LS-DYNA V971 R6.1.2 (R6.85274) released
Filed under: release note
New version of LS-DYNA is released for all common platforms.
Release notes for Version 971 R6.1.2
Herein are summarized most of the new features, enhancements, and significant corrections made since the previous release, 971 R6.1.1.
- Add support for *RIGIDWALL_FORCE_TRANSDUCER.
- Add new *CONTROL_REQUIRE_REVISION keyword. Allows user to control what revisions of the code can be run for a given input deck.
- Add MADCL=741 for new feature of coupling Madymo.
- parallel coupling between two programs
- allow deleted element flags between two codes
- Fix method for energy transfer during particle collision between different gas species (*AIRBAG_PARTICLE).
- Fix output of velocity from CPM sensor - a bug in the averaging method (*DATABASE_CPM_SENSOR).
- Fix uninitialized data in *BOUNDARY_ACOUSTIC_COUPLING_MISMATCH along with *BOUNDARY_USA.
- New implementation for extracting resultant forces due to joint (*CONSTRAINED_JOINT, *DATABASE_JNTFORC) for implicit mechanics.
- Fix for MPP support of *CONSTRAINED_NODAL_RIGID_BODY with release conditions: the special "first node" may not have been properly treated in some cases when the NRB is split across processors.
- decrease memory usage during init for tubular drawbeads
- fix possible memory error for segment based contact
- fix possible performance penalty for *DEFINE_FRICTION
- Correct handling of tied groupable contact with implicit in MPP (*CONTACT_TIED_...).
- Correct "out-of-range forces" problem with *CONTACT_AUTOMATIC_..._TIEBREAK, OPTIONs 7 and 10. Only affects MPP.
- Fix MPP *CONTACT_TIED_SHELL_EDGE_TO_SURFACE_BEAM_OFFSET: solids were in some cases still unstable, particularly if the user specified a large contact thickness (positive or negative) for searching purposes. So switch to using 10% of the characteristic edge length instead of 100% of the "segment" thickness.
- Add MPP support for non-zero birthtime for *CONTACT_SINGLE_EDGE.
- Fix input error when SSTYP=6 in *CONTACT card 1 and there are >9999 parts in the part set defined in SSID.
- Fix for MPP *CONTACT_GUIDED_CABLE.
- Fix for MPP implicit "s=intfor" option (*DATABASE_BINARY_INTFOR) which was causing a memory error.
- Fix SMP and MPP segment-based contact when there are thick shells in contact and also solid element form -1 or -2 in the data. The contact was assuming contiguous connectivity data for the solids and thick shells but this assumption does not work when there are solid form -1 or -2 elements.
- Fix a serious MPP segment-based contact error that occurred when shells with thickness strains are used. The error caused memory corruption. This error dates from update 78150.
- Improve the segment-based single surface contact search for thick segment pairs that are too close together. The code was not working well with triangluar segments. This change affects models with shell segments that have thickness greater than about 2/3 of the segment length.
- Fix a serious issue in the IGAPCL flag of *CONTACT_2D_AUTOMATIC_TIED when adaptive remeshing was used. The result of the bug was a disasterous movement of nodes to close the flag and a terminated job.
- Fix *CONTACT_2D_AUTOMATIC_TIED and _TIED_ONE_WAY in two ways. First, the positioning of slave nodes that are offset from the master surface was not done with sufficient accuracy and the offset distance could change a little. Second, it was possible for master nodes constrained by SPC's to move.
- Fix MPP segment-based contact error that occurred for 2-surface contact in both the initial penetration check and the periodic intersection check. The error was accessing incorrect memory for some segments which probably cause a failure to report intersections with those segments.
- Fix a problem with *CONTACT_2D_AUTOMATIC_TIED and adaptive remeshing where a gap was allowed to grow during remeshing.
- Fix a bug in MPP segment-based contact. A problem was found when contact occurs between segments on different partitions. It was likely that the scale factor used to scale the stiffness for the passed segments was the SFM parameter even if the segment was on the slave side or for single surface contact. If SFM=SFS in the data, then there was not problem.
- Fix *DELETE_CONTACT_2DAUTO when the contact interfaces did not have sequential ID numbers that begin with 1.
- For *CONTACT_2D_TIED and *CONTACT_2D_TIED_ONE_WAY, force penalty constraints if either side of the constraint belongs to a rigid body.
- Fix a memory error in MPP segment-based eroding contact that occurred with thick shells in the contact definition. This error caused a segmentation fault in phase 3 of initialization.
- Fix SMP segment-based contact sliding option and edge-to-edge contact that has been broken since rev 72144. It was crashing during initialization.
- Fix dynamic relaxation bug when IDRFLG=3. When *LOAD_BODY is used and if the rigid body is not in the part set selected for dynamic relaxation, it should not move.
- Fix frequency of output and final output state to spcforc for implicit (*DATABASE_SPCFORC).
- Fix for MPP output of nodal rigid bodies to *DATABASE_RBDOUT file.
- Fix cross section forces output for solid elements when *DATABASE_CROSS_SECTION is used with the PLANE option. The normal direction calculation was broken since R70315. This caused wrong forces as well.
