The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Modeling >> Passive Devices >> Simulation and Modeling of Nonlinear Magnetics
https://designers-guide.org/forum/YaBB.pl?num=1201013958

Message started by TN on Jan 22nd, 2008, 6:59am

Title: Simulation and Modeling of Nonlinear Magnetics
Post by TN on Jan 22nd, 2008, 6:59am

Hello,
I have tried the nonlinear core model by Mark C. Williams, R. S. Vogelsong, and K. S. Kundert in Simulation and Modeling of Nonlinear Magnetics.

Pittingly, I cannot exactly reproduce the simulation results given on pages 10 and 11 of this paper.
But the results from my simulation look rather good.
The model shows in core c1 the hysteresis loop from Figure 4 with the limits from Figure 6.

I'm also wondering about the names of the result curves in the given diagram of Figure 5.
The label 'n4' of curve vnn(i,'n4') appears nowhere in the test circuit.

Can anybody of you acknowledge the results presented in Figures 5 and 6 of the paper?

With thanks in advance and best regards,
Tobias Nähring

Title: Re: Simulation and Modeling of Nonlinear Magnetics
Post by Ken Kundert on Jan 22nd, 2008, 9:55pm

Sorry, I can tell that the results don't seem to correspond precisely with the circuit given. The results were originally generated using Artist, and when the files were converted to pure netlists some of the node names were changed. Also, it appears that damping on the input sources was added as well.  I'm afraid that the details of how to generate the given plots have been lost over the years.

-Ken

Title: Re: Simulation and Modeling of Nonlinear Magnetics
Post by TN on May 29th, 2008, 12:38am

Hello,
the state equation for the irreversible magnetization in [1]  is

dMirr = (Manh - M)/(sign(Hdot)*k - alpha*(Manh-M)) * Hdot

while the formula suggested in [2] is  

dMirr = (Manh - Mirr)/(sign(Hdot)*k - alpha*(Manh - Mirr)) * Hdot.

I assume that this difference results from a modification of the power balance equation from

Mirr * d Beff / dt = Man * d Beff / dt - k* d Mirr / dt * sign(Hdot)

to

M * d Beff / dt = Man * d Beff / dt - k* d Mirr / dt * sign(Hdot).

(Note, that the first version of the power balance is not contained in [2] in this form. But one has to apply the proposed modifications from section "3.4. Domain wall motion of flexible domain walls" to equation (17) or (19) from the paper).
To me it seems that the first version (used by Jiles-Atherton) has a better physical justification since the second version tries to balance the power losses of the irreversible Bloch wall motions with the overall magnetization power which also includes the elastic Bloch wall deformations. But I'm not too much experienced with magnetics. So maybe I am wrong.

Is there some other physical justification for the version from [1]?

I am asking because there are practical cases where the two different formulations deliver significantly different results.

Bibliography:

[1] www.designers-guide.org/Modeling/mag.pdf

[2] D. C. Jiles and D. L. Atherton: Theory of ferromagnetic hysteresis. Journal of Magnetism and Magnetic Materials, Vol 61 (1986), pp. 48-60.

With best regards,
Tobias Naehring


The Designer's Guide Community Forum » Powered by YaBB 2.2.2!
YaBB © 2000-2008. All Rights Reserved.