On 10/26/2011 04:03 PM, Ruth V. Sabariego wrote:
When starting from a geo file, the nodes fit the geometry when doing
refinement by splitting.
This isn't quite true. At least when dealing with geometries which are
defined by extruding transfinite surfaces through a curve, the "refine
by splitting" function tends to generate meshes which don't match the
geometry. As an example, consider the attached .geo model.
Rui Maciel
h = 0.080;
b = 0.046;
t_w = 0.0038;
t_f = 0.0052;
r = 0.005;
l = 2.0;
Point(1) = {-b/2, 0, 0};
Point(2) = {-b/2, t_f, 0};
Point(3) = {b/2, 0, 0};
Point(4) = {b/2, t_f, 0};
Point(5) = {-b/2, h, 0};
Point(6) = {-b/2, h-t_f, 0};
Point(7) = {b/2, h, 0};
Point(8) = {b/2, h-t_f, 0};
Point(9) = {t_w/2+r,t_f,0};
Point(10) = {t_w/2,t_f+r,0};
Point(11) = {-t_w/2-r,t_f,0};
Point(12) = {-t_w/2,t_f+r,0};
// web
Point(13) = {t_w/2+r, h-t_f, 0};
Point(14) = {t_w/2, h-t_f-r, 0};
Point(15) = {-t_w/2-r, h-t_f, 0};
Point(16) = {-t_w/2, h-t_f-r, 0};
// points for the circle center
Point(17) = {t_w/2+r, h-t_f-r, 0};
Point(18) = {-t_w/2-r, h-t_f-r, 0};
Point(19) = {t_w/2+r, t_f+r, 0};
Point(20) = {-t_w/2-r, t_f+r, 0};
// lines
Line(1) = {7, 5};
Line(2) = {5, 6};
Line(3) = {6, 15};
Line(4) = {16, 12};
Line(5) = {11, 2};
Line(6) = {2, 1};
Line(7) = {1, 3};
Line(8) = {3, 4};
Line(9) = {4, 9};
Line(10) = {10, 14};
Line(11) = {13, 8};
Line(12) = {8, 7};
Circle(13) = {13, 17, 14};
Circle(14) = {15, 18, 16};
Circle(15) = {10, 19, 9};
Circle(16) = {12, 20, 11};
// define surface
Line Loop(17) = {12, 1, 2, 3, 14, 4, 16, 5, 6, 7, 8, 9, -15, 10, -13, 11};
Plane Surface(18) = {17};
Transfinite Line{1:16} = 5;
Transfinite Surface{18};
Recombine Surface{18};
Extrude { {0,1,0} , {0.75,0,0} , -Pi/2 }
{
Surface{18}; Layers{5}; Recombine;
}
_______________________________________________
gmsh mailing list
[email protected]
http://www.geuz.org/mailman/listinfo/gmsh
