[sangeet]

Hello,
I am currently trying to do a FSI simulation of blood flow in an artery. I have tried using snappyHexMesh to mesh both the solid and fluid parts, but it seems there are gaps between the meshes at the fluid-solid interface. I am using fsiFoam and it gives an error of "master point addressing not correct" and also a very high GGI interpolation error. I suspect this is due to the gaps between the fluid and solid meshes at the interface.


Could anyone please advice on what would be the optimal way to create solid and fluid meshes for complex geometry for an FSI simulation?


I had also asked this question in the OpenFOAM forum but there were no replies.



Thank You

[ssyadav]

Dear All


I want to share my experience with respect to meshing for fsiFOAM. We modified the default 3dtube tutorial that comes with fsiFOAM so that now it is a full cylindrical pipe instead of a sector.



If we use blockMesh for generating the mesh in solid and fluid regions, the case runs in parallel without any problem. It does not matter in what direction we decompose using simple method.



But if we use snappyHexMesh and generate mesh using an stl file, the case runs well in serial mode but gives floating point exception error while we try to run in parallel.


We have different levels of refinement in snappyHexMesh based mesh with a fine mesh near the fluid-solid interface, but the grid points are common between the fluid and solid regions at the interface.


Does the solver gives error because of different levels of refinements?


If yes, how can we decompose the case so that it becomes aware of the refinement levels?


Thanks in advance.

[LuckyTran]

Originally Posted by ssyadav

But if we use snappyHexMesh and generate mesh using an stl file, the case runs well in serial mode but gives floating point exception error while we try to run in parallel.

Probably a grid quality issue when you decided to use snappy. Run checkMesh on the non-decomposed case for both grids and compare. Probably you will learn something already here.

You can also run checkMesh on the decomposed mesh and see if a particular partition of the mesh is causing issues.

[Hgholami]

Hello Sangeet
Did you solve this problem with snappyHexMesh for fsiFoam?
If you done it, give us your experiment?
Thanks

[ssyadav]

Dear Hojatollah Gholami

We moved to the preCICE library for simulating our FSI problem.

There we are using snappyHexMesh for fluid mesh in OpenFOAM and GMSH for the solid mesh for Calculix.

No problem while running in parallel now....

Check here: https://www.precice.org/