[Gmsh] creation of zero width internal baffles

andrew armitatz at gmail.com
Wed May 11 19:17:52 CEST 2016


I am trying to create a geometry for a 2D problem in openfoam which
has periodic boundary conditions with internal cuts on a surface. That
is a surface with zero width baffles inside (see attached geo file).
The procedure I follow is that I create a perimeter that follows the
cuts that go in and out of the boundary and then I add the internal
lines. Then I create a plane surface and I extrude it, I am asking
three things:

1. Is the way used the correct one? Because I have the feeling that I
cheat somehow by forcing the gmsh to create these zero width cuts

2. I get in the mesh, in various places depending on the settings,
triangles where the mesh crosses the line boundaries. I have seen
these crossings in the perimetrical lines and in the internal ones.
Also I have seen them even when I change the settings. If I change
settings they might be found in a place and with different settings I
see them somewhere else. (see attached jpg files). Delaunay seems to
create the most. Decreasing the size factor seems to make them go away
but only to be found to a different point. And generally I would like
to sure that such a problem would not occur.

3. When I change the Mesh.RandomFactor to 1e-8 I get errors. When I
set it at 1e-6 the errors go away most of the times. But the way I
construct the mesh the minimum distance between lines is bigger than
0.1 which is more than a 1000 times bigger than the scale of 1e-6 or
1e-8. Why does this affects the geometry?

kind rgards
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cross1.jpg
Type: image/jpeg
Size: 197660 bytes
Desc: not available
URL: <http://onelab.info/pipermail/gmsh/attachments/20160511/60cf2aee/attachment-0002.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cross2.jpg
Type: image/jpeg
Size: 133386 bytes
Desc: not available
URL: <http://onelab.info/pipermail/gmsh/attachments/20160511/60cf2aee/attachment-0003.jpg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: se.geo
Type: application/octet-stream
Size: 55962 bytes
Desc: not available
URL: <http://onelab.info/pipermail/gmsh/attachments/20160511/60cf2aee/attachment-0001.geo>

More information about the gmsh mailing list