[Gmsh] Re: Gmsh digest, Vol 1 #61 - 1 msg

Christophe Geuzaine Christophe.Geuzaine at ulg.ac.be
Sat Dec 15 10:28:12 CET 2001


Laurent CHAMPANEY wrote:
> 
> There is mesh mesh generator called Tetgen
> (www.weboo.com/sh/tetgen.htm)
> that uses a tetrahedral quality algorithm.
> It seams that it uses the Shewuck's quality algorithm (??)

Yes, I know this code. Unfortunately, according to the user's guide, it
also suffers from the generation of slivers (flat tetrahedra).

Do you think it would be interesting to integrate it into gmsh as an
alternativa 3D mesh generator?

Christophe


> L. CHAMPANEY
> 
> Shewchuk
> 
> Jean-Francois Remacle wrote:
> 
> >> Hello,
> >>
> > Hello,
> >
> >> I've been trying to use gmsh to construct 3D tetrahedral grids for
> >> our
> >> CFD work. I have run into considerable problems due to the
> >> creation of
> >> very flat tetrahedra when producing grids for complex geometries.
> >> Most
> >> of the
> >> flat tetrahedra are rather envelope-shaped. If you are interested,
> >> I
> >> can send a .geo file which will produce this effect.
> >>
> > The current version of the 3D tet mesh generator does not involve
> > any optimization steps. That means that slivers are often generated.
> > We have a procedure to improve mesh quality by edge and
> > face swapping but I just looked into the code and that part
> > is commented ;-) I guess there are bugs.
> >
> >> This mucks up my calculations due to instabilities.
> >>
> > Yes, I know.
> >
> >> I'm aware that the 3D grid generator is still experimental, but
> >> nevertheless I would like to ask if there exists a sort of
> >> workaround for this. Or is it foreseen that the problem will
> >> be tackled in the near future?
> >>
> > If you can find a tet mesh optimizator, that would help. I'm
> > going to work on that part and see what is wrong in my optimization
> > procedure.
> >
> >> Otherwise I must say that the software looks very nice. I actually
> >> used it for a number of grids, but most of them have proven to be
> >> not
> >> really that useful due to this problem :-(.
> >>
> > Too bad, we're going to work on that.
> > JF
> >
> >> TIA for any answer
> >> Thomas
> >>
> > _______________________________________________
> > Gmsh mailing list
> > Gmsh at geuz.org
> > http://www.geuz.org/mailman/listinfo/gmsh
> >

-- 
Christophe Geuzaine

Tel: 32 (0) 4 366 37 10    http://www.geuz.org
Fax: 32 (0) 4 366 29 10    mailto:Christophe.Geuzaine at ulg.ac.be