Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Torque problems for a 3-phase PMSM motor (2D)
Posted Nov 20, 2022, 5:44 p.m. EST Electromagnetics, Materials, Results & Visualization Version 6.0 7 Replies
Please login with a confirmed email address before reporting spam
Hello!
I am currently modelling a 3-phase PMSM motor in 2D in comsol. It is an outer rotor, inner stator design and everything seems to be working as expected except for the torque calculation.
Due to the motor being part of a project with an NDA agreement I can unfortunately not upload my model but I will try to explain the things I find weird with the torque results as best I can.
I am using both the Arkkio torque method and the normal force calculation. Both result in weird, and wildly different, results. The thing that bothers me the most is that I am getting non-zero torque for both torque calculation methods while supplying zero current to the coils.
The attached image is of the torque when supplying no current to the coils and rotating the rotor with 3000 rpm through a rotating mesh. The green line is the torque from the force calculation and the red line is from Arkkio.
The force calculation is defined over the domain of the rotor and half of the airgap (closest to the rotor.
The Arkkio torque domains are both halfs of the airgap, separated with a boundary pair with continuity.
Another weird effect of the torque is that the conductivity of the stator core has a big impact on it. Lowering the conductivity lowers the torque. If I choose a conductivity that is about two powers lower than the actual conductivity the torque becomes almost zero for zero current.
The entire mesh is extremely fine and with a boundary layer.
Any help would be appreciated as I have been puzzled with this torque issue for quite some time now and tried everything I could think of.
Attachments:
Please login with a confirmed email address before reporting spam
Hello Hendrik,
without knowing the details of your design, as this is a PMSM design with permanent magnets, shouldn't you expect some cogging torque due to the permanent field?
Cheers Edgar
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Please login with a confirmed email address before reporting spam
Hello Edgar and thank you for taking the time to reply!
I agree with you that some cogging torque should be expected. However, a mean of 9 Nm (from Arkkio) of cogging torque with no current seems to be to much for me. The motor is approximately only 100mm long with a radius of 40 mm.
If I apply a sinosoidal 3-phase curent to the coils I get a reduction in Torque, compared to the zero-current case.
Attached is the torque results for a run with 100A peak current supplied. This run is also at 3000 RPM.
Could you think of anything that I could have done wrong or missed while building the model that would result in such results? I know that it is very hard to answer without seeing the model.
Best Regards, Fredrik
Attachments:
Please login with a confirmed email address before reporting spam
As the curve looks rather erratic, is your time stepping fine enough to resolve the expected torque curves?
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Please login with a confirmed email address before reporting spam
Even with finer time stepping I got the same result. However, I think I found a solution to my problem. I found this guide:
https://www.comsol.com/blogs/how-to-analyze-an-induction-motor-a-team-benchmark-model
In which the stator core is given a conductivity of 0 S/m as the core is laminated (and therefore electrically insulated I assume).
If I assign zero conductivity to the stator core in my model the torque starts to behave as expected. Zero current gives almost Zero torque with a little bit of cogging torque. And torque is mostly only influenced by the supplied currrent.
Have I understood it correctly that this is the way to "tell" comsol that the stator core is insulated from the windings?
The Arkkio torque is the one now producing expected results. But the force calculation for torque is not. It feels like they should be giving similar results if the model was working properly.
Please login with a confirmed email address before reporting spam
The lamination prevents eddy currents and thus losses in the core and the low conductivity has a similar effect. The eddy currents (which result from the permanent field, even with zero coil current) produce a damping effect and this is probably what you saw in your model. Regarding the force calculation, it is very mesh dependent. You need to take care particularly in the air gap. I would recommend to do a mesh convergence study.
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Please login with a confirmed email address before reporting spam
Thank you so much for your time Edgar!
The problem seems to have been that I was running the model with a conductivity equal to the material of the core. Which would be like having a solid core without laminations resulting in huge Eddy currents at the speeds I was running it at.
Now I will just have to find the conductivity that represents the stator-core with laminations.
I will look into a mesh convergance study as well to see if I can further improve on my results.
Best regards, Fredrik
Please login with a confirmed email address before reporting spam
Lamination is not exactly easy. You may want to look into this paper:
D. Lin, P. Zhou, et. al. A New Nonlinear Anisotropic Model for Soft Magnetic Materials IEEE Transactions on Magnetics Vol. 42, No. 4, p. 963-966, April 2006
I was using it for earlier work with PMSM motors.
-------------------Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com