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.

How to model transformer in 3D using Comsol 4.0a?

Please login with a confirmed email address before reporting spam

Hello.

I've made transformer model in 2D axial symmetry using 'Multi-Turn Coil' and 'Electrical Circuit' to excite primary winding and induce current in secondary winding. Everything works fine.
But now I would like to model transformer in 3D and I have a big problem with it. I'm using Magnetic and Electric Fields (mef) module and Electrical Circuit. There we have no Multi-Turn Coil tool so I try to connect FEM model and circuit with terminals. Each winding is made of one solid.
I have tried intersect solid with surface and there put terminals and grounds. I've tried also different combinations of Electric and Magnetic Insulation (just like in Spiral Inductor model from Library). But it doesn't compute even using direct solver.
I obtain solution only when I use stationary solver and set DC current. But I have to make my computation in frequency domain.

I would be grateful for any hint.

Przemek

11 Replies Last Post 20 juil. 2016, 12:31 UTC−4

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 déc. 2010, 01:46 UTC−5
Hello Przemek,

attach your model file! This increases chances that somebody can help you dramatically. If you forgot to model something or made a mistake, you woouldn't mention it here. Because if you were aware of the mistake/the thing you forgot, you'd have fixed it already. So somebody can only detect these things and provide you some help by having a look at your model.

Bye
MV
Hello Przemek, attach your model file! This increases chances that somebody can help you dramatically. If you forgot to model something or made a mistake, you woouldn't mention it here. Because if you were aware of the mistake/the thing you forgot, you'd have fixed it already. So somebody can only detect these things and provide you some help by having a look at your model. Bye MV

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 9 déc. 2010, 06:31 UTC−5
Hi,

Of course. You're right. Below I attach one of many test models in 3D which I've done. Also I attach my 2D model. My aim is to obtain in 3D something like that I've obtained in 2D.
My second issue is making simulations of transformer in time domain but I see that there is no possibility to do that with mef module. Maybe I should divide my problem in few parts which I would compute separately?
Now I don't have concrete domain of analysis. I would like to catch transformer's principle of operation using Comsol with Electrical Circuit.

Thanks for any hints.
Przemek
Hi, Of course. You're right. Below I attach one of many test models in 3D which I've done. Also I attach my 2D model. My aim is to obtain in 3D something like that I've obtained in 2D. My second issue is making simulations of transformer in time domain but I see that there is no possibility to do that with mef module. Maybe I should divide my problem in few parts which I would compute separately? Now I don't have concrete domain of analysis. I would like to catch transformer's principle of operation using Comsol with Electrical Circuit. Thanks for any hints. Przemek


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 12 déc. 2010, 10:18 UTC−5
Hi

there are a few comments I would like to add to your 3D "tranformator" model:

1) try to use round (or filelts) Cu coils as by square shapes you have singularities that give quite some wrong results, this applies also to the iron part (in ACDC just as for structural you get singularities at sharpe adges)

2) if you cut open you coils with "air" the way you do your equations are not correct I beleive: how is the current looping around (current conservation) ? (through the electric circuit model, but is that enough for "mef" I'm not sure ?)
I would continue the copper in the gap you have and impose a defined current density perpendicular to the "gap" surface. This would leave at least a closed loop electric circuit in the MEF part.

3) to use or not the multiple turn coil BC's (thats is a question) I do not have a clear answer to propose, but even at 10-20 kHz such bulk coils have rather large skin effects, be aware anyhow they are not there in 3D ;)

4) if you want to have a different material behaviour for your iron (saturation etc) add a specific ampere law and current conservation physics node just for the iron part

--
Good luck
Ivar
Hi there are a few comments I would like to add to your 3D "tranformator" model: 1) try to use round (or filelts) Cu coils as by square shapes you have singularities that give quite some wrong results, this applies also to the iron part (in ACDC just as for structural you get singularities at sharpe adges) 2) if you cut open you coils with "air" the way you do your equations are not correct I beleive: how is the current looping around (current conservation) ? (through the electric circuit model, but is that enough for "mef" I'm not sure ?) I would continue the copper in the gap you have and impose a defined current density perpendicular to the "gap" surface. This would leave at least a closed loop electric circuit in the MEF part. 3) to use or not the multiple turn coil BC's (thats is a question) I do not have a clear answer to propose, but even at 10-20 kHz such bulk coils have rather large skin effects, be aware anyhow they are not there in 3D ;) 4) if you want to have a different material behaviour for your iron (saturation etc) add a specific ampere law and current conservation physics node just for the iron part -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14 déc. 2010, 05:38 UTC−5
Hi Przemek,
I am also doing also same kind of simulation. I was trying to open your model but my comsol is saying there is no model. Which model are you using? if you don't mind can you mail your model in to my Email. My mail id is tummalapranith45@gmail.com. I have a doubt can we use two models like Electrical circuit and magnetic field at a time.
Hi Przemek, I am also doing also same kind of simulation. I was trying to open your model but my comsol is saying there is no model. Which model are you using? if you don't mind can you mail your model in to my Email. My mail id is tummalapranith45@gmail.com. I have a doubt can we use two models like Electrical circuit and magnetic field at a time.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 déc. 2010, 11:14 UTC−5
Hi,

