[askjak]

I am having great difficulties getting rhoSimpleFoam, rhoPisoFoam and rhoPimpleFoam to converge.

At low speeds the solvers converges to a nice solution, but at increasing speeds (but still with max Mach number less than 0.3-0.4 - no shocks!) the solver blows up. I am interested in Mach numbers between 0.1 and approximately 1.

The mesh is fine according to checkMesh and the solution easily converges using incompressible solvers (simpleFoam, pisoFoam). On exactly the same mesh in CCM+ the compressible solver converges with no problems in both the coupled/segregated solver case.

I have tried the following without success.

1) played with relaxation parameters.
2) started with a converged solutions from potentialFoam and simpleFoam.
3) very slowly tried to ramp up the inlet velocity (from 1 to 20 m/s over 40000 steps) using the BC timeVaryingUniformFixedValue.
4) Tried various boundary conditions on inlet and outlet.
5) Slowly ramped up unsteady solvers like rhoPisoFoam and rhoPimpleFoam, but they become unstable at some point.

I have noticed people claiming that rho*Foam are missing density corrections: see http://www.cfd-online.com/Forums/ope...implefoam.html and http://www.cfd-online.com/Forums/ope...-pressure.html. Recently a density predictor was added to rhoPimpleFoam (see git log for 1.6.x). Apparently did not help. Should this predictor not be added to rhoSimpleFoam?

I really hope somebody can help me so I don't have to solve the problem in commercial codes since we need to analyze a very large number of geometries. Maybe somebody can explain why these solvers are so "fragile".

Specific to rhoSimpleFoam: There is no tutorial for this and it needs a scheme for div(U,p) (see hEqn.H). What should I use?

Best regards, Ask