Discussion Forum

Hi Linus,

Those 'unexpected errors' can be cumbersome. I sometimes found that the model file is corrupted and it is not advisable to go ahead with the file that produced the error. I step back to the last version that did not show the error. Save often and use version numbers! Anyway, it is not possible to give any specific advice in your case without seeing the model.

Cheers Edgar

Hi Linus,

Those 'unexpected errors' can be cumbersome. I sometimes found that the model file is corrupted and it is not advisable to go ahead with the file that produced the error. I step back to the last version that did not show the error. Save often and use version numbers! Anyway, it is not possible to give any specific advice in your case without seeing the model.

Cheers Edgar

Alright well if you could take a quick look it would be thankful, but it only happens when i add magnetostriction to the modell otherwise it computes fine.

You have a couple of issues in your model.

• No piezomagnetic domain is defined for the core in the solid physics.
• The solver is not defined
• Compliance matrix, coupling matrix are all zero
• Relative permeability is not defined

There may be other errors. I suggest you look into the nonlinear magnetostriction example in the application library to learn how to properly set up the magnetostriction part.

Cheers Edgar

You have a couple of issues in your model.

• No piezomagnetic domain is defined for the core in the solid physics.
• The solver is not defined
• Compliance matrix, coupling matrix are all zero
• Relative permeability is not defined

There may be other errors. I suggest you look into the nonlinear magnetostriction example in the application library to learn how to properly set up the magnetostriction part.

Cheers Edgar

Hello,

Well first i wanna say thanks, i will for sure take a look into that. But could i ask what do you mean with the solver is not defined?

Also where can i find the expression for Compliance matrix, coupling matrix? And Piezomagnetic coupling matrix, Voigt notation?

Also did you find any other parts missing in the solid mechanics? Like constraints etc?

There wasn't any solver sequence, probably due to the missing piezomagnetic domain and the resulting error. COMSOL stopped before building the solver sequence.

Material data for magnetostriction is a difficult point. You can find data for the material used in the library example (a giant magnetostrictor like Terfenol, used for tranducers), but that is not valid for Iron. Iron and polycrystalline ferrous materials have a more complex magnetostrictive function. Also the current nonlinear magnetostrictive approach in COMSOL is not really suitable for materials like Iron, even if you had a magnetostrictive function. You may get a result but it is probably not correct.

I developed a method for the linear case with DC bias and a small AC perturbation, but this is not suitable for a transformer.

There wasn't any solver sequence, probably due to the missing piezomagnetic domain and the resulting error. COMSOL stopped before building the solver sequence.

Material data for magnetostriction is a difficult point. You can find data for the material used in the library example (a giant magnetostrictor like Terfenol, used for tranducers), but that is not valid for Iron. Iron and polycrystalline ferrous materials have a more complex magnetostrictive function. Also the current nonlinear magnetostrictive approach in COMSOL is not really suitable for materials like Iron, even if you had a magnetostrictive function. You may get a result but it is probably not correct.

I developed a method for the linear case with DC bias and a small AC perturbation, but this is not suitable for a transformer.

Could i use another material for the core then to get it to work?

Transformer cores are made from Iron, so I am afraid you can't.

Transformer cores are made from Iron, so I am afraid you can't.

How about grain oriented electrical steel? Im thinking the silicon steel GO 3%

Do you have its coupling matrix?

Do you have its coupling matrix?

No, will have to research if its possible to find.

Here is some good reading about magnetostriction. It is mostly about magnetostrictive transducers, not about transformers. However, it shows the magnetostrictive function of a typical low carbon steel (page 25) and how to derive a coupling matrix from that. It is good for general understanding, but will be a long way to model magnetostriction in a transformer.

Hirao, M., Hirotsugu, O., Electromagnetic Acoustic Transducers, Springer Series in Measurement and Science Technology, ISBN 978-4-431-56034-0.

Here is some good reading about magnetostriction. It is mostly about magnetostrictive transducers, not about transformers. However, it shows the magnetostrictive function of a typical low carbon steel (page 25) and how to derive a coupling matrix from that. It is good for general understanding, but will be a long way to model magnetostriction in a transformer.

Hirao, M., Hirotsugu, O., Electromagnetic Acoustic Transducers, Springer Series in Measurement and Science Technology, ISBN 978-4-431-56034-0.

Read thru it but still is it even possible to find desired magnetoelastic properties for Silicon Steel GO M-6 Rolling on internet?

Also could i ask you quick do i need some constraints in the solid? Or can i just leave it free

In a time dependent study the solid can be free, in a stationary study the rigid body DOFs must be constrained.

I was doing extended research regarding magnetostrictive properties in the internet, but didn't find much relevant. You may consider to inquire with manufacturers. Of course it is possible to measure magnetostrictive properties of specific materials. Then you still need a nonlinear method that can handle the non-monotonous behavior of Iron based materials.

In a time dependent study the solid can be free, in a stationary study the rigid body DOFs must be constrained.

I was doing extended research regarding magnetostrictive properties in the internet, but didn't find much relevant. You may consider to inquire with manufacturers. Of course it is possible to measure magnetostrictive properties of specific materials. Then you still need a nonlinear method that can handle the non-monotonous behavior of Iron based materials.

Hey,

Well i did research myself and hardly found anything. I will take a consideration tomorrow about what to do but i think its way too big task considering the time i have left on my project.

Also just another question when i tried to solve the FFT, time to frequency. degrees of freedom solved for are zero? I mean I'm using the initial expression and input study as the time dependet. Values of dependent variables (Physics controlled). Should that get the displacement? Even if i have wrong material and getting result that are not correct. the FFT solver operates on dependent variables right? In my case there is already a dependent variable that can be input into the FFT solver, that is the solution variable (magnetic vector potential) from study 1? But still freedom solver for 0 and getting no output? Is it becouse of the solid or what can it be?

This is now a different topic and I recommend to open a new thread. As your model is pretty complex I advise to set up something simple and test the FFT. You may find an example in the application library too.

This is now a different topic and I recommend to open a new thread. As your model is pretty complex I advise to set up something simple and test the FFT. You may find an example in the application library too.

Alright, thanks!