTYPE Type of offset for contact regions. (Integer; Default=0)

0 – Use specified offset value only for single-sided contact (i.e.,NSIDE=1). Use offset of 0.001 for double-sided contact (i.e.,NSIDE=2).

1 – Use specified offset value for either single- or double-sidedcontact.

2 – Half the shell thickness is used for contact regions on shellelements and no offset is used otherwise.

OFFSET Default offset distance for contact regions. (Real;Default=0.0)Note: For contact algorithm XTYPE=0 or 1,individual offset distances can be specified for each contact regionusing the BCRPARA entry to override the default offset distancespecified here.

Contact Algorithm XTYPE=0 or 1

TZPENE Time to eliminate initial penetrations. (Real=0.0; Default=0.0)IfTZPENE=0.0 and INIPENE=0 or 1, then the initial penetrationsare eliminated in the first time step. This may cause convergencedifficulties for certain problems. By using TZPENE > 0.0, theinitial penetrations are eliminated gradually over time TZPENE.

EXTFAC Factor for extending contact surfaces beyond their boundaries.The amount of extension is given by this factor multipliedby the length of the contact segments. (1.0E-6 ≤ Real ≤ 0.1;Default=0.001)

Penalty Contact Algorithm (XTYPE=1)

6-20 NX Nastran 5.0 Release Guide

Page 121: NX Nastran 5 Release Guide

Advanced Nonlinear

Table 6-3. BCTPARA Parameters for SOL 701

Name Description

XKNCRIT Selects the criterion for evaluation of normal penalty stiffness.(Integer; Default=0)

0 - The program calculates the normal penalty stiffness.

1 - The user specifies the normal penalty stiffness (XKN).

XKN Specifies the normal penalty stiffness to be used whenXKNCRIT=1. (Real)

XKTCRIT Selects the criterion for evaluation of tangential penalty stiffness.(Integer; Default=0)

0 - The program calculates the normal penalty stiffness

1 - The user specifies the normal penalty stiffness (XKT).

XKT Specifies the tangential penalty stiffness to be used whenXKTCRIT=1. (Real)

XDAMP Indicates whether damping is used when the penalty explicitcontact algorithm is used. (Integer; Default=0)

0 – Damping is not used, i.e., the XNDAMP parameter is ignored.

1 – Damping is used and XNDAMP is a factor of the criticaldamping, i.e., the damping coefficient is given by XNDAMPmultiplied by the critical damping. This is the recommendedchoice if damping is used.

2 – Damping is included and the damping coefficient is specifieddirectly by XNDAMP.

XNDAMP Specifies the relative or absolute damping coefficient (for normalpenalty stiffness) when the penalty explicit contact algorithm isused and XDAMP=1 or 2. (Real=0.0; Default=0.1)

Rigid Target Algorithm (XTYPE=3)

PENETOL Penetration tolerance which gives the maximum penetrationallowed into a rigid target surface. (Real; Default=1.0E-8)

TCMOD Tangential contact modulus. (Real; Default=0.0)

REMARKS:1. Multiple BCTPARA with the same CSID can be used. The parameters

specified in all BCTPARA with the same CSID will be combined and assignedto the contact set.

2. If duplicate parameters are specified, the last parameter value will be used.

NX Nastran 5.0 Release Guide 6-21

Page 122: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

3. Field 9 on line 1 should be blank. Beginning on the continuation lines, fields2 to 9 can be used for 4 pairs of PARAM/VALUE

6.3 Surface-to-Surface Glue SupportThe option to “glue” element faces together during a 601 solution is available in NX Nastran5. The glue option connects predefined surfaces together and prevents relative motion in alldirections. Predefined regions of element free faces are used to detect where the glue elementsare created.

The regions where the solver creates glue elements are defined using the BSURFS bulk entry. Aregion is a collection of solid element free faces in a section of the model where you expect gluingto occur. The BSURFS entry is defined by its own unique ID and is a list of solid element IDseach followed by 3 grid points defining which face of the 3D element to include in the glue region.

Once regions are created, they must be paired. A glue pair is a way to combine two regions,source and target, in which gluing is analyzed during the solution. The BGSET bulk entry isused to define each glue pair. The GSID field needs to match the value of ‘n’ on the BGSETcase control entry for the solution to recognize this glue definition. The SIDi and TIDi fieldsrefer to regions created by the BSURFS entry, and are used to define source and target regionsrespectively for a pair. The search distance field (SDIST) and penalty (PEN) factor on the BGSETbulk entry are not used by SOL 601. SOL 601 uses a search distance that is equal to the largestelement edge in the source and target regions.

The BGSET case control command, the BGSET bulk entry, and the BSURFS bulk entry areprovided below.

BGSET (Case Control)

Glue Contact Set Selection

Selects the glue contact set.

FORMAT:BGSET=n

EXAMPLES:BGSET=5

DESCRIBERS:

Describer Meaning

n Glue contact set identification of a BGSET Bulk Data entry.(Integer>0)

BGSET (Bulk Entry)

3D Glue Contact Set Definition

Defines glued contact pairs of a 3D set.

6-22 NX Nastran 5.0 Release Guide

Page 123: NX Nastran 5 Release Guide

Advanced Nonlinear

FORMAT:

1 2 3 4 5 6 7 8 9 10BGSET GSID SID1 TID1 SDIST1 PEN

SID2 TID2 SDIST2

-etc-

EXAMPLE:

BGSET 4 1 2

4 3

FIELDS:

Field Contents

GSID Glue set identification number. (Integer > 0).

SIDi Source region identification number for glue pair i. (Integer > 0)

TIDi Target region identification number for glue pair i. (Integer > 0)

SDISTi Search distance for glue surfaces (Real); (Default=100)

PEN Penalty value used in calculations. The value defined on the firstline will be used for all pairs. (Real);(Default=1.0E5)

REMARKS:1. The default search distance will glue all overlapping sections of the source

and target regions. This value can be used in special cases to control whatsubregions are to be glued.

2. The default penalty factor will be sufficient for most cases. Increase it ifseparation of the surfaces is observed. If defined too large, numerical problemsmay occur.

REMARKSRELATED TO

SOL 601:1. SDISTi and PEN are not used by SOL 601. SOL 601 uses a search distance

that is equal to the largest element edge in the source and target region.

BSURFS

3D Contact Region Definition by Solid Elements (SOL 101, 601 and 701 only)

Defines a 3D contact region by the faces of the CHEXA, CPENTA or CTETRAelements.

FORMAT:

1 2 3 4 5 6 7 8 9 10BSURFS ID EID1 G1 G2 G3

EID2 G1 G2 G3 EID3 G1 G2 G3

NX Nastran 5.0 Release Guide 6-23

Page 124: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

1 2 3 4 5 6 7 8 9 10

-etc-

EXAMPLE:

BSURFS 7 10 15 16 20

15 16 17 21 20 17 18 22

11 19 20 24 16 20 21 25

21 21 22 26

FIELDS:

Field Contents

ID Identification number of a contact region. See Remarks 2 and 3.(Integer > 0)

EIDi Element identification numbers of solid elements. (Integer > 0)

G1 - G3 Identification numbers of 3 corner grid points on the face(triangular or quadrilateral) of the solid element. (Integer > 0)

REMARKS:1. The continuation field is optional.

2. BSURFS is a collection of one or more element faces on solid elements.BSURFS defines a contact region which may act as a contact source (contactor)or target.

3. The ID must be unique with respect to all other BSURFS, BSURF, andBCPROP entries.

6.4 Element Birth and DeathAn option to define element birth and death times for a specific set of elements is available in NXNastran 5. The new EBDSET case control command selects an EBDSET bulk data entry to beused in the solution. The EBDSET bulk data entry includes the element IDs along with theirbirth and death times. Another bulk data entry, EBDADD, can be used to combine EBDSET bulkentries such that several of them can be selected using a single, new set ID.

All of these new commands needed for the element birth/death option are included below.

EBDSET

Element Birth/Death Set Selection

Selects the element birth/death set (SOLs 601 and 701)FORMAT:

EBDSET=nEXAMPLES:

EBDSET=5

6-24 NX Nastran 5.0 Release Guide

Page 125: NX Nastran 5 Release Guide

Advanced Nonlinear

DESCRIBERS:

Describer Meaningn Element birth/death set identification of an EBDSET or

EBDADD Bulk Data entry. (Integer>0)

EBDSET

Element Birth/Death Set Definition (SOLs 601 and 701)

Defines element birth and death times for a set of elements.

