[Gmsh] high order meshes and interior elements

[Jean-Francois Remacle]

Le 5 avr. 2011 à 15:41, Christophe Prud'homme a écrit :

> Dear Gmsh dev,
> 
> Is that possible to get with Gmsh a straight edge high order mesh in
> interior elements (not sharing a face with the boundary) ?

normally, this is the usual behavior.

> it seems that interior elements are really curved, is there a reason why ?
> is the option Mesh.SmoothInternalEdges = 1; helping with this ?
> 

this is new stuff indeed. curving elements that have edges or faces on a curved
boundary cannot be sufficient for ensure the validity of the mesh. Some internal
edges/faces may have to be curved as well. I have commited something today that
force the mesh to be valid everywhere.

> with straight edge interior elements we could then affine
> transformations which would be much cheaper computationnally, no ?
> 

our experience in high order methods tell us that this extra cost is not
dramatic. In our computations, we assume all elements to be curvilinear.
We are going to submit a paper in SISC next week about efficient
assembly of high order finite element operators.

JFR 



> Best regards
> C.
> -- 
> Debian Developer - member of Debian Science
> http://wiki.debian.org/DebianScience
> Prof. at Univ. Grenoble in Applied Math.
> http://ljk.imag.fr/membres/Christophe.Prudhomme/
> 
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh

------------------------------------------------------------------
Prof. Jean-Francois Remacle
Universite catholique de Louvain (UCL)
Ecole Polytechnique de Louvain (EPL) - Louvain School of Engineering
Institute of Mechanics, Materials and Civil Engineering (iMMC)
Center for Systems Engineering and Applied Mechanics (CESAME)
Tel : +32-10-472352 -- Mobile : +32-473-909930 

 





-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.geuz.org/pipermail/gmsh/attachments/20110405/ad85be4e/attachment.html>