New version of LS-DYNA is released for all common platforms.
Release notes for Version 971 R5.1.1
Herein are summarized enhancements and significant corrections made since the release of 971 R5.1.0.
Adaptivity
- Fixed error reading adaptivity input file when *DATABASE_TRACER is used with adaptivity.
Airbag
- Fixed bug in *AIRBAG_HYBRID_CHEMKIN_JETTING that, for some platforms, caused a core dump.
- Support added for TITLE and ID options in *AIRBAG_PARTICLE including output of information to abstat_cpm.
- Fixed a problem with contact sensor switch for *AIRBAG_PARTICLE blockage option. The contact forces of vent and fabric nodes are used to determine the airbag blockage factor. If contacts were switched off by *SENSOR option, the nodal
contact forces of slave/master nodes were not reset to zero. The vents/frabric were always blocked after the switch.
- Fixed a problem with multiple definitions of *AIRBAG_PARTICLE with IAIR=2. The initial air particles were not placed at correct location after the first bag.
- Fixed bug in *BOUNDARY_AMBIENT_EOS that was introduced at R3.1. Bug was likely to cause an instability.
Boundary Conditions
- Fixed bug related to VAD=4 for prescribed relative motion between rigid bodies. Motion was in wrong direction.
- Fixed obscure bugs in pore pressure analysis having to do with node behavior at drained/undrained boundaries. This fix prevents the boundary nodes being controlled by the undrained part, even if the user spuriously defined a permeability for the drained part.
- Fixed error termination during input phase when *BOUNDARY_PRESCRIBED_MOTION_RIGID is prescribed for nodal rigid bodies.
Constrained Options
- Made various bug fixes and performance improvements for joints (explicit+penalty, explicit+lagrange, and implicit).
- Enabled input of Contact IDs with more than 8 digits (up to 10).
- Changed how PENOPT>3 in *CONTROL_CONTACT is handled:
- MPP: Input value of PENOPT is now used. - SMP: PENOPT is reset to default value.
- Fixed MPP eroding contact that resulted in wrong node thicknesses in some cases if there were shell elements in the eroding surface.
- Fixed contact warning format affecting contact IDs greater than 5 digits.
- Improved stability of offset tied contacts in MPP.
- Removed extraneous print to screen that slowed performance of *CONTACT_TIED_SURFACE_TO_SURFACE in some cases.
- Fixed beam-to-beam contact (AUTOMATIC_GENERAL) and shell edge contact (SINGLE_EDGE) which did not activate at birth time (BT>0). Affects MPP only.
- Fixed user defined friction so that using more than one forming contact with *USER_INTERFACE_FRICTION would not lead to a crash. Bug affected only SMP.
- Fixed array index problem in user defined friction.
- Fixed the periodic shell segment intersection check of segment based contact (DTPCHK on optional card D of *CONTACT) which failed to report all the intersected segments.
- Fixed a problem with small deck restarts that use parallel consistency and segment based contact. A buffer was not allocated causing an immediate crash.
- Fixed *CONTACT_TIED_SURFACE_TO_SURFACE_FAILURE for MPP which did not exhibit any failure behavior.
- Fixed an error in segment based contact that occurred when FRCENG=1 on *CONTROL_CONTACT and PKP_SEN=0 on *DATABASE_EXTENT_BINARY. Before the fix, this condition would crash the program.
- Fixed segmentation violation problem when viscous damping is enabled for contact, and the slave or master node is massless.
- Enabled nonlinear contact stiffness (FNLSCL and DNLSCL in *CONTACT) to work with implicit, segment-based contact when friction is present. The frictionless case already worked.
- Fixed bug affecting output to ncforc for *CONTACT_SLIDING (MPP only).
- Fixed SMP and MPP *CONTACT_GUIDED_CABLE, for better handling of nodes near the ends of beams. The old code was “sticking” and applying non-physical forces along the beams in some cases.
