How to define hysteresis loop in ANSYS MAXWELL?

[omidfn]

I am trying to simulate an electric motor in ansys maxwell software, but I can't define the whole loop in four quadrants for the material used in the rotor however it is essential to do this since the motor is of hysteresis type and works on the principle of working point moving all around the loop so some output torque can be developed.
any thoughts on that?

 

[Skogsgurra]

Are there no look-up tables where you can put ampereturns and flux density and let the system interpolate?
Have used an ellips approximation in other cases. It works quite well if you are not driving the material into (deep) saturation.
Maxwell (which I never used) should, by the name of it, have an easy-to-use method available.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[omidfn]

Dear Skogsgurra,

Thank you for you quick response.
Unfortunately, there does not seem to be an apparent way of defining a BH loop as the case for BH curve. Yes there is a look up table and I have tried it but when I enter the values for the whole loop, an error pops up which can be seen below

it would only confirm if there is a single corresponding value of B for each value of H like this one

therefore I can't define the BH loop in any closed form including an ellips approximation.

Sincerely

 

[aussiejohn2]

G'day omidfn,
Isn't one branch enough to define the curve?
John.

 

[omidfn]

Dear aussiejohn2,

I'm afraid so, the hysteresis motor works on the principle that in hard magnetic matters the area enclosed by the loop is considerable so the energy conserved in the loop could be used to produce acceptable torque so it's imperative that in fact it should be a loop not a curve.
Although that being said I could be wrong, so if you thing that is the case please do share your insights.

Cordially

 

[aussiejohn2]

Hi omidfn,
I meant: The single-valued hysteresis curve B1(H) that you successfully entered is the upper branch of the hysteresis loop. With few exceptions, hysteresis loops are symmetrical so the lower branch is given by B2(H)=-B1(-H). Perhaps the software assumes this and you only half to enter half the curve.
John.

 

[omidfn]

Dear aussiejohn2,

thank you for your opinion.
I thought so at first but to examine that I defined a transient BH loop plot to see what happens inside the hard magnetic material as the simulation was being conducted, it appeared that it does not make such an assumption and simply goes up and down the predefined curve, apparently not so intelligent after all!
Although for materials of which the BH curve starts from (0,0) for instance steel, it does assume a symmetry with respect to the origin.

Cordially

 

[Skogsgurra]

I bet that Comsol Multiphysics could be a good investment for you. Seems to be able to do anything from cow's digestion to nuclear explosions...

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[omidfn]

Dear Skogsgurra,

Thank you for the offer. I checked that software out and it seemed legit, nonetheless I have to do this simulation with Maxwell, maybe next time I'll give that a try.

sincerely

 

[aussiejohn2]

G'day omidfn,

https://www.ansys.com/-/media/ansys...aswc2014-superconducting-magnet-influence.pdf

on pages 10,15,16,19 seem to have it doing the full loop (for transient simulations).
I wish there were better open source electromagnetic FEM; elmer and openfoam are making progress but last time i looked, not there yet for non-specialists.
John.

 

[omidfn]

Thank you