|
[Sponsors] |
October 8, 2003, 21:01 |
Gambit and parametrized model generation
|
#1 |
Guest
Posts: n/a
|
I am trying to develop a 3D model & mesh that can parameterized in GAMBIT based on size variations. The model itself can be generated with no problem using a parameterized journal file that's a combination of mostly volume creation, and real volume spliting based on volumes and faces, but the issue comes to meshing. There's a couple of tricky volume meshes that have to be done in a certain order of edge/face meshing to work. While this sequence can be programmed in, I'm finding that when I change the sizes of the initial volumes via variables in the journal file, the resulting names for edges and faces don't necessarily remain consistant, such that the operations to link face and edge meshes will fail, and the critical meshing operation above won't work.
Fortunately, I believe that I can actually create the whole 3d shape through creating vertices, then edges, then faces, then volumes, as all but 4 or 6 of the vertices coordinates can be copied straight from parameters of the problem, and the 4 or 6 points can be obtained by used edge projection. But with the model as it is, I have about 50 volumnes, 80 vertices, and 200 faces and edges -- a potentally daunting task but not impossible. Doing it this way will give me full control of the NAMING of the edges/faces/volumes, such that the critical mesh step can then be done without question. Has anyone run into similarly complex problems like that and worked out ways without resorting to the second method to solve it? |
|
|
|