Methodology for turbulence model sensitivity
 

Hi all,

In my research (buoyant air flow in rooms with cold inlet air) I see that the influence of turbulence models is rather large (more important than the mesh sensitivity). I used standard k-e, RNG k-e, realizable k-e and SST k-w.
I do not have validation data available, so have to try to explain the differences by analyzing the theoretical background of the models. I am reading literature and studying the theory, but would still like some help on how to start with analyzing my results.
I would start like this:

• Plot of streamlines
  • Logical flow pattern?
  • Compare flow defining features, like recirculation length, attachment length, separation point location, stagnation point location,.. to quantify differences

• Plots of turbulent quantities: compare and try to explain differences based on the theoretical model (tips and tricks on this?)
  • Turbulent kinetic energy k
  • Turbulent dissipation rate e
  • Turbulent dissipation frequency w
  • Turbulent viscosity μt
  • Wall shear stress tw
  • Production of k
  • Effective viscosity
  • Blending functions for SST k-w (available in Fluent output?)

I realize this is very difficult, but I think here the experience from long-time users is very valuable.
Thanks for your input!

Sarah