[Gmsh] Sphere problem

[Christophe Geuzaine]

Markus Hengsberger wrote:
> Yes, I tried to optimize mesh, but nothing really happens.
> It looks the same after the opt. process.

The optimization pass only changes the volume mesh.

Two solutions:

1) remove all your "Transfinite" commands (the unstructured algorithm will
not produce small elements at the poles)

2) define your sphere using 6 quadrangular patches (like a volleyball)

Christophe


> 
> Markus
> 
> 
> 
> 
>>mhengs at gmx.at wrote:
>>
>>
>>>Hi,
>>>
>>>i am working with 2D structured meshed spheres. Everything works fine,
>>
>>but
>>
>>>i have problem with the small elements on the top of the spheres. My
>>
>>solver
>>
>>>cannot work with this elements.
>>>
>>>Maybe you have a solution how to cope with this problem.
>>>
>>>Markus
>>>
>>>Here is an exmaple of 2x2x2 sphereblock:
>>>
>>>/********************************************************************* 
>>>*
>>>*  
>>>*  source file: example.igs
>>>*  translater version: jIGStoGEO 0.2  
>>>*  
>>>*  output files:
>>>*  example.geo 
>>>*  example.msh
>>>*
>>>*********************************************************************/
>>>
>>>
>>>Function Sphere 
>>>
>>> p1 = newp; Point(p1) = {x,  y,  z,  1.0} ;
>>> p2 = newp; Point(p2) = {x+r,y,  z,   1.0} ;
>>> p3 = newp; Point(p3) = {x,  y+r,z,   1.0} ;
>>> p4 = newp; Point(p4) = {x,  y,  z+r, 1.0} ;
>>> p5 = newp; Point(p5) = {x-r,y,  z,   1.0} ;
>>> p6 = newp; Point(p6) = {x,  y-r,z,   1.0} ;
>>> p7 = newp; Point(p7) = {x,  y,  z-r, 1.0} ;
>>>
>>> c1 = newreg; Circle(c1) = {p2,p1,p7};
>>> c2 = newreg; Circle(c2) = {p7,p1,p5};
>>> c3 = newreg; Circle(c3) = {p5,p1,p4};
>>> c4 = newreg; Circle(c4) = {p4,p1,p2};
>>> c5 = newreg; Circle(c5) = {p2,p1,p3};
>>> c6 = newreg; Circle(c6) = {p3,p1,p5};
>>> c7 = newreg; Circle(c7) = {p5,p1,p6};
>>> c8 = newreg; Circle(c8) = {p6,p1,p2};
>>> c9 = newreg; Circle(c9) = {p7,p1,p3};
>>> c10 = newreg; Circle(c10) = {p3,p1,p4};
>>> c11 = newreg; Circle(c11) = {p4,p1,p6};
>>> c12 = newreg; Circle(c12) = {p6,p1,p7};
>>>
>>>
>>> l1 = newreg; Line Loop(l1) = {c5,c10,c4};
>>> l2 = newreg; Line Loop(l2) = {c9,-c5,c1};
>>> l3 = newreg; Line Loop(l3) = {c12,-c8,-c1};
>>> l4 = newreg; Line Loop(l4) = {c8,-c4,c11};
>>> l5 = newreg; Line Loop(l5) = {-c10,c6,c3};
>>> l6 = newreg; Line Loop(l6) = {-c11,-c3,c7};
>>> l7 = newreg; Line Loop(l7) = {-c2,-c7,-c12};
>>> l8 = newreg; Line Loop(l8) = {-c6,-c9,c2};
>>>
>>>
>>> s1 = newreg; Ruled Surface(newreg) = {l1};
>>> s2 = newreg; Ruled Surface(newreg) = {l2};
>>> s3 = newreg; Ruled Surface(newreg) = {l3};
>>> s4 = newreg; Ruled Surface(newreg) = {l4};
>>> s5 = newreg; Ruled Surface(newreg) = {l5};
>>> s6 = newreg; Ruled Surface(newreg) = {l6};
>>> s7 = newreg; Ruled Surface(newreg) = {l7};
>>> s8 = newreg; Ruled Surface(newreg) = {l8};
>>>
>>>Transfinite Line {c5,c10,c4} = 15 Using Progression 1.0;
>>>Transfinite Line {c9,-c5,c1} = 15 Using Progression 1.0;
>>>Transfinite Line {c12,-c8,-c1} = 15 Using Progression 1.0;
>>>Transfinite Line {c8,-c4,c11} = 15 Using Progression 1.0;
>>>Transfinite Line {-c10,c6,c3} = 15 Using Progression 1.0;
>>>Transfinite Line {-c11,-c3,c7} = 15 Using Progression 1.0;
>>>Transfinite Line {-c2,-c7,-c12} = 15 Using Progression 1.0;
>>>Transfinite Line {-c6,-c9,c2} = 15 Using Progression 1.0;
>>>
>>>Transfinite Surface {s1} = {p4,p3,p2};
>>>Transfinite Surface {s2} = {p7,p3,p2};
>>>Transfinite Surface {s3} = {p7,p6,p2};
>>>Transfinite Surface {s4} = {p4,p6,p2};
>>>Transfinite Surface {s5} = {p4,p5,p3};
>>>Transfinite Surface {s6} = {p4,p6,p5};
>>>Transfinite Surface {s7} = {p7,p6,p5};
>>>Transfinite Surface {s8} = {p7,p3,p5};
>>>
>>>Recombine Surface {s1};
>>>Recombine Surface {s2};
>>>Recombine Surface {s3};
>>>Recombine Surface {s4};
>>>Recombine Surface {s5};
>>>Recombine Surface {s6};
>>>Recombine Surface {s7};
>>>Recombine Surface {s8};
>>>
>>>Return
>>>
>>>x = 0;
>>>y = 0;
>>>z = 0;
>>>r = 0.5;
>>>
>>>
>>>For w In {1:2}
>>>For u In {1:2}
>>>For t In {1:2}
>>>
>>>
>>>x += 1.0;
>>>
>>> Call Sphere ;
>>>
>>>
>>>EndFor
>>>x = 0;
>>>y += 1.0;
>>>
>>>EndFor
>>>z += 1.0;
>>>x = 0;
>>>y = 0;
>>>
>>>
>>>EndFor
>>>
>>> 
>>>
>>
>>
>>
>>Have you tryed to optimize the mesh ? It takes time but
>>gives very good enhancement.
>>
>>JFR
>>
>>
>>
>>
>>-- 
>>Prof. Jean-François Remacle
>>Department of Civil Engineering
>>Center for Systems Engineering and Applied Mechanics
>>Université catholique de Louvain, Belgium
>>Tel : +32 10 472082
>>Fax : +32 10 472210 
>>
> 
> 


-- 
Christophe Geuzaine
Applied and Computational Mathematics, Caltech
geuzaine at acm.caltech.edu - http://geuz.org