FORMAT:

1 2 3 4 5 6 7 8 9 10

EBDSET BDID TBIRTH TDEATH EID1 EID2 EID3 EID4 EID5

CONTINUATIONFORMAT 1:

EID6 EID7 EID8 -etc-

CONTINUATIONFORMAT 2

(“THRU”OPTION IS

ONLYAVAILABLE

ON ACONTINUATION

LINE):

EID6 “THRU” EID7 “BY” INC

EXAMPLE:

EBDSET 3 0.2 0.5 101 201 300 400 450

25 THRU 33

FIELDS:

Field Contents

BDID Element birth/death set identification number. See Remark 2. (Integer> 0)

TBIRTHElement birth time. (Real ≥ 0.0; Default = 0.0)

TDEATHElement death time. (Real > TBIRTH; Default = 1.0E+20)

EIDi Element identification numbers. See Remark 3. (Integer > 0)

NX Nastran 5.0 Release Guide 6-25

Page 126: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

REMARKS:1. The continuation line is optional.

2. The “THRU” option is only available on a continuation line.

3. BDID may be selected by Case Control command EBDSET. If other elementbirth/death sets are defined, the EBDADD entry must be used to combine allthe EBDSET entries.

4. Element birth/death may be used with all elements, i.e., with all elementsdefined with CROD, CONROD, CBAR, CBEAM, CQUAD, CQUAD4, CQUAD8,CQUADR, CQUADX, CTRIA3, CTRIA6, CTRIAR, CTRIAX, CHEXA,CPENTA, CTETRA, CELAS1, CELAS2, CDAMP1, CDAMP2, CMASS1,CMASS2, CGAP, or CBUSH1D entry.

EBDADD

Element Birth/Death Set Combination (SOLs 601 and 701)

Defines an element birth/death set as a union of element birth/death sets definedon EBDSET entries.

FORMAT:

1 2 3 4 5 6 7 8 9 10

EBDADD BDID BD1 BD2 -etc-

EXAMPLE:

EBDADD 10 1 2 3

FIELDS:

Field Contents

BDID Element birth/death set identification number. (Integer > 0)

BDi Identification numbers of element birth/death sets defined via EBDSETentries. (Integer > 0)

REMARKS:1. To include several element birth/death sets defined via EBDSET entries in a

model, EBDADD must be used to combine the element birth/death sets. BDIDin EBDADD is then selected with the Case Control command EBDSET.

2. BDi must be unique and may not be the identification of this or any otherEBDADD entry.

6.5 Bolt Preload CapabilityWhen components of an assembly are bolted together, a specified torque translates into anaxial bolt preload. Bolts should be properly preloaded in this way before service conditions are

6-26 NX Nastran 5.0 Release Guide

Page 127: NX Nastran 5 Release Guide

Advanced Nonlinear

applied to the assembly. When analyzing preloaded bolts, you may be interested in obtainingthe stresses due to the preload condition alone, or due to a combination of the bolt preload andservice load. You can manually determine the preloaded bolt condition by using equivalentthermal loads, although using this method is approximate and typically requires many solutioniterations when multiple bolts exist.

NX Nastran 5 offers an automated method to simplify this multisolution process. A bolt analysiscan be selected by including the new case control command BOLTLD=n, where "n" selects thenew BOLTFOR bulk data entry containing the bolt preload value. The bolts should be modeledas beam elements, and included in the new BOLT bulk entry. A single beam may be usedto represent a bolt.

The new BOLTLD case control command, and the new BOLTFOR and BOLT bulk entries areprovided below. The bolt preload capability is not supported in SOL 701.

New BOLTLD Case Control Command

Bolt Load Set Selection

Selects the BOLTFOR Bulk Data entry for bolt pre-load processing.

FORMAT:BOLTLD=n

EXAMPLES:BOLTLD=5

DESCRIBERS:

Describer Meaning

n Selects the BOLTFOR Bulk Data entry for bolt pre-loadprocessing. (Integer>0)

REMARKS:1. Bolt pre-loading is supported in SOLs 101, 103, 105, 107 through 112 and 601.

New BOLTFOR Bulk Entry

Preload Force on Set of Bolts

Defines preload force on a set of bolts.

FORMAT:

1 2 3 4 5 6 7 8 9 10

BOLTFOR SID LOAD B1 B2 B3 B4 B5 B6

B7 THRU B8

B9 B10 -etc-

EXAMPLE:

BOLTFOR 4 1000.0 12 THRU 21

1 4 6 9 10 26 32 34

NX Nastran 5.0 Release Guide 6-27

Page 128: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

37 43 51

FIELDS:

Field Contents

SID Bolt preload force set identification number. (Integer>0)

LOAD Magnitude of the preload force. (Real)

Bi Bolt identification numbers defined by bulk entry BOLT (Integer>0),or using “THRU” (B7<B8 for THRU option).

REMARKS:1. Multiple BOLTFOR entries with the same SID can be used and the data will

be combined.

2. SID is selected by BOLTLD Case Control command.

3. Entering the same bolt id multiple times for the same SID will produce anerror.

BOLT

Bolt definition

Selects the CBEAM or CBAR elements to be included in the bolt pre-loadcalculation.

FORMAT:

1 2 3 4 5 6 7 8 9 10

BOLT BID ETYPE EID1 EID2 EID3 etc.

EID7 “THRU” EID8

-etc-

ALTERNATEFORMAT:

1 2 3 4 5 6 7 8 9 10

BOLT BID ETYPE EID1 “THRU” EID2

EXAMPLE 1

SOLs 101, 103, 105, 107 through 112 Example (single EID defines bolt):

BOLT 4 1 11

EXAMPLE 2

SOL 601 Example (all EIDs required to define bolt):

BOLT 4 1 11 8 2 1 20 14

15 16 28 30 33

6-28 NX Nastran 5.0 Release Guide

Page 129: NX Nastran 5 Release Guide

Advanced Nonlinear

FIELDS:

Field Contents

BID Bolt identification number. (Integer>0)

ETYPE Element type (Integer=1 is required in this field until other optionsbecome available in the future).

EIDi Element identification numbers of CBEAM or CBAR elementsto include in the bolt pre-load calculation.(Integer>0), or using“THRU” (EID7<EID8 for THRU option).

REMARKS:1. One BOLT entry is required to define each physical bolt.

2. Bolt preload is supported in SOLs 101, 103, 105, 107 through 112 and 601.

3. In a SOL 101, 103, 105, 107 through 112, only one EID is required to definea bolt even if it was modeled as several elements. Any additional entrieswill be ignored.

4. In a SOL 105, both the bolt preload and service load will be scaled to determinethe buckling load.

REMARKSRELATED TO

SOL 601:1. All CBEAM or CBAR elements representing a physical bolt must be identified

in the BOLT entry.

6.6 Shell Thickness Result RecoveryA new shell thickness result output option is available when using advanced nonlinear solutionsto solve large strain problems. The new case control command SHELLTHK is needed to requestthe shell thickness results. Since this capability is only supported for large strain solutions, youmust include PARAM,LGSTRN,1 in the input file if thickness results are requested. The newSHELLTHK case control command is included below.

SHELLTHK Case Control Command

Shell Thickness Output Request

Requests the form of shell thickness output (SOLs 601 and 701).

FORMAT:

EXAMPLES:SHELLTHK(PLOT)=ALL

NX Nastran 5.0 Release Guide 6-29

Page 130: NX Nastran 5 Release Guide

Chapter 6 Advanced Nonlinear

DESCRIBERS:

Describer Meaning

PRINT The printer will be the output medium.

PLOT Computes and puts shell thickness results in op2 file only.

ALL Shell thickness results will be output for all applicable shellelements. See remark 1.

NONE Shell thickness results will not be output.

REMARKS:1. Shell thickness results are output only for large strain analysis, i.e.,

PARAM,LGSTRN,1.

