[Gmsh] Second Order Elements on Curved Surfaces

[Christophe Geuzaine]

Jacques Kools wrote:
> Hallo Christophe,
> 
>  
> 
> Hope you are doing fine. I had a question about generating second order 
> triangles on curved surfaces (something like sphere.geo in the demo 
> section is perfect to demonstrate). I noted that the mid-edge nodes are 
> located on the geometrical surface rather than on the triangles, i.e. 
> they are not really mid-edge. If you then do a 3D mesh, you get tets 
> that are not really tets near the surface. I hope I make myself clear.
> 
>  
> 
> I understand that this is a desirable feature for some people, but for 
> the particular application we have in mind, we would rather have higher 
> order tets and tris where the mid-edge points are really halfway between 
> the vertices. Would it be possible to have the option to generate such 
> elements? Don’t worry if it is too much of a hassle, it would be quite 
> easy for us to generate the additional points when reading the mesh.
> 

Hi Jacques - we could indeed add an option to simply subdivide the
edges/faces instead of generating the new vertices on the geometry. I'll
add this to our TODO list.

>  
> 
> I also noted that applying “optimize quality” on a  second order mesh 
> generates again a first order mesh.
> 

Yes, that's due to the way the optimizer works. The best is to optimize
the mesh first, then create the 2nd order nodes.

Take care,

Christophe


>  
> 
> Hope this is useful,
> 
>  
> 
> Jacques Kools
> 
> 
> ------------------------------------------------------------------------
> 
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh


-- 
Christophe Geuzaine
Assistant Professor, Case Western Reserve University, Mathematics
http://www.case.edu/artsci/math/geuzaine