[Gmsh] Issues with OpenCascade

Christophe Geuzaine cgeuzaine at ulg.ac.be
Thu Oct 26 21:05:04 CEST 2017

Hi Alessandro,

> On 23 Oct 2017, at 12:34, Alessandro Vicini <alessandro.vicini at sitael.com> wrote:
> Hullo everybody,
> I find the following issues when using OpenCascade:
> 1) I start from the geometry of the attached test_00.geo: one cube, with a cubic hole;
> 2) Now if I want to apply a rotation to the inner cube (test_01.geo), I obtain that only the edges are rotated, and the original cube is conserved. This does not happen when the built-in geometry kernel is used (the applied rotation does not modify the surface loops that define the volume);

Indeed. With OpenCASCADE all operations actually create new shapes; for transformations we try to re-bind the new shapes to the old tags when possible - it's not the case here since the old shapes are still used in the volume.

> 3) On the other hand, if I first apply the rotation (test_02.geo) and then define the volume (test_03.geo), the inner surface loop is not recognized as a closed volume.

That's a limitation of the way we currently perform the transformation operations (we do it "shape by shape") : this creates duplicate entities (here lines) on the boundaries. Currently you'll need to apply 

  BooleanFragments{ Surface{1:6}; Delete; }

to get rid of them. Then you'll be able to define a closed volume.


> Is there some way to work this out? Especially point 2) is an issue for me, as for my application I'd need to append the rotation command at the end of the geo file (i.e. after volume has been defined).
> A.
> <test_00.geo><test_01.geo><test_02.geo><test_03.geo>_______________________________________________
> gmsh mailing list
> gmsh at onelab.info
> http://onelab.info/mailman/listinfo/gmsh

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

Free software: http://gmsh.info | http://getdp.info | http://onelab.info

More information about the gmsh mailing list