Hi everybody, I just have a
 

Hi everybody,

I just have a 3D spherical sliding interface around a rotating object. The case is working using moveDynamicMesh. All schemes are default and boundary conditions are just symmetric and 1 inflow and 1 outflow, so very basic. Reynolds number is about 200....

The problem is: in controlDict I've put maxCo to 0.5, but icoDyMFoam doesn't seem to care. While iterating the timestep reduces such that meanCo is about 1e-5 and maxCo about 1e-3......

Has anyone observed similar problems, could it be related to boundary conditions or to the choice of schemes?

Regards, Frank

Hi Frank Good Evening I
 

Hi Frank

Good Evening

I have been facing the similar problem with my sliding interface.

The first guess is that the timeStep, meanCo amd maxCo are mesh size dependent. Haven't tried the same setup with a coarser mesh so cannot establish the dependence.

If you are using blockMeshDict and can coarsen you mesh quickly then keep the same solver settings as now and try different mesh resolutions.

Please let me know if you figure out the dependence.

With Best Regards
Jaswinder

Hi Jaswinder! I don't use b
 

Hi Jaswinder!

I don't use blockMesh, and I am aware that a moving mesh with smaller cells leads to smaller timesteps and larger Courant numbers.....

The problem is that I define maxCo in controlDict and the solver doesn't even come close to that value. It doesn't diverge either, the adjustable timestep just becomes very very small and the maxCo far below the desired value in controlDict.

Very strange, anyone any ideas?

Frank