WhitneyAVHarmonicSolver for health protection

Numerical methods and mathematical models of Elmer
Post Reply
pziegler
Posts: 3
Joined: 14 Jul 2022, 09:47
Antispam: Yes
Location: Switzerland
Contact:

WhitneyAVHarmonicSolver for health protection

Post by pziegler »

I would like to calculate the magnetic flux in a welding machine to ensure that the health protection limits for people nearby are complied with.
2d problem.png
(96.69 KiB) Not downloaded yet
I used to do this in commercial software but I want to make this available for small companies without those possibilities...
First I am struggling with a very trivial example in 2d. The current flows through two parallel rectangular conductors and back through a round conductor.

I get "trivially zero" from the MagnetoDynamicsCalcFields solver.
In the result file I see that the current density is applied but there are no fields.
Here's the complete ElmerGUI Project:
ElmerGUI.zip
(282.92 KiB) Downloaded 56 times
I'm sure there is something I did not correctly understand...but what :x
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: WhitneyAVHarmonicSolver for health protection

Post by raback »

Hi,

The Solvers are run in the order they are numbered (except if the "exec solver" keyword determines otherwise). So here you have CalcFields being run as solver 1 and the actual computation of vector potential as solver 2. If you switch these, you may have a shot.

-Peter
pziegler
Posts: 3
Joined: 14 Jul 2022, 09:47
Antispam: Yes
Location: Switzerland
Contact:

Re: WhitneyAVHarmonicSolver for health protection

Post by pziegler »

Thanks Peter

Oh, that was a bad mistake happened when I switched between GUI and non-GUI ;-)
I have now at least some joule heating but still no fields and the trivially zero.
I've read a lot about the BCs and I am very confused (2d and 3d) what is the best way to set the normal component of the flux to 0 on all outer faces of the surrounding air. I found some statements (I think from you) to not use the flux BCs but the potential ones. I also found examples with initial conditions of 0 to all Potentials and examples where they are not defined.
Is my BC correct for 2d?
When I set "Calculate Nodal Fields" to False, I don't have the trivially zero message but still no fields. The GUI sets "Skip Nodal Fields"...I assume this should have the same effect as Calculate Nodal Fields...?
pziegler
Posts: 3
Joined: 14 Jul 2022, 09:47
Antispam: Yes
Location: Switzerland
Contact:

Re: WhitneyAVHarmonicSolver for health protection

Post by pziegler »

I found the problem in the mesh.
Even if i don't understand it because in ElmerGUI it looked well. But there was no field calculated, only currents...
The mesh was created in Salome from a geometry created by the new (very nice!) shaper workbench.
I now exported it to the geometry module to create a partition consisting only of 2d components...that's until now not possible in the shaper WB.
Here's the working project with DC (MagnetoDynamics2D):
WeldingMachine2d.zip
(251.65 KiB) Downloaded 59 times
With AC:
If I calculate it in harmonic, then one can see the current distribution as expected (skin & proximity effect)
Current distribution at 100 Hz.png
Current distribution at 100 Hz.png (61.98 KiB) Viewed 683 times
but the field surrounding decreases very much with increasing frequency. That's not logic from my understanding of the physics.
If I change the material of the conductors from copper to air, then the field gets high as expected.
What's that :shock: ?
Working project with conductors made of copper at 100 Hz (MagnetoDynamics2DHarmonic):
WeldingMachine2d - AC.zip
(98.42 KiB) Downloaded 56 times
xiejinzhi
Posts: 22
Joined: 16 Jun 2023, 09:45
Antispam: Yes

Re: WhitneyAVHarmonicSolver for health protection

Post by xiejinzhi »

hi,
thank u for sharing your project file "WeldingMachine2d - AC.zip",but when i run "start solver",it shows an error: SectionContents: Mismatch of declared and given dimension for keyword "angular frequency". Ignored input: * 6.28

can u help me ,plz
kevinarden
Posts: 2237
Joined: 25 Jan 2019, 01:28
Antispam: Yes

Re: WhitneyAVHarmonicSolver for health protection

Post by kevinarden »

changing it to
Angular Frequency = Real MATC "100.0 * 6.28"

worked for me
xiejinzhi
Posts: 22
Joined: 16 Jun 2023, 09:45
Antispam: Yes

Re: WhitneyAVHarmonicSolver for health protection

Post by xiejinzhi »

kevinarden wrote: 29 Jun 2023, 15:29 changing it to
Angular Frequency = Real MATC "100.0 * 6.28"

worked for me
thank u
Post Reply