DSPRelated.com
Forums
Online · Live + on demand
February 9-11, 2027
Price goes up Jan 1: register now to lock in today's rate
Register Now →

Modeling Hysteresis

Started by Tim Wescott ●July 18, 2012
On 7/18/2012 6:58 PM, Jim Thompson wrote:
> On Wed, 18 Jul 2012 11:00:36 -0500, Tim Wescott <tim@seemywebsite.com> > wrote: > >> Anyone have any pointers to mathematical models that they like which >> model hysteresis phenomena? Any hysteresis phenomena? > > If you find one that doesn't cause convergence issues in a simulator, > let me know. I'd like a general purpose hysteresis model myself. > > Magnetic hysteresis adds the modeling difficulty of being variable > with starting point :-(
The John Chan model of magnetic hysterisis built into LTSpice seems to work, and it's easy enough to extract the fitting parameters from the ferrite manufacturers' data sheets. The LTSpice help file credits it to "John Chan et la. in IEEE Transactions On Computer-Aided Design, Vol. 10. No. 4, April 1991 but extended with the methods in United States Patent 7,502,723". -- Bill Sloman, Nijmegen
W dniu &#347;roda, 18 lipca 2012 18:00:36 UTC+2 u&#380;ytkownik Tim Wescott napisa&#322;:
> Anyone have any pointers to mathematical models that they like which > model hysteresis phenomena? Any hysteresis phenomena? > > I&#39;m feeling kind of constrained by NDA at the moment, but all of the > hysteresis behaviors that I know of (gear backlash, magnetic, stress in > piezoelectric crystals) seem to share some common behaviors, so just > about _any_ model that&#39;s mathematically tractable would be helpful. > > I also know that I&#39;ve seen pictures out there in internet-land of the de- > Gaussing process plotted on a B-H curve that showed that multiple small > hysteresis cycles tend (at least in some magnetic materials) to relax the > stored energy due to hysteresis -- I want to run that by my customer and > ask them if it happens in their material. > > Thankee in advance. > > -- > My liberal friends think I&#39;m a conservative kook. > My conservative friends think I&#39;m a liberal kook. > Why am I not happy that they have found common ground? > > Tim Wescott, Communications, Control, Circuits &amp; Software > http://www.wescottdesign.com
A back-of-the-envelope model, which behaves quite well: y[n] = y[n-1] + (x[n] - y[n-1])^5 It is pure hysteresis, without saturation. It is stable for roughly -0.25 < x,y[0] < 0.25. The idea behind the formula is that when y is far from x, y will quickly tend to x. But when they become close, y stops chasing x, leaving some distance. This is the meaning of the fifth power. You can include some scaling the powered component (it will work as a coupling coeff), also the exponent can be tweaked but must be odd. WP