can you help me about reactive power harmonics measurement by FFT

alpere wrote:
> Hi all, thanks for your relevance.
> 
> I can measure current RMS value, voltage RMS value, active power, reactive
> power, power factor on three phase systems.

What results do your reactive power and power factor measurements give 
for non-sinusoidal current waveforms?

> But i can't measure current harmonics, voltage harmonics, active power
> harmonics, reactive power harmonics.

Why? You said you are using a Fourier transform. What more is needed?

> For harmonics measurement we must use the DFT. And to do it fastly, we
> must use the FFT.

Fine.

> This subject is realm of DSP.

True.

> But the question is:
> What way do I use to measure the harmonics?

The output of the FFT. What else? There are various ways to reduce the 
calculations by tuning the FFT bins to be exactly on the harmonic lines 
of the power waveform. You appear to have the compute power to do an FFT 
with a lot of samples. That's fine.

> And i want to know: 
> Why energy analyzer shows only 1., 3., 5., ..., 19. (odd harmonics)
> harmonics value?

This is common, and a little bogus. Traditional loads would only produce 
odd harmonics in the current waveform. There really wasn't much point in 
  looking at the even harmonic frequencies. Modern electronic loads can 
be much more creative in the current waveform they produce. Monitoring 
the even harmonics makes sense these days, but power industry practice 
changes very slowly.

> I have samplings of current and voltage.
> Samplings form is array of 16 bits integer that array_current(256),
> array_voltage(256).

I assume the 256 here is the number of samples per mains cycle, right?

> And, i have to convert from those arrays to amplitude_1stHarmonic_current,
> amplitude_3stHarmonic_current etc. with FFT.

> And, i have to convert for voltage harmonics, current harmonics. Active
> and reactive power is derived from current and voltage.

Active power is usually calculated as the dot product of the DC 
corrected raw voltage and current samples. No Fourier transform is 
involved there. What you do for reactive power is debatable, as people 
argue about what reactive means for non-sinusoidal waveforms. There are 
IEC metering specs which say you delay the voltage waveform by 90 
degrees at the fundamental, and take the dot product of that and the 
current waveform. Again, no Fourier transform is involved.

Maybe what you are actually looking for is the harmonic by harmonic 
active power.

Regards,
Steve

Jerry Avins wrote:
(snip)

> Most nonlinearities in power circuits arise from magnetic saturation.

Or with non-linear loads such as discharge lamps.
(Which are often used with magnetic saturation ballasts for
voltage regulating reasons.)

> In the absence of DC, such nonlinearities are symmetric for positive and 
> negative excursions. Waveforms like that have only odd harmonics. 
> (Consider a square wave.)

> In balanced three-phase circuits, harmonic currents divisible by three 
> cancel in the neutral leg of wye-connected systems. They cannot exist at 
> all in a delta-connected system.

With three-phase wye, the fundamental neutral current can't be more
than the phase current for any phase.  The third harmonic will add
in the neutral, such that the neutral current can be larger, and
require a larger wire gauge in the case of non-linear loads.
One common case is discharge lamps, for example, on the top of a pole.
In addition to the neutral currents being in phase, magnetic circuits
also cause a low power factor increasing the current required for
a given load power, and requiring larger wire.

-- glen

glen herrmannsfeldt wrote:
> Jerry Avins wrote:
> (snip)
> 
>> Most nonlinearities in power circuits arise from magnetic saturation.
> 
> Or with non-linear loads such as discharge lamps.
> (Which are often used with magnetic saturation ballasts for
> voltage regulating reasons.)

Even such loats treat positive and negative half cycles similarly. 
Common causes of asymmetry are low-power SCR and diode-in-series dimmers.

>> In the absence of DC, such nonlinearities are symmetric for positive 
>> and negative excursions. Waveforms like that have only odd harmonics. 
>> (Consider a square wave.)
> 
>> In balanced three-phase circuits, harmonic currents divisible by three 
>> cancel in the neutral leg of wye-connected systems. They cannot exist 
>> at all in a delta-connected system.
> 
> With three-phase wye, the fundamental neutral current can't be more
> than the phase current for any phase.  The third harmonic will add
> in the neutral, such that the neutral current can be larger, and
> require a larger wire gauge in the case of non-linear loads.
> One common case is discharge lamps, for example, on the top of a pole.
> In addition to the neutral currents being in phase, magnetic circuits
> also cause a low power factor increasing the current required for
> a given load power, and requiring larger wire.

My mistake. Harmonics divisible by three add, while others cancel.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������