What's New

View new features for OptiStruct 2021.2.

Altair OptiStruct 2021.2 Release Notes


  • Sine on Random Fatigue
  • User material and MATMDS for shells
  • Electrical Analysis
  • 1D Gasket
  • SENSOR for Transient Heat Transfer analysis
  • Marlow hyperelasticity material model

New Features

Stiffness, Strength, and Stability
1D Gasket
1D gasket (thickness-direction behavior only) option can be activated via the GASK1D field on PGASK Bulk Data for CGASK6/8/12/16 elements.
Enhanced Nonlinear Restart
The Nonlinear Restart functionality has been enhanced with the new RMDX option on the RESTARTW entry, which allows all the model contents, including the Subcase Entries, to be saved in a restart file (new *.rmdx file).
With this option, simple restart case would require only RESTARTR Case Control Entry in entire restart input file. Here, additional entities are needed which are not available in input file with RESTARTW run, they can simply be added in the restart input file (RESTARTR).
CORNER option as default in Surface to Surface (S2S) Contact
CORNER treatment is turned on by default (with 30 degree as break angle) for Surface to Surface Contact.
Material user subroutine (MATUSR) and Integration with Multiscale Designer (MATMDS) for Shells
MATUSR and MATMDS are now supported for shell elements. The property of shells could be PSHELL, PCOMPP, or PCOMP(G).
Marlow material model
The Marlow model is a hyperelastic material model which directly defines the potential based on the experiment test data. There are no mathematical expressions based on the deformation tensors’ invariants or the deformation stretches for the potential. The isochoric deformation potential is determined by TAB1, TAB2, or TAB4. Only one test can be specified. D1, TABD, or Poisson’s ratio (NU) can be defined to specify the volumetric behavior. Either D1 or TABD can be specified, but not both simultaneously.
  • If Poisson’s ratio is specified, it is used to determine the volumetric behavior (D1 and TABD are ignored).
  • If D1 or TABD is specified, but Poisson’s ratio is blank, the volumetric behavior is determined by D1 or TABD.
  • If D1, TABD, or Poisson’s ratio are not defined, the default Poisson’s ratio of 0.495 is used to determine volumetric behavior.
Temperature dependent Marlow material is also supported with MATTHE Bulk Entry.
H3D output for preloaded Linear analysis following Nonlinear analysis as preloading
New option, MESHNLPTB is available on the SYSSETTING entry.
Parameter to adjust the stiffness calculation for small displacement NLSTAT preloaded linear analysis
PARAM,KSMNL4PL has been added.
Tangential stiffness in linear analysis for closed contact
CONTPRM,KTLIN has been added.
Explicit Dynamic Analysis
Interpolation, SMOOTH, option for Table data (TABLED/TABLEG)
SMOOTH has been added for Y-axis for smooth interpolation of the table data.
Heat Transfer Analysis
OP2 file support for Transient Heat Transfer Analysis
Supported results include temperature, thermal gradient, and fluxes.
Easier definition of convection from fluid to surrounding structure using CAFLUID
New Bulk Data Entry CACONV is now available for automatic generation of CONV and CHBDYE entities, which are directly linked to CAFLUID advection nodes. CACONV inputs are the main and secondary side of models, similar to CONTACT setup.
If PARAM,CACONV,YES is defined in the input deck, a dedicated *_CONV.fem file will be generated, that contains internally generated CONV, CHBDYE elements for convection definition.
Temperature SENSOR
Temperature sensor functionality is available with SENSOR Bulk Entry. Currently, the temperature SENSOR is available for Transient Heat Transfer analysis. This functionality allows the activation/deactivation of loading according to the desired range of response that the user defines. For transient heat transfer application, this functionality would simulate thermostat, where the flux or volumetric heat will be adjusted in order to satisfy the desired temperature range (lower and upper bound) during the heat transfer analysis.
Electrical Analysis
Electrical Conduction Analysis is now supported as a new solution type. Coupling between Electrical analysis and heat transfer analysis is also supported through Joule heating. Temperature-dependent electrical material is also supported. Supported element types in Electrical analysis:
  • Shells, Solids
  • CGAP(G)PGAPEC as property
  • CONTACT - PCONTEC as property
Supported Loading and Boundary Conditions:
Loading/BCs Type
Bulk Entry
Nodal Current
Current Density
Supported Material:
Material Type
Bulk Entry
Isotropic Electrical Material
Anisotropic Electrical Material
Temp Dependent Material
Supported output:
Output Type
Output Request
Joule Loss Density
Current Density
Electric Field
Conductivity and Resistivity
Grid point Current
Applied Nodal Current
NDIV=3 for Composite results (CSTRESS/CSTRAIN/CFAILURE)
NDIV=3 is now the default for composite results (CSTRESS/CSTRAIN/CFAILURE).
Cuntze criteria update
Cuntze’s failure criterion has been updated according to the publication: The predictive capability of failure mode concept-based strength conditions for laminates composed of unidirectional laminae under static triaxial stress states, Journal of Composite Materials, September 2012.
Cuntze’s failure criterion determines five potential failure modes and combines them when determining the resultant stress effort, which is considered as failure index. In addition to the basic strength allowables, the criterion includes three material parameters. The values suggested by Cuntze for pre-dimensioning, are used as default.
Damage due to Sine-On-Random (SOR) vibration
  • In this schema, sinusoidal vibration (excitation) is superimposed on Random vibration.
  • Input load is represented by mixed mode: PSD of random load and amplitude of sine load.
  • A Random Response analysis and a Frequency Response analysis are underlying analyses for Sine-on-Random fatigue. FATLOADs exclusive to Random Response analysis loadcase and Frequency Response analysis load case are referenced in the FATEVNT card.
