[Selinus]

Hi everyone,
I have to simulate a steady state compressible problem with OpenFoam.
I generated new meshes for a grid independence analysis (coarse,medium,fine).
rhoSimpleFoam solver gave me many problems to obtain convergence with this new meshes.
I decided so to use rhoPimpleFoam with high CFL in order to eliminate the transient term in the equations.


Usually I started with CFL=1 and then gradually increase the value (50-100)
It worked fine for coarse and medium runs.
Instead I have big problem to obtain convergence with fine mesh.


My questions:
1) What do you think about using rhoPimpleFoam with high CFL value?
2) Do you have any ideas could help me to overcome the problem and obtain convergence in fine simulation?


Thank you



Best regards

[cryabroad]

1) It all depends on what you want. I've been running cases with high CFL (10-50) as initial guess for the more time-consuming LES I will do later. In that sense, I don't need a very accurate outcome, so high CFL is okay. If you are running a serious case (meaning that you want to get actual results out of if), small CFL is desired. Or, at least you should run the case with smaller CFL to see if you are missing some information.
2) This depends on your coarse and medium mesh cases, did the results get improved as the mesh is refined? If yes, changing the spatial schemes may be a good start.
A more general question though, are you sure that your case is truly steady state? My experience is that, if rhoSimpleFoam is very very hard to converge but rhoPimpleFoam converges easily, there is some form of unsteadiness in the system.