Problem with Mixing plane
 

Hello there,
I am trying desperately to run a simulation of a centrifugal compressor with ideal gas CO2. The machine is running at operating condition (around 80 bar) with ideal gas.

I managed to have a converged solution with density based solver, Implicit formulation and second order upwind schemes for all the fluid properties. However, I am using Axial AREA Averaged mixing plane.

If I try to restart the simulation switching to Mixed-out mixing plane (as recommended by the fluent manual), even with very small CFL and under relaxation parameters, the simulation blows up.

If I look at the results just after applying the mixing-out average mixing plane it seems like a very high pressure is generating around the mixing plane.

Does anyone have any experience using mixed-out mixing plane? I do not see a solution to this problem: The mixed-out mixing plane seems very unstable and as soon as an adverse pressure gradient is built the simulation crashes because of backflow!!! Please, help