[jiaodanuma]

Hello, all

I encounter a problem which puzzles me a lot. I generated the trimmed mesh in star-ccm and then converted it to openfoam. I checked the quality of mesh using checkMesh in openfoam and everything seems fine. However, the case diverged suddenly during a few time steps, when it ran about 13.5 round trips in the computational domain. Before that, the CFL number is about 1.2~1.3 and the results is promising. The geometry is a simple cube.

Then, I enlarged the first refinement box trying to include the whole recirculation region behind the cube and keep everything else exactly the same. The difference between the mesh can be seen in the attached pictures. In the common sense, the mesh quality should be improved after the enlargement of refinement box. Unfortunately, the case diverged even more quickly... less than a round trip

Does anyone have any ideas about what's going on? Thank you very much.

Best wishes