[Gmsh] periodic BCs
cgeuzaine at ulg.ac.be
Tue Jun 2 22:34:00 CEST 2009
Danny Holstein wrote:
> You mean to imply that since the 2 surfaces are related through a rotation/translation, that they will be meshed identically?
> That's a neat trick if that's the case.
Yes, all the "Extrusion" commands have the option to also extrude the
mesh in addition to the geometry (with the "Layers" options).
It's usually used to generate the volume mesh (cf. tutorial/t3.geo);
here we subvert the command to only keep the "top" and "bottom" surface
> ----- Original Message -----
> From: "Christophe Geuzaine" <cgeuzaine at ulg.ac.be>
> To: "Danny Holstein" <dgholstein at embarqmail.com>
> Cc: gmsh at geuz.org
> Sent: Tuesday, June 2, 2009 10:20:41 AM GMT -05:00 US/Canada Eastern
> Subject: Re: [Gmsh] periodic BCs
> Danny Holstein wrote:
>> Let me try again. How difficult would it be to modify GMSH to create a "Periodic Plane"? Either a boolean switch could be set while doing a translate/rotate or the 2 planes could be grouped as a physical entity, which would be 2-D meshed once with the mesh elements copied to it's "periodic" brother.
>> Certainly, many problems have 180 degree rotational symmetry, and many RF devices (such as helix and coupled cavity TWTs and klystrons) have axial symmetry, magnetrons/CFAs have rotational symmetry to the vanes. Mechanical structures with spokes would be greatly simplified with this symmetry.
>> It seems the switch used with a new "Periodic Plane" would be best, the translation matrix could be stored on generation and used immediately after mesh generation.
> Hi Danny - Here's a little example on how you could do this using
> extruded meshes.
>> gmsh mailing list
>> gmsh at geuz.org
Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science
More information about the gmsh