First of all thanks to Ivar for response. But there is still a problem with 3D transformer simulation.
Maybe this is result of physics laws but we can't execute Time Dependet simulation using Magnetic and Electric Fields (mef) module. Another desired feature for me is possibility of connection with Electrical Circuit (cir) module. In the future I would like to model 3-phase transformer with variations of circuits (e.g. Y-?). In 2D these things are fulfilled with "Multi-turn Coil Domain" from Magnetic Fields (mf) module. It works fine with Electrical Circuit. But in 3D mf is even unable to connect with cir module.
My idea is to use Electric Currents(ec) module and Magnetic Fields(mf) module but I don't know how to couple them together.
Or maybe is there possibility to apply my own expressions to connect mf with cir module in 3D the way how it is made in 2D mf module?
I know that Comsol is able to compute almost everything but I can't find any working example of transformer with the
exception of untypical planar transformer.



@pranith

Probably You're using earlier version of Comsol than 4.0a. However, if You have Comsol 4.0a You should patch it to version 4.0.0.993 using following link: www.comsol.com/support/updates/comsol40ap/. Only with the same Comsol version You can open my files.
In 2D Electrical Circuit and Magnetic Fields works fine together. You can connect Your model with the circuit using Multi-Turn Coil Domain and select as coil excitation - Circuit. While in Electrical Circuit place 'External I Vs. U' device in suitable positions of circuit (where coils should be) and change User Defined voltage to Total voltage(mf/mtcd...).
Hi, First of all thanks to Ivar for response. But there is still a problem with 3D transformer simulation. Maybe this is result of physics laws but we can't execute Time Dependet simulation using Magnetic and Electric Fields (mef) module. Another desired feature for me is possibility of connection with Electrical Circuit (cir) module. In the future I would like to model 3-phase transformer with variations of circuits (e.g. Y-?). In 2D these things are fulfilled with "Multi-turn Coil Domain" from Magnetic Fields (mf) module. It works fine with Electrical Circuit. But in 3D mf is even unable to connect with cir module. My idea is to use Electric Currents(ec) module and Magnetic Fields(mf) module but I don't know how to couple them together. Or maybe is there possibility to apply my own expressions to connect mf with cir module in 3D the way how it is made in 2D mf module? I know that Comsol is able to compute almost everything but I can't find any working example of transformer with the exception of untypical planar transformer. @pranith Probably You're using earlier version of Comsol than 4.0a. However, if You have Comsol 4.0a You should patch it to version 4.0.0.993 using following link: http://www.comsol.com/support/updates/comsol40ap/. Only with the same Comsol version You can open my files. In 2D Electrical Circuit and Magnetic Fields works fine together. You can connect Your model with the circuit using Multi-Turn Coil Domain and select as coil excitation - Circuit. While in Electrical Circuit place 'External I Vs. U' device in suitable positions of circuit (where coils should be) and change User Defined voltage to Total voltage(mf/mtcd...).

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 16 déc. 2010, 01:57 UTC−5
Hi

for both of you get the 4.1 and update to latest patch (to correct a formula error in previous mef external currents modules) then probably it will work better.

Your comment to link SPICE is certainly very pertinent and I'm sure COMSOL developpersare working on that, it's just that in 3D you must also respect th physics and cutting your coils in your way changes the physics (via the change in topology as you are missing the full curent "loop"

--
Good luck
Ivar
Hi for both of you get the 4.1 and update to latest patch (to correct a formula error in previous mef external currents modules) then probably it will work better. Your comment to link SPICE is certainly very pertinent and I'm sure COMSOL developpersare working on that, it's just that in 3D you must also respect th physics and cutting your coils in your way changes the physics (via the change in topology as you are missing the full curent "loop" -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 7 mars 2012, 11:13 UTC−5

Hello.

