Page 1 of 1

Conformal mesh for magnetostatic calculation?

Posted: 26 Oct 2014, 12:48
by aether
Hi,
Do I need to use conformal meshing for magnetostatics?

For example, consider a current-carrying wire surrounded by an air/vacuum box. My workflow is FreeCAD->NetGen->Elmer to generate two bodies in Elmer. Do the two meshes have to be conformal in order to run magnetostatics?

TIA
John

Re: Conformal mesh for magnetostatic calculation?

Posted: 27 Oct 2014, 10:52
by raback
Hi

There is a developing machinery to enforce continuity also for nonconforming meshes. You can see this in some tests using "Mortar" keywords. However, you always loose some accuracy when you join meshes with mortar projectors instead of having just one continuous mesh. Sometimes this is a necessity, like in the case of rotating machines. Also the equality constraints lead to more problematic linear systems. So if you have no physical reason I would recommend using conformal meshes.

-Peter

Re: Conformal mesh for magnetostatic calculation?

Posted: 27 Oct 2014, 20:16
by aether
Thanks, Peter.