B-H loops / Hysteresis

Discussion about new features and development support

Moderator: thorsten

Post Reply
frankst
Posts: 32
Joined: Thu 21 Jul 2016, 11:36

B-H loops / Hysteresis

Post by frankst » Wed 07 Jun 2017, 10:24

Hi Thorsten,

Do you believe it would be possible to implement "realistic" B field <-> H field relations, e.g. to simulate saturation of magnetic materials?
While full B-H loops, i.e. with hysteresis, would be most welcome, I guess these would be difficult to implement.
May be simpler (and still helpful) would be just a (bijective) functional relation of B and H.
The function could be specified by the user.

Cheers
Frank

thorsten
Posts: 1438
Joined: Mon 27 Jun 2011, 12:26

Re: B-H loops / Hysteresis

Post by thorsten » Sat 10 Jun 2017, 08:49

Hi Frank,

sorry I missed this topic, I think my notifications are still off for this new sub-forum...

To answer your question. No I do not think that this is possible in FDTD. Any non-linear material is very very difficult and I'm not sure if possible at all.
Even a functional relation between H/B would be close to impossible (I think) as this would require to adept the coefficients during simulation? I have no idea.
You may of course do some research about FDTD and these kind of materials. But I have my doubts...

regards
Thorsten

frankst
Posts: 32
Joined: Thu 21 Jul 2016, 11:36

Re: B-H loops / Hysteresis

Post by frankst » Sat 10 Jun 2017, 21:16

Hi Thorsten,

Thank you for the answer.
I already suspected that this would be difficult.
Without having studied the topic in detail, my guess would be that a simple functional relation would be possible (by adapting properties for every time step) but pretty speed limiting.
Hysteresis is certainly a lot more complicated.
Though I do not believe these things are impossible.
They are just difficult ;-)

Cheers
Frank

frankst
Posts: 32
Joined: Thu 21 Jul 2016, 11:36

Re: B-H loops / Hysteresis

Post by frankst » Mon 12 Jun 2017, 10:32

Hi Thorsten,

Just for reference, I stumbled upon this:
https://accelconf.web.cern.ch/accelconf ... PAT040.PDF
It seems that in principle hysteresis can be modeled in FDTD codes.
Obviously, I have no clue how much work the implementation would be in openEMS.

Cheers
Frank

Post Reply