I have a question regarding the choice of support radius using rbf-compact-tps-c2. I have a mapping setup, where the fluid writes force to the solid mesh and reads displacement from the solid mesh. See config visualization below.
The solid mesh is coarser than the fluid mesh. Do I have to specify different support radii per mapping? If the support radius is applied to the solid nodes for reading the displacement, doesn’t it have to be smaller than the support radius for writing the forces to the solid mesh? I drew up my understanding of the method below. First mapping from the fluid mesh (yellow) to the solid mesh (blue) would require a larger radius to include enough solid nodes. Mapping from the solid mesh to the fluid mesh would be a smaller radius, right?
FLUID TO SOLID
SOLID TO FLUID
You generally need to consider enough mesh points to construct the RBF. If you want to map from the fluid to the solid (i.e., you know the values on the fluid, but not on the solid), then you need a radius that includes several fluid points.
If you want to map from the solid to the fluid, then you need a radius that includes several solid points.
You create the interpolum on the other mesh, then you evaluate your points on that interpolum.
By the way, do you know about this tool?
Thank you for this explanation. It seems I misunderstood the documentation before.
I do indeed know this tool, but since I was using rbf-compact and not rbf-gaussian, I hadn’t been using it before. I know it might be hard to make a blanket statement, but would you recommend one mapping method over the other?
There is an extensive section about the performance of different mapping methods in the preCICE v2 paper https://open-research-europe.ec.europa.eu/articles/2-51/v2
Thanks! I will read up on it