Problems with extrapolation of pressure in BC

Sergio

Hi,

I am running a Jameson-type of code (cell-centered finite volume, Runge-Kutta time stepping, Jameson scheme for convection, residual smoothin, etc.) for a turbulent flow calculation around a wingbody configuration. I have a problem with the no-slip boundary condition. Since the code is cell-centered, the pressure must be linearly extrapolated from the center of a boundary cell to the wall in order to compute momentum fluxes at the wall. This procedure is very instable in separated zone at the base of a vehicle. Why is that? In these regions velocity is low and pressure is nearly constant, so linear extrapolation of pressure should not be a problem.

Best regards,

deltaquattro