Did anyone succesfully conduct an FSI simulation using preCICE with an overset grid in OpenFOAM?
Hi,
as far as I am aware, no one has yet. Theoretically it should be possible. preCICE only knows information about the coupled interface (coordinates of the vertices, values on the surface etc.). Whether there is some kind of grid overlap does not/should not concern preCICE. With most overset grid methods, there is a background mesh, and another mesh around an object that is at least a few cells “deep”. Essentially preCICE would only need to know about the interface points on the object, which should stay constant. As long as the only cells that are altered are the ones far away from the interface, then that should be fine.
Derek Risseeuw used the overset grid functionality of OpenFOAM (v1712) in his Master’s Thesis Fluid Structure Interaction Modelling of Flapping Wings (TU Delft, 2019). A summary is also presented in a paper of the same at the ECCOMAS Coupled Problems 2019.
I am not aware of anyone else at the moment, but maybe someone in our community is currently active on this. I am marking this as “resolved”, but please feel free to add more resources here in the future.
(later update: I was wrong about overset being used in Derek’s work, but Julian Schliessus did make it work.)
Hi, is there any benchmark tutorial for preCiCE with the OpenFOAM overset meshing? Any help regarding this will be appreciated.
Best regards,
Chandan
No, but we definitely do need one: Needed: Tutorial with OpenFOAM and overset mesh · Issue #335 · precice/tutorials · GitHub
Feel free to add ideas in the issue!
Hi,
I don’t think it’s too hard to implement fluid-structure coupling using the overlapping grid functionality OF OF, but precice seems to support only a few solvers, for example displacementLaplacian supports precice, but, Openfoam multiSolidBodyMotionSolver solver does not support precice.
You can refer to the article DOI 10.1088/1748-3190/ac1cee, which appears to implement the fluid-structure coupling using overset grid.
Of course, I’ve also been testing overset grids recently, and if that goes well, I’ll post a result for your reference.
1 Like
Wanting to use the rbfMeshMotionSolver with preCICE, @philip realized that it was actually not that hard to modify to give it the same interface as displacementLaplacian. I don’t think it is a matter of fundamentally not supporting one or another mesh motion solver, it is mostly a matter of technical integration, which can be fixed at different layers (e.g., in the mesh motion solver, or in the adapter).
If there is a particular need for the multiSolidBodyMotionSolver, I guess we could try together to make it work. But is it necessary to use this mesh motion solver when running simulations with overset meshes? I don’t have any experience with that so far.