Newbie:Turbulence Modelling

Rajil Saraswat

Hi all, I am trying to model the metal flow in a continuous caster(very much like flow in a tank). From the papers it has been very much established that the flow is turbulent and people have used k-epsilon model for modelling it. I am trying to repeat the same thing but i am not sure about the right approach. Can you tell whether this is correct. 1. I first turned on laminar flow with just the Energy equation(default options for solver,first order upwind etc)using steady state. The solution converged quickly. 2. Then i turned on the fluid flow with unsteady solution but still under laminar. The solution was running fine but then i started getting reverse flows at the outlet after some time. 3. Can i now turn on the turbulent k-epsilon model?

Is this the correct approach? Should i turn the k-epsilon model in step 1 itself? Thanks for bearing with me. Being new into cfd, i very much need your help.