momentum equation (Normal to wall)
 

Dear colleagues:

It is well-known that no-slip BCs (u,v,w =0) has been typically applied to stationary wall boundary.

For those of the tangential momentum equation, we can apply the laminar/wall function approach to simulate the effect of wall on the tangential velocity.

How about for the normal momentum equation? I am wondering abt the discretization of momentum equation (normal to the wall) at cell adjacent to the wall boundary, in a cell-centered SIMPLE solver.

The reason I am asking this question is: As I include the Buoyancy term into the z-momentum equation (parallel to gravitational acceleration), i can see some "rising stream" near the ceiling (due to T>Tavg) and "descending stream" near the floor (due to T<Tavg), which is physically non-existence after I have simulated the problem using commercial code. I have quite doubtful about my implementation of my z-momentum discretization at the wall..

Please help..