CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Community Contributions (https://www.cfd-online.com/Forums/openfoam-community-contributions/)
-   -   [waves2Foam] waveFoam coupled with AMI sliding interface (https://www.cfd-online.com/Forums/openfoam-community-contributions/208044-wavefoam-coupled-ami-sliding-interface.html)

surfer November 17, 2014 11:34

waveFoam coupled with AMI sliding interface
 
Dear waveFoamers,

I have a problem in using waveFoam solver coupled with AMI interface. I would like to analyse how a wave influence the rotation of a simple laminar plate around a fixed point in terms of its rotation velocity; the plate is partially submerged in water. I create a simple mesh made by two regions: an external region which is fixed, and an internal region which is allowed to rotate around a fixed point which is placed at the interface between water and air; the rotating plate is placed into the rotating region. On the interfaces between these two regions I imposed "cyclicAMI" boundary condition in order to allow the solution to be interpolated and the wave to pass from one region to the other. Unfortunately the simulation blows up: after some iterations, when the wave starts entering the rotating mesh region, I note a sudden increase in velocity: this increase continues inside the moving region till unphysical values.

I made some tests, and the problem arises even if the rotating mesh region does not move, that means even if I use waveFoam instead of waveDyMFoam and in dynamicMeshDict I set the staticFvMesh option.

Furthermore, I made another test (with interFoam), where I imposed a rotating fixed velocity for the plate, and I assumed the water as completely at rest at the initial time without simulating any wave propagation. AMI set-up is the same as in the previous simulations, generating a sliding interface between the static and the rotating mesh region. This simulation runs perfectly without blowing up.

So it seems to me that the problem arises when I set up sixDoFRigidBodyMotion options in dynamicMeshDict and I couple it to waveFoam solver.

Does anybody of you have similar problems? Does any of you have any idea about where I am going wrong?

I than you a lot in advance for your help.

Kind regards,

surfer

ngj November 17, 2014 15:21

Good evening,

It is really hard to know, but you could try with a damBreak coming from outside through the AMI-interface.

Another thing. I cannot remember whether I ever made support for topological changes in the mesh, and you might need that with AMI? It should be fairly simple to check, if you look in the relaxationShape objects.

This, however, cannot be the problem, since it also crashes under static conditions. There are a lot of possibility: Mesh quality, courant numbers, schemes not compatible with both rotating meshes and waves, etc. With the rather limited information on your system it is hard to help.

Did you solve the problem with 2.3.x?

Kind regards,

Niels

surfer November 18, 2014 03:29

Dear Niels,

I thank you a lot for your quick answer and support.

Yes, I run the case with 2.3.x, but this is due to the fact that with the 2.3.0 version crashes too in the same manner. I found a thread where they talk about a commit that was done for correcting a bug in OpenFOAM files where AMI is implemented: the bug concerned the behavior of the interface tracked by interFoam solver across cyclicAMI patches. And so I decided to pass to the version 2.3.x, in order to include this commit.
But nothing has changed.

I will try to increase the mesh quality in order to see if this is the problem. I hope so. I let you know about the results.

Thanks a lot again.

Kind regards,

surfer


All times are GMT -4. The time now is 21:06.