Bugs in LS-DYNA and user-material models

[Jorgen]

If you are developing user-material models for ls971 (both implicit and explicit), then you should be aware of the two following bugs :eek: that I recently discovered and reported to LSTC.


LS-DYNA only forwards a maximum of 48 material parameters to user-material subroutines. If the input file specifies more than 48 material parameters, then all parameters after the first 48 are initialized to zero. This happens without any warning!

If the input file contains multiple *Mat_User_Defined_Material_Models, then the nhv parameter of the first is used for all user material models. This is a serious problem for models that use the deformation gradient.


Let's hope LSTC fixes these problems soon. I'll keep you updated...

[Jorgen]

Here's another ls971 bug :eek::


If you are using a user-material model that is basing its stress calculation on the deformation gradient, and you are running an explicit simulation, then you cannot specify mass scaling using *Control_Timestep. If you do, then the simulation will crash on the first increment!

A workaround is to specify the mass scaling by simply changing the density on the *Mat_User_Defined_Material_Models card.