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 The information provided may be out of date.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.

Trouble modeling magnets

Please login with a confirmed email address before reporting spam

Hello, I am modeling a tool that is two opposing magnets (i.e., north faces north or south faces south), which have a solenoid (copper wire wrapped around a soft iron core) between them.

I have attached my model. (I surrounded my device by rectangles/cylinders of air so I can implement a moving mesh.)

My magnets correspond with Ampere's laws 2 and 3 under the Magnetic Field physics, and their magnetic fields are described by the constitutive relation of remanent flux density. My questions are:

1. How can I model them so the like poles face on another? I've been doing it by making one B_r positive and the other negative.

2. My material is anisotropic, but for mur, when I select "User defined" and then "Anisotropic" and then click on any other node in the Model Builder window, the selection changes back to "Isotropic". Why can't I model them as being anisotropic magnets? Do I have to change that matrix? Sorry, I don't understand what that is.

Thanks! Any input would be much appreciated.


7 Replies Last Post Apr 22, 2013, 3:17 a.m. EDT
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 Jun 19, 2012, 12:59 a.m. EDT
Hi

I mostly use the magnetisation as material properties for permanent magnets, you need to define one node per magnet orientation, except if there are easy ways to define the relative orientation i.e. with a sign(y) operator to inverse the one one ht positive Y side w.r.t the negative Y side. With anisotropy you mostly should work with one node per orientation.

As I have not yet received the 4.3 CD (it's in the post somewhere) I cannot open your model

--
Good luck
Ivar
Hi I mostly use the magnetisation as material properties for permanent magnets, you need to define one node per magnet orientation, except if there are easy ways to define the relative orientation i.e. with a sign(y) operator to inverse the one one ht positive Y side w.r.t the negative Y side. With anisotropy you mostly should work with one node per orientation. As I have not yet received the 4.3 CD (it's in the post somewhere) I cannot open your model -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 19, 2012, 9:10 a.m. EDT
Thanks for your input, Ivar! One question though:

If i'm just given the properties:
- Max Energy Product (BH)max
- Residual Induction Br
- Coercive Force Hc
- Intrinsic Coercive Force

along with some physical properties, do you know how I can use these to realize the magnetization in Comsol?

Thanks!
Thanks for your input, Ivar! One question though: If i'm just given the properties: - Max Energy Product (BH)max - Residual Induction Br - Coercive Force Hc - Intrinsic Coercive Force along with some physical properties, do you know how I can use these to realize the magnetization in Comsol? Thanks!

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 Jun 19, 2012, 10:03 a.m. EDT
Hi

as first guess I would try to use the Coercive Force Hc in [A/m] as the magnetisation value, provided you have a high quality Nd or SmCo type magnet with a straight Hc-Br line

Then check the Br you get when you put the magnet almost in short circuit via an unsaturated iron in a small 2D model, you might want to consider to use a mur of Br/Hc too

By the way, somebody else has a better approach ?


--
Good luck
Ivar
Hi as first guess I would try to use the Coercive Force Hc in [A/m] as the magnetisation value, provided you have a high quality Nd or SmCo type magnet with a straight Hc-Br line Then check the Br you get when you put the magnet almost in short circuit via an unsaturated iron in a small 2D model, you might want to consider to use a mur of Br/Hc too By the way, somebody else has a better approach ? -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jun 19, 2012, 10:37 a.m. EDT
Thank you for the quick reply!

Since I have two Alnico magnets, which don't have a similarly straight demagnetization curves as Nd or SmCo magnets, it would be less accurate to use Hc as the magnetization right? Could you please explain the process of making a 2D model to check the Br? I don't really understand the set-up or the motivation.

Also, do you know why I might be getting errors when I change my material type of the magnets (in the Ampere's Law node) from the default "non-solid" to "solid"? The error is saying that it failed to evaluate variables mod1.Aphiz, mod1.mf.BR , and mod1.mf.HR.

