Tom

August 21, 2003 06:03 
Boundary conditions for KEPS
Hi, I'm implementing a standard twoequation ke turbulence model in a former code that I've written which simulates laminar supersonic flow over a flat plate at low Reynolds number (1000). The code has successfully captures a shock wave over the plate. I want to now change the dimensions of the plate and inflow conditions to simulate subsonic, fully turbulent flow (at high Reynolds numbers). At the wall I'm using a standard wall function approach which solves for k as a function of tauw (shear at the wall) using loglaw. Can someone suggest to me what are the boundary conditions for k and eps at the inflow and outflow and initial values throughout the computational domain before running the code? (I assume there are some explicit values for k and eps that must be specified or obtained by extrapolation from inside the domain. I know that specification of k and eps can be done directly, so how to do this? Basically I want my code after each iteration to calculate velocity, density etc in the computational domain and then using these known values, I will calculate the scalar quantities k and eps and then use them to calculate my turbulent eddy viscosity and then update my modified viscosity term mu=(mu_laminar + mu_turb + mu_artif) which is found in the viscous terms of the momentum and energy equations. Is this a good approach? I'm more interested in writing my own programs at the moment so that when I start to use commercial CFD software, I know what the fundamentals are. Any help will greatly be appreciated! Kind regards, Tom
