[MarcoC501]

Dear all,

I am a newbie in foam-extend 5.0 and I would like to simulate the NASA 37 stage (rotor + stator), which I attach here for those who might not be familiar with that.

As standard practice to solve this steady-state problem, we can model only one blade passage and apply an interface treatment (e.g. mixing plane) between rotor and stator (to notice that the interface does not have the same area). In CFX, with the geometry attached and a good mesh the solution is well-converging and it leads to reasonable results. It is also important to point out that the stage is transonic and highly turbulent.

I wanted to simulate the same problem in OpenFOAM: after successfully simulating only the rotor, I now would like to add the stator and I tried several methods (all are using k-epsilon method and a robust PISO algorithm):
- steadyCompressibleMRFFoam with ggi interface: it diverges after around 100 iterations and the source of divergence seems to be the interface itself;
- steadyCompressibleMRFFoam with MixingPlane: it is crashing very fast, i also tried very low / 0 rotational speed but no improvement.

Have you ever had experience with that solver? What the problem might be in your experience?
If needed I could attach my BCs and the other files.

Thank you so much for your help :=)

[MarcoC501]

Apparently the source of all my problems is the pitch difference between stator and rotor. In fact I tried a simulation with the full annulus and both interface modeling approaches are woking.
It apprears strange to me that this problem does exist not only for the ggi interface but also for the mixing plane.
According to your experience, do you confirm that the mixing plane does not work with different pitches / number of blades?
Thanks a lot for your help and support

[Yann Scott]

Hi Marco,
I am facing a similar issue. Did you find a solution back then?

In my case I am modelling (in openFOAM2412 with foam-extend add-in) one rotor blade and guideVane of an axial fan that originally has 6 blades and 13 guideVanes. Therefore the areas at the mixingPlane interface don't match (rotor_outlet area stator_inlet).

Any clues on how to run my case? What interface boundary condition should I try next?

Downunder the error msg of decomposePar:

--> FOAM FATAL ERROR: (openfoam-2412 patch=250814)
MixingPlane interface error: (rotor_outlet, stator_inlet)
Detecting at least one mixingPlane ribbon or one mixingPlane patch facet without any neighbours.
This is a fatal error. Some possible solutions:
- Review your mixingPlane interface definition
- Adjust your mesh in the area of the mixingPlane patches
- Adjust your user-defined profile to cover the full extent of the mixingPlane patches in the stacking direction
- Adjust your runtime parameters like the GGIFeatureCosTol tolerance factor


From virtual void Foam::mixingPlanePolyPatch::calcPatchToPatch() const
in file OpenFOAM/meshes/polyMesh/polyPatches/constraint/mixingPlane/mixingPlanePolyPatch.C at line 297.

Update:
Use similar layer discretization before and after the mixingPlane, so the first ribbons have elements on both sides (if the option bothPatches was sellected). Otherwise, map the ribbons according to the coarser mesh side.