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 The information provided may be out of date.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.

Modeling DC magnet

Please login with a confirmed email address before reporting spam

Hi,

I try to simulate the magnetic field between two magnet poles and the velocity-induced currents in a metal plate below.
For symmetry reasons, only the air region around one magnet pole is modeled. The current-carrying top block is used to initiate the magnetic field between the poles by a voltage difference.

The simulation shows that the electric currents do not build closed loops in the plate as I would expect but come to the air region and the top block instead. How can I avoid that? Decreasing the electrical conductivity of air brings convergence issues.

I also tried to use the magnetic field boundary on the air region top side (H in y direction) and to delete the upper block. In this case it seems necessary to define 'ground' elsewhere, but I can't find an appropriate boundary for it.

Any hints are kindly appreciated




3 Replies Last Post 3 févr. 2011, 13:20 UTC−5
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 3 févr. 2011, 06:42 UTC−5
Hi

I do not have access to COMSOL for a few days, but if you are in MEF you need some conductivity in all domains to solve correctly for V, avalue of 1-100 for air is mostly OK. Corollairy to not have any comsol access: I connot open your model and can provide only some ideas of the type "best guess"

Another point, in 4.1 the Lorentz term has not been implemented yet in 2D axi (but OK in 3D and 2D), so for 2D-axi you must write out the equtions by hand as an external current source based on the fields and currents obtained (you can alsways use the equations of the 2D case and adapt them)

in MEF do not forge to define enough BC's for both dependent variables V and A, these are subnodes to be added, that one often forget about.

I have done recently a test of a "falling magnet in a Cu tube", it worked out OK both in 2D with existing BCs and then in 2d-Axi with home-brew equations.

Another point MEF does not accept transient / time series analysis, only stationary, you can get the transient response of a system by adding your governing equation (i.e. gravity versus eddy current induced forces) as an global ODE and solve the ODE in transient mode, MEF in stationary and link both phyics in a segregated way. This is really nice and straightforward in 4.1

--
Good luck
Ivar
Hi I do not have access to COMSOL for a few days, but if you are in MEF you need some conductivity in all domains to solve correctly for V, avalue of 1-100 for air is mostly OK. Corollairy to not have any comsol access: I connot open your model and can provide only some ideas of the type "best guess" Another point, in 4.1 the Lorentz term has not been implemented yet in 2D axi (but OK in 3D and 2D), so for 2D-axi you must write out the equtions by hand as an external current source based on the fields and currents obtained (you can alsways use the equations of the 2D case and adapt them) in MEF do not forge to define enough BC's for both dependent variables V and A, these are subnodes to be added, that one often forget about. I have done recently a test of a "falling magnet in a Cu tube", it worked out OK both in 2D with existing BCs and then in 2d-Axi with home-brew equations. Another point MEF does not accept transient / time series analysis, only stationary, you can get the transient response of a system by adding your governing equation (i.e. gravity versus eddy current induced forces) as an global ODE and solve the ODE in transient mode, MEF in stationary and link both phyics in a segregated way. This is really nice and straightforward in 4.1 -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 3 févr. 2011, 06:56 UTC−5
Hi Ivar,

thanks for your hints.

The electrical conductivity in the air region is 1, in the plate ~1e+7 and in the upper conductor around 1000.
I use ground and electric potential on the upper conductor to get a current and also a magnetic field distribution in the air and the plate.
In older versions of COMSOL the electric currents that form due to the presence of a velocity and the magnetic field in the plate were closed loops as expected. Up to v4.0a I could set the normal component of current density to zero at the plate boundaries (electric insulation). In v4.1 electric insulation is a subnode of magnetic insulation which is obviously wrong in my case.

Due to the electric potential distribution the currents escape from the plate to the upper conductor which is unphysical.

How can I solve this in v4.1?

Best regards, Marcel
Hi Ivar, thanks for your hints. The electrical conductivity in the air region is 1, in the plate ~1e+7 and in the upper conductor around 1000. I use ground and electric potential on the upper conductor to get a current and also a magnetic field distribution in the air and the plate. In older versions of COMSOL the electric currents that form due to the presence of a velocity and the magnetic field in the plate were closed loops as expected. Up to v4.0a I could set the normal component of current density to zero at the plate boundaries (electric insulation). In v4.1 electric insulation is a subnode of magnetic insulation which is obviously wrong in my case. Due to the electric potential distribution the currents escape from the plate to the upper conductor which is unphysical. How can I solve this in v4.1? Best regards, Marcel

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 3 févr. 2011, 13:20 UTC−5
Hi

I'm not fully sure, these are the issues I typically send to support.

There have indeed been different changes recently, one of the reasons, explained to me, was to have clean current loops.

What I understand now is that you have found a case that is not "clean" ;)

--
Good luck
Ivar
Hi I'm not fully sure, these are the issues I typically send to support. There have indeed been different changes recently, one of the reasons, explained to me, was to have clean current loops. What I understand now is that you have found a case that is not "clean" ;) -- 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.