Propagation of results on multibody ElectroMagnetic models

Numerical methods and mathematical models of Elmer
Post Reply
gabrisg
Posts: 2
Joined: 01 Apr 2021, 19:20
Antispam: Yes

Propagation of results on multibody ElectroMagnetic models

Post by gabrisg »

Hello all,
I'm working on a model using MGHarm solver and at first, I've used one metal part inside a gel environment. The mesh was pretty simple, done in GMSH, and the surfaces between the part and the gel were shared so I got the results as expected.
Now, I want to put an assembly of 6 parts inside this gel, but I can't get the mesh on Gmsh with compound parts due to its complexity. So it is meshed independently for each part and when I include the geometry on Elmer, I get null results for electrical/magnetic fields and heating on the inner parts (It seems the meshes are completely disconnected), even though the disconnect bodies' option is checked on the solver. Is there any way to "propagate" the results to the inner parts, or somehow connect the surfaces of different parts so there is a flux of electric/magnetic field inside the parts?
PS: I've changed the inner parts material properties to the same properties as the gel, so the electrical field should just be uniform on the whole enviroronment+parts geometry, and the results show no electrical field on the parts.
EM_results.PNG
(699.41 KiB) Not downloaded yet
Thanks
Gabriel
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Propagation of results on multibody ElectroMagnetic models

Post by raback »

Hi

The best way to deal with multibody problems is during the meshing. You can deal with some problems in the solution phase but this will cause a lot of unnecessary numerical trouble. For rotating interfaces, for example, it cannot be avoided. Otherwise, I recommend avoiding.

If the parts have a conforming mesh you can eliminate the redundent nodes with "-merge" in ElmerGrid. However, I fear they are not.

The discontinuous bodies is related to visualization of fields such that you maintain the discontinuities only at the interface. It bears no connection to the solution.

-Peter
gabrisg
Posts: 2
Joined: 01 Apr 2021, 19:20
Antispam: Yes

Re: Propagation of results on multibody ElectroMagnetic models

Post by gabrisg »

Thank you for you reply, raback.
I've tried to merge nodes on ElmerGrid but the problem was that there were no redundant nodes, I had to rework all the parts to delete assembly small gaps and combine the surfaces of the parts on gmsh. It took lot of time to generate the mesh but it worked in the end =D
Post Reply