2. Shell thickness results are output at nodes on elements.

6.7 Iterative Solution ImprovementA new 3D-iterative solution option is now available in NX Nastran 5 for SOL 601 to efficientlysolve large models containing mainly higher order 3D solid elements (e.g., 10-node CTETRA,20-node CHEXA, and so on). The new iterative solver is invoked when SOLVER=2 in theNXSTRAT bulk entry. In addition to the higher order 3D solid elements, models can contain theother advanced nonlinear supported element types (for example, shells,rods, beams, rebars, andso on), and contact conditions. The SOL 601 3D-iterative solver option is effective in linear ornonlinear static analysis and in nonlinear dynamic analysis.

6-30 NX Nastran 5.0 Release Guide

Page 131: NX Nastran 5 Release Guide

Chapter

7 Linear Contact Enhancements

NX Nastran 5.0 Release Guide 7-1

Page 132: NX Nastran 5 Release Guide

Chapter 7 Linear Contact Enhancements

7.1 New Options to Improve Contact Solution AccuracyNX Nastran uses pre-defined source and target regions of element free faces to detect contactconditions in the model. More specifically, the solver uses the element faces from a source regionto project normals, then checks if these normals intersect with other element faces on a targetregion. A contact element is created if:

• NX Nastran finds an intersection between element faces, and

• The distance between the two faces is equal to or less than a distance that you specify.

The number of contact elements created and their distribution determines the accuracy of thecontact solution. Accuracy is particularly important when pressure results at the contactinterface are required.

Two new contact solution parameters, INTORD and REFINE, are now available in NX Nastran5 on the BCTPARM bulk entry to improve the accuracy of the contact solution. The updatedBCTPARM bulk entry is repeated below for convenience.

Previously, the number of locations where normals were projected (contact evaluation points)from the source region was dependent on the element type. For example, the linear triangle facewould always project a single normal, while the parabolic quad would project 4 normals.

In NX Nastran 5, the number of contact evaluation points is dependent on the value assigned tothe new INTORD parameter, and on the element face type. The following table summarizes howthe INTORD value adjusts the number of contact evaluation points for a particular element face:

Number of Contact Evaluation PointsFace Type INTORD=1 INTORD=2 (default) INTORD=3Linear Triangle 1 3 7Parabolic Triangle 3 7 12Linear Quad 1 4 9Parabolic Quad 4 9 16

INTORD allows for control over the number of contact elements created over a given elementface. Use of INTORD=2 (default) is recommended for increased accuracy with minimal increasein computational cost, where accuracy is defined as computing smoother stress and contactpressure results. This recommendation applies for flat and curved contact surfaces as wellas load path determination and detailed stress determination analysis scenarios. When theINTORD=1 is used, the number of contact evaluation points is the same as in previous releases.

The new REFINE parameter increases the number of contact evaluation points by refining themesh on the source region. Part of the refinement process is to project element edges and gridsfrom the associated target region back to the source region. The resulting refinement on thesource region is then more consistent with the target side, which then gives a better distributionof contact elements. The refined grids and elements are only used during the solution. Thecontact results are transferred back to the original mesh for post processing results.

The use of the contact region refinement option (REFINE=1) is primarily recommended foranalysis where detailed stress results are required since an increase in the computational costmay occur. Smoother results are obtained for flat contact regions versus curved. The refinementprocess provides the added benefit of bringing the target region accuracy closer to the accuracylevel of the source region. In general, the accuracy level of the source region results tends to begreater than that of the target region.

7-2 NX Nastran 5.0 Release Guide

Page 133: NX Nastran 5 Release Guide

Linear Contact Enhancements

By default, REFINE=1 and mesh refinement occurs. REFINE=0 will turn off the refinementcapability.

Since both INTORD and REFINE increase the number of contact elements, an increase insolution times may result.

Updated BCTPARM Bulk Entry

BCTPARM

Contact Parameters (SOLs 101, 103, 111 and 112).

Control parameters for the SOL 101 contact algorithm.

FORMAT:

1 2 3 4 5 6 7 8 9 10

BCTPARM CSID Param1 Value1 Param2 Value2 Param3 Value3

Param4 Value4 Param5 Value5 -etc-

EXAMPLE:

BCTPARM 1 PENN 1.0 CTOL 0.001

FIELDS:

Field Contents

CSID Contact set ID. Parameters defined in this command apply tocontact set CSID defined by a BCTSET entry. (Integer > 0)

PARAMi Name of the BCTPARM parameter. Allowable names are given inthe parameter listing below. (Character)

VALUEi Value of the parameter. See Table 7-1 for the parameter listing.(Real or Integer)

Table 7-1. BCTPARM Parameters

Name Description

PENN Penalty factor for normal direction. (Default=10.0) See Remark 2.

PENT Penalty factor for tangential direction. (Default=1.0) See Remark2.

CTOL Contact force convergence tolerance. (Default=0.01)

MAXF Maximum number of iterations for force a loop. (Default=10)

MAXS Maximum number of iterations for a status loop. (Default=20)

NCHG Allowable number of contact changes for convergence. (Default=0)

NX Nastran 5.0 Release Guide 7-3

Page 134: NX Nastran 5 Release Guide

Chapter 7 Linear Contact Enhancements

Table 7-1. BCTPARM Parameters

Name Description

MPER Minimum contact set percentage. (Default=100)

SHLTHK Shell thickness offset flag.

0 - Includes half shell thickness as surface offset. (Default)

1 - Does not include thickness offset.

RESET Flag to indicate if the contact status for a specific subcase is tostart from the final status of the previous subcase

0 - Starts from previous subcase. (Default)

1 - Starts from initial state.

AVGSTS Determines the averaging method for contact pressure/tractionresults.

0 - The averaging of Pressure/Traction values for a contact gridwill include the results from ALL contact elements attached tothe grid regardless of whether they are active or inactive in thecontact problem (Default).

1 - The averaging of the Pressure/Traction values for a contactgrid will exclude those contact elements which are not active inthe contact solution and thus have a zero Pressure/Traction value.

INIPENE Use when the goal is for a pair of contact regions to be initiallytouching without interference, but due to the faceted nature offinite elements around curved geometry, some of the element facesmay have a slight gap or penetration.

0 or 1 - Contact is evaluated exactly as geometry is modeled. Nocorrections will occur for gaps or penetrations (Default).

2 - Penetrations will be reset to a new initial condition in whichthere is no interference.

3 - Gaps and penetrations are both reset to a new initial conditionin which there is no interference.

Remarks:

See “Understanding the Contact Control Parameters Used inSOL 101 - BCTPARM” in the NX Nastran User’s Guide for moreinformation on the BCTPARM options.

7-4 NX Nastran 5.0 Release Guide

Page 135: NX Nastran 5 Release Guide

Linear Contact Enhancements

Table 7-1. BCTPARM Parameters

Name Description

REFINE Determines if the mesh on the source region is refined duringthe contact solution.

0 - Do not refine the contact source region based on target surfacedefinition.

1 - Refine the contact source region based on target surfacedefinition (default).

INTORD Determines the number of contact evaluation points for a singleelement face on the source region. The number of contactevaluation points is dependent on the value of INTORD, and onthe type of element face. See the table in Remark 1 for specificvalues.

1 – The reduced number of contact evaluation points is used.

2 - Use an increased number of contact evaluation points (default).

3 – Use a high number of contact evaluation points.

REMARKS:1. A higher number of contact evaluation points can be used to increase the

accuracy of a contact solution. Inaccuracies sometimes appear in the formof nonuniform contact pressure and stress results. There may be a penaltyassociated with using more evaluation points since the time for a contactproblem to converge may be longer. The table below shows how the number ofcontact evaluation points is dependent on the element type, and how it can beadjusted using the INTORD option. The “Face Type” column applies to shellelements, and to the solid element with the associated face type.

Number of Contact Evaluation PointsFace Type INTORD=1 INTORD=2 INTORD=3Linear Triangle 1 3 7Parabolic Triangle 3 7 12Linear Quad 1 4 9Parabolic Quad 4 9 16

2. Penalty factors have units of 1/(length). A physical interpretation is that it isequivalent to the axial stiffness of a rod (K=e*E*dA) with area dA, modulus E,and length 1/e. The defaults for the penalty factors generally work well, but inthe event that meshes have very large or very small edge lengths, adjustmentsmay be necessary. See the chapter Surface Contact for SOL 101 in the NXNastran User’s Guide for tips on adjusting penalty factors.

