Inductance Extraction (MagDyn2D, MagDynPost, Circuits & Dynamics)

General discussion about Elmer
Post Reply
greenlinux
Posts: 9
Joined: 15 Aug 2020, 18:11
Antispam: Yes

Inductance Extraction (MagDyn2D, MagDynPost, Circuits & Dynamics)

Post by greenlinux »

I am attempting to evaluate the inductance of a magnetic component using simulated with MagetoDynamics 2D and the Circuits and Dynamics modules.

One approach on this forum is to use the (Electro)Magnetic Field Energy which is output from MagnetoDynamics 2D on the console as the solver runs. However, this produces the sum of the electric field energy and the magnetic field energy (Elmer models manual pp. 116 version 3/8/20). To use Wm = ½ L * I^2 (where Wm is the magnetic field energy, L is the inductance and I is the current) only the magnetic component of the solver’s output is required but it does not split them up into electric and magnetic parts.

I have explored the save scalars module for more than a week of late-evenings (that is a valid unit of time when you have children) but, have been unable to determine how to construct what I need using it, although I am fairly sure it will be possible.

I have considered (and am still considering) finding a method to load the VTU into Matlab and asking it to integrate B . H and then use Poynting’s approach to provide the stored magnetic energy. I can’t help feeling that this is not the best way and it should be possible to obtain this data using SaveScalars rather than post-processing the mesh results in another application.

My final line of attack is to use a transient simulation and a known ideal resistance in series with the FE component (inductor) to obtain the time constant, then work backwards to get L. It should be possible to obtain the inductance from a static analysis though. I have checked the same geometry in FEMM and it provides an answer that is sufficiently close to my analytical calculations to convince me that I’m not doing anything stupid.

The inductance is close to 150x10^(-6) H. The setup is cartesian 2D. In the sif there are 2 component regions representing one winding around a ferrite core everything else is air. Current flows into the "page" in one half of the winding and out of the "page" in the other. There is another winding (2 regions) but it is unused presently and is air and has no component sections associated with it. The current density, flux density and magnetic field all seem fine in the VTU. J is about the value I expect and B is about the value I expect (except in the corners of the core where there is saturation), H is big in the air regions between the ferrite blocks that make up the core and is low elsewhere. The VTU data has been removed because it causes the zip to exceed the max file size. I'm running Elmer 8.4 Release on Windows 7 64 bit.

My questions is, can I get the inductance using savescalars? If so, could you point me in the right direction, please?

Many thanks,

James
Attachments
InductorCircuitsAndDynamics.zip
(220.77 KiB) Downloaded 285 times
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Inductance Extraction (MagDyn2D, MagDynPost, Circuits & Dynamics)

Post by raback »

Hi

SaveScalars is agnostic to physics. It just saves stuff that is provided by other solvers. We have a pretty dirty way of communicating values to SaveScalars: We add strings with prefix "res:" to the Model % Simulation list.

If you want to understand where the Energy is being computed look "Energy" in the this module:
https://github.com/ElmerCSC/elmerfem/bl ... Fields.F90

Indeed it seems that Energy consists of two parts. It would be rather miminal coding to separate it to its two components.

-Peter
greenlinux
Posts: 9
Joined: 15 Aug 2020, 18:11
Antispam: Yes

Re: Inductance Extraction (MagDyn2D, MagDynPost, Circuits & Dynamics)

Post by greenlinux »

Hi Peter,

Many thanks for replying. You obviously put a lot of your time into working on and supporting ELMER. I am very grateful.

Presuming I make the required addition to the source file to list the parts of the energy separately. Would I then compile the source with elmerf90 as Pavel did with the rotating machine kinematics module on pp 61 of VTT-R-02819-17 (induction machine tutorial report) and as documenter in chapter 18 of the Elmer Solver Manual?

Presumably I should make the change and submit a pull request to make the adjustment available in future versions or is this too niche / small to be bothered with?

Thanks,

James
mika
Posts: 230
Joined: 15 Sep 2009, 07:44

Re: Inductance Extraction (MagDyn2D, MagDynPost, Circuits & Dynamics)

Post by mika »

How about just setting "Permittivity = 0.0"? Then the total electromagnetic energy computed by the solver would be the magnetic energy.

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

Re: Inductance Extraction (MagDyn2D, MagDynPost, Circuits & Dynamics)

Post by raback »

Since two minutes there is a keyword for CalcFields solver:

Separate Magnetic Field = Logical True

This will enable to print out the parts separately. This could be a default but I worry that there are some uses of this out there. This should be backward compatible.

-Peter
greenlinux
Posts: 9
Joined: 15 Aug 2020, 18:11
Antispam: Yes

Re: Inductance Extraction (MagDyn2D, MagDynPost, Circuits & Dynamics)

Post by greenlinux »

Hi Peter,

Many thanks for adding that in. Much more professional than me hacking away at it. No Fortran for me since my PhD days.

I'll download the nightly build from 24/9/20 tomorrow and fire it up.

James
Remo
Posts: 1
Joined: 24 Jun 2022, 12:22
Antispam: Yes

Re: Inductance Extraction (MagDyn2D, MagDynPost, Circuits & Dynamics)

Post by Remo »

In case somebody stumbles over this, it's

Code: Select all

Separate Magnetic Energy = Logical True
Elmer Version: 9.0
Post Reply