[jenksson]

Hello everyone,

I could not solve the solids4foam FSI overset tutorial (beaminCrossFlow) in parallel. I'm getting mpi segmentation fault right after the initialization of the newSubsetMotionSolverFvMesh (tried simple, scotch and metis decomposition methods). It works in serial without problem.
Does anybody have some experience or any information regarding this issue?

Thank you for any help.

Cenk.

[bigphil]

Hi Cenk,

As you noted, this is also discussed here: Support thread for "Solid Mechanics Solvers added to OpenFOAM Extend".

I have just added a new issue to the solids4foam repository issues page so that it doesn't get forgotten: https://bitbucket.org/philip_cardiff...do-not-work-in

If it works in the fluid-only cases then it should be possible to get it to work.

Philip

[Tianyz]

Originally Posted by bigphil

Hi Cenk,

As you noted, this is also discussed here: Support thread for "Solid Mechanics Solvers added to OpenFOAM Extend".

I have just added a new issue to the solids4foam repository issues page so that it doesn't get forgotten: https://bitbucket.org/philip_cardiff...do-not-work-in

If it works in the fluid-only cases then it should be possible to get it to work.

Philip

Hi, Philip.
Many thanks for your contribution to this open-source package for FSI, which makes my work much easier. I'm now doing a FSI simulation with solids4Foam, but got some problem.
My case is turbulent flow around long vertical flexible riser, with deformable overset mesh. To accormadate deformation of solid part, newSubsetMotionSolverFvMesh is imployed according to the tutorial case flexible beam (overset version).
But as you found, this method doesn't work well in parallel, which will take much more time to run if in serial, since the huge number of cells of my case.
So, I'm wondering is there any other avaliable model to replace this newSubsetMotionSolverFvMesh? Or can we try to fix it?
Thank you again for work!

Best,
Tian