[Gmsh] Possible Bug with Orientations

Christophe Geuzaine geuzaine at acm.caltech.edu
Tue May 13 07:09:00 CEST 2003


Archi - it's a known problem (that we have to fix...). In the meantime,
you can simply change the definition of the Physical Surface to control
the orientation, e.g.,

Physical Surface(surfPhys) = {s1,-s2,s3,s4,-s5,s6};

Christophe


Archi Forshop wrote:
> Hi,
> 
> I am having problems making meshed surface elements
> oriented the same way as the physical surface entity
> they belong to. I'm attaching below a simple script
> that shows it clearly when both Geometry-Normals and
> Surface-Normals are displayed at the same time.
> Different mesh algorithms produce sometimes different
> orientations, all basically wrong ones.
> 
> Maybe I'm doing something wrong myself.
> 
> Archi
> 
> P.S. I'm using the latest version gmsh-1.44.0
> 
> ---start cut here ---
> 
> /*********************************************************************
> 
>  *
>  *  brick.geo
>  *
> *********************************************************************/
> 
> lc = 1.0;
> 
> xa    = 3.0;
> yb    = 1.0;
> zc    = 2.0;
> 
> p1 = newp; Point(p1) = {0.0, 0.0, 0.0, lc};
> p2 = newp; Point(p2) = {xa , 0.0, 0.0, lc};
> p3 = newp; Point(p3) = {xa , yb , 0.0, lc};
> p4 = newp; Point(p4) = {0.0, yb , 0.0, lc};
> p5 = newp; Point(p5) = {0.0, 0.0, zc , lc};
> p6 = newp; Point(p6) = {xa , 0.0, zc , lc};
> p7 = newp; Point(p7) = {xa , yb , zc , lc};
> p8 = newp; Point(p8) = {0.0, yb , zc , lc};
> 
> l1  = newreg; Line(l1)  = {p1,p2};
> l2  = newreg; Line(l2)  = {p2,p3};
> l3  = newreg; Line(l3)  = {p3,p4};
> l4  = newreg; Line(l4)  = {p4,p1};
> l5  = newreg; Line(l5)  = {p1,p5};
> l6  = newreg; Line(l6)  = {p2,p6};
> l7  = newreg; Line(l7)  = {p3,p7};
> l8  = newreg; Line(l8)  = {p4,p8};
> l9  = newreg; Line(l9)  = {p5,p6};
> l10 = newreg; Line(l10) = {p6,p7};
> l11 = newreg; Line(l11) = {p7,p8};
> l12 = newreg; Line(l12) = {p8,p5};
> 
> llp1 = newreg; Line Loop(llp1) = {l1,l6,-l9,-l5};
> llp2 = newreg; Line Loop(llp2) = {l2,l7,-l10,-l6};
> llp3 = newreg; Line Loop(llp3) = {l3,l8,-l11,-l7};
> llp4 = newreg; Line Loop(llp4) = {l4,l5,-l12,-l8};
> llp5 = newreg; Line Loop(llp5) = {l9,l10,l11,l12};
> llp6 = newreg; Line Loop(llp6) = {-l1,-l4,-l3,-l2};
> 
> s1 = newreg; Plane Surface(s1)={llp1};
> s2 = newreg; Plane Surface(s2)={llp2};
> s3 = newreg; Plane Surface(s3)={llp3};
> s4 = newreg; Plane Surface(s4)={llp4};
> s5 = newreg; Plane Surface(s5)={llp5};
> s6 = newreg; Plane Surface(s6)={llp6};
> 
> slp1 = newreg; Surface Loop(slp1) =
> {s1,s2,s3,s4,s5,s6};
> 
> surfPhys = newreg; Physical Surface(surfPhys) =
> {s1,s2,s3,s4,s5,s6};
> Printf("surfPhys# = %g",surfPhys);
> 
> vbrick = newreg; Complex Volume(vbrick) = {slp1};
> 
> brick = newreg; Physical Volume(brick) = {vbrick};
> 
> ---end cut here ---
> 
> 
> __________________________________
> Do you Yahoo!?
> The New Yahoo! Search - Faster. Easier. Bingo.
> http://search.yahoo.com
> _______________________________________________
> gmsh mailing list
> gmsh at geuz.org
> http://www.geuz.org/mailman/listinfo/gmsh
> 
> 


-- 
Christophe Geuzaine

Tel: (626) 395-4552    http://www.geuz.org
Fax: (626) 578-0124    mailto:geuzaine at acm.caltech.edu