Hi,

I am trying to simulate transient flow around an exhaust valve and I am studying the tutorial Fluid Structure Interaction and Mesh Deformation. I am thinking to use Junction Box Routine in my case. Analogous with the description in the tutorial, I tried to create several meshes in ICEM-CFD, but I can't grasp which file types are in the meshes directory (CheckValve.0 - 19).

I am not sure, if the mesh files must be of the same type as in the tutorial, but I am not that good in FORTRAN, so I am thinking, that way I won't have to modify the FORTRAN subroutines too much.

I am thanking you in advance for any suggestion or explanation.

Hi,

Why are you trying to do FSI on an exhaust valve? If you are trying to model the valve motion to open and shut the valve you explicitly define the valve position using mesh motion. There is no need for FSI.

Glenn Horrocks

Thanks for your response. I am not trying to do FSI, I just thought to automize the mesh replacing during the calculation. With mesh deformation, I have to stop the calculation before the mesh colapses(or it gets too bad), load another mesh manually and use results from previous step as initial conditions for the next one. So I thought it would be easier to prepare a set of meshes beforehand and and reload them for each time step using a Junction Box Routine (JBR)as described in the tutorial. If I misunderstood the procedure, please correct me. Is it possible to combine a mesh deformation and JBR?

Hi,

What you describe can be done with pure mesh motion as long as the mesh topology does not change. Have a look at the moving mesh tutorial problem as it shows doing this sort of stuff either by CEL expressions and user fortran. If you can do it purely by CEL expression it is much easier.

A hint on the moving mesh: Generate your meshes at the squished up condition and stretch them out rather than generating the meshes stretched out and you squish them up. The moving mesh algorithm by default stretches meshes more reliably than it squishes it up.

Glenn Horrocks

Hi,

I tried just what you suggested, I stretched the mesh from the "closed" valve position, but at some point quality of the mesh gets too poor to describe physical phenomena correctly, so the mesh should be reloaded with new quality mesh at some point (new position of the valve).

And that's the thing, I don't know how to automize this. It would be great, if I could combine mesh deformation and JBR, but I think it can't be done. And now I can't even use the JBR, because I can't grasp what is the type of the mesh files (CheckValve.0 -.19) in the sub-directory meshes of the tutorial 20.

Hi,

You cannot do remeshing with a new mesh topology with a JBR. You need to stop the simulation, remesh and restart the simulation interpolating the results onto the new mesh.

This can be automated in a batch file.

Glenn Horrocks

I thought so.

Thank you for you time, Glenn.