I started working on the "Direct Power Control of Grid Connected Voltage Source Converters", everywhere in the literature, authors state that one of the major drawbacks of this control technique is the presence of active and reactive power ripples due to the use of Hysteresis controllers.

In the other hand I've already worked on Power Harmonics, active power filtering and UPQCs, and i know that harmonics must be multiple of the fundamental frequency.

My questions are the following :

  • Are "power ripples" cause by hysteresis controllers considered "harmonics"?
  • What are the frequencies of those power ripples?
  • Do they affect the value of THD?
  • \$\begingroup\$ Ripples - are unwanted oscillations. Harmonics are the Fourier components of any signal. \$\endgroup\$
    – Eugene Sh.
    Commented Jun 27, 2018 at 14:49
  • \$\begingroup\$ Discovering what a hysteresis controller is, will probably help answer the question. \$\endgroup\$
    – user16324
    Commented Jun 27, 2018 at 14:54

2 Answers 2


Are "power ripples" cause by hysteresis controllers considered "harmonics" ?

No, not all of them, harmonics are specific frequencies that are resonant, all noise ripples such as switching noise or hysteresis add up to the total noise 'ripple' of the circuit. Ripple is less applicable in AC systems because the signal (the AC waveform itself) is much higher than the noise. To see the noise in an AC system you need to look at the frequency content or subtract out the first harmonic (ie 60 or 50Hz)

enter image description here
Source: http://www.compliance-club.com/archive/keitharmstrong/design_techniques6.html

What are the frequencies of those power [harmonics]?

Typically multiples of the carrier, if your on 60Hz, you get one at double and triple so 120Hz and 180Hz and so on.

Do they affect the value of THD ?

THD only takes into consideration the harmonics to the fundamental:

$$ %THD = \frac{\sqrt{v_2^2+v_3^2 \ddots v_n^2}}{v_1^2} $$

  • \$\begingroup\$ First, thank you for answer, sec : can you re-answer the second question What are the frequencies of those power [harmonics]? ! i was actually asking about the frequency of the ripples ! i know that the frequency of the harmonic is n*f (where n is the rank and the f is the fund freq 50/60Hz), but what about the Ripples ? \$\endgroup\$ Commented Jun 27, 2018 at 17:57
  • 1
    \$\begingroup\$ The ripples from nonlinear sources can show up as harmonics, but usually the harmonics are the greatest source of error. If it doesn't show up as a harmonic, it will show up as a noise passband, in all of the frequency plots I've seen or on instruments any other source of noise is much lower than the harmonics \$\endgroup\$
    – Voltage Spike
    Commented Jun 27, 2018 at 18:02

For GTI’s it can be the same thing.or not. Ripples are often the filtered response to pulse noise.

The power noise current in grid tied inverters (GTI) is somewhat complex.

The purpose of a GTI is to supply a constant power output in sync with grid such that it produces pure sinusoidal power with pf=1.0 .

However the non-linear positive feedback due to hysteresis if a control system produces current phase noise from harmonics which after filtering looks like ripple.

Attempting to design a GTI as a digitally controlled power source as a very low impedance voltage source has inherent grid noise from other GTI ‘s. The proper method is to use a filtered switching current source that detects line voltage and output current and regulates current to perfectly match frequency, phase, voltage and current with an offset for constant AC power at the fundamental with e.g. < 1% THD . Hysteric voltage control of any of these variables will amplify phase noise ( broad band harmonics with near band jitter on each ).

Therefore the design must factor these specs to avoid harmonics at low output impedance (caps) and have full specs for all 4 scattering parameters as well. (Sorry if this is to complex at the moment) this means output forward + reverse impedance and transfer function)

This can cause distribution protection fault issues.

For example , I have recent reports that have been observed recently in SC, USA on a 1MW utility Solar farm with 30 Huawei GTI’s . That while putting out perfect 40A current for over a year until other 3rd party Solar farms nearby came online. Then it started to blow 90 fuses which were each supplying 40 Amps forward to the grid while other farms were introducing noise in harmonics in the reverse direction Both adding >> 60A by up to 35% due low impedance between both GTI’s at harmonics measured as power phase noise.

Hysteresis can certainly be controlled but system control stability factor also becomes more complex with the network fighting against each other from low impedance s parameters where Zo <<10%


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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