Please help! I am trying to finish up this project so that I can graduate in June! I am modeling half an aircraft at incompressible flight speeds and have tried to get the Spalart-Allmaras model to work for higher angle of attack cases, but my solution diverges after 9-10 iterations each time. I have ran an inviscid solution on the same grid and got a converged solution after a little more than 400 iterations that seemed reasonable. I have a feeling that it may be caused by some highly skewed cells present in the mesh. Is this correct? I have tried to minimize the number of highly skewed cells as much as possible, but can't seem to completely get rid of them. Would anyone be willing to look at my mesh in GAMBIT and give me advise? I would greatly appreciate it. I unfortunately do not have access to TGrid. Please email me for more details on the problem and things that I have tried. Thank you so much for any help!!!

Use size functions throughout your domain and also resolve the boundary layer with hex/quad (not tri) elements.

I'm using a curvature size function for the aircraft body surfaces and a meshed size function for the flow volume with the aircraft faces as sources. I fixed the skewness problem early this morning (found a ver small sliver that contained a bunch of skewed elements), so hopefully that will help. Any suggestions on using GAMBIT for the boundary layer versus adapting the grid in FLUENT? Any advice for either approach? Thank you!!

If using a fixed BL, do not exceed an aspect ratio of 1 : 1.2.