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 The information provided may be out of date.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.

Total magnetic Energy

Gerhard Edmound Stebner

Please login with a confirmed email address before reporting spam

Hello,

im Using COMSOL 4.1 and I have got heavy problems simulating the cogging of a synchronous machine. The calculated Torque using the stress tensor is correct. But if I use the method of virtual work something ist going terribly wrong . If I use a stationary solver and add "Parametric" to the solver the magnetic energy has the wrong periodic time and is not sinusoidal (the cogging torque is more or less sinusoidal). Using the timedependant simulation tends to result in a sinusoidal form after a ramp.

What can I do to improve the calculation of the total magnetic Energy? And why are the results so different?


5 Replies Last Post 04.05.2013, 11:51 GMT-4
COMSOL Moderator

Hello Gerhard Edmound Stebner

Your Discussion has gone 30 days without a reply. If you still need help with COMSOL and have an on-subscription license, please visit our Support Center for help.

If you do not hold an on-subscription license, you may find an answer in another Discussion or in the Knowledge Base.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 05.07.2011, 09:56 GMT-4
Hello Gerhard,

I had and still have trouble calcualting magnetic energy, too.
The problem with comsol is, that magnetic energy is calculated with a linear approach by default!!!
I do not know, which version of comsol you are using right now, but with version 4 and later you can check this fact by activating the equation view and search in the corresponding Ampere's Law node for variable mf.Wm. As told before it normally is calculated by mf.Wm = 0.5*(mf.Bx*mf.Hx+mf.By*mf.Hy+mf.Bz*mf.Hz), which is only correct for linear magnetic materials. For any non linear magnetic Material it is necessary to define a function (interpolation table) for magnetic energy density. You can calcualte the the magnetic energy density out of the BH - or HB - curve of the corresponding material by integrating H dB. Then you need to integrate the energy density over all domains of the model to get the total amount of magnetic energy. So far seems to work fine, but proving forces by virtual work approach I get forces wich are less than half of comsol results!?! Testing forces by virtual work with just linear materials works perfekt.
I got the feeling, that there might be a bug in the software. I will contact comsol support again help and advice....

Best regards
Oliver Vogel


I attached a file, so you can explore it yourself.
Hello Gerhard, I had and still have trouble calcualting magnetic energy, too. The problem with comsol is, that magnetic energy is calculated with a linear approach by default!!! I do not know, which version of comsol you are using right now, but with version 4 and later you can check this fact by activating the equation view and search in the corresponding Ampere's Law node for variable mf.Wm. As told before it normally is calculated by mf.Wm = 0.5*(mf.Bx*mf.Hx+mf.By*mf.Hy+mf.Bz*mf.Hz), which is only correct for linear magnetic materials. For any non linear magnetic Material it is necessary to define a function (interpolation table) for magnetic energy density. You can calcualte the the magnetic energy density out of the BH - or HB - curve of the corresponding material by integrating H dB. Then you need to integrate the energy density over all domains of the model to get the total amount of magnetic energy. So far seems to work fine, but proving forces by virtual work approach I get forces wich are less than half of comsol results!?! Testing forces by virtual work with just linear materials works perfekt. I got the feeling, that there might be a bug in the software. I will contact comsol support again help and advice.... Best regards Oliver Vogel I attached a file, so you can explore it yourself.


Gerhard Edmound Stebner

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 05.07.2011, 11:27 GMT-4
Hello Oliver,

unfortunately I am not at work... so I cant test your model. But incredible... you are right it is a linear approach. I will test my torque calculations on synchronous machines with the method described. Thanks a lot. I will share my new results as soon as possible. One question... did you multiply the result of the integral with the length?! Your equation would be right for 1 m of depth...
Hello Oliver, unfortunately I am not at work... so I cant test your model. But incredible... you are right it is a linear approach. I will test my torque calculations on synchronous machines with the method described. Thanks a lot. I will share my new results as soon as possible. One question... did you multiply the result of the integral with the length?! Your equation would be right for 1 m of depth...

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 11.07.2011, 03:42 GMT-4
Hello again,

I guess your question assumes a 2D-model...
My test model is actually a axisymmetric 2D model, so there is no depth!

I also tried to simulate a small synchronous without success. I did not get the clue how to model the interface betwee stator and rotor correctly. I would be very very pleased, if you could provide a working model.

I mentioned that there still might be bug in energy caclualtion in comsol, so I am definetly interestet for your results.
I am also preparing a big request for comsol support about that. Kepp you informed when it is answered.

Best regards
Oliver
Hello again, I guess your question assumes a 2D-model... My test model is actually a axisymmetric 2D model, so there is no depth! I also tried to simulate a small synchronous without success. I did not get the clue how to model the interface betwee stator and rotor correctly. I would be very very pleased, if you could provide a working model. I mentioned that there still might be bug in energy caclualtion in comsol, so I am definetly interestet for your results. I am also preparing a big request for comsol support about that. Kepp you informed when it is answered. Best regards Oliver

Gerhard Edmound Stebner

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 28.07.2011, 06:36 GMT-4
Hello again,

I was on vacation so i didnt answer. Sorry for that. I'm not allowed to post my models. But if you give me your email-adress I will send a modified geometry of my recent project. This will help you to understand how synchronous machines can be simulated.

Best regards,

Gerhard
Hello again, I was on vacation so i didnt answer. Sorry for that. I'm not allowed to post my models. But if you give me your email-adress I will send a modified geometry of my recent project. This will help you to understand how synchronous machines can be simulated. Best regards, Gerhard

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 04.05.2013, 11:51 GMT-4
Hi,

The same problem for me:

I am performing some simulations of a synchronous motor. Now I am dealing with cogging torque:

I'm using nonlinear material (Iron)

I checked the mesh to be symmetrical in the air gap, where I perform the stress tensor integration (I built a middle line in the air gap so as the mesh to be symmetrical: three stripes of element inner and outer it)

Then I integrate the magnetic energy in the rotor domains but what I get is a very irregular line (what I expect is a theta-derivative not higher than 0.01 Nm .. )

Could anybody help me?
Hi, The same problem for me: I am performing some simulations of a synchronous motor. Now I am dealing with cogging torque: I'm using nonlinear material (Iron) I checked the mesh to be symmetrical in the air gap, where I perform the stress tensor integration (I built a middle line in the air gap so as the mesh to be symmetrical: three stripes of element inner and outer it) Then I integrate the magnetic energy in the rotor domains but what I get is a very irregular line (what I expect is a theta-derivative not higher than 0.01 Nm .. ) Could anybody help me?

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.