Hi

I am running a incompressible channel flow simulation and the NS solver is by the fractional step method. The discretization is by finite volume method. I gave the velocity field periodic boundary conditions

my question is:

should the pressure boundary conditions also be periodic in the same way?( set boundary points at each side be equal to the first and last calculated point)

Strictly speaking, a periodic BC on the pressure would be no pressure gradient at all, and so no flow. Perhaps you could implement something like the pressure upstream is the downstream pressure plus a constant (the pressure difference across your channel segment).

It depends on the channel forms, it is uniform shape, or it is non uniform shape, and also it depends on the flow itself, it is pressure driven flow, or cavity flow like plate driven flow, such as wind driven flow, in that case, for the special configuration, you can use periodic pressure boundary condition.

Kim

Thanks

Sorry, I do not describe the question clearly. I simulated turbulent channel flow by DNS, and in the x momentum equation I gave it a pressure source term. The pressure source term is obtained by assuming that the pressure gradient balenced the shear stress. So, I should Not use the periodic boundary conditions as I mentioned before or ...

In fact you do have to put a periodic boundary condition also on the pressure term !

But since you have no pressure gradient, it's true that you have to introduce a source term in the x-direction to maintain the flow.

Alexandre Chatelain

PhD Student - LES of Heat Transfer CEA Grenoble - LEGI/MoST Team FRANCE