John C. Chien
|
January 16, 2001 15:23 |
Re: meshing questions
(1). A good question. (2). Everyone can design his own mesh and run the code to obtain a solution. (3). So, how good is the solution obtained? That's the hard question to answer, because it depends on the mesh and the solver (and turbulence models,etc.). (4). The company which obtain the right solution is likely to survive and the company which obtain the wrong solution is likely to die. So, it is not just an issue of which meshing scheme to use. (5). As the first step, I think, you should try to obtain the "mesh independent solution" first. In other words, make sure that your solution obtained using your meshes is independent of the mesh you used. (6). "Optimum meshing scheme" is the one which is unlikely to create troubles for the solver and the particular problem you are trying to solve. For example, unstructured tet/tri cell meshes are not suitable for wall boundary layer flows. You will still get a solution, but you will have a hard time in getting the mesh independent solution. (7). Well, if you are not after the mesh independent solution, then any mesh is a good mesh, as long as there is a solution.(it's a joke.)
|