NX Nastran 5.0 Release Guide 7-5

Page 136: NX Nastran 5 Release Guide

Chapter 7 Linear Contact Enhancements

7.2 Support of Contact in SOLs 103, 111, and 112In NX Nastran 5, a contact definition can be included in a normal mode solution (SOL 103), andin an optional dynamic response calculation (SOLs 111 and 112).

In the normal mode solution, NX Nastran adds the contact stiffness result from the end ofthe converged linear statics contact solution. The contact stiffness values in the normal modesolution represents the final contact condition of the structure around the contact interface. Thus,it will appear that the resulting contact surfaces are attached during the normal mode analysis.

Since the calculated normal modes include the final contact interface conditions, the responsecalculation (SOLs 111 and 112) which uses these normal modes automatically includes thesame conditions.

The inputs for the normal mode solution are consistent with differential stiffness solutions whichrequire a linear statics subcase. The difference is that the linear statics subcase should includethe BCSET case control command. When defining the normal modes subcase, a STATSUBbulk entry must be included to reference the subcase id containing the contact definition. Thecontact solution in the linear statics subcase must fully converge before moving to the normalmode portion of the run.

7.3 Superelements with ContactIn NX Nastran 5, the use of superelements is now permitted in solution sequences which supportcontact (SOLs 101, 103, 111 and 112). The only requirement is that the contact definition mustoccur in the residual structure.

7.4 Shell Element Z-Offset with ContactShell elements such as the CQUAD4 can be offset relative to the mean plane of their connectedgrid points using the ZOFFS option. In NX Nastran 5, the linear contact solution can includethe shell element ZOFFS when evaluating the contact surfaces. The value assigned to the newZOFFSET option on the BCTPARM bulk entry determines if the contact solution recognizes theZOFFS value. By default, ZOFFSET=0 and the ZOFFS value is used when evaluating the contactsurfaces. When ZOFFSET=1, the ZOFFS value is not used when evaluating the contact surfaces.

7-6 NX Nastran 5.0 Release Guide

Page 137: NX Nastran 5 Release Guide

Chapter

8 Strength Ratio Outputfor Composites

NX Nastran 5.0 Release Guide 8-1

Page 138: NX Nastran 5 Release Guide

Chapter 8 Strength Ratio Output for Composites

8.1 Laminate Strength Ratio OutputStrength ratio is now output together with the failure index when using the PCOMP bulk entry.Strength ratio is a more direct indicator of failure than failure index since it demonstrates thepercentage of applied load to the failure criteria. Strength ratio is defined as:

Strength Ratio (SR) = Allowable Stress / Calculated Stress

For example, a SR = .75 not only indicates that a failure has occurred, but also indicates that theapplied load is 25% beyond the allowable. A FI = 1.25 on the other hand does not represent apercentage of failure; only that a failure condition exists.

NX Nastran provides strength ratio output for the same four commonly used definitions of thefailure surface: Hill’s Theory, Hoffman’s Theory, Tsai-Wu Theory, and maximum strain theory.Since the strength ratio calculation is based on the equations for failure index, both of theirdetails are presented here.

Failure indices assume a value of one on the periphery of a failure surface in stress space.

• If the failure index is < 1, the lamina stress is interior to the periphery of the failure surfaceand the lamina is assumed to be safe.

• If the failure index is > 1, the lamina stress is exterior to the periphery of the failure surfaceand the lamina is assumed to have failed.

These failure indices represent a phenomenological failure criterion in that only an occurrence ofa failure is indicated and not the mode of failure. In the present context, concern is with theanalytical definition of a failure surface in stress or strain space for use with laminae underbiaxial loading.

Selecting a Failure Criterion or Strength Ratio

If requested, NX Nastran will calculate a failure index and a strength ratio for each ply. Thisfailure index is obtained by considering the failure criteria for unidirectional fiber compositesas in the commonly used failure theories. You can select one of the following failure criteriafor composites:

• Hill’s Theory

• Hoffman’s Theory

• Tsai-Wu Theory

• Maximum Strain

Note: If you specify a failure theory (using the FT option in field 6 of the PCOMP entry), you mustalso specify the allowable shear stress of the bonding material (using the SB option in field 5).

In the analysis of isotropic materials, strength is independent of the orientation of the body underload and one may compare the largest computed principal stress with an allowable stress toestablish the integrity of the structure. Laminated composites, on the other hand, are orthotropicmaterials and may exhibit unequal properties in tension and compression. Thus, the strength ofthese orthotropic laminae is a function of body orientation relative to the imposed stresses.

8-2 NX Nastran 5.0 Release Guide

Page 139: NX Nastran 5 Release Guide

Strength Ratio Output for Composites

As the evaluation of the matrices of material moduli for laminated composites provides sufficientinformation to determine the actual stress field sustained by the material, the determinationof structural integrity will depend on the definition of an allowable stress field. The basicingredient of this definition is the establishment of a set of allowable stresses or strengths inthe principal material directions.

Xt = Allowable tensile stress in the principal x (or 1) direction of the material.

Xc = Allowable compressive stress in the principal x (or 1) direction of the material.

Yt = Allowable tensile stress in the principal y (or 2) direction of the material.

Yc = Allowable compressive stress in the principal y (or 2) direction of the material.

S = Allowable shear stress in the principal material system.

The strength ratio (SR) is calculated for each of the four failure theories by solving the quadraticequation with FI =1, and replacing the applied stress with (SR · applied stress). The SRcalculation detail is discussed below for each theory.

Hill’s Theory (HILL)

Hill’s failure theory for orthotropic materials that have the same strength in tension andcompression, i.e., xt = xc and yt = yc can be expressed as:

X is allowable stress in 1-directionY is allowable stress in 2-directionS is allowable stress in shear

and X = Xt if σ1 is positive or X = Xc if σ1 is negative and similarly for Y and σ2. For theinteraction term σ1σ2/X2, X = Xt if σ1σ2 is positive or X = Xc if σ1σ2 is negative.

A plot of the above equation obtained by setting the failure index to 1 on the σ1-σ2 plane yieldsan ellipse and is the anisotropic yield criterion of Hill (modified later by Tsai, and hence alsosometimes known as the Tsai-Hill theory). Therefore, if the failure index so calculated is lessthan 1, the ply stresses are inside the yield ellipse and the ply is said to be “safe”; conversely, if thefailure index is greater than 1, the ply stresses are outside the yield ellipse and the ply has failed.

Replacing the applied stress with (SR · applied stress), the Hill Failure Criteria can be rewrittenin terms of a strength ratio:

The above equation can be rearranged into quadratic equation format:

giving:

NX Nastran 5.0 Release Guide 8-3

Page 140: NX Nastran 5 Release Guide

Chapter 8 Strength Ratio Output for Composites

The value of a in the quadratic equation format matches the value for the Hill failure index (FI).

Substituting the values of a = FI, b = 0, and c = -1 into the general quadratic equation solution:

gives the Hill strength ratio:

Hoffman’s Failure Theory (HOFF)

Hoffman’s theory for an orthotropic lamina in a general state of plane stress with unequaltensile and compressive strengths is given by

The failure index is obtained by evaluating the left-hand side of the above equation.

Note that this theory takes into account the difference in tensile and compressive allowablestresses by using linear terms in the equation.

To calculate the strength ratio, the following terms are defined:

F1 =

F2 =

F11 =

F22 =

F66 =

Substituting above terms into Hoffman FI equation and setting FI = 1:

Replacing the applied stress with (SR · applied stress), the Hoffman Failure Criteria can berewritten in terms of a strength ratio:

8-4 NX Nastran 5.0 Release Guide

Page 141: NX Nastran 5 Release Guide

Strength Ratio Output for Composites

Rearranging into quadratic equation format:

Using the general quadratic equation solution:

The Hoffman strength ratio is calculated using the following roots:

Tensor Polynomial Theory of Tsai-Wu (TSAI)

The theory of strength for anisotropic materials proposed by Tsai and Wu specialized to the caseof an orthotropic lamina in a general state of plane stress with unequal tensile and compressivestrengths is

where:

F1 =

F2 =

F11 =

F22 =

F66 =

and F12 is to be evaluated experimentally.

