Force calculations on permanent cube magnet and coil

Please login with a confirmed email address before reporting spam

Dear All,

I had a quick question about my model attached,

I have a cube permanent magnet above a coil. The system is trying to model a cube magnet that has north and south faces along the x-axis, and is fixed at the center and free to rotate (z-axis) left or right based on the polarity changes of the coil when current direction is changed. I have three questions:

1- When you run the model, I would expect there to be a negative y-axis force, no/negligible z-axis force (as the magnet and coil are symmetrical) and an x-axis force either in +/- based on the polarity of the coil. I.e. the coil is trying to pull/rotate the cube magnet left or right based on the polarity of it. However, I seem to get a z-axis force, and the x-axis doesn't seem to mirror the force on the cube equally in the +- when the polarity of the coil changes. why is that? shouldn't I get a relatively equal force on the cube magnet in x-axis but mirrored based on the polarity of the coil?

2- In my 1-D line plots, why are the plots not continuous and are jumpy?

Really appreciate all your support!

Best, F



5 Replies Last Post Jul 6, 2024, 11:59 a.m. EDT
Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jul 6, 2024, 5:57 a.m. EDT

A few comments:

  1. The mesh on and in the magnet is far too coarse for a good force calculation. Apply a dense custom mesh on the boundaries of the magnet. The coil may be ok but a higher mesh density won't harm. You will need to do a mesh convergence study to verify the mesh.

  2. Instead of using such a huge air box, make it smaller and apply infinite element domains at the outer boundaries of the box. This way you can make the box much smaller and save much memory.

Cheers Edgar

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
A few comments: 1. The mesh on and in the magnet is far too coarse for a good force calculation. Apply a dense custom mesh on the boundaries of the magnet. The coil may be ok but a higher mesh density won't harm. You will need to do a mesh convergence study to verify the mesh. 2. Instead of using such a huge air box, make it smaller and apply infinite element domains at the outer boundaries of the box. This way you can make the box much smaller and save much memory. Cheers Edgar

Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jul 6, 2024, 6:08 a.m. EDT

Thank you so much Edgar,

I am quite new with COMSOL, and I havent fully grasped the steps required to do better meshing optimization.

I tried doing a infinte element mesh for the air box, but doesn't seem to work. I tried the layer around the air box and used the infinite domain element, but the run time took so long and didn't seem to work for me.

I tried doing extra fine for the air box domain, and 0.1mm maximum size for the magnet but seems to take forever to build the mesh.

Could you please show me the steps I need to take? and how to do a mesh convergence study?

Does the air domain have to be very fine too?

Thank you so much for your support!

Best, F

Thank you so much Edgar, I am quite new with COMSOL, and I havent fully grasped the steps required to do better meshing optimization. I tried doing a infinte element mesh for the air box, but doesn't seem to work. I tried the layer around the air box and used the infinite domain element, but the run time took so long and didn't seem to work for me. I tried doing extra fine for the air box domain, and 0.1mm maximum size for the magnet but seems to take forever to build the mesh. Could you please show me the steps I need to take? and how to do a mesh convergence study? Does the air domain have to be very fine too? Thank you so much for your support! Best, F

Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jul 6, 2024, 6:28 a.m. EDT

There are examples in the application library showing the infinite element domains. For a mesh convergence study you define a parameter that multiplies with the mesh size parameters. You can then run a parametric sweep over this parameter.

In the forum I am happy to give hints, but I cannot afford the time to do the detail work.

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
There are examples in the application library showing the infinite element domains. For a mesh convergence study you define a parameter that multiplies with the mesh size parameters. You can then run a parametric sweep over this parameter. In the forum I am happy to give hints, but I cannot afford the time to do the detail work.

Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jul 6, 2024, 6:39 a.m. EDT
Updated: 4 months ago Jul 6, 2024, 6:39 a.m. EDT

Thank you, I totally understand, I really appreciate your hints.

Whats the benefit of a mesh convergence study?

If i try to do a mesh sweep on the cube would that work? and I can make the element sizes very small? Is a mesh sweep with one of the square faces a good way to do things?

Can I mesh the faces of the cube differently from the inside and body of the cube? is it worth doing that?

So is my inaccurate and misleading force calculations definitely a cause of meshing rather than something else I did incorrectly?

Best, F

Thank you, I totally understand, I really appreciate your hints. Whats the benefit of a mesh convergence study? If i try to do a mesh sweep on the cube would that work? and I can make the element sizes very small? Is a mesh sweep with one of the square faces a good way to do things? Can I mesh the faces of the cube differently from the inside and body of the cube? is it worth doing that? So is my inaccurate and misleading force calculations definitely a cause of meshing rather than something else I did incorrectly? Best, F

Edgar J. Kaiser Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 4 months ago Jul 6, 2024, 11:59 a.m. EDT

You need the mesh convergence study to make sure the mesh is good enough and not too good to avoid high computational costs. See if the results remain (almost) constant at a certain mesh density. Regarding meshing details, you will need to try a few things. Check the documentation and examples.

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
You need the mesh convergence study to make sure the mesh is good enough and not too good to avoid high computational costs. See if the results remain (almost) constant at a certain mesh density. Regarding meshing details, you will need to try a few things. Check the documentation and examples.

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

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.