I've been trying to get a multi-surface plasticity model running, without finally succeeding.
I am coding the algorithm Simo & Hughes give in "Computational Inelasticity" in Box 5.2a-c into an Abaqus-umat written in fortran. I apply it for perfect plasticity (at the moment), therefore it is simplified from the original.
Unfortunately, my code does not provide proper results even in a quite simple example (two surfaces only) that can be more or less followed.
Nevertheless, the algorithm does work if only one surface is active. These results are verified by a single-surface model. I assume therefore, that the algorithm is OK in principle, only the combinatorial terms are somehow not working properly. I have also rewritten the code in mathematica which yields the very same results - wrong ones in case of multiple active surfaces.
Questions to the community are:
- Is the algorithm given in Simo & Hughes working at all?
- Has someone experienced some troubles with this as well? Were there any minor/major/obvious/hidden misinterpretations or errors in understanding the procedure? Any hints what one can make wrong?
- Does anyone know of a similarly well-documented algorithm for multi-surface plasticity?
- Any other information that may be helpful?