CFD Online Discussion Forums - Doubt about BC for k and epsilon

Hi, everybody,
I'm trying to do set up a 2D RANS simulation of a methane flame using OpenFOAM.
Actually, my domain is rectangular with, on one side, an inner inlet for fuel and an outter inlet for coflow.
For my case, I needed to impose a profile for the BC of velocity for both fuel inlet and coflow inlet (as shown in the attached image). Because I want to use the k-e model for turbulence, I need also to define boundaries for both k and epsilon. And here is my problem: maybe the simplest way is to put constant values for both of them; but in my case I'm simulating turbulent combustion so, since I have a profile for mean velocity, I must have also a profile for k and then for epsilon.
So, epsilon could be defined from k, and k could be defined from the normal stresses (u'u', v'v', w'w') of the reynolds tensor.
Finally, is there a way to estimate those stresses from my well defined mean velocity profile? Is there an approximated correlation for that?

Thank you in advance, ask me if something isn't clear in the explanation!