B-H Curve MagnetoDynamics2D

[Bavragor]

Hello Elmer-Users and developers,

I have a problem with the MagnetoDynamics2D-Solver. I have to compute a simple round coil with an iron core. The iron have a nonlinear B-H Curve. In the Models Manuel it is written, that in the Materials part you have to define the B-H Curve as cubic spline with the property Cubic Real. The Problem is that i have no idea how to set the B-H curve for the material.
Could someone give me a short example how the materials part with a B-H curve should look like?
If it is necessary i will give you the files of my project.
Thank you in advance! and sorry for the bad english

Michael

[raback]

Hi

See test case:
http://sourceforge.net/p/elmerfem/code/ ... /mgdyn_bh/

-Peter

[Bavragor]

Thank you very much Peter

Is there in the Documentation a hint how to do this and I fail to notice it?

[Bavragor]

Hi again,

I try to include the B-H-Curve like in the example, but got errors:

Code: Select all

Program received signal SIGSEGV: Segmentation fault - invalid memory reference.

Backtrace for this error:
#0  6f61143e
#1  6f68ec2f
#2  00401261

Anyone an idea what I'm doing wrong? I attached my sif-file, the geometry data and the b-h-curve.

Edit: my OS is Windows 7 Professional

Thanks in advance.
Michael

[akirahinoshiro]

I tried your code on win7 and Linux.
On Win7 I got some errors like

Code: Select all

ERROR:: Model Input: For property name:[target variable complex]

.
On Linux there was a convergence problem. After changing the bh to relative permeability 1 it worked on Linux.
So I think you have at least two problems.
Your Windows have the latest svn version compiled?
And the second one should be your case file with that I cannot help you.

[Bavragor]

Hi,

Thanks for the input. i use the latest svn-version. Maybe i will try it again on Linux.
I know that my case-file worked with relative permeability = const. But with an nonlinear curve i have no idea how to do that.
In the models manuel it is written that the input of an nonlinear curve should work with B-H Curve = Cubic Real ... in the style of a cubic spline. When I look on the other test with the MagnetoDynamics-Solver the style is H-B Curve = Real with some points of this curve. With the first method i only get errors, that the program doesn't know B-H Curve or Cubic Real. With the second method i got the errors above mentioned.
Is there no way to run the program in the style which is mentioned in the models manual for MagnetoDynamics2D-solver?

Maybe I'm just to stupid to get an understanding of the program and too unexperienced with FEM. But unfortunately i have to work with it^^

[raback]

Hi Bavragor

Can you run the test case mgdyn_bh?

-Peter

[Bavragor]

Hi,

Yes, this test works.
Maybe i should try to use MagnetoDynamics as in the example and not the MagnetoDynamics2d solver?

[akirahinoshiro]

If you need fast a 2d only simulation maybe give femm a try.
It is much easier to understand but elmer is much more powerful.

But i found another thing you should look at.
From the elmer models manual:

Code: Select all

B-H Curve Cubic Real
The B-H curve must be given as a cubic spline. This enables that the derivative of the curve
is computed analytically from the spline coefficients.

I don't think the whitney solvers are the right ones for 2d.

For the 3d solution maybe look at this:
viewtopic.php?f=3&t=2053

[Bavragor]

That is the problem i have. In which way i can told elmer that this is a cubic spline? Which form should it have?

FEMM is compared to Elmer very easy. But for my bachelor thesis i have to use elmer.