Thanks!
Thank you for the quick reply! Since I have two Alnico magnets, which don't have a similarly straight demagnetization curves as Nd or SmCo magnets, it would be less accurate to use Hc as the magnetization right? Could you please explain the process of making a 2D model to check the Br? I don't really understand the set-up or the motivation. Also, do you know why I might be getting errors when I change my material type of the magnets (in the Ampere's Law node) from the default "non-solid" to "solid"? The error is saying that it failed to evaluate variables mod1.Aphiz, mod1.mf.BR , and mod1.mf.HR. Thanks!

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 Jun 19, 2012, 4:24 p.m. EDT
Hi

indeed I suspect you need then to define your magnetisation with a lookup table, which means also making your model highly non linear.

When I have a doubt I often make a simple model, i.e. the one of a Hc to Br curve measurement device, by looping my magnet poles around with a perfect iron core + a small air gap that I then increases just to verify that I get a flux response with reasonable values w.r.t a ideal magnet Hc to Br load curve

Your error seems strange, it could come from some update errors internally to COMSOL (when you change the physics), sometimes its worth to reset the moel save it and reload it (useful in 4.2a at least, could still be valis in v4.3 ;)

--
Good luck
Ivar
Hi indeed I suspect you need then to define your magnetisation with a lookup table, which means also making your model highly non linear. When I have a doubt I often make a simple model, i.e. the one of a Hc to Br curve measurement device, by looping my magnet poles around with a perfect iron core + a small air gap that I then increases just to verify that I get a flux response with reasonable values w.r.t a ideal magnet Hc to Br load curve Your error seems strange, it could come from some update errors internally to COMSOL (when you change the physics), sometimes its worth to reset the moel save it and reload it (useful in 4.2a at least, could still be valis in v4.3 ;) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Apr 21, 2013, 3:07 p.m. EDT
Hi,

I have 2 questions:


1-What is the difference between solid and non solid ?
Does non solid means that it can not get deformed?

2- For the domain which we define the coils(using multi-turn coil domain), should we also define cooper and assign it as the material for that domain? This is what have done in the above example, but I highly doubt it!

Thanks,
Ehsan
Hi, I have 2 questions: 1-What is the difference between solid and non solid ? Does non solid means that it can not get deformed? 2- For the domain which we define the coils(using multi-turn coil domain), should we also define cooper and assign it as the material for that domain? This is what have done in the above example, but I highly doubt it! Thanks, Ehsan

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 Apr 22, 2013, 3:17 a.m. EDT
Hi

the "material type", from what I understand is particularly useful in "Solid" and FSI, normally you can leave it to its default with little risk, and in ACDC alone, I would say no risk ;)

I agree it could be better explained in the doc. Probably the best place to read is a search for " "material type" feature "
all three words with inner """

It defines on which of the material or spatial frames the material is locked to. These two frames dissociate in Solid when non-linear geometry is selected (by default or by the user) and for FSI or when mixed with CFD modules

For the multi turn domain it's easier to give it a material definition, by default you will see that epsilon_r and mu_r are read from the material node, but NOT the resistivity or conductivity sigma, that is defined in the multi-turn BC node, you can also set the other material properties to user defined whence no need for materials for that domain, or you can leave "air" also for the domain, the multi-turn BC node will pick the desired values for the ACDC current calculation part

--
Good luck
Ivar
Hi the "material type", from what I understand is particularly useful in "Solid" and FSI, normally you can leave it to its default with little risk, and in ACDC alone, I would say no risk ;) I agree it could be better explained in the doc. Probably the best place to read is a search for " "material type" feature " all three words with inner """ It defines on which of the material or spatial frames the material is locked to. These two frames dissociate in Solid when non-linear geometry is selected (by default or by the user) and for FSI or when mixed with CFD modules For the multi turn domain it's easier to give it a material definition, by default you will see that epsilon_r and mu_r are read from the material node, but NOT the resistivity or conductivity sigma, that is defined in the multi-turn BC node, you can also set the other material properties to user defined whence no need for materials for that domain, or you can leave "air" also for the domain, the multi-turn BC node will pick the desired values for the ACDC current calculation part -- 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.