GGI parallel run
 

Dear OpenFOAM users,

I am doing a study on computational capabilities of GGI, AMI and MRF using a small wind turbine. I have started this project with a test case and so far I am successful in running the simulation 'pimpleDyMFoam' with single processor i.e., Serial run. Thanks to the entire community.

Now I am trying to simulate the same case to save the computational time using cluster, and as part of simulation it went very smooth, but while reconstructing the case I am getting the following error:

Face zone name cylinderInletZone for GGI patch cylinderInlet not found. Please check your GGI interface definition.

I have seen all the posts regarding this error for solution but I could not get the solution. I have checked my boundary file as well as my setBatch file. Please let me know the possible errors which you came across. I followed the procedure as stared below.

1. Created seperate mesh regions for bot rotating and non rotating parts and merged them together using mergeMeshes.

2. Renumbered the case and changed the patch names to GGI wherever required.

3. Then the toposet has been run using setSet -batch setBatchGgi (This and boundar files has been attached) and the zones are defined with setsToZones -noFlipMap

4. Decomposed the case using simple method.

5. potentialFoam and pimpleDyMFoam has been used.

6. reconstructParMesh here I am getting the error

Mesh motion detected. Reconstruct motion points


--> FOAM FATAL ERROR:
Face zone name cylinderInletZone for GGI patch cylinderInlet not found. Please check your GGI interface definition.

From function label ggiPolyPatch::zoneIndex() const
in file meshes/polyMesh/polyPatches/constraint/ggi/ggiPolyPatch.C at line 815.


The boundary, change dictionary and Topo files are

Attachment 61738

Attachment 61739

Attachment 61740

It would be really of a great help if you can give me some guidance regarding the above issue.

Thank you all.

Hi Metikurke,

did you add the GGI faceZones that you created with the setBatch utility in your decomposePar dictionary? Why do you use the simple method instead of for example scotch decomposition?

For what specific reason do you want to reconstruct only the mesh? Typically I only use reconstructPar to reconstruct my results.

A good reference for rotating machinery simulations is the rotating machinery training at the OpenFOAM workshop of Prof. Nilsson from Chalmers:

https://pingpong.chalmers.se/public/...ingOFW11-1.pdf

I hope this already helps.

Kind Regards,

Andreas

Hi Andreas,

Thank you very much for the reply

1) did you add the GGI faceZones that you created with the setBatch utility in your decomposePar dictionary? Why do you use the simple method instead of for example scotch decomposition?

Yes I have added all the zones as globalFaceZones,

globalFaceZones
(
cylinderInletZone
cylinderOutletZone
circumferenceZone
cylinderInletInnerZone
cylinderOutletInnerZone
circumferenceInnerZone
);

The simulation was not started with scotch and I got highly nonphysical courant number and velocity magnitude

2)For what specific reason do you want to reconstruct only the mesh? Typically I only use reconstructPar to reconstruct my results.

reconstructPar was not working without doing reconstructParMesh during serialRun.

I have followed this tutorial during my study. But thank you very much.

It would be really nice if you can share more information on the above error

Regards,

Kanaada Metikurke