  • The spectral moments calculated include contribution from both random (PSD) loading and sine tones (sinusoidal excitation).
  • The damage estimation procedure is similar to Random Vibration Fatigue where Probability Density Function (PDF) are generated from the spectral moments. Dirlik is the recommended PDF.
  • The FATLOAD card is enhanced for continuation line with keyword HARMO. This is used to refer FRF analysis loadcase and list the frequencies with their amplitude factors.
  • The damage output is factored with input exposure time on FATSEQ card.
Damage due to multiple sine tones
  • In this schema, multiple sine tones (sinusoidal excitations) acting simultaneously are considered as random vibration without exclusive random vibration load (PSD) input.
  • Input load is represented by multiple sinusoidal excitations (Sine tones).
  • A Frequency Response analysis (FRF) is the underlying analysis for fatigue due to multiple sine tones. FATLOAD referencing Frequency Response loadcase is specified in the FATEVNT card.
  • The spectral moments calculated include contribution from all the sinusoidal excitation considered in the input.
  • As multiple sine tones acting simultaneously is considered as Random Vibration load, the damage estimation procedure is similar to Random Vibration fatigue.
  • The FATLOAD card is enhanced for continuation line with keyword HARMO. This is used to refer FRF analysis loadcase and list the frequencies with their amplitude factors.
  • The damage output is factored with input exposure time on FATSEQ card.
Damage due to sine sweep on random vibration
  • In this schema, the Swept sine tone (sinusoidal excitation) is considered on a random signal (PSD) without superimposing the two. The loading is considered as a series of single sine tone on top of random vibration. Swept-sine on-random signal can be discretized into intervals based on the time duration (T), the input sine tone spent on moving from one frequency to another.
  • The Damage is calculated for each interval of time (T). Total damage (D) per sweep due to swept sine tone on random vibration is the summation of damages during each time duration (T).
  • Input load is represented by mixed mode: PSD of random load and amplitude of sine load together with sweep rate. The sweep rate is defined on the FATLOAD continuation line following SWEEP keyword.
  • A Random Response analysis and a Frequency Response analysis are underlying analyses for Sine sweep on Random fatigue. FATLOADs exclusive to Random Response analysis loadcase and Frequency Response analysis (FRF) loadcase are referenced in the FATEVNT card.
  • Spectral moments are calculated for each time interval (T) which has contribution from both Random (PSD) loading and sine tone corresponding to the time interval (T).
  • The damage estimation procedure is similar to Random Vibration fatigue where Probability Density Function (PDF) is generated from the spectral moments.
  • Output damage is factored by the input number of sweeps on FATSEQ card.
DTPL dependent MATINIT
MATINIT through DOPTPRM can be used to adjust the initial volume fraction for topology (DTPL) and free-size (DSIZE). MATINIT is now available as an option in continuation line in DTPL, DSIZE which would allow different MATINIT for different DTPL/DSIZE in the model. MATINIT defined in DTPL/DSIZE would take precedence over DOPTPRM,MATINIT.
Optimization support for Direct Frequency Response
Optimization is now supported for Direct Frequency Response in addition to Modal Frequency Response. Supported design variables are size shape (including free-shape). All responses, except FRFLOW and FRERP are supported for Modal Frequency Response base optimization are also supported for Direct Frequency Response optimization.
Restart support for LEVELSET topology optimization
Restart is now supported for LEVELSET topology optimization.
Modal effective mass fraction in .mvw file
Modal effective mass fraction is now available in .mvw file with OUTPUT,HGEFFMASS output request. This value is identical to the one in .out file.
LOADLIB enhancement for user material
  • No LOADLIB: In this case, OptiStruct assumes it is default library, which is umat.dll (umat.so on Linux). Environment variables are used to indicate the location of such default library.
  • LOADLIB but no path specified (for example, LOADLIB....,test.so): Environment variables are used to indicate the location of such default library.
Real value support for TB and TP filed on RLOAD1/2
The TB and TP fields on RLOAD1/2 now accept Real value input (instead of table data) as a constant data over loading frequencies.
Accurate mass printing in residual run using CMS .h3d file
Mass results will be accurate in residual run for superelement application in case the reduced matrices are available with .h3d file.
User-defined error data in user subroutine
User error data is now available for user material (usermaterial, usermatht, smatusr). Additional arguments (userdata, datalen, ierr) are added in these subroutines for this purpose.
-1: information
0: default
1: error
A grid SET can now be referred to on PARAM,ZAERO for normal modes data output associated with grids in the SET.
MPC/Rigids contribution in the .gpf file
GPFORCE output in ASCII format (OPTI) contains F-MPC (total contribution from MPC and any rigids), as well as each rigid element contribution as separate items (for example, RBE or MPC). This is the default behavior.
With SYSSETTING,GPFMPC= FMPC, only F-MPC will be printed in the .gpf file.
With SYSSETTING,GPFMPC= SEPAR, only the individual separate items (RBE, MPC, and so on) will be printed in the .gpf file.
MUMPS Solver default for Linear Analysis
MUMPS Solver is now the default solver for Linear Analysis.
  • HDF5 (.h5) output format is now available for part superelements.
    Table 1. Supported Results and Analysis Types for Part Superelements Summary
    Result Type Linear Static Normal Modes Transient/

