[Gmsh] Adjustment for mesh configuration for quasi structured quadrilateral mesh

Christophe Geuzaine cgeuzaine at uliege.be
Tue Nov 19 20:15:16 CET 2019

> On 19 Nov 2019, at 14:56, Maurice Rohracker <maurice.rohracker at gmx.net> wrote:
> Dear GMSH users and developers,
> For a project at the University Erlangen-Nürnberg, I would like to mesh some geometry with randomly generated inclusions in a box (as you can see in the attachment) to get a relatively structured quadrilateral mesh with good quality using this then for FE simulations. We are using therefor the python API.
> Right now, we get as a result in most cases these diamond elements, which some of them have very large angles up to 180 degrees or very acute angles. How can I avoid these and get a much better quality mesh?
> Here are the adjustments we’re using right now. 
> Mesh.Algorithm: 9 -  Packing of Parallelograms 
> Mesh.RecombineAll: 1 - quad elements
> Mesh.SubdivisionAlgorithm: 1 - all quads
> Mesh.RecombinationAlgorithm: 2 - simple quad (the mesh result didn’t get even better for different adjustments)
> We specify the mesh size at the points. Especially at the four corner points (0.3) and at the 3 points on the circles (0.12), where the size of the box is 10x10 and the radius 0.5.
> Do have any tips to improve the quality, are there adjustments missing (like transfinite) or anything else. It doesn’t have to be fully structured, but we try to get as far as possible, especially we don’t want to have large or small angles.

I would use this, which seems fine :

-------------- next part --------------
A non-text attachment was scrubbed...
Name: input_Geom_new.geo
Type: application/octet-stream
Size: 6952 bytes
Desc: not available
URL: <http://onelab.info/pipermail/gmsh/attachments/20191119/6cb969aa/attachment.geo>
-------------- next part --------------

> If needed I also can share the .msh file.
> Thank you very much in advance.
> Best regards
> Maurice Rohracker
> Master student Computational Engineering FAU Erlangen-Nürnberg
> <input_Geom.geo><geometry.png><mesh.png>_______________________________________________
> gmsh mailing list
> gmsh at onelab.info
> http://onelab.info/mailman/listinfo/gmsh

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

More information about the gmsh mailing list