CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   FLUENT (https://www.cfd-online.com/Forums/fluent/)
-   -   Oddity in meshing half a torus (https://www.cfd-online.com/Forums/fluent/33220-oddity-meshing-half-torus.html)

 Maarten March 12, 2004 09:04

Oddity in meshing half a torus

I am running into a slight oddity while meshing half a torus using Gambit 1.3. The torus has been cut in half radially, along its primary axis. (Think of it as if you were trying to create a sandwich from a doughnut.) I want to use the Cooper tool on such a shape, and for that the torus has to be cut in half yet again, but this time along its other axis. There's only one way to choose the source faces in this case: I've labelled those 'sides' or 'side faces' in the following paragraphs.

The problem is that I noticed that the mesh is not continuous at the point where the sides meet: it has a slight clover-shape. After careful inspection, I've come to the conclusion that the mesh projection messes up slightly: instead of straight lines from one side to the other, they tend to 'sag'. Think of it as a two necklaces around someone's neck: one is pulled tight and resembles a choker, but is otherwise straight; the other one is more loose and sags in the middle. For some reason, Gambit insists on introducing slack, thus ruining the mesh. Equiangle skews are quite high for obvious reasons.

I have not come up with a solution to remedy this problem. Cutting the shape in quarters or even eights does help somewhat, but this is not a feasible solution in view of the specifications of the remainder of my geometry. Is anyone familiar with the problem? If so, can you offer a solution? I realise things are difficult to explain in words, but I'm more than happy to send you a small picture.

 All times are GMT -4. The time now is 08:01.