- fix hanging for SMS with tet form 13
- Correct orientation check for cohesive elements (solid forms 19,20) in MPP, which was broken since r63986.
- Fix deformation and stresses in shells when using *DAMPING_PART_STIFFNESS with *CONTROL_BULK_VISCOSITY type -2.
- Fix a bug that could occur when part set for shell thickness update is used (*CONTROL_SHELL).
- Fix parsing problems with *SECTION_BEAM_AISC.
- Fix minor problem with tet form 13 volumetric strain calculation.
- Fix tet form 4 (*SECTION_SOLID: ELFORM=4). The combination with *DAMPING_PART_STIFFNESS was not working correctly, leading to wrong stresses or even error termination.
- Fix issues with *ELEMENT_BEAM_PULLEY:
- correct history update after element passes through pulley (swap)
- correct swap calculations for cable elements with *MAT_071 and E<0
- Fix bug in input processing of the DIRECT option for *ELEMENT_SEATBELT_SLIPRING, which was broken in r69902.
- Fix bug when both 4-noded and 8-noded EFG parts are used.
- Fix the use of *LOAD_BODY with *ELEMENT_MASS_MATRIX. The applied load was scaled using the nodal mass without the contribution of the added mass matrix, so the load was too small.
- Fix strain output of shell formulations 13-15 when 1-point integration is used or shell formulation 13 when 2x2 integration is used -and- the 2nd order objective stress update is active (*CONTROL_ACCURACY). The output to d3plot and elout was not correct for elements that rotate.
- Fix tet formulation 4 section force. The sign was wrong on the forces and moments.
- Fix parallel consistency for warped beam formulation 11. It was giving wrong results for all but small problems and could cause convergence failure of implicit problems.
- Fix the section force calculation to work with the implicit solid form 3. It was reporting nonsensical values.
- Fix the option to use a negative connection ID on *DEFINE_CONNECTION_PROPERTIES with solid weld assemblies. This option caused the failure function to be evaluated but the weld is not failed.
- Fix MPP hex spot weld assemblies by *DEFINE_HEX_SPOTWELD_ASSEMBLY when more than 8 elements are in the assembly. There was an error in the decomposition code that caused and error terminatation due to bad structured data.
- Fix a potential segmentation fault in thick shell form 3 (*SECTION_TSHELL) when used with isotropic materials in an implicit run.
- Fix the volume calculation for thick shell form 3. Prior to the fix, the calculated volume was 2x the correct value. This caused a wrong pressure calculation for equation of states 2, 3, and 14, which use the element volume in the denominator of the pressure calulation. It also caused the mass scaling calculation to use 2x the added mass that was needed for effective mass scaling.
- Fix mass scaling of thick shell 5 which added 8x the mass that was needed.
- When using *DATABASE_HISTORY_BEAM and *INITIAL_AXIAL_FORCE_BEAM with *CONTROL_SPOTWELD_BEAM, convert *DATABASE_HISTORY_BEAM to *DATABASE_CROSS_SECTION and *INITIAL_AXIAL_FORCE_BEAM to *INITIAL_STRESS_CROSS_SECTION for the spotweld beams that were converted to hex spotwelds.
- Fix a bug in *SENSOR_DEFINE_FORCE that could occur when vid is larger than 100000.
- Fix bug for *SENSOR_DEFINE when used with long format (long=s) in MPP.
- Fix a bug that could occur when *SENSOR_CONTROL is used to control *BOUNDARY_PRESCRIBED_MOTION.
- Fix a bug that could occur when *SENSOR_DEFINE_..._SET refers to a node/element set containing a large number of nodes/elements.
- Implement combination of *SENSOR_DEFINE_..._SET with *SENSOR_DEFINE_CALC-MATH.
- Add option to *MAT_058 whereby load curves for rate dependent strain values can be defined on new optional card 9. Abscissa values can optionally be natural log of strain rate (input negative curve IDs). Add optional transverse shear damage to *MAT_058.
- Add option to *MAT_058 whereby load curves for rate dependent strengths and a rate averaging flag can be defined on new optional card 8.
- Fix error in stress transformation via *INCLUDE_TRANSFORM for *MAT_058.
- Fix incorrect added mass for mass scaling when using *MAT_COHESIVE_MIXED_MODE.
- Fix output of triaxiality (history var #12) for *MAT_224 shell elements.
- Avoid division by zero in shell materials *MAT_024 (VP=1), *MAT_123 (VP=1), *MAT_224, *MAT_225, and *MAT_251 (VP=1). This was possible if table option was used with identical starting points in associated yield curves.
- Add *MAT_124 as potential weld partner material for PROPRUL=2/3 of *DEFINE_CONNECTION_PROPERTIES.
- Fix *MAT_VISCOELASTIC (*MAT_006) for type 13 tet elements. Before, it was not really working at all.
- Enable varying temperatures over shell thickness for *MAT_106, which fixes bug described as "*MAT_106 with shells does not bend under differential temperature loading".
- Fix spurious element deletion when using *MAT_81/*MAT_PLASTICITY_WITH_DAMAGE with *MAT_ADD_EROSION.
