I’m running a fairly simple setup in a 3D box that uses a world builder input file and the viscoplastic material model. I’m currently trying to use AMR in the model but when I have periodic boundary conditions on the sides I get a cycle of constraints error when testing in debug mode. The AMR will work if periodic conditions are false, so I’m wondering if it’s an issue with my setup?
Could you post the header of your log.txt file that contains the versions of ASPECT and so on, and an example parameter file that reproduces the issue? Does the issue disappear if you do not use the world builder? I know we had a problem with constraints at periodic boundary conditions in one of the developer versions of ASPECT (independent of the world builder) a few years ago, but I was not aware it is still an issue.
Here is the header and parameter file. The issue still occurs if I don’t use the world builder. I also tried changing the viscous flow law and still got the same issue.
original (1).prm (4.8 KB)
I think I encountered the same problem in a different (unrelated) issue:
This is a bug in deal.II that is tracked here: https://github.com/dealii/dealii/pull/10604
I currently do not have a fix, sorry. But with your reminder, I hope to have this in the next deal.II release.
I am not sure what the best way to work around it is at this point: adaptive refinement across a periodic face in 3d seems to be the problem.