hello

there is some problem in giving boundary condition on w for internal flows. can anybody give reference of some literature on boundary condition for w in near wall turbulent flow.

For example, see Menter's article with his k-w SST model F.R.Menter, Two-Equation Eddy-Viscosity Turbulence Models for Engineering Applications// AIAA J. v.32 N 8, 1994, pp. 1598 - 1605 where he gives the following boundarty condition for w at the solid wall: w = const_1/(const_2*delta*delta), where "delta" is the distance to the next grid point away from the wall.

See Wilcox's book, Turbulence Modeling, 1998, 2nd edition.

I´ve found that the surface roughness model as presented in Wilcox's book (see post by Steve Ciesla) works best. It is more stable to use than the theoretically correct 1/y^2 asymptotic behaviour and gives as good or better results. If you use a sand grain roughness kr+ which is less than say 4 you will get a "smooth wall" result. I've used this type of boundary condition for many turbine blade heat transfer simulations. All my experience is for very well resolved walls though, there are a few extra things to think about if you have poor wall resolution.