Stranger than Fiction

[Kumar Kurambakurash]

Dear Jorgen and all,

Recently I have been working on a subroutine (VUMAT) for finite strain elastoplasticity which utilizes generalized strain and stress measures.

I have done my 1 element analysis with error-free results. Then I have started to solve more complicated ones.

For 3D analysis, although the code works quite well until certain deformation levels, there comes a point where the analysis looses stability and crashes.

As an example, for a 3D necking problem (triggered by a geometrical imperfection in the case of Simo and colleagues), I have quite good results
which preserves symmetry, however then after certain point the symmetry is lost, as far as the distribution of equivalent plastic strains are concerned.

I also see that the loss of symmetry is always in a certain direction. I mean if the circular bar has a x-section located on 1-2 axis and longitudinal axis is 3, the accumulation of equivalent plastic strains occur on axis-1 at the face.

What do I have to suspect at this point. Can you help me with solutions or can you suggest me some benchmarks?

Thank you for your time,

Kumar

[Jorgen]

Stranger than Fiction - I kind of liked that movie

How distorted are the elements at the time you reach the instability?
Does it help to use smaller time increments?

I would try different elements types, e.g. plane strain, and displacement BCs.

Unfortunately, debugging a VUMAT can be quite frustrating

-Jorgen

[Kumar Kurambakurash]

Dear Jorgen,

Thank you for your interest. I have solved the problem.

I have recently found that the error was in one of the shear component of the push-front transformation of the Second PK tensor. This error is hidden quite well that it really was frustrating to debug .

Best

Kumar

ps: Stranger than Fiction is a nice movie, I agree