k-omega: cold streak

Christian Fuchs

Hi all,

I have a quesion about how the turbulent kinetic energy counts to the Navier-Stokes (total internal) Energy budget.

My code is a RANS solver, and I implemented the Wilcox k-omega model into it. Now when k-omega is switched on, the temperature in the boundary layer drops, effectiveley producing a cold streak. During further convergence, heat flux will make this streak get weaker and eventually dissapear, but somehow I feel that this multipies the convergence time.

The reason for the temperature drop seems to be based on my decision, how to account for tke (turbulent kinetic energy); the total internal energy is e = cv T + u^2/2 + tke (all times rho of course). When turbulent production kicks in, tke increases, but e is kept unchanged - effectively there remains lesser energy for T, so temperature drops.

I am not sure if this is right; I assumed that tke feeds on e, not on p/rho (which is not included in the total internal energy budget e), and I also assumed that the total internal energy remains conserved when tke grows.

Is this correct, or should the budget be different?

Thanks a lot Christian