[kyle]

Would anyone be willing to share their experiences using OpenFOAM for automobile aero or underhood simulations? I have been evaluating this software for the better part of two months trying to get a feel for if we can replace Star-CCM+. Everything works great on our validation cases, but more complex geometry is troublesome. SnappyHexMesh is working for me very well. The mesh is at least as good at what I can get out of CCM+ and can be generated an order of magnitude faster. I can get very good results with simpleFoam and kOmega SST, but my main gripe is that I end up babysitting every run to convergence. I have to fiddle with solver settings and monitor max values for U/p/k and remove cells as they get out of control before they pollute the entire simulation. Any fvSolution/fvSchemes combination I get to work with one mesh seems to immediately diverge on the next mesh, no matter how similar the mesh.

I have spoken with engineers at other companies that have very robust OpenFOAM setups, but of course they were unable to give me any detailed information on how they got to that point. I am hoping someone here is able to share a little more. I'd love to hear about different workflows that are being used successfully and if possible even see some fvScheme/fvSolution/snappyHexMeshDict/other files that are being used in the wild. Failing that just let me know whether or not I have unrealistic expectations of this software!

Thanks for any information you can provide!