How to calculate the mutual inductance of a coil pair

[millim]

Hello again,

as follow up to my post above. I tried it with boolean body difference in FreeCAD resulting in a double surface interface which is wrong. Netgen doctor does not help either. It only reduces the double surface to one again, meaning no interface. The clue is how to apply one surface to (in my case) two bodies? I have no clue! Maybe Mike can help here since he accomplished that in his example?

Regards

[mzenker]

Hi,

this is a problem of all free meshers (that I know of, e.g. netgen, gmsh).
If you have two bodies one inside each other, I would advise you to export just the outer body with the hole from your CAD and define a volume inside the hole using the inner surfaces of the outer body. AFAIK, a 3D-CAD program cannot do that, but a mesher should be able to do it (gmsh, for example, can do it). Then you have just one interface, and the mesh will be conformal.

HTH,

Matthias

[millim]

Hi,

meanwhile I have found a way how to apply a surface interface to two parents (bodies). The good thing is that the step-file can be used directly from the CAD tool without doing some boolean modifications. The bad thing, I loaded the step file into netgen, then using the netgen step file doctor under Geometry -> IGES/STEP Topology Explorer.., enabling Split partitions ONLY (important!), then meshing, and exporting in Gmsh2 format. Then using elmergrid to perform the rest.

If it works right, elmergrid gives an info > Found xxx side elements that have two partners <, so the conformal surface interface of two bodies will be created. Unfortunately, the Elmer export from netgen has a bug doing that. It reduces the interface to one body only. The weird thing is that volumes will be numbered starting with 100001? These are all netgen issues. If this is a problem to Elmersolver, an additional renaming step has to be inserted. I added the Elmer mesh files ,too.

Regards

[millim]

Hi, numbering resolving with:
>elmergrid 14 2 file.msh -bulkorder -boundorder

[millim]

Finally, I managed to calculate the self and mutual impedances using magnetic energy approach. Unfortunately, the electromagnetic solvers are not frequently used here; also the forum support was somewhat weak. No I have to check the results against the results of numerical approximations and the validity of my approach used.

Electric potential V

ElmerGUI postprocessor_2012-05-11_11-18-49.jpg (38.03 KiB) Viewed 4835 times

Flux density B

ElmerGUI postprocessor_2012-05-11_11-33-03.jpg (69.99 KiB) Viewed 4835 times

But, finally thx to CSC for this great piece of software and for the great documentation!

Regards

[raback]

Hi millim

Nice work! It would be nice to see your full case and how you finally computed the mutual impedance. I think that the Whitney solver has great promise even though not so much effort has been put to it (~2 PMs by Juha). Electromagnetics users seem to be active in pushing the limits of Elmer forward in this area which is really appreciated. We hope to have some more projects in this area so that the features could be also further developed.

-Peter

[orcaja]

Yes, It would be nice to see the full case, congrats!

joc

[federico]

Hi, I am new on Elmer, i would like to see the sif file of this project.

Regards
federico