splitMeshRegions - Splits mesh into multiple regions.
Splits mesh into multiple regions (detected by walking across faces).
Each region is defined as a domain whose cells can all be reached by cell-face-cell walking without crossing
- boundary faces
- additional faces from faceset (-blockedFaces faceSet).
- any face in-between differing cellZones (-cellZones)
- volScalarField with regions as different scalars (-detectOnly)
- mesh with multiple regions and mapped patches. These patches either cover the whole interface between two region (default) or only part according to faceZones (-useFaceZones)
- mesh with cells put into cellZones (-makeCellZones)
- Specify additional region boundaries that walking does not cross by specifying a faceSet
- Split additionally cellZones off into separate regions
- Use cellZones only to split mesh into regions; do not use walking
- Do not write mesh
- Only write region containing POINT
- Only write largest region
- Place cells into cellZones instead of splitting mesh
- Try to match heuristically regions to existing cell zones
- Use faceZones to patch inter-region faces instead of single patch
- Overwrite the existing mesh files
- Run the utility in parallel
-roots "(DIR1 [...DIRN])"
- Directories through which the data are distributed
- Execute the command on the case directory DIR. If not provided, use the current directory
- Skip the execution of the functionObjects
- Display the help and exit
- cellZonesOnly does not do a walk and uses the cellZones only. Use this if you don't mind having disconnected domains in a single region. This option requires all cells to be in one (and one only) cellZone.
- cellZonesFileOnly behaves like -cellZonesOnly but reads the cellZones from the specified file. This allows one to explicitly specify the region distribution and still have multiple cellZones per region.
- useCellZonesOnly does not do a walk and uses the cellZones only. Use this if you don't mind having disconnected domains in a single region. This option requires all cells to be in one (and one only) cellZone.
- Should work in parallel.
- cellZones can differ on either side of processor boundaries in which case the faces get moved from processor patch to directMapped patch. Not the faces get moved from processor patch to mapped patch. Not very well tested.
- If a cell zone gets split into more than one region it can detect the largest matching region (-sloppyCellZones). This will accept any region that covers more than 50% of the zone. It has to be a subset so cannot have any cells in any other zone.
- If explicitly a single region has been selected (-largestOnly or -insidePoint) its region name will be either
- name of a cellZone it matches to or
- largestOnly respectively insidePoint or
- polyMesh::defaultRegion if additionally -overwrite (so it will overwrite the input mesh!)
- writes maps like decomposePar back to original mesh:
- pointRegionAddressing : for every point in this region the point in the original mesh
- cellRegionAddressing : for every cell in this region the cell in the original mesh
- faceRegionAddressing : for every face in this region the face in the original mesh + "turning index". For a face in the same orientation this is the original facelabel+1, for a turned face this is -facelabel-1
- boundaryRegionAddressing : for every patch in this region the patch in the original mesh (or -1 if added patch)
This utility is used in the following tutorials: