[amtri]

Hello,

An OpenFOAM mesh is stored under the polyMesh directory. I'm trying to understand the precedence for which polyMesh directory is to be used in an OpenFOAM file structure.

I have the following directories, each with the files it contains:

1) constant/polyMesh:

blockMeshDict boundary faces neighbour owner points

2) processor0/constant/polyMesh:
boundary boundaryProcAddressing cellProcAddressing faceProcAddressing faces neighbour owner pointProcAddressing points

processor1/constant/polyMesh:
boundary boundaryProcAddressing cellProcAddressing faceProcAddressing faces neighbour owner pointProcAddressing points

3) processor0/0/polyMesh:
boundary cellLevel cellZones faceZones faces neighbour owner pointLevel pointZones points refinementHistory

processor1/0/polyMesh
undary cellLevel cellZones faceZones faces neighbour owner pointLevel pointZones points refinementHistory

My questions:

a) I assume that constant/polyMesh is the first polyMesh in the model, but not necessarily used. It was probably constructed prior to domain decomposition/refinements.

b) The polyMesh directories in case (2) above reflects the domain decomposition: I can see the communication files XXXProcAddressing in these directories.

However, I was under the impression that if a polyMesh directory exists under a time step directory - 0/polyMesh - then all results in that time step will refer to this mesh.

But this mesh does not have any of the communication files XXXProcAddressing. This is, however, a much finer mesh and I can see from the results files under the processor[i]/0/ directory that they refer to this mesh.

But there are no communication files available. Shouldn't there be communication files for this mesh? Aren't these files necessary for any polyMesh directory under processor[i]? If not, how is the communication information determined?

Thanks.