[besh]

Hi OFers
I have some questions about wall function implementations in OF.

a) For the Launder-Spalding wall function, first we calculate u_* from k , then we solve epsilon/omega equation fixing the near war cell value, then solve k and momentum equation. I see this in OF where epsilon is fixed

epsEqn().boundaryManipulate(epsilon_.boundaryField());

For the standard wall functions where u_* is not calculated from k, but iteratively solved from log-law equation we need to fix both k and epsilon. But I don't see k being fixed , why ?

b) What if we fix the values at the boundary cell (not the first near war cell). It seems to me that this avoids
- The need to add contributions from two or more adjacent walls
- Manipulation of eps and k-equations to fix values
It may give bad results with coarse mesh though.

c) Finally are the LES wall functions same as RANS? I could not find wall functions directory in LES. What is the basic difference between nut and nusgs at near wall cells when both are modelled?

Thanks for any help