@aurelien

That's all ? So simple ? I thought the common surfaces between two regions need to be meshed exactly the same way (same points)... actually I thought they were the same mesh named in two different ways.

@antonessiu
I think you can try lower your refinement level requirement.
In the sHMDict I posted yesterday the setup is quite coarse...

Daniele

In "wonderfull land", you should use only structured mesh (the exact same surface mesh for both fluid region and solid region) to get the best results.

But in real life since I was not able to deal with SHM, an unstructured mesh is a first solution. If you use the same cell size for both regions, you will be able to limit the error from the meshes differences.

I have already checked OpenFOAM can deal with it on a simple test case. With SHM, The only problem that could appear is from the feature edge option. Sometimes SHM is a bit crafty. But with an other mesh tool, it would be easy.