Variation between kEpsilon and realizableKE
 

Hi,
I am trying to simulate a incompressible turbulent round jet impingement with different RANS models.Case1: Fig1 shows the velocity magnitude which is on par with the results that I need and the turbulence model used here is the standard kEpsilon. Case2: Now, I have kept everything else same and changed the RAS model to realizableKE and the resulting velocity magnitude is shown in the Fig2. The solution converges with 24,000 iterations in the first case and the solution converges within 7000 iterations in the second case. I could go about changing the residuals and the fvSchemes but, I wanted to understand, how for the exact mesh and boundary conditions just a change in the RAS model could cause in such huge variation in results and what could possibly cause them.The solver used was buoyantBoussinesqSimpleFoam with version 2.1.x. I have tried other models including kOmega, kOmegaSST also but results were similar to case 2. Any thoughts would be appreciated.

Thanks.

Any ideas at all???