ENH: Adding subMesh option to momentumError and div FOs

• Adding subMesh capabilities to momentumError and div FOs.
  • A subMesh is created from cellZones to compute FOs.
  • The operators (div, etc) are only calculated in the subMesh.
  • Zero field is written in the non-subMesh cells.
• Optionally, halo cells can be added to the cellZones.
• New helper class to handle the subMesh creation and field mapping.
• TUT: add an example to tutorials/compressible/rhoSimpleFoam/aerofoilNACA0012

changed milestone to %v2106

added development label

added 1 commit

• cf8521f2 - ENH: Chaning fvSchmes for momErr calculation.

Compare with previous version

added 1 commit

• 4c1457f9 - STY: SQUASH IT BEFORE MERGING

Compare with previous version

added 1 commit

• 694f77d5 - ENH: Fixing haloToCell for parallel.

Compare with previous version

@Mattijs , I can get away without using interpolated values on the new patches. But the default value on the new patches is pretty random and sometimes zero which makes the operation on that field crash. We need to set a value different from zero. At the moment, don't think any application uses it apart from my FO. But, as I am creating layers on around the cellZone the wrong values on the cell next to the patch are deleted in the mapping. So, the error created by the wrong values on the patches is not seen.I had to assigned an arbitrary value of one for surf/vol field to avoid crashing the operators.

@mark , Small change in haloToCell for parallel. See last commit please. As well optionally have 0 steps rather than 1 as minimum.

• please change subsetMesh app back to use empty by default. My change to subsetMesh was just a hack to demo the new fvMeshSubset functionality
• do any assignment to the exposed fvPatchFields either outside the interpolate function (since only the driving utility knows what value it should be) or have a generic dictionary constructor as for the setCellSubset.

added 2 commits

• e67346a9 - ENH: Adding sub-mesh capabilities to div FO.
• ad030021 - ENH: Adding zoneSubSet on functionObject tools folder

Compare with previous version

added 59 commits

• ad030021...30da494e - 52 commits from branch develop
• 6db2078f - ENH: Adding zone sub-mesh option for momentumError calculation
• a5becd2c - ENH: Update tutorial aerofoilNACA0012 for using momErr in SubMesh
• ef540b76 - STY: Minor style changes and tutorial update
• 254903ef - ENH: Chaning fvSchmes for momErr calculation.
• dea9a86c - ENH: Fixing haloToCell for parallel.
• 4311e001 - ENH: Adding sub-mesh capabilities to div FO.
• 42bcf34c - ENH: Adding zoneSubSet on functionObject tools folder

Compare with previous version

had been doing a clean compilation - any ideas?

-/src/functionObjects/field/lnInclude/fieldExpression.H:74:24: fatal error: zoneSubSet.H: No such file or directory
 #include "zoneSubSet.H"

log.linux64GccDPInt32Opt.gz

I can see that src/functionObjects/tools/zoneSubSet/zoneSubSet.H is there.

Also, a .swp file crept into the commit as well - we might need to remove it? !463 (42bcf34c)

added 1 commit

• 48721c10 - STY: Deleting .swp file

Compare with previous version

@kuti Pushed to clean the .swp file. I think you need to do a full clean of the function objects first. zoneSubSet.H should be in the tools/lnInclude folder as well.

hmm.. but note that I had been compiling the branch on a new git clone - so it was clean.

added 1 commit

• c81c03cf - BUG: Reverting subMesh modifications

Compare with previous version

@Mattijs . Done

@kuti . the ./Allwmake in src/functionObject should run a lnIncude for the tools folder.

@Sergio, any chance for you to double-check the compilation, please?

Prashant has encountered the same error as well (forwarded the e-mail to you).

His additional comment:

using multiple processors for compilation can trigger this issue

Build report
============

Build problems encountered:
/scratch/pss/1515045/OpenFOAM/OpenFOAM.feature-momErrSubMesh.int32DP/src/functionObjects/field/lnInclude/fieldExpression.H:74:24: fatal error: zoneSubSet.H: No such file or directory