John C. Chien

December 22, 2000 03:30 
Re: Request
(1). Based on your configuration,the Reynolds number of the cavity flow is slightly over 1,000,000. (2). Assuming that you are solving the laminar flow problem and assuming that the steadystate solution exist, then the boundary layer on the walls will be relatively thin. (3). In this case, you will need very fine mesh near the walls to resolve the thin boundary layer. (4). If you use coarse mesh in your formualtion, then the solution is not going to be correct. (5). My suggestion is: refine the mesh near the wall, and plot the computed velocity gradient at the middle of the moving wall, in terms of the total number of mesh points or cells as you refine the mesh. You have to obtain the mesh independent first in order to address the accuracy of the result. (6). If you are relatively new to CFD, then, my suggestion is: run the cavity at lower Reynolds numbers, say RE=10, 100, 400, 1000, etc. In this way, you can get a better feeling about the CFD solutions.