- *MAT_CABLE (and possibly some other material types) did not initialize correctly with Staged Construction. Problem occurred when starting from a non-zero time.
- Add support to *MAT_219 (*MAT_CODAM2) for negative AOPT values which point to coordinate system ID's.
- Fix deformable beam spot weld failure option 9 (*MAT_SPOTWELD) for a 3 sheet weld.
- Fix damage calculation of *MAT_221, *MAT_ORTHOTROPIC_SIMPLIFIED_DAMAGE. The strain calculation used to determine the damage and failure was incorrect causing failure to occur at too small a strain value.
- Fix RCDC damage of *MAT_082_RCDC. The internal parameter Ad was set to 1.9999 even in tensile cases when the first and second principal stresses are equal. Now this case is checked and Ad=1.9999 only when the the max princiapl stress is closer to zero than the other two, in other words with compression. When Ad=1.9999, damage growth is nearly zero so this fix should possiblly speed damage growth a little bit.
- Fix SMP and hybrid parallel consistency when using shell *MAT_239.
- Fix the use of *MAT_ADD_EROSION with thick shell forms 1 and 2. The messages reported include element IDs and either those IDs were were wrong, or a segmentation fault occurred when writing the messages.
- Fix type 6,7,9,and 10 hourglass controls for *MAT_076, *MAT_GENERAL_VISCOELASTIC. These hourglass types all use material constants and for brick elements of *MAT_076, a failure to define the shear modulus caused a divide by zero and NaN's.
- Fix a conflict in *MAT_FABRIC when form 14 is used with the negative ID for unload curves, and Autoliv venting is being used with a Wang Nefske control volume. The conflict caused a bad venting calculation leading to NaN pressure on the control volume.
- Add nodal contact stiffness for nodes from SPH element -- MPP only (*CONTACT_..._NODES_TO_SURFACE).
- Fix bug for SPH tracer particle output (*DATABASE_TRHIST) for SMP. Output was disabled when binary output is on.
- Properly support SPH moving box option in MPP (*DEFINE_BOX_SPH).
- Fix MPP bug whereby initial velocities were not initialized during transient phase when IDRFLG=5 in *CONTROL_DYNAMIC_RELAXATION despite presence of *INITIAL_VELOCITY_GENERATION.
- Fix long time error in the MPP implicit treatment of superelements (*CONTROL_IMPLICIT_MODES).
- Fix rigid bodies with prescribed motion combined with implicit dynamics. Basically do not place dynamic effects on rigid body dofs that are constrained.
- Fix computation of inertia for superelements created by *CONTROL_IMPLICIT_MODES.
- Add support for using explicit in dynamic relaxation and implicit in transient simulation.
- Add support for *DEFINE_FUNCTION and *DEFINE_FUNCTION_TABULATED to implicit for prescribed motion constraints.
- Fix the use of eigenmodes with *CONTROL_IMPLICIT_MODES especially using the Craig-Bampton approach for constructing superelements.
- Correct the generation of the header for the d3eigv, d3iter, and d3mode files for the case of MPP with predecomposition.
- Correct hand off to implicit for *BOUNDARY_PRESCRIBED_ORIENTATION when explicit-implicit switching is active.
- Fix *MAT_ADD_THERMAL_EXPANSION for thick shell form 5. It was off by a factor of the time step so the strain was totally wrong. Also, for thick shell forms 3 and 5, added support for themal expansion of total Lagrangian materials that use the deformation gradient.
- MPP decomposition: distribute massless nodes among all processors, instead of putting them all on processor 0
- Fix dynamic relaxation initialization with m=drdisp.sif in MPP if rigid body is shared among processors.
- Make important bug fixes and improvements for trimming.
Discrete Element Method
- Fix bug involving frictional forces between discrete spheres -- MPP only (*ELEMENT_DISCRETE_SPHERE)
- Remove input error message complaining of undefined beam # when there are discrete elements and BEAM=1 in *DATABASE_BINARY_D3PLOT.
- Fix bug affecting encryption when model includes adaptivity.
- Fix problem of not being able to disable plastic strain mapping in *INCLUDE_STAMPED_PART.
- Prevent output of *DEFINE_FUNCTION expression in case of encryption.
- Do not echo parameters of *MAT_ADD_EROSION in case of encryption.
- Use correct offset (IDSOFF) from *INCLUDE_TRANSFORM for *ELEMENT_SEATBELT_RETRACTOR option SBRNID<0.
- Fix combination of *DEFINE_ELEMENT_DEATH_... and *INCLUDE_TRANSFORM. Offset for EID/SID (ideoff/idsoff) was not correct.
- Fixed a bug in reading in rail track data (*RAIL) for MPP.
- Fix for some *DEFINE_CURVE_FUNCTION options that use model rotational velocities (WX,WY,WZ) that were broken in r70763.
- Fix *SET_SHELL_GENERAL input error if number of elements generated is >250.
- Improve the initial rotational velocity calculation of rigid bodies in the single precision version by promoting the moment arm terms to double precision. This probably will not be noticed except when the body is given an initial pure translational velocity so that the rotational velocity is zero. Before this update, a small rotational velocity was calculated.