Problem/ bug in mesh.cellCells() ??
 

I'm just working on a general possibility to refine locally on a selection of the mesh (blockMesh with refinementIterator - no hexref8). Since I have higher refinement levels I need to lay some buffer-layers around the selection in order to ensure 2:1 refinement. For this purpose I use the cellCells() function in multiple iterations.

My problem is that cellCells() does not recognize correctly all neighbours of already refined cells (The refinement history I have implemented works fine and has no bugs) .

I'll give a little example:

- 8 cells (= refinement level 3) on the west side of a simple refined cell (= 2 cells f.e. cell A and B).
- this should give cell A as eastern neighbour for four of the tripple refined cells and cell B for the other four cells
- instead of this cellCells returns only cell A as eastern neighbour for all eight cells!!

And now the really confuseing problem:

This bug is not reproduceable for all meshes in all cases, since the correct recognition of all neighbours works fine in the mesh of forwardStep (tutorials/compressible/rhoCentralFo...), but not in the mesh of counterFlow2D (tutorials/combustion/reactingFoam...) :mad:. I have no idea if the bug is caused by different mesh setups (I did not find any significant differences)?!?

I would really appreciate it if anyone who has deeper insight into the cellCells() - function could give me a hint.

Rudi T.


Please do not puzzle any longer!
Problem is solved: I forgot to update the primitive mesh after each buffering iteration :D

Is there already a solution to this problem? I have experienced too a similar issue.

Greetings Vitor,

daZigeiner did edit the post and add this to the end of it:
Therefore... whichever problem you're having, if that information isn't enough to solve your problem, then we (anyone on the forum) will need more information in order to diagnose the problem you're having.

Best regards,
Bruno