AC Drive China Forum

Forum » General Discussion » Saturation on the Iron losses for PM machines

Topics: **Saturation on the Iron losses for PM machines** on General Discussion
# Saturation on the Iron losses for PM machines

Hello
Everyone I would like to know if there is some R&D done on the
impact of Saturation on the Iron losses (impact on harmonic contents
then impact on iron losses) for PM machines. Thanks a lot

#1

09-23-2013 09:32 AM

Top #2

Dear Abedelhadi. You do not need an R&D study. Any good reputable Motor/Generator manufacture has this data today. So contact, ABB, GE, Regal Beloit, Siemens, WEG or TECO Westinghouse. Studies are nice, but nothing beats good hard factual data to global standards.

09-23-2013 12:08 PM

Top #3

I belive NEMA and some steel companies have done this

09-23-2013 02:15 PM

Top #4

Michel, while it may be true of some steel companies, motor and drive manufactures are the BEST source for real data. They have actual, factual test data. Especially if they play in API markets and applications.

09-23-2013 05:06 PM

Top #5

Hi Abedelhadi,

Can you look into Somaloy 550 properties & related activities in motors

Can you look into Somaloy 550 properties & related activities in motors

09-23-2013 07:13 PM

Top #6

thank you all,

I was looking for some analytical models. My objective is to obtain high accuracy loss models for PMSM. 3D FEM is ok when material data is accurate but even in this case from a sample (motor) to another you may have some time up to 10 or 15 % error. you need to calibrate the model once again using remeasured values.

I'm thinking now of a grey box modelling (circuit oriented model(some knowledge of the PM) + Measurement based model). do you know some papers on this topic?

Concerning the data from steel companies, when the iron sheets are manufactured into stator or rotor elements the properties change and so the iron losses.

I was looking for some analytical models. My objective is to obtain high accuracy loss models for PMSM. 3D FEM is ok when material data is accurate but even in this case from a sample (motor) to another you may have some time up to 10 or 15 % error. you need to calibrate the model once again using remeasured values.

I'm thinking now of a grey box modelling (circuit oriented model(some knowledge of the PM) + Measurement based model). do you know some papers on this topic?

Concerning the data from steel companies, when the iron sheets are manufactured into stator or rotor elements the properties change and so the iron losses.

09-23-2013 09:57 PM

Top #7

If you are looking for realistic analysis, then I would suggest; to characterize the stamping material for magnetization characterization and use the test results as an input to your 3D FEA. You can do BH characterization by manufacturing toroids of the material having same heat number that of the material you will be actually using in the product. Also BH characterization can be done at the same temperature at which the cores of PMSM will be stabilizing in the real application.

09-24-2013 12:51 AM

Top #8

BH characteristics change both with heat exposure and plastic deformation. I think the most accurate model feasible for a given motor would involve a series of BH models that cover the various local properties of the metal in the motor as built, together with a FEM treatment that adjusts the local BH model to best fit measured characteristics.

It might be helpful to know which motor characteristics present the greatest challenge. Is it just a matter of losses, efficiency and stall torque, or is electrical noise a more pressing concern. Perhaps vibration originating from harmonics is the main challenge. Each issue would present a different concern in modeling accuracy.

It might be helpful to know which motor characteristics present the greatest challenge. Is it just a matter of losses, efficiency and stall torque, or is electrical noise a more pressing concern. Perhaps vibration originating from harmonics is the main challenge. Each issue would present a different concern in modeling accuracy.

09-24-2013 03:29 AM

Top #9

yes I had done tests of plastic deformation effects on the material but at motor design stage. the complication of my present experiment is to not "open" the motor. the idea is to consider the motor as a black box and the target is to obtain a measurement based model. first a global accurate modelling (PMSM loss model) then an electrical behaviour model taking into account the harmonic contents effects. DC copper loss is the first to be measured, then the mechanical loss (with iron loss due to speed) then the iron loss due to current.separation between mechanical loss and iron loss is not done because we do not want to open the motor and extract the PM elements and run the no load tests for mechanical losses. we consider the mechanical losses constant with current variations.

don you know about papers presenting some theory about linking the input PWM voltage to the losses (specially iron losses) and all this only from measuring currents and voltages at the motor terminals and the torque and speed?

Thank you !

don you know about papers presenting some theory about linking the input PWM voltage to the losses (specially iron losses) and all this only from measuring currents and voltages at the motor terminals and the torque and speed?

Thank you !

09-24-2013 06:23 AM

Top #10

I can't direct you to a paper, but I can give some ideas for how to do the analysis you desire. If the motor is series wound, or if current in each winding is separately measurable, copper losses can be estimated by I^2R, where R is the winding resistance. Mechanical losses can be estimated by passively turning the motor (with another motor) open circuit, in a dynamometer setting where speed and torque can be measured.

We can determine Iron losses as <input power>-<output power>-<copper losses>-<mechanical losses>. This is not very accurate if iron losses are small (loss of precision from cancellation>.

Another way to estimate iron losses is the Steinmetz equation P=k*B^<alpha>,

where P is iron (hysteresis) losses, B is the maximum Magnetic Flux, and k and alpha are empirically derived constants (possibly derived from the manufacturer).

http://power.thayer.dartmouth.edu/WebCoreLoss.html

We can determine Iron losses as <input power>-<output power>-<copper losses>-<mechanical losses>. This is not very accurate if iron losses are small (loss of precision from cancellation>.

Another way to estimate iron losses is the Steinmetz equation P=k*B^<alpha>,

where P is iron (hysteresis) losses, B is the maximum Magnetic Flux, and k and alpha are empirically derived constants (possibly derived from the manufacturer).

http://power.thayer.dartmouth.edu/WebCoreLoss.html

09-24-2013 08:46 AM

Top #11

If the motor has a parallel winding, there are possibly currents circulating in opposite directions, cancelling at the nodes where the motor power is applied. It may be difficult or impossible to measure these currents in some intact motors.

Perhaps it is possible to measure magnetic flux, or use a clip-on ammeter in your motor, to resolve the parallel winding question.

Perhaps it is possible to measure magnetic flux, or use a clip-on ammeter in your motor, to resolve the parallel winding question.