Hi all

I am implementing a k-epsilon turbulence model in my general purpose CFD code. Here is the situation. The code engine employs an ADI-Brian solution to a general advective-diffusive transport equation. It is general because I port to the solver a vector of states and sources to derive a solution.

Q = [u,v,w,T,P] with sources, [P_x, P_y, P_z, Q, B] and respective diffusion coefficients

The pressure solution is derive by a parabolized extention of the PPE.

With this formulation I plan to port the additional [k, epsilon] states. I think I can figure out the source terms (they appear to be well documented). However the question is boundary conditions for [k, epsilon] states. My code engine can take on either standard Dirichlet or Neumann type BCs. I appear to be reading conflicting information on the type of BC for [k e].

I must stress that I am looking for something simple. Let's keep this easy. I have low Re numbers incompressible flows over obstructions. Generally I have defined velocity BCs for an inflow region (and zero for solid walls) and defined P-BC for an exit region. Dirichlet BCs for velocity force a Neumann BCs for pressure and vice versa. Given that what would you recommend for the BC formulation for k and e.

Thanks in advance.

Look for Reichardt law.

Or if can read french, I could send you Bc's and an algorithm for calculating the BC's.

(1).For turbulence modeling, please see David C. Wilcox book, "Turbulence Modeling for CFD". (2). Make sure that you understand the definition of k and epsilon. There are a couple of versions of epsilon being used. Also the nature of the equations require special attention in terms of solution procedure.

Hi... Perron... I have same problem with Schrage, so if you don't mind please send me your Bc's and an algorithm for calculating the BC's. thank you...!!!

ssutikno