Hi,
I am working with FSI coupling between SU2 and a structural mesh, using the mapping:rbf-thin-plate-splines . The mesh movement/deformation seems to be only around the interface and the fluid elements close to the inteface look pretty distorted, as in this picture:
The mesh movement in a similar case with OpenFOAM (such as the vertical flap tutorial) seems to me much more “diffused” and the elements less distorted. I would like to ask:
- Are there differences in the mapping with different solvers?
- which kind of mapping would be better in my case? I have read something about rbf:gaussian and/or the rbfShape.py file but I am not sure how to use it
Thank you
Claudio
Hi,
I could have found the answer myself. There are grid movement parameters in the SU2 config file that can be tweaked to make the mesh stiffer: in particular, afaik, DEFORM_COEFF (and maybe the DEFORM_STIFFNESS_TYPE) can change considerably how the SU2 mesh deforms.
Claudio
Hi Claudio,
the mapping in preCICE is not related to the mesh motion in the SU2 solver. The mapping configuration only tells preCICE how to map data between non-matching mesh interfaces. In the case of a displacement boundary condition, one solver will compute the displacement on the structural solver, and preCICE will map this displacement to the interface of the fluid solver (in this case SU2). How the solver moves the internal mesh is done within the solver and not preCICE.
To find which mesh motion options would be better for this case, perhaps the SU2 forum on CFD Online would be a better place to look. You can have a look here.
The mapping only really plays a role when you have a much coarser mesh in the structural solver that the fluid solver. In this case, nearest-neighbor mapping results in a very blocky interface in the fluid solver. This is when nearest-projection of RBF is better.
Regards
Kyle
Hi Kyle,
thank you very much. That confirms what I discovered.
Claudio
