Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

Moving Mesh for electromagnetics

Please login with a confirmed email address before reporting spam

I am trying to simulate the moving magnet in time domain to see the effect of eddy force on the aluminium plate as the magnet moves up and down. I modelled the same problem in 3.5 and it simulated but the problem was it was giving warning of inverted mesh. However, I am trying to model this same problem in comsol 4.2. I have attached the file. I would highly appreciate it if someone can look at it and help me out.


3 Replies Last Post 08.10.2012, 16:34 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 08.10.2012, 02:23 GMT-4
Hi

A few suggestions:
Personally I do not trust 100% the translations, I know COMSOL has put a lot of effort to keep it backwards compatible, and this is, for me, handy to understand how to rebuild your model, but there are so many things that are different between the two v3 and v4 versions, not to say variable names etc, that I systematically remake my models in v4. It goes anyhow so quick now to construct a model that the time you do it is well spent to refresh the structure and clean up your model, I spend now 4-8 times more time to document all the useful results from a run than to set it up, solving anyhow is separate as this can go from a few minutes to a few days ;)

If you go for steady state oscillations, use rather the harmonic development frequency domain. (i.e. add to your model a frequency domain study in addition to your time series. Then go back to the MF (EMQA in 3.5) node and check then equation, while flipping between Time Dependent and Frequency Domain, you will notice then that the latter has a term more the Att (2nd time derivative of the magnetic vector potential) So for me it is more reliable to use Frequency domain for such a steady state oscillation. Note: this is not the case for a transient effect, then other consideration must be taken. Another reason, if you switch later to MEF you have no "time dependent" study option, only the frequency domain and there you do not even have any omega^2 term. I understand the reason is linked to the ACDC hypothesis versus RF hypothesis related to the model and wave length size, but there might well be other more subtle reasons also.

Then you apply force calculations on your aluminium part. But Aluminium has a mur of "1" hence as Air, if you have no difference in mu-r you have no magnetic flux direction change on the boundaries, and then for me (again) you cannot generate any magnetic forces. On the other hand, aluminium is a good conductor, when you move it in a magnetic field you generate Eddy currents in the conducting metal, this again generates a magnetic field that interacts with the static one and gives you Lorentz forces. But to resolve induced currents you need to solve for V and this is only done in MEF physics.

Somewhere in the models there are the falling magnet in a copper tube example (if you read French you have the description of the experience here
www.iai.heig-vd.ch/fr-ch/Enseignement/Supports/Physique%20électricité%20et%20magnétisme%20(PHY2)/Aimant_dans_un_tube.pdf)

I believe I published somewhere either on the Forum I think this example probably in 4.1, but here we do not use ALE but the trick of linear sliding boundary

--
Good luck
Ivar
Hi A few suggestions: Personally I do not trust 100% the translations, I know COMSOL has put a lot of effort to keep it backwards compatible, and this is, for me, handy to understand how to rebuild your model, but there are so many things that are different between the two v3 and v4 versions, not to say variable names etc, that I systematically remake my models in v4. It goes anyhow so quick now to construct a model that the time you do it is well spent to refresh the structure and clean up your model, I spend now 4-8 times more time to document all the useful results from a run than to set it up, solving anyhow is separate as this can go from a few minutes to a few days ;) If you go for steady state oscillations, use rather the harmonic development frequency domain. (i.e. add to your model a frequency domain study in addition to your time series. Then go back to the MF (EMQA in 3.5) node and check then equation, while flipping between Time Dependent and Frequency Domain, you will notice then that the latter has a term more the Att (2nd time derivative of the magnetic vector potential) So for me it is more reliable to use Frequency domain for such a steady state oscillation. Note: this is not the case for a transient effect, then other consideration must be taken. Another reason, if you switch later to MEF you have no "time dependent" study option, only the frequency domain and there you do not even have any omega^2 term. I understand the reason is linked to the ACDC hypothesis versus RF hypothesis related to the model and wave length size, but there might well be other more subtle reasons also. Then you apply force calculations on your aluminium part. But Aluminium has a mur of "1" hence as Air, if you have no difference in mu-r you have no magnetic flux direction change on the boundaries, and then for me (again) you cannot generate any magnetic forces. On the other hand, aluminium is a good conductor, when you move it in a magnetic field you generate Eddy currents in the conducting metal, this again generates a magnetic field that interacts with the static one and gives you Lorentz forces. But to resolve induced currents you need to solve for V and this is only done in MEF physics. Somewhere in the models there are the falling magnet in a copper tube example (if you read French you have the description of the experience here http://www.iai.heig-vd.ch/fr-ch/Enseignement/Supports/Physique%20électricité%20et%20magnétisme%20(PHY2)/Aimant_dans_un_tube.pdf) I believe I published somewhere either on the Forum I think this example probably in 4.1, but here we do not use ALE but the trick of linear sliding boundary -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 08.10.2012, 11:03 GMT-4
Hi Ivar,

Thank you for your suggestions. I could not find the "Magnet falling in copper tube example" that you suggested. If you still have it can you please post it again. I looked all over the forum but couldn't find it. I would highly appreciate it as it will reduce my headache of formulating a solution to this problem.


Regards,

Pratik
Hi Ivar, Thank you for your suggestions. I could not find the "Magnet falling in copper tube example" that you suggested. If you still have it can you please post it again. I looked all over the forum but couldn't find it. I would highly appreciate it as it will reduce my headache of formulating a solution to this problem. Regards, Pratik

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 08.10.2012, 16:34 GMT-4
Hi

I know that COMSOl has one too, you can try also that way via your rep, as I'm on travel, and certianly no longer sure where I have hidden that one among the many models I havemade ;)


But certainly, I willdo, if I find it back ...
--
Good luck
Ivar
Hi I know that COMSOl has one too, you can try also that way via your rep, as I'm on travel, and certianly no longer sure where I have hidden that one among the many models I havemade ;) But certainly, I willdo, if I find it back ... -- Good luck Ivar

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.