[nks]

I have developed k-epsilon solver for channel flow. Now I have to implement the v2-f model for the problem. I have been trying for quite some time but with no success. Please help me in this regard. The main issues are:
1.boundary conditions: At inlet, how to decide value of v2,k and epsilon? For the elliptic function 'f', at the wall I used -20nu2v2/(eps*y4). This results in quite large negative values of f.As a result v2 and f become unstable.
2.I then tried the modified model (Hanjalic 2001)such that f=0 at the wall. But now k becomes unstable, also the pressure field becomes unstable and code diverges after some time.
3. Source terms: How to linearize (in form S=Sc+Sp*phi p)them to get stable solution.
For example the k equation has source term G-rho*epsilon
in k-epsilon solver, I took Sc=G*deltax*deltay and Sp=(-rho2*Cmu*k/muturbulent)*deltax*deltay, since muturbulent =rho*cmu*k2/epsilon.
How to do in v2f model.
Please help me in this regard.