Magentostatic Problem

Numerical methods and mathematical models of Elmer
Post Reply
soup_dog
Posts: 2
Joined: 12 Oct 2018, 16:02
Antispam: Yes

Magentostatic Problem

Post by soup_dog »

hello,

I have been working at setting up a 3D magnetostatic problem to test out the functionality of the software. This problem involves two cylindrical magnets with axial magnetization. for this problem, I used gmsh to construct the mesh.

When running the elmer solver, The system converges and appears to give an answer, but when viewing in either elmer post or paraview, the data is either not there, not displaying, or all 0 magnitude.

Not being entirely familiar with this software, I am hoping someone can catch my error in the setup of this problem.

gmsh file:
two cylinders.geo
(445 Bytes) Downloaded 247 times
sif file:
case.sif
(3.95 KiB) Downloaded 228 times
mzenker
Posts: 1999
Joined: 07 Dec 2009, 11:49
Location: Germany

Re: Magentostatic Problem

Post by mzenker »

Hi,

I am not familiar with magnetostatic simulation.

To rule out a postprocessing problem: Your simulation does write a file case.vtu, which ElmerPost cannot display. For ElmerPost you would have to change the output file extension to .ep in Model>Setup.
But ParaView is the recommended postprocessing tool anyway.
If you launch ParaView from within ElmerGUI and don't see anything, try to open the case.vtu file from within ParaView. Then click "Apply" in the properties panel and change the coloring from "solid color" to something else, e.g. the variable you want to view.

HTH,
Matthias
soup_dog
Posts: 2
Joined: 12 Oct 2018, 16:02
Antispam: Yes

Re: Magentostatic Problem

Post by soup_dog »

Matthias,

A magnetostatic problem simply refers to static magnetic fields caused by either constant currents or permanent magnets.

I was aware of the difference between the .ep and the .vtu file type in the output of the elmer simulation. For both elmer post and paraview, the geometry loads fine and will display correctly, but the data for the magnetic field strength, AV, and the flux density do not appear to be calculated correctly. The field in all spots comes back as "0". Interestingly, the field for the magnets themselves also appear to also be 0 (the model in para viewer shows the nodes as also being "0") . This has led me to concluded that the problem might originate from either how I am defining the material properties, how the solver I chose to use in elmer is solving the simulation, or I am or how the mesh connects the objects in the simulation.

-Andrew
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Magentostatic Problem

Post by raback »

Hi Andrew

You should use MagnetoDynamicsCalcFields to get the physical fields. The primary field is the vector potential living in Hcurl space which you cannot directly visualize.

Further, if you have AV appering as a field that would hint that your binary is not terribly fresh.

There are plenty of examples among tests which could show you how your .sif file should look like.

-Peter
Post Reply