Interface for MRF
 

I have a question about the interface geometry between the moving volume and the stationary volume in a turbomachinery analysis I am working on.

I have an axial fan, and I have modeled a structured hexahedral mesh around the blades aligned along the fluid flow direction. Now I have to section the mesh into a rotating volume (enclosing the blades) and a non-rotating volume for inlet and outlet bulk volume.

Does the interface between these two volumes has to be a plane surface or it can be any surface as long as the nodes at the interface are connected?

It depends on the solver you are using. Most CFD codes will interpolate the solution from one set of faces to another. In OpenFOAM this would be accomplished by using an AMI interface where appropriate weighting factors are computed for interpolation. A perfect weight would be a case with a one-to-one point/face matched interface and no interpolation would be necessary. You should always try to generate the interfaces so they are as close as possible to one another (exact if you can) in terms of cell size, shape, and distribution.

Thank you for your response Mr. Carrigan. I understand how the exchange of data between the two faces takes place.

I have a structured mesh in my problem (as mentioned earlier) and in my mesh the elements at the interface are exactly similar on the adjoining surfaces. I will add a picture here to clarify the problem.

As visible in the attached image, the surface between the two attached domains (yellow and pink) is not a plane face. Would that create a problem for the solver because one of these domains is moving and the other is not?

My solver is Acusolve.

Ah, I think I understand the problem now. Yes, this is not a candidate for an MRF simulation. The interfaces (both stationary and rotating) must be coincident throughout the simulation. You'll need to pick a different interface or adjust your computational domain.

Thank you again for replying :)