Hello all,
we are working on a forming simulation with high degrees of forming. In order to work out the right settings and the correct procedure, I have first created a planar abstracted model.
The challenge is that the mesh distorts so much during forming that regular remeshing is needed during forming. Although the simplified model converges even without remeshing, the results show a strong distortion of the mesh. You can find the results attached in this and the following posts. I have packed a video (Forming.ogv/Forming.mp4) and the simulation result PlaneContact.rmed into the attachment.
We assume that the distortions in the 3D model will get worse and therefore rely on remeshing.
I have already done first tests for the remeshing with HOMMARD. These are based on the test case SSNP158. This was a super template and I have extended this test case for the use of nonlinear contact. I would also like to share this simulation model with you. I have put it in the Attachment. The associated files are named PlaneContact.*
Also this simulation with remeshing is already running quite well. Currently, however, I wonder if HOMMARD can accomplish the kind of remeshing we need in this case.
It seems to me that HOMMARD only ever splits the existing elements when the criterion of the chosen error estimator is exceeded. Moreover, this seems to happen on the undeformed mesh alone.
It is true that the distorted mesh also makes the values of the error estimator worse and the mesh would be refined at the appropriate points. However, at this time it does not appear to me that this would make the mesh less distorted. But even if that were the case, the mesh would become finer and finer with this approach and that wouldn't be effective.
What we actually need is simply a remeshing of the deformed geometry with the same mesh refinement as the initial mesh.
Here are my questions:
-> Do anyone of you have experience in code_aster with the type of remeshing described above? Would she or he then share his experience with me ;-)
-> Does anyone have any idea how we can implement the above type of remeshing with HOMMARD?
-> I'm considering if we can use another mesher (e.g. Salome or directly Netgen) instead of HOMMARD at the same place in the COMM-File or use the Python library. Does anyone have experience with this and would share it with me?
I think it is a very exciting topic and I would be happy to discuss the possibilities here.
Many thanks in advance,
Tobias
