CFD Online Discussion Forums - hel (turbulent viscosity ratio limited) for supersonic combustion problem

hey all,

Problem description:

I'm trying to model the supersonic flow in a 2D scramjet combustor. I used a configuration as a mid-injection of H2 at sonic conditions (M=1) and a Mach 2 freestream conditions. The inlet inputs can be found at the end. I used the k-epsilon turbulence model since the turbulent kinetic energy and dissipation rate are given in the experiment ( where I got the information of the inlet from ). I used a non-premixed combustion with specifying all the data and calculating the PDF table for the present problem. The mesh is really fine ( I think fine enough ) and do not think the problem relies there, so I think my turbulent BCs are somewhat wrong. I also specified the exit plane as a pressure outlet, but got a reversed flow there too. There are many issues with the simulation so if someone has good ideas of a way to get rid of the errors please let me know. Many thanks and please find the inputs below.

So my questions now are:

1- How to get rid of the turbulent viscosity ratio?
2- What is the best turbulence model for the imposed problem?
3- Shall I change the turbulent BCs to something like Hydraulic diameter and intensity ?
4- How to get rid of the reversed flow? I'm still not clear of how to specify the backflow conditions.
5- Shall I increase the limit of the viscosity ratio ? or is that unrealistic?
6- How can I calculate the mean mixture fraction of both fuel and air? The user manual does not provide a formula for this.

I need some help please. Thanks to all.

Air / Hydrogen
Mach number 2 / 1
u m/s) 730 / 1200
T (K) 340 / 250
P (10^5 Pa) 1 / 1
density (kg/m^3) 1.002 / 0.097
mass fraction (O2) 0.232 / 0
mass fraction (N2) 0.736 / 0
mass fraction (H2O) 0.032/ 0
mass fraction (H2) 0 / 1
k (m^2/s^2) 10 / 2400
epsilon (m^2/s^3) 650 / 10^8