- Modified MPP tied contact initialization: when detecting which nodes should be tied and which not, consider the distance to the actual segment, not including any extension. This better honors any user specified tying criteria.
- Fixed problem of MPP contact initialization via full deck restart if the original problem used dynamic relaxation.
- Added MPP support for SPOTSTP (*CONTROL_CONTACT) which involves deformable spot weld nodes which are not tied to a master surface.
- Fixed MPP implicit problem with penalty tied contacts that was introduced in revision 63515.
- Enabled the KEEPCS parameter of *CONTROL_SHELL to work with segment-based contact.
- Fixed MPP bug whereby nodes were not constrained by *CONTACT_CONSTRAINT_NODES_TO_SURFACE when SHLTHK=1 in *CONTROL_CONTACT.
- Fixed segmentation fault during MPP initialization when *CONTACT_TIED_SURFACE_TO_SURFACE_CONSTRAINED_OFFSET is used.
Damping Options
- Made changes to *DAMPING_FREQUENCY_RANGE:
- fixed bug related to damping of nodes containing *ELEMENT_INERTIA, - added option to damp motion relative to a rigid body.
Database/Output
- Fixed a bug which caused incorrect stress to be output to eloutdet forshell type 21 (Wilson linear element).
- Fixed stress output with the CMPFLG=1 for brick materials 2, 21, 23, 40, 91, 92, 128, and 132. It’s been broken since January 10, 2011.
- Fixed bug whereby certain combinations of output options set in *DATABASE_EXTENT_BINARY caused shell strains to be output incorrectly to d3plot.
- Fixed bug affecting d3part database that includes failed tshells.
- Fixed output of the LEGEND to the LSDA/binout jntforc file, which was not correct for some joint types.
- Fixed use of load curve for d3part output.
- Fixed l2a crash problem when extracting eloutdet file from binout database.
Define Options
- Fixed segmentation fault problem when *DEFINE_CURVE_ENTITY is used.
Elements
- Fixed negative internal energy bug of beam formulation type 9 (spotweld beams) first noticed in using *INITIAL_AXIAL_FORCE_BEAM.
- Fixed error termination problem during input when *ELEMENT_MASS_MATRIX_NODE_SET is used.
- Fixed broken laminated shell theory in ELFORM 3 tshell elements.
- Fixes made to mapping in *INCLUDE_STAMPED_PART. It could happen that a crash part far away from the stamping part got wrong shell thickness and strains, where it should get default thickness and zero strains.
Implicit
- Fixed bug causing intermittent eigenvalue analysis to fail in MPP.
- Fixed indexing for implicit *CONSTRAINED_LINEAR_GLOBAL.
- Fixed problem of IDRFLG=-999 in *CONTROL_DYNAMIC_RELAXATION switching implicit solution to explicit solution.
- Fixed implicit mechanics output to spcforc file.
- Fixed problem with multiple prescribed motion of rigid bodies in local coordinates for implicit mechanics.
- Fixed implicit’s treatment of prescribed relative displacement between two rigid bodies (VAD=4), including honoring constraints.
- Fixed recent problem with implicit using *DAMPING_GLOBAL or *DAMPING_PART_MASS.
- Placed limit on non-convergence warning messages (20) to prevent unnecessarily large d3hsp and messag files.
Initial Conditions
- Fixed error termination when *INITIAL_VELOCITY_GENERATION_START_TIME is used.
- Fixed bug whereby nodal rigid bodies defined with the INERTIA option sometimes had their initial velocities inappropriately overwritten by *INITIAL_VELOCITY_GENERTATION.
- Fixed bug affecting *INITIAL_VELOCITY_GENERATION wherein nodal rigid body velocity definitions were removed based on the node set ID and not the part ID. This resulted in standard parts being removed from the velocity definition.
Loads
- Made parameters defined in *USER_LOADING available in subroutine loadsetud which pertains to *USER_LOADING_SET.
- Fixed SMP bug whereby loading a nodal rigid body with a moment in a model containing only solid elements caused negative volumes.