The magnitude of F12 is, however, constrained by the following inequality called a “stabilitycriterion”:

The necessity of satisfying the stability criterion, together with the requirement that F12 bedetermined experimentally from a combined stress state, poses difficulties in the use of thistheory. Narayanaswami and Adelman (Narayanaswami, R., and H. M. Adelman, “Evaluation

NX Nastran 5.0 Release Guide 8-5

Page 142: NX Nastran 5 Release Guide

Chapter 8 Strength Ratio Output for Composites

of the Tensor Polynomial and Hoffman Strength Theories for Composite Materials,” Journal ofComposite Materials, Vol. II, 1977, p. 366.) have suggested that F12 be set to zero and that useof Hoffman’s theory or Tensor Polynomial theory with F12 = 0 is a preferred alternative to theexperimental determination of F12. If you have a value for use with F12 in the theory, you caninput that value in the MAT8 Bulk Data entry; otherwise, with this failure theory, NX Nastransets F12 to 0.0. The left-hand side of the above equation will be evaluated as the failure indexby this theory.

NX Nastran calculates the failure index of bonding material as the maximum interlaminar shearstress divided by the allowable bonding stress. NX Nastran writes the failure indices for all theplies into the OEFIT (Output Element failure index Table) output if stresses are requested. Thefailure index for the element is the largest value of the failure indices for all plies of the element.

Replacing the applied stress with (SR · applied stress), the Hoffman Failure Criteria can berewritten in terms of a strength ratio:

Rearranging into quadratic equation format:

Using the general quadratic equation solution:

The TSAI-Wu strength ratio is calculated using the following roots:

Maximum Strain Theory (STRN)

The midplane strains and curvatures are available in the element coordinate system. Fromthese, the stresses and strains in each individual lamina along the fiber direction and transversedirection can be easily calculated. You can use the STRAIN case control command to request theoutput of lamina strains.

The maximum strain criteria has no strain interaction terms. The strain allowables specifiedon the MAT8 entry for each lamina include

Xt,XcAllowable strains in tension and compression, respectively, in thelongitudinal direction.

Yt,YcAllowable strains in tension and compression, respectively, in thetransverse direction.

S Allowable strain for inplane shear.

8-6 NX Nastran 5.0 Release Guide

Page 143: NX Nastran 5 Release Guide

Strength Ratio Output for Composites

The failure index is calculated using

and

i.e., where 1, 2,γ12 are the elastic strains (totalstrains minus thermal strains).

Because you must understand which mode of failure index is critical; i.e., longitudinal (1),transverse (2), or shear (12), NX Nastran prints the mnemonic 1, 2 or 12 alongside the FPvalue to indicate the critical direction.

Note: There is no change in the way the failure index is calculated for interlaminar shear stresses.

When you use the maximum strain theory, you may want to specify lamina stress allowablesinstead of strain allowables on the MAT8. To do this, leave the STRN field on the MAT8 entryblank.

For this case, NX Nastran calculates the failure indices using

and

that is,

To calculate the Maximum Stress (Strain) strength ratio, the failure index which is defined asFI = Calculated Stress / Allowable Stress is set to unity, and the applied stress is replaced with(SR · applied stress). The result is:

NX Nastran 5.0 Release Guide 8-7

Page 144: NX Nastran 5 Release Guide

Chapter 8 Strength Ratio Output for Composites

Inputs for Strength Ratio Output

Since strength ratio is output together with failure index, the output requirements are thesame. The only exception is that the new SRCOMPS parameter has been created to turn on thestrength ratio output. By default, the parameter is set to “NO” and strength ratio is not output.

SRCOMPS Default = NO

SRCOMPS controls the computation and printout of ply strength ratios. IfSRCOMPS = YES, ply strength ratios are output for composite elements thathave failure indices requested.

See the PCOMP bulk entry in the QRG for information on the failure index and strength ratioinput requirements.

8-8 NX Nastran 5.0 Release Guide

Page 145: NX Nastran 5 Release Guide

Chapter

9 Miscellaneous Enhancements

NX Nastran 5.0 Release Guide 9-1

Page 146: NX Nastran 5 Release Guide

Chapter 9 Miscellaneous Enhancements

9.1 Thermal Expansion of Rigid ElementsThe option to include rigid elements in thermal expansion calculations is available in NXNastran 5. The new case control command, RIGID, should be assigned to “LAGRAN” to use thenew method. In addition, the coefficient of thermal expansion (ALPHA) has been added as anadditional entry on the rigid elements: RBAR, RBE1, RBE2, RROD, RTRPPLT. This capabilityis supported in solutions 101 through 112.

When the RIGID case control command is set to the default of “LINEAR”, the original linearelimination method is used. This method treats rigid elements as an MPC equation withoutthermal loading effects. The new RIGID case control command is duplicated below forconvenience.

New RIGID Case Control Command

RIGID

Rigid Element Method

Selects the rigid element processing method for RBAR, RBE1, RBE2, RROD andRTRPLT elements.

FORMAT:

EXAMPLES:RIGID=LAGRAN

DESCRIBERS:

Describer Meaning

LINEAR Selects the linear elimination method.

LAGRAN Selects the Lagrange multiplier method.

REMARKS:1. The RIGID command must be above the SUBCASE level.

2. The LAGRAN method allows for the thermal expansion of the rigid elements.

3. The RIGID command can only be used in SOLs 101 through 112. For all othersolution sequences, RIGID command is ignored and RIGID=LINEAR is used.

4. If the RIGID command is not specified, RIGID=LINEAR is used.

5. LINEAR processing will not compute the thermal loads. Also, in SOLs 103through 112, LAGRAN method must be used to compute the differentialstiffness due to the thermal expansion of the rigid elements.

6. When using RIGID=LAGRAN, K6ROT must be defined as non-zero.

9-2 NX Nastran 5.0 Release Guide

Page 147: NX Nastran 5 Release Guide

Miscellaneous Enhancements

9.2 Bolt Preload AnalysisWhen components of an assembly are bolted together, a specified torque translates into anaxial bolt preload. Bolts should be properly preloaded in this way before service conditions areapplied to the assembly. When analyzing preloaded bolts, you may be interested in obtainingthe stresses due to the preload condition alone, or due to a combination of the bolt preload andservice load. You can manually determine the preloaded bolt condition by using equivalentthermal loads, although using this method is approximate and typically requires many solutioniterations when multiple bolts exist.

NX Nastran 5 offers an automated method to simplify this multisolution process. A bolt analysiscan be selected by including the new case control command BOLTLD=n, where "n" selects thenew BOLTFOR bulk data entry containing the bolt preload value. The bolts should be modeledas beam elements, and included in the new BOLT bulk entry. One BOLT entry is required todefine each physical bolt. The service loads can also be included and are selected using the LOADcase control command. Superelements are permitted with preloaded bolts but the elementswhich define the bolts must be in the residual structure.

NX Nastran uses the following two solution process to automate the preloading of bolts:

• The beam elements which represent the bolts are reduced in stiffness by the value of theparameter BOLTFACT which makes their stiffness insignificant.

• The bolt preloads are applied at the ends of the bolts in the axial direction.

• A linear statics solution runs to get the relative displacements, U2 and U1, for each pair ofgrids.

• The bolt strains are calculated as:

Bolt Strain = – (U2-U1)/L – P/AE

where U1 and U2 are the deflections at the ends of the bolts, P is the preload, and A isthe area of the bolt.

In the final solution step, the bolts are treated as they were modeled (beam elements), thenthe calculated bolt strains are applied plus any service loads (if defined).

The new BOLTLD case control command, and the new BOLTFOR and BOLT bulk entries areprovided below for convenience.

New BOLTLD Case Control Command

BOLTLD

Bolt Load Set Selection

Selects the BOLTFOR Bulk Data entry for bolt preload processing.

FORMAT:BOLTLD=n

EXAMPLES:BOLTLD=5

NX Nastran 5.0 Release Guide 9-3

Page 148: NX Nastran 5 Release Guide

Chapter 9 Miscellaneous Enhancements

DESCRIBERS:

Describer Meaning

n Selects the BOLTFOR Bulk Data entry for bolt preloadprocessing. (Integer>0)

REMARKS:1. Bolt preloading is supported in SOLs 101, 103, 105, 107 through 112 and 601.

