my problem is how to incorporate the boundary conditions rather how to specify them in case of low re k-e model. the the original paper by Launder and Sharma says that dissipation rate at wall is zero but dissipation is finite (obviously) what does this mean..since you solve through the viscous sub-layer, k is taken as zero for yplus <10. my question is should u solve for k if yplus <10 and how to give the boundary condition for e. can any body give good refrence.

What is best to do depends on your numerics. In a cell-centered finite volume code you usually solve for k and epsilon in all cells, including those below yplus 10 or whatever. On the wall, which corresponds to a cell-face, you set k equal to zero in the flux computation - your cell closest to the wall gets whatever value you compute, you just use the wall value to compute a flux. What you should set epsilon to on the wall depends on which Low-Re model you are using. Many models solve for a modified epsilon, which is zero on the wall. Other models solve for the real epsilon, which goes as nu*d^2k/dy^2. This can often be approximated by 2*nu*(d(sqrt(k))/dy)^2, which is simpler to compute and more stable, so use this! In any case you should also solve epsilon in all cells down to the wall.

The Launder-Sharma model that you mention solves for a modified epsilon, which should be zero on the wall.

The best review article that I know about for classical Low-Re models is written by Patel, Rodi, Scheuerer - "Turbulence Models for Near-Wall and Low Reynolds Number Flows: A Review" AIAA Journal, Vol 23, No 9, pp 1308-1319. It is a bit outdated by now though.