- Blast ID’s from *LOAD_BLAST_ENHANCED are now written to the BLSTFOR database.
Material Models
- Fixed bug in *MAT_224 which could lead to zero yield stress and therefore a divide by zero.
- Fixed *MAT_224 with EOS: corrected triaxiality calculation. Before this fix, the use of LCF (load curve defining plastic failure strain as a function of triaxiality) could give wrong results (e.g. element failure) when *MAT_224 was used together with an equation of state.
- Improved algorithm in *MAT_223 to prevent numerical underflow (very rare).
- Fixed XMFLAG option on *MAT_162 (composite_msc_dmg).
- Fixed *MAT_255 which gave nonsensical results for both SMP/MPP and explicit/implicit as compared to *MAT_004.
- Fixed bug for *MAT_187 (SAMP-1). The fade out of the stresses defined by EPFAIL and DEPRPT was not following the correct path, leading to the wrong equivalent plastic strain at rupture.
- Fixed bugs associated with new option LMCA for reading more material parameters than afforded by LMC in *MAT_USER_DEFINED_MATERIAL_MODELS.
- Made improvements to *MAT_272 (RHT) with more to follow. *MAT_272 is still in a state of development.
- Upgrade arrays in *MAT_098 shell subroutine to double precision to stablize certain crash problems.
- Made fix for combination of *MAT_ADD_EROSION and triangular elements. There was a possible memory clobber if ELFORM=16 and ESORT=1.
- Fixed the moment calculation that is done for the brick weld assembly failure calculation. The accuracy of the moment calculation was improved, and the TRUE_T parameter now correctly reduces the moment.
- Fixed bug affecting *MAT_082 with the ORTHO option.
- Fixed laminated shell therory (by LAMSHT=3 or 5) for shells of *MAT_058.
- Updated *MAT_ADD_EROSION damage model GISSMO:
- New option for LCSRS. LCSRS > 0: scale ecrit with strain rate scale factor LCSRS < 0: don't scale ecrit with strain rate scale factor - Use original strain rate curve instead of the discretized curve. - Scale ecrit also with the strain rate scale factor, not only epsf.
- Fixed bug in *DEFINE_ADAPTIVE_SOLID_TO_SPH which resulted in floating point exception on some platforms.
- Minimum time step for SPH particle deactivation modified to consider time step scale factor TSSFAC. Deactivation criterion is invoked using ERODE=1 and DTMIN.
- Made change to eliminate bucket sort memory warning when there are multiple SPH parts.
- Added ID and TITLE options for *DEFINE_ADAPTIVE_SOLID_TO_SPH.
- Fixed a failure to write matsum data for tshell parts that use user defined integration, either by *INTEGRATION_SHELL or *PART_COMPOSITE_TSHELL.
Thermal
- Fixed segmentation fault for thermal problems when BINARY=3 in *DATABASE_TPRINT.
General
- Fixed bug affecting nested *INCLUDE_TRANSFORM commands wherein all the expected transformations were not being invoked.
- Fixes made to *INCLUDE_TRANSFORM affecting:
- offset of CSID in *INITIAL_STRESS_SECTION: idroff -> idpoff - offset of ID in *ELEMENT_MASS_PART_SET: idpoff -> idsoff
- Fixed bug that caused improper printout of memory in some cases.
- Corrected solve storage estimate for superelement problems.
- Enhanced superelements:
1. allow single precision version to read binary direct matrix input 2. allow single precision version to handle larger problems
- Fixed problem in Madymo 7.x MDCAL coupling MPP decomposition didn’t correctly handle solid parts in the LS_DYNA/Madymo extended coupling. The coupling job failed right away.
- Increased the accuracy of part masses printed to d3hsp.
- Made change to not prepend jobid on the MPP optional command line “n=” file.
- Fixed corrupt d3plot produced by *CONTROL_MPP_DECOMPOSITION_SHOW when running MPP double precision with environment variable LSTC_BINARY set to 32ieee.
