Understanding Transformer Losses and Harmonics: A Guide

[hisham.i]

Studying the transformer shows that if the secondary has no load then hystersis phenomena will make a deformation in the shape of the current leading a losses, due to the presence of harmonics in the wave.
Why harmonics make losses?

 

[Bob S]

The nonlinear shape of the magnetization current in a transformer is due to the change in the permeability (B = μμ₀H) of the transformer core at high field (over 1 Tesla). Transformer core losses are due to the area of the ∫BH hysteresis curve. If the transformer primary and secondary are tightly coupled, the secondary voltage wave should closely match the primary voltage curve. The nonlinearity of the core magnetization current primarily affects the reactive power needed to excite the transformer.
Bob S

 

[hisham.i]

But i don't know in what form this loss it is?

 

[Bob S]

But i don't know in what form this loss it is?

∫BH curve (hysteresis) losses and eddy current losses cause heating of the transformer core.
Bob S

 

[hisham.i]

What is the effect of the resulting harmonics on the power lost in the transformer?

 

[Bob S]

What is the effect of the resulting harmonics on the power lost in the transformer?

The power loss for hysteresis loss (I believe) is relatively independent of frequency for low frequencies, but the eddy current losses in laminations for harmonics scale as frequency squared. See the eddy current lamination loss equation in
http://en.wikipedia.org/wiki/Eddy_current
Note that the power loss shown in the equation is the power loss per kilogram.
Bob S

 

[sophiecentaur]

I think you are after an answer that doesn't really exist unless you can analyse the emerging waveform into its particular harmonics. This will tell you the amplitude of the wanted fundamental, which will tell you the overall losses. If the load is not frequency sensitive (i.e. mainly resistive), then you can add all the harmonic powers together, assuming the same load resistance at all frequencies. This total, plus the fundamental power will still be less than the input power due to the power dissipated internally.
What I mean is this:
Input Power = Wanted Power + (Harmonic Power + Dissipated Power)

If the load is very frequency sensitive, then you can't tell how the quantities in the parentheses are shared.