For calculating the plasticity, damage (and other models with a criterion in a point), you should not use linear elements. In a linear element, the derivatives of the displacements (strains and calculated from them stress) are constant throughout the element, and all values calculated from their values at the point (yield limit and etc) are mesh-dependent.

I am not sure whether it is correct. However, in my opinion, you may require to address the characteristic length in order to avoid mesh-dependency.

Dear Simun,

The problem seems to be due to change in mesh along the vertical direction (with the assumption that the model size and feed is same in both cases). If we use mesh1 (same mesh along vertical surface) with increase in number of elements along thickness, the simulation should run. Let us know the outcome.

Of couse, we can have one element, but need to respect aspect ratio as well. In the case of mesh1, reduce the thickness of the workpiece to the size of the first layer element (~ 1 aspect ratio).

Regards,

Chandrashekhar

First, tank you all for the response!

Dear Alexander I understand that linear elements can prescribe linear displacement so deformation is constant so I didn't used them in this model.

Dear Chandrashekhar, I will try this and let you know what are results.

Dear Adsam, can you explain more briefly how should I define characteristic length?

Do you have some literature or rule of thumb advice??

Many thanks!