Modelling skin and proximity effects of current flow in two rectangular bodies

The graphical user interface of Elmer
Post Reply
sbaker
Posts: 2
Joined: 08 Jan 2020, 13:55
Antispam: Yes

Modelling skin and proximity effects of current flow in two rectangular bodies

Post by sbaker » 13 Jan 2020, 14:23

Hi, I'm new to ELMER and I'm not sure I'm setting up my model correctly as I'm not getting the results I'm expecting to see.

I'm trying to model the skin and proximity effects of A/C current flow through two rectangular bodies, where the current flow is in the opposite direction through each of the bodies and the two bodies are parallel to each other.

The frequency is 150 kHz. To apply the current through the bodies I have used the same method as in tutorial 7 (Magnetic field induced by harmonic current in a wire). I've attached the Sif file to show how I have applied the current.

When viewing the results in ParaView I'm expecting to see the current distributed in an hour glass shape but this isn't happening when looking at the current density of the results. I've attached a diagram of how the current density should roughly look across the cross section of the two bodies.

I created the mesh in gmsh so have also attached the .geo file as the Elmer mesh files were too large to attach.

Sam
Attachments
Cross section.PNG
Cross section.PNG (6.58 KiB) Viewed 69 times
Geometry.geo
(3.61 KiB) Downloaded 4 times
case.sif
(4.03 KiB) Downloaded 8 times

kevinarden
Posts: 384
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Modelling skin and proximity effects of current flow in two rectangular bodies

Post by kevinarden » 13 Jan 2020, 15:59

Can you post the mesh.header file so I can see what element types you are meshing?

sbaker
Posts: 2
Joined: 08 Jan 2020, 13:55
Antispam: Yes

Re: Modelling skin and proximity effects of current flow in two rectangular bodies

Post by sbaker » 13 Jan 2020, 18:59

When converting my mesh to Elmer format it wasn't converting from gmsh to elmer so I saved the mesh as a .unv file and then converted that to elmer format.

I've attached the mesh.header file as well.
Attachments
mesh.header
(60 Bytes) Downloaded 4 times

kevinarden
Posts: 384
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Modelling skin and proximity effects of current flow in two rectangular bodies

Post by kevinarden » 14 Jan 2020, 21:55

Seeing improved results with higher frequency. You are using 150. Seeing better results at 5000. The tutorial uses 1.E5.

kevinarden
Posts: 384
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: Modelling skin and proximity effects of current flow in two rectangular bodies

Post by kevinarden » 15 Jan 2020, 21:43

Improved results with higher mesh density and frequency.
Not sure if you need that specific geometry, but the rectangles and them being close together drives down the element size causing a very large mesh for the required mesh density. From my studies I believe your issues are caused by mesh density and frequency range.
2wires.png
(238.67 KiB) Not downloaded yet

Post Reply