CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM (https://www.cfd-online.com/Forums/openfoam/)
-   -   Keeping a cellZone on the same CPU when decomposing (https://www.cfd-online.com/Forums/openfoam/229776-keeping-cellzone-same-cpu-when-decomposing.html)

MikeWorth August 24, 2020 06:58

Keeping a cellZone on the same CPU when decomposing
 
I'm running a moving mesh case, with layer addition/removal. Things are initially looking good, but when most of the layers have been removed from processor0 things go haywire.

Suddenly velocities jump orders of magnitude in a single timestep (I've seen an increase by 12 orders of magnitude over a single timestep!), when I look closer checkmesh is giving me warnings about very high aspect ratio cells. I think this is because the last layer of cells on the processor (there are plenty more, but on processor1) is refusing to be deleted. Trying to reconcile having those cells, but there being no space is I think resulting in terribly mangled meshing.

If I can force decomposePar to keep all of the layer removal cells on the same processor, then my problem should go away. Does anyone know how I can achieve that?

Thanks,
Mike

fumiya August 30, 2020 06:49

manual method of decomposePar
 
Hi,

You might want to try 'manual' decomposition method of decomposePar utility.

Hope this helps,
Fumiya

olesen September 7, 2020 12:19

Decomposition constraints might be what you are looking for.
Eg in


- tutorials/lagrangian/reactingParcelFoam/filter/system/decomposeParDict
- tutorials/preProcessing/decompositionConstraints/geometric/system/decomposeConstraints


etc


All times are GMT -4. The time now is 09:22.