[Gmsh] Strange behaviour with extrude and physical geometry

[Christophe Geuzaine]

Jean Vis wrote:
> Hello,
> I have got problem with this .geo file.
> I do not understand why the elements related with the Physical Line 1011,12,13 are not in the .msh file.
> 
> The difference I see is:
> - elements 3,6,9,12 are created from the command extrude
> - elements 1,2,5,8 are not
> 
> Is this normal?

Jean - unfortunately this "feature" is there for backward compatibility
when region numbers are forced in the extrude commands. I would
recommend not forcing the region numbers: in this case everything should
just work as expected.

Here is a modified version of your file that uses the normal numbering
of entities during extrusions:

Point(1)={0,0,0,0.1};
Point(2)={0.99,0,0,0.1};
Point(3)={0,0.99,0,0.1};

Circle(1) = {2,1,3};

Extrude Line {1,{0,1,0},{0,0,0},Pi/2}
              {Layers{30,1}; };
Extrude Line {2,{0,1,0},{0,0,0},Pi/2}
              {Layers{30,1}; };
Extrude Line {5,{0,1,0},{0,0,0},Pi/2}
              {Layers{30,1}; };
Extrude Line {8,{0,1,0},{0,0,0},Pi/2}
              {Layers{30,1}; };
Coherence;

Physical Line(10) = {3};
Physical Line(11) = {6};
Physical Line(12) = {9};
Physical Line(13) = {12};


Physical Point(1)={2,4,5,6};
Physical Point(2)={3};

Physical Surface(100)={4,7,10,13};
Physical Line(101) = {1,2,5,8};


Cheers,

Christophe


> 
> Thanks
> 
> 
> Point(1)={0,0,0,0.1};
> Point(2)={0.99,0,0,0.1};
> Point(3)={0,0.99,0,0.1};
> 
> Circle(1) = {2,1,3};
> 
> Extrude Line {1,{0,1,0},{0,0,0},Pi/2}
>              {Layers{30,111,1}; };
> Extrude Line {2,{0,1,0},{0,0,0},Pi/2}
>              {Layers{30,112,1}; };
> Extrude Line {5,{0,1,0},{0,0,0},Pi/2}
>              {Layers{30,113,1}; };
> Extrude Line {8,{0,1,0},{0,0,0},Pi/2}
>              {Layers{30,114,1}; };
> Coherence;
> 
> Physical Line(10) = {3};
> Physical Line(11) = {6};
> Physical Line(12) = {9};
> Physical Line(13) = {12};
> 
> 
> Physical Point(1)={2,4,5,6};
> Physical Point(2)={3};
> 
> Physical Surface(100)={111,112,113,114};
> Physical Line(101) = {1,2,5,8};
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh
> 


-- 
Christophe Geuzaine
Assistant Professor, Case Western Reserve University, Mathematics
http://www.case.edu/artsci/math/geuzaine