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.

Lift on Magnets due to Eddy Currents

Please login with a confirmed email address before reporting spam

Hello

I’m relatively new to Comsol and I’m doing a project where I want to simulate the lifting force on a magnet Halbach array as it rotates over a conducting surface (disk) of aluminum. Eddy currents induced in the disk by the rotating array will generate an opposing magnetic field, which will then generate lift.

The simulation that I’ve made is 3D using the Magnetic and Electric Fields (mef) physics interface and stationary study. Instead of rotating the magnet array I have opted for using the Velocity (Lorentz Term) to emulate the conductive disk rotating. To get the lifting force I use Lorentz force contribution, z component integrated all over the disk.

I then do a parametric sweep, where the rotational speed of the disk is incremented. Then the force acting on the disk is plotted as a function of the angular velocity. The results end up being a polynomial, but I would expect the force to be linearly proportional to the angular velocity. Conducted test also back this up - showing a linear relationship.

How do I do this right? I want to simulate the force on the disk as it is being repelled by the magnetic field of the rotating Halbach array. I hope my problem is understandable.

Thanks in advance.

Mads Larsen



4 Replies Last Post 25 ott 2017, 14:20 GMT-4

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 25 ott 2017, 03:54 GMT-4

Hallo Mads,

from my first quick look I do not see any problems. Why do you expect it to be linear? I think the force should be an even function of spped because when rotating in opposite direction the force should be still repelling. So around zero it looks quadratic. At higher speeds the force will be smaller due to skin effect and the curve might look linear. Try to compare skin depth (with correct pole multiplicity) with the disk thickness.

Regards,

Jens

Hallo Mads, from my first quick look I do not see any problems. Why do you expect it to be linear? I think the force should be an even function of spped because when rotating in opposite direction the force should be still repelling. So around zero it looks quadratic. At higher speeds the force will be smaller due to skin effect and the curve might look linear. Try to compare skin depth (with correct pole multiplicity) with the disk thickness. Regards, Jens

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 25 ott 2017, 07:34 GMT-4

Hello Jens

Thanks a lot for your reply.

You might be right that the relationship is not linear but I have another problem. The data from the experiments show the force to be much higher than what the simulation predicts. The force measured in the experiment are approximately 4 to 5 times larger than the forces given in the Comsol simulation. I did a quick plot of the simulation and the experiment data in the attached file. Do you have any idea about why the simulation results show much weaker force than the data from the experiment?

Best regards

Mads

Hello Jens Thanks a lot for your reply. You might be right that the relationship is not linear but I have another problem. The data from the experiments show the force to be much higher than what the simulation predicts. The force measured in the experiment are approximately 4 to 5 times larger than the forces given in the Comsol simulation. I did a quick plot of the simulation and the experiment data in the attached file. Do you have any idea about why the simulation results show much weaker force than the data from the experiment? Best regards Mads


Nirmal Paudel Tribhuvan University/Electrical Engineering

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 25 ott 2017, 09:58 GMT-4
Updated: 7 years ago 25 ott 2017, 09:58 GMT-4

Hi Mads,

I have couple of comments regarding this.

  1. The mesh refinement is needed to be sure that the calculation is accurate from the simulation. The Maxwell Stress tensor force obtained from "Force Calculation" node is based on the surface integration on the Disc, therefore, more sensitive to mesh size. However, the Lorentz force is a volumetric integration and are usually more accurate even with relative coarse mesh. Eventhough you are calculating force via Lorentz method, I would still recommend you to refine the mesh in the disc and recompute the solution.

  2. The other major issue here is the the vecocity vector under "Velocity (Lorentz Term)" should have "-2piRPMy" & "2piRPMx". See the attached document of a Homopolar generator examples on page # 15. https://www.comsol.com/model/download/384861/models.acdc.homopolar_generator.pdf

Hope this helps. Best Regards, Nirmal Paudel

Hi Mads, I have couple of comments regarding this. 1. The mesh refinement is needed to be sure that the calculation is accurate from the simulation. The Maxwell Stress tensor force obtained from "Force Calculation" node is based on the surface integration on the Disc, therefore, more sensitive to mesh size. However, the Lorentz force is a volumetric integration and are usually more accurate even with relative coarse mesh. Eventhough you are calculating force via Lorentz method, I would still recommend you to refine the mesh in the disc and recompute the solution. 2. The other major issue here is the the vecocity vector under "Velocity (Lorentz Term)" should have "-2*pi*RPM*y" & "2*pi*RPM*x". See the attached document of a Homopolar generator examples on page # 15. https://www.comsol.com/model/download/384861/models.acdc.homopolar_generator.pdf Hope this helps. Best Regards, Nirmal Paudel

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 25 ott 2017, 14:20 GMT-4

Hello Nirmal

Thank you very much for the help. That seems to have solved the problem.

Thanks a lot.

Best regards Mads Larsen

Hello Nirmal Thank you very much for the help. That seems to have solved the problem. Thanks a lot. Best regards Mads Larsen

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.