2 questions - magnetodynamics / whitney elements

[madstamm]

Hello All,

I have put together a "reference" case, with which to test the magnetodynamics/whitney elements solver.
Elmer project folder: http://dl.dropbox.com/u/1102403/elmer_p ... ore.tar.gz
Salome study: http://dl.dropbox.com/u/1102403/elmer_p ... oncore.hdf

It is a classic "iron core with airgap" case, using the following parameters:
- a linear! B-H iron core model with ur=1000
- core has outer dimensions of 10cm, 2cm "thickness"
- airgap is 1mm
- 2 turn copper coil, 4mm diameter and about 24cm long, giving about 0.38mOhm
- 0.38mV applied to coil, giving 1A current
- air cube with 50cm sides

To check the accuracy of the solver, I added a simple surface inside the iron core, so I can check the flux density or magnetic field inside the core.

From hand calculations, i expect the flux inside the core to be about 1500T (unrealistic since core is a linear model), and the mag field inside the core to be 1.5At/m.
Resistance and joule heating is calculated exactly to what I expect.

Now my questions:


1)
The solver, using the SIF file attached, gives me 2.45E-5 At/m for mag field through BC4 (my "measurement" boundary)... I expect 1.5At/m... what am I doing wrong?

2)
the SaveScalars solver also gives the identical value for "total joule heating" and "effective resistance"... I think the "joule heating" value is accidentally written twice.


Thank you for any help - much appreciated in advance!

Mike

PS:

SaveScalars output:
--------------------------------------------------------------
-2.485450933473E-005 -3.174803583010E+000 3.213939506523E+000 3.754303291024E-004 3.754303291024E-004 2.848053946269E-021 1.393910653993E-006

File started at: 2012/01/24 21:58:42

Variables in columns of matrix: magscalars.txt
1: convective flux: magnetic field strength over bc 4
2: min convective flux: magnetic field strength
3: max convective flux: magnetic field strength
4: res: total joule heating << same as 5
5: res: effective resistance << same as 4
6: res: eddy current power
7: res: magnetic field energy
--------------------------------------------------------------


PPS: pictures

[raback]

Hi Mike

Impressive pictures again!

I had a quick look at the scalars created by StatCurrentSolver - there was a bug. The resistance should come out now correctly. Thanx for reporting.

-Peter

[madstamm]

Hi Peter,

Thank you for fixing the stat current solver Bug..

any idea why the flux values dont seem to be correct?

Mike

[raback]

Hi Mike

Note that the "convective flux" gives the total integral over area. Perhaps you should divide it by the area?

-Peter

[madstamm]

I still cannot get the result, but need to do some experiments first to see how we can home in on the problem.

in the meantime, the result with uCore=10 (this is realistic by the way)