Need help with a compressible flow problem
 

Hello Everyone,

I need to model flow / temperature / pressure in a long, sealed off (closed from everywhere), air-containing cylinder. There is a propeller fixed inside the cylinder, which stirs up the air to subsonic (but close to 330 m/s) speeds. Pressure and temperature effects are important, and the flow is modeled as compressible.

The system was meshed in GAMBIT; mesh quality is high ( << 0.7 skewness).

In FLUENT, I tried both Segregated and Coupled solvers in implicit formulation. Enabled Energy equation and then tried Viscous --> Spalart-Allmaras (standard) with default values and vorticity-based production OR Viscous --> k-epsilon (standard) with standard wall functions both times viscous heating was enabled.

To produce compressibility, the fluid was modeled with ideal-gas law.

Problems start when I have to set operating pressure. I am unsure how to define it in a closed, turbulent system. Trying to set it to zero brings many problems - the solution is unsteady (or diverges), even though I tried decreasing the Courant number to 1 as well as decreasing all the URFs and choosing 'First order upwind' in all options. My limits are quite reasonable too.

If I set the operating pressure to 1atm, the solution converges with solution parameters at their default values and 'second order upwind' enabled. That's OK, but I want to be sure in my results... and the 1atm operating pressure makes me uncomfortable. Besides, most residuals converge to below 1e-3, but keep oscillating (for many, many cycles) between 1.5e-3 and say, 9e-4.

Please let me know what do you think; I really appreciate your help.