0
\$\begingroup\$

I'm looking for a software which can compute the losses (or just the AC resistance of the wires) in a transformer due to proximity and skin effect in function of the different parameters of a transformer (number of layers, core size, interleaved or not, current waveform applied to the transformer, etc). It would be nice if it is based on the Dowell's equations or even better as the Dowell's equation are based on sinusoidal waveform current.

Otherwise, is there someone who knows the equivalent equations of the Dowell's equations for a SMPS waveform ?

Thank you very much,

\$\endgroup\$
6
  • 1
    \$\begingroup\$ I use quickfield student edition (free) but be prepared to pull your hair out a tad when first setting up an analysis. I did have the full blown version many years ago when designing magnetics for metal detectors so, if you have the cash (and the patience).... \$\endgroup\$
    – Andy aka
    Mar 23, 2020 at 10:53
  • \$\begingroup\$ Ok ^^ I will take a look :D Thank you very much ! \$\endgroup\$
    – Jess
    Mar 23, 2020 at 10:59
  • 1
    \$\begingroup\$ Give yourself 4 hours to get your first basic simulation actually producing results. Thereafter it gets easier. The student edition is limited on finite element nodes so this can be an annoyance but, you can set up a 2 D model of an infinitely long wire(s) and use the inbuilt analyser to deliver "per metre" values of which power loss is one as well as AC resistance and inductance (plus many more). Start simply and use square section wires to make the finite elements fewer. It's hard work but will pay off if you press hard. I've had a couple of patents on the back end of using quickfield. \$\endgroup\$
    – Andy aka
    Mar 23, 2020 at 11:03
  • \$\begingroup\$ I will follow your recommendations :) Thank you ! \$\endgroup\$
    – Jess
    Mar 23, 2020 at 13:42
  • 1
    \$\begingroup\$ Remember it's a 2D model and assumes that objects you create are infinitely long into and out of the screen. That won't be a problem if you then realize that any analysis is based on 1 metre length of that infinitely deep object. Having not used it for over ten years I had need to reuse it and it still hurt me re-acquainting myself with the jargon but it's worth it for the recent problem we have discussed. It makes awesome field plots (eventually LOL). \$\endgroup\$
    – Andy aka
    Mar 23, 2020 at 14:06

1 Answer 1

1
\$\begingroup\$

I believe that you will not find any commercially available (or free) software that has been written to do just that. Some manufacturers develop programs like this, but only for their own use.

Often, for a number of reasons, the easiest way is to use the Finite Element Method. There are excellent paid programs, but also free. In this work here, we used FEMM 4.2 to calculate the parameters of resistance and inductance (which also vary with frequency) in function of frequency.

If you train this way, at very high frequencies, the time needed for calculation can become very high because of the fine mesh needed to model the depth of field. The best alternative we found was to use the complex permeability that makes the mesh used independent of the frequency (FEMM 4.2 can do this).

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.