[fxzf]

Hello,

I am currenlty using rhoSimpleFoam to solve my case which is body with very hot jet flow at rear (temperature of jet flow is around 800K) under low speed flow field Mach < 0.1. rhoSimpleFoam can get results. But residual and CL, Cd are very oscillating.

The p is only get to 0.01 level. The problem is Cl, Cd quantaties is oscillating about 5% which make results hard to use.

The solver for p I used is GAMG with preconditioner. The relaxation I used for rho is 0.01, for p is 0.3.

Does anyone has better idea to improve the convergence, at least to reduce the Cl, Cd oscillation?

Thanks very much.

[kZaq]

Hi,

I recently fought with rhoSimpleFoam too. Oscillations happen pretty much always, because very few compressible cases have a steady-state solution. So your results can ever be only so accurate.

What I did was find the timestep when the solution is close to convergence, but just before it starts to oscillate, and at that timestep drop the relaxation factors significantly. For example, 0.01 for p and rho fields and 0.1 for all equations. This gives results that can be compared reliably from case to case.

The relaxation factors can be modified on the fly

[fxzf]

Originally Posted by kZaq

Hi,

I recently fought with rhoSimpleFoam too. Oscillations happen pretty much always, because very few compressible cases have a steady-state solution. So your results can ever be only so accurate.

What I did was find the timestep when the solution is close to convergence, but just before it starts to oscillate, and at that timestep drop the relaxation factors significantly. For example, 0.01 for p and rho fields and 0.1 for all equations. This gives results that can be compared reliably from case to case.

The relaxation factors can be modified on the fly

Great. Thank very much for your reply. I tried this strategy and it improved a bit but then need a long time to converge.

By the way, do you have any reference paper about rhoSimpleFoam. I want to figure out how this solver works.