    Frequency Response

    Random Response
    VELOCITY     #  
    ACCELERATION     #  
    ELFORCE   #  
    MPCFORCE   #  
    STRESS #  
    STRAIN #  

    # These result types are now available in HDF5 (.h5) format for transient analysis based on regular finite element based (without superelement) models.

  • Element force output is supported for CFAST elements in linear static and normal modes analysis.
  • File size reduction switch is available through the OUTPUT,HDF5,CMP entry.
  • Cylindrical and spherical coordinate systems are supported.
  • PCOMP objects are now available in the .h5 file.


Model Files Removed from Installation Package
To reduce the footprint size, the following tutorial model and demo model files are no longer included in the local installation. You can now find zipped tutorial model files and demo model files on Altair One via the Altair Community, Altair Marketplace, and Altair Connect sites. Altair recommends that you create an Altair One account and use it as your primary portal to access product documentation, a Knowledge Base, and customer support.
Tutorial Model Files
Note: Model files are also available in local installation for this release only.
Demo Model Files
OptiStruct Examples

Resolved Issues

  • Threaded bolt feature with CLRNC Bulk Entry no longer leads to Error 14,Missing CLRNC.
  • Element force output with ELFORCE for OPTI format returns results when H3D output is disabled.
  • Internal superelement now does not require METHOD entry for reduction subcase when the reduction is static. In previous releases, Internal Superelements used to require METHOD in the subcase which was supposed to be reduced out. This has been enhanced so that the subcase without METHOD is assumed to be static reduction.
  • When unsymmetric or complex DMIG is used with PFPATH, the job no longer terminates abruptly.
  • Sensitivity output with OUTPUT,ASENS has been improved to avoid "***", if the element/grid IDs are longer than eight digits.
  • MATFVE with creep test input will no longer cause the job to terminate without any error.
  • FLUX output with 2nd order shells in Heat Transfer analysis has been corrected.
  • In HyperMesh v2021.1, a new hm comment, HMCOMP was added to better handle direct/indirect property assignment. Legacy comment HMMOVE has been replaced with this new comment in HyperMesh v2021.1. The new hm comment, HMCOMP is now properly supported in OptiStruct, so that the component organization accurately shows up in the .h3d file.
  • The required disk space for random response has been reduced in case the elemental random response outputs are requested.
  • The performance and disk space required for GPSTRESS has been improved.
  • When TIE contact is applied on 2nd order solids in Linear analysis, the programming error elqc_util.F no longer occurs.
  • Large displacement analysis with RBAR no longer leads to a programing error.
  • Shape optimization that includes contact no longer leads to the programming error, cnt_contacts.F.
  • Convergence has been improved for CELAS1 in small displacement nonlinear analysis.
  • After free-size optimization with composites, the HM.ent.tcl file the extra backslash at the end of each element list has been eliminated, which prevented the script from running properly.