Hello, I'm trying to simulate a highspeed transsonic (fluid domain including areas with with local Mach Number greater and areas with local Mach Number lower than speed of sound) air compressor. Until now fluent was not able to reach convergence and altough not able to reach continuity of massflow in and out. Anybody who knows something that could help, please write your ideas down or contact me via email: tfroebel@gmx.de. Thank you and please excuse my bad english skills.

Here is some data to specify my simulation:

grid: 1,2 mio cells Solver Fluent6 3d, coupled,explicit Boundaries: rotor: total pressure inlet->mixing plane between rotor an stator-> stator: static pressue outlet mixing plane: conserve total enthalpy Fluid: air, normal conditions, 288.15K, 101325 Pa turbulence model: SST k-w initial guess: possible average values between inlet and outlet

Try segregated solver, with under-relaxation for Pressure set to 0.5 and Momentum to 0.3. Start 1st order, converge, then go to 2nd order. Also set your pressure to Second order and Converge. Then switch to Coupled.

Although not prefered, segregated should also work for transonic flows - but is very diffusive and does not capture shock locations well.

Thank you Riaan, I'll try to set up your ideas. If it works I'll report to you. Perhaps could you explain to me why Fluent does not capture shock locations well ? (You have to understand that even this feature seems to be very important for my application) Thanks, Tobi.

The segregated solver is very diffusive for shocks - the coupled solver captures it more accurately.

Check www.fluentusers.com under the Anual user group meetings for 2005 - there is a very good presentation on segregated/coupled solver use for transonic/supersonic flow

Thank you for that link, I found that document and am now trying to work with the recommended solver and URF. It seems to solve my problem more accurately and especially the solving process converges.

Thanks Tobias.