[ViktorKL]

Hello,

I am modelling a multi region mesh consisting of a solid and a fluid part in OpenFoam.
It is supposed to depict a duct structure.
The interfaces between solid and fluid part (INTERFf & INTERFs) are of "regionCouple" boundary type and transfered to zones by (eg. for solid region):

setSet -batch solid.setSet -region solid
setsToZones -noFlipMap -region solid

with solid.setSet:

faceSet solidFluid new patchToFace INTERFs

The same goes for the fluid region.

For using a customized conjugate solver for conform meshes on this model, owner and neighbour cells of the defined zone must be processed on the same processor.
So, I wanted to use the preserveFaceZones constraint in decomposeParDict like explained in:
https://github.com/OpenFOAM/OpenFOAM...composeParDict

I used:

 
preserveFaceZones (fluidSolid solidFluid);

in both the decomposeParDict of the fluid region and the decomposeParDict of the solid one.

When now decomposing the mesh automatically (eg. using simple, hierachical or scotch) the preseveFaceZone does not have any influence on the decomposition.
The single decomposed meshes are still inconsistently distributed.
As an example the zones for one processor are depicted on the following picture, where red shows the solid and blue the fluid zone.
For my purpose, these zones should be overlaying.



I executed the decomposition first with fe4, then of4, fe32 and later of5 without any improvement.
Unfortunately, I also did not find any instructions or similiar topic on a forum explaining the preserveFaceZones constraint in detail. Therfore I opened this thread.

Is there any possibility using an automatic method for my purpose?
Maybe I did some severe mistakes in defining the zones or executing the decomposition.
Right now I am doing the decomposition manually with the setFieldsDict.
Thus, an automized decomposition would save me a lot of time in the future.


Thank you for your help in advance.


Kind regards,
Viktor

[Saleh Abuhanieh]

Hi Viktor,


Did you solve it?

[ViktorKL]

Hey,
Unfortunately no! I am still reliant on the manual decomposition of my multi-region meshes via the setFieldDict scheme.
Best regards,
Viktor

[Saleh Abuhanieh]

Hi Vektor,


Could you please explain how you are decomposing manually?


I am facing the same problem, I opened a thread while ago:


Segmentation fault initiated from a region-to-region BC

[shanyeyun]

Hi,

Have you solved the problem?
May I ask that how do you view mesh in each processor? After I decompose the mesh, when I tried to view in paraFoam it failed.


Thanks a lot

[ViktorKL]

Originally Posted by shanyeyun

Hi,

Have you solved the problem?
May I ask that how do you view mesh in each processor? After I decompose the mesh, when I tried to view in paraFoam it failed.


Thanks a lot

Unfortunately I haven't found a solution yet.
You can view single segements of the decomposed mesh by executing parafoam in the respective processor# subfolder.