New BOLTFOR Bulk Entry

BOLTFOR

Preload Force on Set of Bolts

Defines preload force on a set of bolts.

FORMAT:

1 2 3 4 5 6 7 8 9 10

BOLTFOR SID LOAD B1 B2 B3 B4 B5 B6

B7 THRU B8

B9 B10 -etc-

EXAMPLE:

BOLTFOR 4 1000.0 12 THRU 21

1 4 6 9 10 26 32 34

37 43 51

FIELDS:

Field Contents

SID Bolt preload force set identification number. (Integer>0)

LOAD Magnitude of the preload force. (Real)

Bi Bolt identification numbers defined by bulk entry BOLT(Integer>0), or using “THRU” (B7<B8 for THRU option).

REMARKS:1. Multiple BOLTFOR entries with the same SID can be used and the data will

be combined.

2. SID is selected by BOLTLD Case Control command.

9-4 NX Nastran 5.0 Release Guide

Page 149: NX Nastran 5 Release Guide

Miscellaneous Enhancements

New BOLT Bulk Entry

BOLT

Bolt definition

Selects the CBEAM or CBAR elements to be included in the bolt pre-loadcalculation.

FORMAT:

1 2 3 4 5 6 7 8 9 10

BOLT BID ETYPE EID1 EID2 EID3 etc.

EID7 “THRU” EID8

-etc-

ALTERNATEFORMAT:

1 2 3 4 5 6 7 8 9 10

BOLT BID ETYPE EID1 “THRU” EID2

EXAMPLE 1

SOLs 101, 103, 105, 107 through 112 Example (single EID defines bolt):

BOLT 4 1 11

EXAMPLE 2

SOL 601 Example (all EIDs required to define bolt):

BOLT 4 1 11 8 2 1 20 14

15 16 28 30 33

FIELDS:

Field Contents

BID Bolt identification number. (Integer>0)

ETYPE Element type (Integer=1 is required in this field until other optionsbecome available in the future).

EIDi Element identification numbers of CBEAM or CBAR elementsto include in the bolt pre-load calculation.(Integer>0), or using“THRU” (EID7<EID8 for THRU option).

REMARKS:1. One BOLT entry is required to define each physical bolt.

2. Bolt preload is supported in SOLs 101, 103, 105, 107 through 112 and 601.

NX Nastran 5.0 Release Guide 9-5

Page 150: NX Nastran 5 Release Guide

Chapter 9 Miscellaneous Enhancements

3. In a SOL 101, 103, 105, 107 through 112, only one EID is required to definea bolt even if it was modeled as several elements. Any additional entrieswill be ignored.

4. In a SOL 105, both the bolt preload and service load will be scaled to determinethe buckling load.

REMARKSRELATED TO

SOL 601:1. All CBEAM or CBAR elements representing a physical bolt must be identified

in the BOLT entry.

9.3 Improvements to Surface Glue

Element Iterative Solver Support of Surface GlueIn NX Nastran 4.1, a new option to “glue” element faces together during a solution becameavailable. Glue definitions, which are supported in all solution sequences except for SOL 701,create stiff springs to connect the pre-defined surfaces. Glue elements connect componentstogether in such a way as to prevent all relative motion. Glue definitions in a SOL 153 heattransfer analysis are treated as near infinite conductivity connections.

The surface-to-surface glue capability was supported by the global iterative solver in NX Nastran4.1, but not by the element iterative solver. Now in NX Nastran 5, the surface-to-surface gluedefinitions can be included when using the element iterative solver.

To utilize the element based iterative solver, add the keyword ELEMITER=YES to the Nastrancard, so that you have:

NASTRAN ITER=YES, ELEMITER=YES $

When using glue definitions with the element iterative solver in NX Nastran 5, it is recommendedyou use the primal preconditioner which was introduced in NX Nastran 4. It can significantlyreduce the number of iterations required, thus decreasing the solution time. To turn this optionon, enter ‘PRIMAL’ in the PRECOND field on the ITER bulk data card.

Glue Accuracy ImprovementNX Nastran uses pre-defined source and target regions of element free faces to detect glueconditions in the model. More specifically, the solver uses the element faces from a sourceregion to project normals, then checks if these normals intersect with other element faces ona target region.

A glue element is created if:

• NX Nastran finds an intersection between element faces, and

• The distance between the two faces is equal to or less than a distance that you specify.

The number of glue elements created and their distribution determine the accuracy of the gluedinterface.

Two new glue solution parameters, INTORD and REFINE, are now available in NX Nastran 5on the BGPARM bulk entry to improve the accuracy of the glue solution. The updated BGPARMbulk entry is repeated below for convenience.

9-6 NX Nastran 5.0 Release Guide

Page 151: NX Nastran 5 Release Guide

Miscellaneous Enhancements

Previously, the number of locations where normals were projected (glue points) from the sourceregion was dependent on the element type. For example, the linear triangle face would alwaysproject a single normal, while the parabolic quad would project 4 normals.

Now in NX Nastran 5, the number of glue points used is dependent on the value assigned to thenew INTORD parameter, and on the element face type. By default, INTORD=2 and the numberof glue points increases as compared to the previous release. When INTORD=1, the numberof glue points is the same as the previous release. The following table summarizes how theINTORD value adjusts the number of glue points for a particular element face:

Number of Glue Points Used in Glue Element EvaluationFace Type INTORD=1 INTORD=2 (default) INTORD=3Linear Triangle 1 3 7Parabolic Triangle 3 7 12Linear Quad 1 4 9Parabolic Quad 4 9 16

The new REFINE parameter increases the number of glue points by refining the mesh on thesource region. Part of the refinement process is to project element edges and grids from theassociated target region back to the source region. The resulting refinement on the source regionis then more consistent with the target side, which then gives a better distribution of glueelements. The refined grids and elements are only used during the solution. The glue results aretransferred back to the original mesh for post processing results.

By default, REFINE=1 and mesh refinement occurs. REFINE=0 turns off the refinementcapability.

NX Nastran 5.0 Release Guide 9-7

Page 152: NX Nastran 5 Release Guide

Chapter 9 Miscellaneous Enhancements

New BGPARM Bulk Entry

BGPARM

Glue Parameters.

CONTROLPARAMETER

FOR THEGLUE

CONTACTALGORITHM.

FORMAT:

1 2 3 4 5 6 7 8 9 10

BGPARM GSID Param1 Value1

EXAMPLE:

BGPARM 4 INTORD 2

FIELDS:

Field Contents

GSID Glue set ID. Parameters defined in this command apply to glue setGSID defined by a BGSET entry. (Integer > 0)

PARAMi Name of the BGPARM parameter. Allowable names are given inthe parameter listing below. (Character)

VALUEi Value of the parameter. See below for the parameter listing. (Realor Integer)

Table 9-1. BGPARM Parameters

Name Description

ZOFFSET Determines if the shell element z-offset is included in the gluesolution.

0 - Includes the shell z-offset when determining the glue surfaces(Default)

1 - Does not include the shell z-offset when determining the gluesurfaces (as in NXN4.1 and prior)

INTORD Determines the number of glue points for a single element face onthe source region.

1 - Low order

2 - Medium order (default)

3 - High order

9-8 NX Nastran 5.0 Release Guide

Page 153: NX Nastran 5 Release Guide

Miscellaneous Enhancements

Table 9-1. BGPARM Parameters

Name Description

REFINE Determines if the mesh on the source region is refined duringthe glue solution.

0 - Do not refine the glue source region based on target surfacedefinition.

1 - Refine the glue source region based on target surface definition(default).

Shell Element Z-offset with Surface GlueShell elements such as the CQUAD4 can be offset relative to the mean plane of their connectedgrid points using the ZOFFS option. In NX Nastran 5, the surface glue solution can includethe shell element ZOFFS when evaluating the glue surfaces. The value assigned to the newZOFFSET option on the BGPARM bulk entry determines if the glue solution recognizes theZOFFS value. By default, ZOFFSET=0 and the ZOFFS value is used when evaluating the gluesurfaces. When ZOFFSET=1, the ZOFFS value is not used when evaluating the glue surfaces.

