Getting GGI-ready mesh from conversion
 

I am trying to use Pointwise to create structured meshes for a GGI case (which I run unstructured grids with successfully). The openfoam export does not seem to be ready for ggi (includes 'boundary','faces','neighbor','owner','points'). So I try to export it to fluent and convert it.

My question is, has anybody converted a fluent mesh for ggi?

Yes. What exactly is your problem?

Regards Bastian

Yes I have converted some fluent-meshes to Foam with fluent3DMeshToFoam.

But take a look on your converted patches - maybe you have to change them (for example from "wall" / "patch" to "ggi") in OpenFoam!

Regards

Thanks guys for the responses.

So whenever I convert the mesh it doesnt seem to be ggi-ready and im not sure what to do.

I noticed that 'cellZones' is created but it is empty. Is this an essential file for ggi?

you can get "CellZones" with the command "splitMeshRegions -makeCellZones".

In your boundary-file you have to define the desired ggi-patches - take a look on the validation test cases to get an overview of setting up a case!

http://openfoamwiki.net/index.php/Si...ion_test_cases

In the "centrifugal pump" there is described how to create ggi patches correctly.

Thanks guys for the replies. I have been able to make progress with you all's help, but still I face a problem. The following is the procedure I used to generate the mesh case: (ggi, there exists a rotor and stator)

1. Pointwise 3D mesh OpenFoam export for 2 grids (rotor and stator), 5 files each (boundary, points, faces, neighbor, owner)
2. mergeMeshes
3. splitMeshRegions -makeCellZones
4. setSet -batch setBatch (file 'setBatch' contains: (line 1) 'faceSet insideZone new patchToFace insideSlider' (line 2) 'faceSet outsideZone new patchToFace outsideSlider' (line 3) 'quit')
5. setsToZones -noFlipMap

Then i run it using 'decomposePar' and 'foamJob -s -p turbDyMFoam'.

It runs for a while, but then things start going crazy on the readout, because i have not set it up for ggi to work properly, please see the attached images of the ggi interface.

Anybody know what I am doing wrong? I will be trying to debug this and try splitMeshRegions on a Pointwise export with both rotor and stator in one mesh (differently defined cell 'sets') to see if that works.

do you set up the "dynamicMeshDict" file?
please post your boundary- and the dynamicMeshDict-file and the log-files of checkMesh and your run , so that we can get an overview of your problem!
What happens,if you run the solver with only ONE CPU??

Attached are the 3 files you requested. I will try serial run and let you know how that goes.

The case is an o-grid external flow (so it has a circular outer boundary).

Did you adapt your field-files correctly (in the field-files there are "boundary-fields" too, and they must have the same entries as your boundary-file!!)?

In the cellZones-file, the ROTATING ZONE MUST have the name "movingCells" - ckeck this out! (If you use a transient solver,you do not need to create a cellZones file, mixerGGI... will do this for you!Delete the cellZones-file,start the solver and check the things I described above).


...and please post the log of your calculation!!