I've made transformer model in 2D axial symmetry using 'Multi-Turn Coil' and 'Electrical Circuit' to excite primary winding and induce current in secondary winding. Everything works fine.
But now I would like to model transformer in 3D and I have a big problem with it. I'm using Magnetic and Electric Fields (mef) module and Electrical Circuit. There we have no Multi-Turn Coil tool so I try to connect FEM model and circuit with terminals. Each winding is made of one solid.
I have tried intersect solid with surface and there put terminals and grounds. I've tried also different combinations of Electric and Magnetic Insulation (just like in Spiral Inductor model from Library). But it doesn't compute even using direct solver.
I obtain solution only when I use stationary solver and set DC current. But I have to make my computation in frequency domain.

I would be grateful for any hint.

Przemek


Dear Przemek,

Were you finally able to build a working simulation model of a AC transformer?

Best regards,
Iker.

[QUOTE] Hello. I've made transformer model in 2D axial symmetry using 'Multi-Turn Coil' and 'Electrical Circuit' to excite primary winding and induce current in secondary winding. Everything works fine. But now I would like to model transformer in 3D and I have a big problem with it. I'm using Magnetic and Electric Fields (mef) module and Electrical Circuit. There we have no Multi-Turn Coil tool so I try to connect FEM model and circuit with terminals. Each winding is made of one solid. I have tried intersect solid with surface and there put terminals and grounds. I've tried also different combinations of Electric and Magnetic Insulation (just like in Spiral Inductor model from Library). But it doesn't compute even using direct solver. I obtain solution only when I use stationary solver and set DC current. But I have to make my computation in frequency domain. I would be grateful for any hint. Przemek [/QUOTE] Dear Przemek, Were you finally able to build a working simulation model of a AC transformer? Best regards, Iker.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 8 mars 2012, 07:05 UTC−5
Hi,

Yes, I know how to build both 2D (out of plane) and 3D model of transformer (one or more phases) with external circuit for transient simulation. Now I'm using mf module instead of mef module.

In 2D model without axial symmetry You have to use Multi-Turn Coil Domain. One winding consists of two MTCD interfaces. For the first part of the winding in MTCD select circuit as a "coil excitation" and for the second write by hand variable name which carry total current of the first part but with reverse sign, for example: -mf.Itot_1. In electrical circuit (cir) module in External I Vs. U don't use predefined Total voltage but write by hand an algebraic sum of voltages from two winding parts, for example: mf.Vtot_1-mf.Vtot_2. This ensure a real current and voltage coupling in the model. It's a little bit tricky method and may cause errors during computation. Then You should use predefined total voltages in External I Vs. U, next start the computation, let COMSOL compile equations, stop the computation and back to algebraic sums I mentioned earlier. Maybe it looks chaotic but should works.

In 3D model You have to use Lumped Port feature. Look at inductor_3d model in the Model Library (ACDC Module -> Tutorial Models -> inductor_3d) how does Lumped Port work. I've discoverd that it can be utilized also for windings consisted of one solid region (like MTCD). For modeling turns number You have simply to divide "Width of lumped port" by this number. I didn't made comprehensive verification of this approach but looks to work correct.

With Regards,
Przemek
Hi, Yes, I know how to build both 2D (out of plane) and 3D model of transformer (one or more phases) with external circuit for transient simulation. Now I'm using mf module instead of mef module. In 2D model without axial symmetry You have to use Multi-Turn Coil Domain. One winding consists of two MTCD interfaces. For the first part of the winding in MTCD select circuit as a "coil excitation" and for the second write by hand variable name which carry total current of the first part but with reverse sign, for example: -mf.Itot_1. In electrical circuit (cir) module in External I Vs. U don't use predefined Total voltage but write by hand an algebraic sum of voltages from two winding parts, for example: mf.Vtot_1-mf.Vtot_2. This ensure a real current and voltage coupling in the model. It's a little bit tricky method and may cause errors during computation. Then You should use predefined total voltages in External I Vs. U, next start the computation, let COMSOL compile equations, stop the computation and back to algebraic sums I mentioned earlier. Maybe it looks chaotic but should works. In 3D model You have to use Lumped Port feature. Look at inductor_3d model in the Model Library (ACDC Module -> Tutorial Models -> inductor_3d) how does Lumped Port work. I've discoverd that it can be utilized also for windings consisted of one solid region (like MTCD). For modeling turns number You have simply to divide "Width of lumped port" by this number. I didn't made comprehensive verification of this approach but looks to work correct. With Regards, Przemek

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 8 mars 2012, 08:31 UTC−5
Thanks a lot Przemek. I'll check your approach. I've been trying with the mf module using air conductivity of 1 S/m. That's the only way I can get simulations running. In addition, Iron conductivity must also be very low to avoid extrange results in the magnetic flux density inside. But I have not checked the results.
Thanks a lot Przemek. I'll check your approach. I've been trying with the mf module using air conductivity of 1 S/m. That's the only way I can get simulations running. In addition, Iron conductivity must also be very low to avoid extrange results in the magnetic flux density inside. But I have not checked the results.

