[Gmsh] How to improve the quality of mesh generation

[Christophe Geuzaine]

Dear Jun,

Did you check where these bad quality elements were located (e.g. with Tools->Mesh and select a quality range)? They might be constrained by the geometry. If you can share your geometry we can have a look.

Christophe
 


> On 29 Aug 2018, at 21:11, Junhui Liu <junhui.liu at nrl.navy.mil> wrote:
> 
> Hi All,
> 
> It appears that the quality of mesh generation depends on the version you use. I use Mesh.Algorithm3D = 1, and Mesh.Algorithm =5.
> 
> For example, 
> 
> if I use gmsh-3.0.3, I got
> Info    : Opti : (7594757,0,400672) =  1.34590E+03 QBAD  6.40230E-03 QAVG  6.92222E-01 ( 650.191 sec)
> Info    : Opti : (7649463,0,408083) =  1.34590E+03 QBAD  6.40230E-03 QAVG  6.92316E-01 ( 965.826 sec)
> Info    : Opti : 4 ill-shaped tets are still in the mesh
> Info    : Opti : 0.00 < QUAL < 0.10 :         7 elements
> I
> 
> But if I use gmsh-3.0.6, I got  
> Info    : Opti : (7588653,0,400922) =  1.34590E+03 QBAD  1.05131E-02 QAVG  6.92189E-01 ( 561.789 sec)
> Info    : Opti : (7643427,0,408262) =  1.34590E+03 QBAD  1.05131E-02 QAVG  6.92284E-01 ( 821.088 sec)
> Info    : Opti : 4 ill-shaped tets are still in the mesh
> Info    : Opti : 0.00 < QUAL < 0.10 :         5 elements
> I
> 
> If I use the newest stable version gmsh-4.0.0, I got 
> Info    : Opti : (5126023,0,163893) =  1.34579E+03 QBAD  6.22727E-04 QAVG  7.99073E-01 ( 556.652 sec)
> Info    : Opti : (5145672,0,165902) =  1.34579E+03 QBAD  6.22727E-04 QAVG  7.99114E-01 ( 772.064 sec)
> Info    : Opti : 6 ill-shaped tets are still in the mesh
> Info    : Opti : 0.00 < QUAL < 0.10 :         9 elements
> 
> Although the newer version gave overall better quality, but it generated some very bad elements. Is there a way that I can improve the quality of the mesh, especially to eliminate the elements of poor quality?
> 
> Thanks a lot for your help.
> 
> Jun
> 
> 
> _______________________________________________
> gmsh mailing list
> gmsh at onelab.info
> http://onelab.info/mailman/listinfo/gmsh

— 
Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science 
http://www.montefiore.ulg.ac.be/~geuzaine

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