9.4 Parameter Specification ImprovementsNX Nastran 5 allows the specification of PARAM statements as follows:

1. Parameter values can be specified in the nast5rc (UNIX) and nast5.rcf (WINDOWS).

2. Parameters can be assigned a user defined keyword. The new keyword can then be used tospecify a value for the parameter on the command line or in the nastran resource file. Thekeywords can be defined in the “nastran.params” file in the architecture directory.

Specifying Parameters in the Resource (rc) fileParameters can be specified using the following syntax in the rc file:

PARAM,name,value

For example:

PARAM,POST,-2

PARAM,AUTOSPC,NO

Using the nastran.params fileThe parameter defining keywords are assigned in the nastran.params file located at:

UNIX: install_dir/nxn5/arch_dir/nastran.params

Windows: install_dir\nxn5\i386\nastran.params

You can override this default file using the “0.params” keyword on the command line. Forexample, if you have a custom defined parameter file named “my_specs.param”, the commandline submittal would be:

NX Nastran 5.0 Release Guide 9-9

Page 154: NX Nastran 5 Release Guide

Chapter 9 Miscellaneous Enhancements

nastran input_file.dat 0.params=full_path_name/my_specs.param

where “full_path_name” is the fully qualified path name of the file “my_specs.param”.

The parameters can be assigned to keywords in the nastran.params file as follows:

keyword name : param name : param value

where

• Keyword name – the name to be assigned as a mnemonic for the parameter name. Thisshould be a single word without embedded white space. Also, this name should not be thesame as any of the internal keywords. The keyword name can be the same as the parametername.

• Parameter Name – the actual name of the parameter, for example, “autospc”.

• Parameter Value – used for syntax checking when the keyword name is used in the commandline. The parameter value is either a “number” (integer/real) or an acceptable value list.For example, “yes”,”no”.

Comments can be included in the nastran.params file by starting the line with “#”, “$” or “;”as the first character.

Below is an example nastran.params file:

The punch file algorithm has been corrected in NX Nastran 5 to interpret data values correctly,thus data entries which are character, real or integer will each be written to the punch output assuch. The new system cell 210 has been created to control the punch formatting. By default,system cell 210 is set to “1” which correctly interprets data values. To revert back to the old,incorrect method of punch output, the system cell can be set to “0”.

## example nastran.params filemypost : POST : numbernxautospc: AUTOSPC : “yes”, “no”alpha1: alpha1 : number,number

In the nastran submittal, these keywords can be used to assign values to the parameters:

nastran input_file.dat mypost=-2 nxautospc=no alpha1=0.1,0.2

The order of processing the parameter values follows the same precedence rules as in the “rc” file.That is, parameters in the input file have the precedence over the command line specificationwhich has precedence over the rc file specification.

9.5 Punch Output CorrectionsIn previous releases of NX Nastran, the algorithm which writes punch files assumed that thefirst item in an entry was either an integer or real, and that all consecutive items were real.These were sometimes invalid assumptions since there are occasions in which the first item in adata entry is a character string, and there are also occasions in which some consecutive itemsare integers and not real.

This has been corrected in NX Nastran 5. The example below demonstrates how an incorrectlywritten punch might look, and how it is corrected in NX Nastran 5:

9-10 NX Nastran 5.0 Release Guide

Page 155: NX Nastran 5 Release Guide

Miscellaneous Enhancements

The characters highlighted in red were incorrectly written using the “old” punch algorithm. Thecharacters in green demonstrate how the “new” algorithm interprets and writes the correctdata type.

The system cell 210 is available to revert back to the “old” punch algorithm.

System cell 210 settings are summarized as follows:

System Cell 210 Value Result0 “OLD” Punch algorithm used1 “NEW” Punch algorithm used (default)

NX Nastran 5.0 Release Guide 9-11

Page 156: NX Nastran 5 Release Guide

Chapter 9 Miscellaneous Enhancements

9.6 PARAM K6ROT UpdateThe formulation of the CQUAD4 and CTRIA3 elements are based on the Mindlin-Reissner shelltheory. These elements do not provide direct stiffness for the rotational degrees-of-freedom whichare normal to the surface of the element. This zero stiffness results in a singular stiffnessmatrix, preventing NX Nastran from solving.

To avoid this problem, you can:

• Constrain the rotational degree-of-freedom either manually with an SPC entry orautomatically with the AUTOSPC parameter.

• Apply an artificial stiffness term to the degrees of freedom using PARAM K6ROT.

Assigning PARAM K6ROT to 0 may overly constraint, particularly when curvature exists, orwhen shell and solid elements are connected to common grids. To avoid overly constraining shellrotations, it is better to assign a penalty stiffness value to PARAM K6ROT versus the fixedcondition when 0 is used. Assigning PARAM K6ROT to 100 has shown to be a good, generalpenalty stiffness value which will not adversely affect results, yet will properly constrain therotation degrees-of-freedom.

Therefore, the default for parameter K6ROT has been modified from 0 to 100 in NX Nastran 5.

In addition to the default update, K6ROT will now be assigned a value of 0 when a model ismade entirely of membrane elements. The recommendation to not use K6ROT with membraneelements has always existed. This change simply makes the recommendation a requirement.

K6ROT Default = 100.0K6ROT specifies the stiffness to be added to the normal rotation for CQUAD4and CTRIA3 elements. This is an alternate method to suppress the grid pointsingularities, and is intended primarily for geometric nonlinear analysis. Avalue between 1.0 and 100.0 is recommended to suppress singularities. A largevalue may be required in nonlinear and eigenvalue analyses. This parameteris ignored for CQUADR, CTRIAR, CQUAD8, and CTRIA6 elements. K6ROTis forced to 0 when only membrane elements exist.

9.7 AUTHQUEUE UpdateWhen an NX Nastran job fails because of a failed license request, an option to have the jobretried automatically became available in NX Nastran 4.1. NX Nastran will retry a failed licenserequest job every minute up to the value of the AUTHQUEUE keyword. The AUTHQUEUEkeyword default was 20 minutes in NX Nastran 4.1. Based on customer feedback, the default hasbeen modified to 0 in NX Nastran 5. You must now explicitly define the AUTHQUEUE keywordin the nast5rc/nast5.rcf files to take advantage of this capability.

authqueue authqueue=number Default: 0

9.8 Documentation ImprovementsThe following documentation improvements have occurred in NX Nastran 5:

• The NX Nastran documentation is now available in both PDF and HTML format. A singlebookshelf.html provides the links to the PDF and the HTML documentation.

9-12 NX Nastran 5.0 Release Guide

Page 157: NX Nastran 5 Release Guide

Miscellaneous Enhancements

• HTML search is available. You can search a single, or all HTML documents by selectinglocated at the top left corner of the bookshelf page. The web browser must have a Java plugininstalled, and the Java plugin option must be turned on in the browser by checking the Javaplugin box under: Tools->Internet Options->Advanced.

The pdf documents can be searched using the Adobe Reader search options when a PDF isopened. To search across multiple PDF documents, the pdf file must open directly within theAdobe Reader, not from within a web browser, and the advanced search options selected.To reconfigure your preferences in Adobe Reader 7.0 to launch files outside of the webbrowser, select on the Reader menu Edit—>Preferences—>Internet, then deselect DisplayPDF in browser.

• The new NX Nastran Parallel Processing User’s Guide is included with the NX Nastrandocumentation. This guide provides information on NX Nastran parallel methods for Linuxand UNIX systems including parallel processing basics, detailed computational methods, anoverview of execution methods, and examples using various solution sequences.

• The NX Nastran Numerical Methods User’s Guide has been rewritten in this release. It nowreflects the software improvements since NX NASTRAN 1.

NX Nastran 5.0 Release Guide 9-13

Page 158: NX Nastran 5 Release Guide
Page 159: NX Nastran 5 Release Guide

Chapter

10 Upward Compatibility

See the books NX Nastran 5 DMAP Updates and Additions and NX Nastran DMAPProgrammer’s Guide for the updated formats.

