exhaust flow from noozle
 

Hi, I am trying to solve this problem from past couple of months. Please help me on this flow problem.
Constant velocity inlet - 125 m/s at 850K
Pressure far field or opening - 1 atm at 300K
Velocity inlet is the exhaust from the turbine which passes through lobed nozzle into atmosphere. So i am using single domain and ideal gas as fluid.
I am using 1mm thickness on each wall and 30 degree periodic faces. so max quality is 0.25. Using cfx to solve this, but results are not converging. Should use k-omega SST model. At first tried using fluent but failed due to temperature divergence. Sometimes results converge to 1e-4 at 150 iterations. but results do not match. and sometimes residuals straighten out. Do i need to change anything or did i miss out any BC?
If you see the result file ISO region i got is lengthier than original. I think the hot flow is not drawing in cooler air into flow path. What should i do to make iso region shorter?