[Gmsh] Problem with a multiblock mesh

Christophe Geuzaine cgeuzaine at ulg.ac.be
Wed Aug 4 11:01:53 CEST 2010

On 25/06/10 23:38, Felix Salazar wrote:
> I created a simple mesh to test a finite element model in a CFD code. I
> wanted to test in my formulation for tetrahedral elements, in a simple
> poiseuille flow. I build a circular cylinder, and added volumes near the
> wall, so I could mesh them with a transfinite algorithm and have a
> refinement near the wall. However, my solution is not good. I am having
> a sink of mass within my numerical domain, and I think that in some of
> my volumes (I group all of them at the end in one physical volume), the
> numbering of the nodes is different that in the others, so I have
> negatives cell volumes, and negative jacobians for the elements. There
> is a way to check this within Gmsh? Calculate the cell volumes, or the
> jacobians for the elements. If I identify the problematic region, I
> think I could fix it changing the sign of that region in the Physical
> Volume command. Am I right?

The problem probably came from the fact that you saved your mesh in 
.mesh format: we did not enforce positive jacobians in that format. This 
should be fixed in SVN.

> I am attaching my geometry script
> Thanks,
> --
> Félix Salazar
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh

Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science