[femm] Re: Problems with non-linear B-H-curves

[David Meeker <dmeeker@xxxxxxxxxxxxxxxxx>]

Kuepferling Michaela wrote:

> In our group at the Vienna University of Technology we measure magnetic
> materials with magnetic pulse fields. I wanted to compare a
> hysteresis-measurement of a Ni-sphere with a radius of 4.85mm with the
> results of FEMM. The following questions arised: I got with FEMM B~8.3T
> inside the sphere (see the attached file NiBr2.EMF). But with the applied
> field of ~5.2T and J(Ni)~0.6T, when Ni is saturated, B becomes about 5.8T.
> Why is the calculated B so high? The other problem is, that we expected a
> decay of B inside the sphere because of the induced eddy currents. But with
> FEMM I found an increase of B with decreasing radius (see the attached file
> NiBrki.EMF).
> I welcome any comments!
> Michaela Küpferling

It's hard to tell what is going on w/out seeing the .fem file that led to the
results. Could you send it to me (perhaps just to me at dcm3c@xxxxxxx)?

Anyhow, one thing to remember is that although magnetostatic problems can use a
nonlinear B-H curve, the harmonic solver is limited to linear materials properties.
For harmonic problems, it just uses the values for linear permeability entered in the
mu_r and mu_z boxes in the Material Property dialog. At some point,I might kludge in
support for nonlinear materials in harmonic problems, but the problem is that a
nonlinear harmonic formulation probably wouldn't be that accurate -- for nonlinear
materials, you don't get a purely harmonic response from a sinusoidal applied current
or field.

Dave.
--
http://members.aol.com/dcm3c