[christian_kunkelmann]

Dear all,

I would like to apply leastSquares schemes for grad(gamma) when I use the VOF solvers of OpenFOAM because I realized that they do a pretty good job on the mesh that I use. However, when I use the adaptive mesh refinement feature (dynamicRefineFvMesh) with interDyMFoam I do not get good results. The simulation runs well until the mesh is updated for the first time (after 10 timesteps in my case).

When I let the solver write the gradient field and visualize it in paraview I can see that the arrows point in the right direction in the beginning of the simulation. After the refinement, the arrows point randomly in almost any possible direction which causes the interface to become unstable.

Is it possible that the leastSquares schemes need some kind of reinitialization when the mesh has changed. If yes, how could this be implemented. I have alread tried calling mesh.fvSchemes::read() after mesh.update() but the results did not change.

Thanks in advance for your support.

Regards,
Christian

[christian_kunkelmann]

I just made an interesting observation:

When I stop the simulation after the first refinement and restart it from that time (i. e. with the refined mesh) the gradient are calculated correctly...

... which again lets me think that the leastSquares scheme apparantly needs some kind of reinitialization when the mesh has been refined.

I am looking forward to your comments/suggestions!

Thanks again!
Christian

[hjasak]

Fixed in 1.5-dev a long time ago.

Hrv