10.1 Updated and New DMAP Modules

Updated ModulesBCDR

DDRMM

DSCMCOL

EFFMAS

EMG

FOCOEL

FOELCS

FRLG

GP1

GPSTR2

MATMOD

MTRXIN

NXNADAMS

PARAML

RANDOM

SDR2

SDRCOMP

SDRX

SOLVIT

XYTRAN

New ModulesADDVM

NX Nastran 5.0 Release Guide 10-1

Page 160: NX Nastran 5 Release Guide

Chapter 10 Upward Compatibility

BOLTFOR

BOLTSF

DDR2N

EMR

GP1LM

LRFORCE

NXNRFI

10.2 Updated and New Datablocks

Updated DatablocksCASECC

CONTACT

DSCMCOL

DYNAMIC

OEE

OEF

OES

OPG

OQG

OUG

New DatablocksOCCORF

OCPSDF

OSDISP2

OSHT

10.3 Updated and New SubdmapsThe following is the list of subdmaps where calling arguments and/or the input/output datablocks have changed:

attmods.dat (new)

cforce.dat

cforcerd.dat

cforcrd1.dat (new)

cforcrd2.dat

10-2 NX Nastran 5.0 Release Guide

Page 161: NX Nastran 5 Release Guide

Upward Compatibility

conattm.dat (new)

conmods.dat

ddamexe.dat (new)

ddampost.dat (new)

ddampre.dat (new)

disopts.dat

disprs.dat

genmodes.dat (new)

ifpl.dat largrgb.dat (new)

mkpxout.dat (new)

modcon.dat (new)

phase1a.dat

phase1b.dat

phase1c.dat

phase1d.dat (new)

phase1dr.dat

phase1e.dat (new)

phase1f.dat (new)

poolit.dat (new)

resddam.dat (new)

sdr2stat.dat

sdree1.dat (new)

sedrcvr.dat

semg.dat

semg1.dat

skewmat.dat (new)

super1.dat

super3.dat

tranmat.dat (new)

vdr1.dat

xform.dat (new)

xmtrxin.dat

NX Nastran 5.0 Release Guide 10-3

Page 162: NX Nastran 5 Release Guide
Page 163: NX Nastran 5 Release Guide

Chapter

11 System Description Summary

Table 11-1. System Description – HP9000 – HP-UXItem Description

Supported Model(s) PA-RISC

Configurations for InstalledTiming Constants

250, 710, 712, 715, 720, 730, 735, 778, 800, 819, 889, 2200, 2600, 2733, 3700, 4000, 4900, 6750

Build Operating System PA-RISC: HP-UX B.11.00Other Supported OperatingSystems

HP-UX B.11.11

Word Length 32 bits

Build Type LP-64, LP-64 DMP

MPI required for DMP HP MPI 2.0.2 (comes with OS)

Table 11-2. System Description – Intel Itanium HP-UXItem Description

Supported Model(s) Intel-Itanium-HP-UX

Configurations for InstalledTiming Constants

4900, 5300, 5400, 5600

Build Operating System HP-UX B.11.23

Other Supported OperatingSystems

Word Length LP-64: 32 bits; ILP-64: 64 bits

Build Type LP-64, LP-64 DMP, ILP-64, ILP-64 DMP

MPI required for DMP HP MPI 2.0.2 (comes with OS)

Table 11-3. System Description – Intel – Windows (32-bit)Item Description

Supported Model(s) Intel and Intel-compatible

Configurations for InstalledTiming Constants

Pentium II 400 MHz, P4 1.5GHz, Pentium Pro., P4 3 GHz

Build Operating System Windows 2000, SP3

Other Supported OperatingSystems

WXP SP2, WXP-64 SP1 (on EM64T/Opteron)

Word Length 32 bits

Build Type ILP-32

NX Nastran 5.0 Release Guide 11-1

Page 164: NX Nastran 5 Release Guide

Chapter 11 System Description Summary

Table 11-4. System Description – Intel (EM64T/Opteron) – Windows (64 bit)Item Description

Supported Model(s) Intel and Intel-compatible

Configurations for InstalledTiming Constants

Pentium II 400 MHz, 733 MHz Itanium 32 Bit, P4 1.5GHz, Pentium Pro.

Build Operating System Windows Server 2003 – 64 SP1Other Supported OperatingSystems Windows XP–64 SP1

Word Length 32 bits

Build Type LP-64

Table 11-5. System Description – Intel – LinuxItem Description

Supported Model(s) Intel and Intel-compatible

Configurations for InstalledTiming Constants P4 2.8Ghz

Build Operating System Redhat 7.3Other Supported OperatingSystems

Suse 9.2, Redhat 9, Redhat EL 3.0, Redhat EL 4.0

Word Length 32 bits

Build Type ILP-32, ILP-32 DMP

MPI required for DMP HP MPI 2.0.2 (included with NX Nastran install)

Table 11-6. System Description – X86_64 Linux (AMD Opteron/EM64T)Item Description

Supported Model(s) X86-64 Linux

Configurations for InstalledTiming Constants

8664

Build Operating System Suse 9.0Other Supported OperatingSystems

Suse 9.1, Suse 9.3, Redhat EL 3.0, Redhat EL 4.0

Word Length LP-64: 32 bits; ILP-64: 64 bits

Build Type LP-64, LP-64 DMP, ILP-64, ILP-64 DMP

MPI required for DMP HP MPI 2.0.2 (included with NX Nastran install)

Table 11-7. System Description – Intel Itanium LinuxItem Description

Supported Model(s) Itanium II

Configurations for InstalledTiming Constants IA-64 800 Mhz & 733Mhz

Build Operating System Redhat EL3.0Other Supported OperatingSystems Redhat EL4.0

Word Length LP-64: 32 bits; ILP-64: 64 bits

11-2 NX Nastran 5.0 Release Guide

Page 165: NX Nastran 5 Release Guide

System Description Summary

Table 11-7. System Description – Intel Itanium LinuxBuild Type LP-64, LP-64 DMP, ILP-64, ILP-64 DMP

MPI required for DMP HP MPI 2.0.2 (included with NX Nastran install)

Table 11-8. System Description – Sun SPARC – SolarisItem Description

Supported Model(s) UltraSPARC

Configurations for InstalledTiming Constants UltraSPARC (75 & 95)

Build Operating System UltraSPARC: Solaris 8Other Supported OperatingSystems

Solaris 9, Solaris 10

Word Length 32 bits

Build Type LP-64

Table 11-9. System Description – IBM RS/6000 – AIX (64 bit)Item Description

Supported Model(s) Power3, Power4, Power5

Configurations for InstalledTiming Constants

303, 320H, 370, 375, 390, 397, 530, 530h, 550, 560, 570, 580, 590, 591, 950, 980, 990, 4316, 9133

Build Operating System AIX 5.1Other Supported OperatingSystems

AIX 5.2, AIX 5.3

Word Length LP-64: 32 bits; ILP-64: 64 bits

Build Type LP-64, LP-64 DMP, ILP-64, ILP-64 DMP

MPI required for DMP POE 3.2.0.0 (add on from IBM)

Table 11-10. System Description – SGI R8K, R10K, R12K – IRIX64Item Description

Supported Model(s) R8K, R10K, R12K, R16K

Configurations for InstalledTiming Constants

IP7, IP19, IP20, IP21, IP22, IP27, IP28, IP30, IP35, 240, 510

Build Operating System IRIX 6.5.21

Other Supported OperatingSystems

IRIX 6.5.24m, IRIX 6.5.27m

Word Length 32 bits

Build Type LP-64, LP-64 DMP

MPI required for DMP MPI comes with OS, version depends on OS level

Table 11-11. System Description – SGI AltixItem Description

Supported Model(s) SGI-ALTIX

NX Nastran 5.0 Release Guide 11-3

Page 166: NX Nastran 5 Release Guide

Chapter 11 System Description Summary

Table 11-11. System Description – SGI AltixConfigurations for InstalledTiming Constants

6402

Build Operating System SGI ProPack 3 sp3

Other Supported OperatingSystems SGI ProPack 4

Word Length LP-64: 32 bits; ILP-64: 64 bits

Build Type LP-64, LP-64 DMP, ILP-64, ILP-64 DMP

MPI required for DMP SGI MPT 1.10 (comes with OS)

11-4 NX Nastran 5.0 Release Guide