SimpleFoam Race Car case, solution diverges
I've been using simpleFoam to analyse our FSAE car and then later develop an aero kit for it. I'm using a k-w SST model and see below for problem outline. I have 7 patches/ .obj's, 4 wheels, chassis+driver+engine, nose and side panels, and suspension. They are refined (6 7) except for the chassis which is (5 6) and each cell outside refinement region is 1m cubed.
The omega is wrong in the image, using 0.622, reduced size of boundaries as well to x=60m, y=10m, z=8m. Refinement box is also slightly reduced. So I have run it at 12.5m/s and it worked ok, you can see the residual below. I've tried many variations and simplifications of geometry, however the oscillations don't disappear. I believe this is cause by vortexes being shed from the wheels.
The blue line represents 12.5 m/s, and the red 15 m/s - basically I want Cd, Cl, and Cm values, however I want to ensure they are velocity independent which is why I'm running for velocities above and below 12.5.
The settings are identical for both except @ 20m/s, k = 0.135, w= 0.745. my fvSolution file is based off of the motorBike tutorial, and the only changes are relTol = 0.001, and tolerance is at 1e-09. fvSchemes is untouched and is identical to the motorbike case.
as you can see the solution blows up, and it happens 9 times out of 10, with different settings, I've no idea why this is happening or how I can prevent it, any help would be appreciated. I am running this case in parallel and used the scotch method for decomposing. Using 4 cores.
Right so i narrowed it down and it could be the mesh causing this, I get an error with skewed faces, any help in solving, or fixing this mesh would be appreciated.
|All times are GMT -4. The time now is 10:24.|