[universez]

Hi, all.
I am trying to use the modified SIMPLE method by Anil W. Date with curvilinear grids. First, I do not know if the method is good or not. Please let me know your suggestion if it is not a good one.


I have several questions.

1, what is the boundary condition for p'. In the book, "Introduction to Computational Fluid Dynamics". It was suggested that dp'/dn=0, with some part of boundary p'=0.

2, for the boundaries what coeffecients should be ? e.g. for cell p, the west side is a wall or a symetrical boundary, how to define Aw?

3 for a block, extraplated pressure values should be values at cell surface or at cell center.

I did two different tests,
a, ractangular domain, left side input with velocity boundary, right side is outlet boundary, top is wall and bottom is a symmetrical boundary. For pressure i have p'=0 at the right side. during iteration, when the flow reach to the right side, pressure at the right boundary deverges.

b, a domain with 2 block region, there is ZigZag problem, which starts from a corner of one of the block, then spread to the whole domain.

I am not sure if it is because of wrong values the pressure at boundaris, or corners points.

I appreciate any help.


Yu