I am trying to simulate radial compressor with compressible ideal gas in StarCD 3260.

Does anyone has any suggestions regarding turbulence model, time step, boundary conditions, initialisation values.....

My simulations are crashing due to negative densities.

I tried mass flow b.c. for inlet and pressure b.c. for outlet, but it keeps crashing. Also I tried pressure b.c. for inlet and nothing.

Target compressor speed is 140000 rpm.

I think that it is not a mesh problem, since my model works fine up to 90000 rpm.

I also tried restarting from pst file for 90000 rpm to increase the speed but always the same problem, negative density.

I would appreciate any suggestions.

Thanks in advance.

Bojan

Hi,

May be that you need to reduce the time step size after 90000 rpm case. Since it happened to me once, I could solve the problem this way. Hope it works !!

Thanks, but I have already tried that. Even I tried with variable time step to keep Courant number below 2. I think I have to increase rotational speed gradually in order to avoid negative densities.

You should use the underrelaxation of the piso algorithm. (algo piso node 20 0.15 0.5) or lower. You can also try with "switch 60 on".

Thanks Blake, but I also tried using 0.5 for underrelaxation factor and it didn't help.

I think that the time step is too large. you should keep the Courant number below 1 or less.

I tried that also, even with variable time step to keep Courant number below 1, but simulation crashed. I have sent my model to adapco support and I can't wait to see results.

Thanks for help!!!!

Hi,Bojan Stojanovic!

When you get the method,I hope you can tell us. I have also come across the same problem.

Thanks!!