Detection is key - Harmonic detection methods for active power filter applications

In order to improve the filtering efficiency and solve the issues of classical passive filters, active power filters (APFs) are implemented to minimize the harmonic distortion from nonlinear loads. There are different harmonic detection methods for APF applications which include the frequency-domain method mainly identified with Fourier analysis, rearranged to provide the result as fast as possible with a reduced number of calculations, to allow a real-time implementation in DSPs. A discrete Fourier transform (DFT) is a mathematical transformation for discrete signals. It gives both the amplitude and phase informations of the selected harmonic. A reconstruction back in the time domain of the reference signal for the inner controller is done when the harmonics are detected and isolated in the frequency domain. Another method is time-domain methods which give increased speed and less calculations compared to the frequency-domain methods. The method eliminates some challenges by means of implementing the instantaneous power theory which determine the harmonic distortion by calculating the instantaneous power in a three-phase system. Other methods that are less popular due to their difficulty during implementation include sinusoidal subtraction, filtering, nonactive-power-related theories, online estimation algorithms used for time-varying harmonics applications, neural networks used for special applications and wavelet filtering used for compensation of rapidly changing harmonics.