Please login with a confirmed email address before reporting spam

Posted: 10 years ago 1 mai 2015, 20:48 UTC−4

Hi,

Yes, I know how to build both 2D (out of plane) and 3D model of transformer (one or more phases) with external circuit for transient simulation. Now I'm using mf module instead of mef module.

In 2D model without axial symmetry You have to use Multi-Turn Coil Domain. One winding consists of two MTCD interfaces. For the first part of the winding in MTCD select circuit as a "coil excitation" and for the second write by hand variable name which carry total current of the first part but with reverse sign, for example: -mf.Itot_1. In electrical circuit (cir) module in External I Vs. U don't use predefined Total voltage but write by hand an algebraic sum of voltages from two winding parts, for example: mf.Vtot_1-mf.Vtot_2. This ensure a real current and voltage coupling in the model. It's a little bit tricky method and may cause errors during computation. Then You should use predefined total voltages in External I Vs. U, next start the computation, let COMSOL compile equations, stop the computation and back to algebraic sums I mentioned earlier. Maybe it looks chaotic but should works.

In 3D model You have to use Lumped Port feature. Look at inductor_3d model in the Model Library (ACDC Module -> Tutorial Models -> inductor_3d) how does Lumped Port work. I've discoverd that it can be utilized also for windings consisted of one solid region (like MTCD). For modeling turns number You have simply to divide "Width of lumped port" by this number. I didn't made comprehensive verification of this approach but looks to work correct.

With Regards,
Przemek


Hi Przemek
Very interesting and useful thread. I am trying to simulate a transformer model in 3D. In the model library of COMSOL (inductor_3D.mph) it works at 1 MHz frequency. But if I try to go down in frequency at 60 Hz, it do not work. I badly need it to simulate a transformer intended to use in the transmission line. Any help would be greatly appreciated.
Thanks.
Md. Saimoom Ferdous
[QUOTE] Hi, Yes, I know how to build both 2D (out of plane) and 3D model of transformer (one or more phases) with external circuit for transient simulation. Now I'm using mf module instead of mef module. In 2D model without axial symmetry You have to use Multi-Turn Coil Domain. One winding consists of two MTCD interfaces. For the first part of the winding in MTCD select circuit as a "coil excitation" and for the second write by hand variable name which carry total current of the first part but with reverse sign, for example: -mf.Itot_1. In electrical circuit (cir) module in External I Vs. U don't use predefined Total voltage but write by hand an algebraic sum of voltages from two winding parts, for example: mf.Vtot_1-mf.Vtot_2. This ensure a real current and voltage coupling in the model. It's a little bit tricky method and may cause errors during computation. Then You should use predefined total voltages in External I Vs. U, next start the computation, let COMSOL compile equations, stop the computation and back to algebraic sums I mentioned earlier. Maybe it looks chaotic but should works. In 3D model You have to use Lumped Port feature. Look at inductor_3d model in the Model Library (ACDC Module -> Tutorial Models -> inductor_3d) how does Lumped Port work. I've discoverd that it can be utilized also for windings consisted of one solid region (like MTCD). For modeling turns number You have simply to divide "Width of lumped port" by this number. I didn't made comprehensive verification of this approach but looks to work correct. With Regards, Przemek [/QUOTE] Hi Przemek Very interesting and useful thread. I am trying to simulate a transformer model in 3D. In the model library of COMSOL (inductor_3D.mph) it works at 1 MHz frequency. But if I try to go down in frequency at 60 Hz, it do not work. I badly need it to simulate a transformer intended to use in the transmission line. Any help would be greatly appreciated. Thanks. Md. Saimoom Ferdous

Please login with a confirmed email address before reporting spam

Posted: 8 years ago 20 juil. 2016, 12:31 UTC−4
sir plz email me your planar transformer model. i also work on transformer but some error occurs
sir plz email me your planar transformer model. i also work on transformer but some error occurs

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.