turbulence in compressible cases
 

Hi all,
I've been having some trouble with understanding the turbulence models in the compressible case, I have a cell that is essentially a cuboid with a central pillar cut out of its central (long) axis, so a slice through it would look like:


Excuse the ASCII art. So the central pillar patches are defined as walls with zeroGradient boundary conditions for the pressure and noSlip conditions for velocity, while the outer top and side patches are open boundaries, with zeroGradient boundary conditions for both the pressure and velocity.

So, the solver I've been running is essentially mhdFoam combined with rhoPisofoam, as in the paper by Carlos Xisto (https://doi.org/10.1016/j.jcp.2014.07.009), so there are magnetic fields applied to generate the velocities.
The issue I've been having is that the velocities always diverge: all of the turbulence properties I set in the /0/ folder (k,nut,omega for RAN omegaSST and k,nut for LES) seem to be calculated, so their values immediately starts to decrease from the internalField values I set on the first iteration and turbulence seems to have less of an effect. For the walls I have a kqRWallFunction, nutkWallFunction, omegaWallFunction set for k, nut, and omega respectively, and all have zeroGradient set for the outer patches, with internal fields set to 1e5, 9e-3, 1.11e7 values for the internalField of k,nut,omega respectively.

My question is how do I increase the effect of the turbulence on the velocities to damp these instabilities? I don't really understand where to set constants to increase the effect of turbulence since every variable (k,nut,omega) seems to be calculated, should I be setting the outer patches to fixedValue? Should I just increase the viscosity mu in thermophysicalProperties?
Thank you.

bumping for response