How to model "LITZ" or "STRANDED" coils and windings

[zand10000]

TL;DR Summary

In Simulation of transformer and electrical motors in software like ansys maxwell we usually have to model conductors, but does it matter how to draw them?

In Simulation of transformer and electrical motors in software like ansys maxwell we usually have to model conductors, but does it matter how to draw them?

I use Ansys software for finite element simulations. Maxwell, Simplorer, RMXpert and other toolboxes. but I always have a doubt on how to model a coil?! should we use a single geometry as a whole coil or we should draw wires one by one! but even if we could, what should we do about twisted litz wires? it is almost impossible to model them in real shape.
what consideration should have in modeling?

I hope electrical engineers help to have a useful conclusion.

 

[Baluncore]

The technique will depend on why you are modelling the conductor.
Any approximation will be a compromise.

You could use a single solid conductor to model the Litz bundle. To correctly model the skin effect advantage of Litz, you must lower the resistivity of the conductor material.

With Litz, the high-frequency current flows throughout the area of the conductor bundle, like a DC current would flow in a solid conductor. That will change the inductance of the bundle. There is a parameter in the inductance formula that allows for the difference in current distribution, between the high-frequency tubular, or low-frequency cylindrical model. That parameter takes a value between 0.75 and 1.
https://en.wikipedia.org/wiki/Inductance#Practical_formulas

 

[zand10000]

The technique will depend on why you are modelling the conductor.
Any approximation will be a compromise.

You could use a single solid conductor to model the Litz bundle. To correctly model the skin effect advantage of Litz, you must lower the resistivity of the conductor material.

With Litz, the high-frequency current flows throughout the area of the conductor bundle, like a DC current would flow in a solid conductor. That will change the inductance of the bundle. There is a parameter in the inductance formula that allows for the difference in current distribution, between the high-frequency tubular, or low-frequency cylindrical model. That parameter takes a value between 0.75 and 1.
https://en.wikipedia.org/wiki/Inductance#Practical_formulas

A very good point ! why we model the conductor.
Due to your answer, practically we should classify models by their frequency and importance of leakage or inductance characteristics. for example :
1- electro motors simulation ( low frequency, not sensitive to inductance precision)
2- power transformer simulation ( low frequency, sensitive to inductance precision)
3- high-freq power electronics transformer simulation ( high frequency, very sensitive to inductance precision)

the important question is what do we compromise if we model the high-freq power electronics transformer using a simple stranded coil without litz or considering eddy effect? would it be negligible? in which frequency range it is negligible?

for example in fly-back convertors the leakage of ferrite core transformer is really important for other parts.

 

[Baluncore]

You need to gather baseline data.
Model the converter with solid wire.
Build two prototype circuits, one with a solid wire, the other with a Litz wire transformer.
Measure the critical parameters of both prototypes.
Compare the measured parameters with the model.

 

[Windadct]

Kinda my point about simulation - and especially so in magnetics. Engineering CAN be an art, this is one of those cases.

 

[anorlunda]

Kinda my point about simulation - and especially so in magnetics. Engineering CAN be an art, this is one of those cases.

Huh? The procedure suggested by @Baluncore is almost universally valid, and objective. To find out if any factor is significant, run it both ways. That doesn't sound very arty.

Build two prototype circuits, one with a solid wire, the other with a Litz wire transformer.

 

[tech99]

I presume the models will take into account the capacitance between the conductors. This capacitance usually limits the advantage of Litz to below about 1 MHz. For big Litz cables it is important how the termination is made, so that current is shared equally among the conductors.

 

[zand10000]

I have simulated a sample of single geometry coil using stranded coil or model all conductors. The sample is a BLDC motor and of course with low frequency. the results was very close. so in low frequency excitation the difference is negligible. You can find the simulation conditions here: