Ben Axelrod wrote:
I am a little unclear on what this is and what it is used for. Is
this to visualize the triangular grid for things like Finite Element Analysis
(FEM) and Computational Fluid Dynamics (CFD)?
In can be, yes -- though that's really only the part that renders the mesh. Contouring on a triangular mesh can be used to visualize data on such a mesh, as generated by FE or FV codes, but can also be used to contour and visualize data on any arbitrary points -- it's an alternative to interpolating to a regular grid before contouring.
Doing the contouring directly on a triangular mesh is more efficient and bit more robust, particularly near the bounds of the data.
What kind of format is the data in?
It would take simple x,y,z vectors for the contouring routines. There is some discussion of the API in that thread. I'd like to see a standard class that holds a triangular mesh while we're at it -- to make it easier to separate the generation of the mesh from the use of the mesh.
> Are there any standards for this type of thing?
I wish.
There is a netcdf standard for unstructured meshes being developed as we speak, primarily for oceanographic model, but that's all I know of -- and MPL doesn't (and shouldn't) support netcdf directly anyway.
There are common ways to express a triangular mesh -- ans we should use them:
ie. a NX3 array of indexes, where each row in the array is the three indexes into the point coordinate array.
some systems also store indexes to the neighbors of each triangle, etc. I think that should all be hidden in the mesh class.
> Do you have some example code or images?
There were some links in that thread. Ian, maybe you should post them again.
-Chris
···
--
Christopher Barker, Ph.D.
Oceanographer
Emergency Response Division
NOAA/NOS/OR&R (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception
Chris.Barker@...236...