Turbulence Models.
 

I have been asked to use a "Realizable k-eplsilon" model in my industrial project for my masters. Does anyone know what is the fundamental difference between this and the standard "k-eplsilon" model? All comments appreciated. Cheers.

Re: Turbulence Models.
 

Hello,

The name Realizable k-epsilon model is used by Fluent for the Shih, Zhu and Lumley evolution of the k-epsilon model. It refers to several NASA/ICOMP papers. The final and "really" realizable version of this model is describe in the following paper "Developments in Computational Modeling of the Turbulent Flows" - Tsan-Hsing Shih - NASA Contractor Report 198458 - ICOMP-96-04; CMOTT-96-03 - February 1996.

The main features of this model are the following :

- uu = 2/3*k - 2*c_nu*k^2/eps*dU/dx can NOT be negative since c_nu is no more a constant.

- the production term of the epsilon equation is equal to c_1*eps*S instead of c_eps_1*c_nu*k*S^2, where c_1 is NOT a constant.

This last item gives a better stability to the turbulence model.

Hope that helps,

Sylvain

Re: Turbulence Models.
 

simply speaking, realizable k-e models impose some conditions, usually derived based on some turbulence physics arguments, to prevent k and e (or some parameters derived from k and e) from going to wrong values...