A power harmonics detection approach based on the least squares energy minimization principle

A new approach to power harmonic detection and a fast algorithm for computation of the total harmonic distortion (THD) are proposed based on the least squares (LS) method and the Parseval theorem. This approach is applied to detect a certain order harmonic component of stationary signals or nonstationary signals with little computational cost in comparison with other common techniques such as the Kalman filter and the FFT, thus it can track the time-varying harmonics and can be used to reconstruct the compensation reference current template real time in harmonic control systems especially in active power filters. The theoretical analysis results of the harmonics estimation error are given also. The inter-harmonics are taken into account in the THD algorithm. The THD is calculated based on the sum of harmonics energy estimated in terms of Parseval theorem and the reconstructed fundamental component by using the LS scheme, thus the tracking of time-varying THD becomes possible. Experimental results both in the ideal condition and noisy condition associated with the LS, the Kalman filters and the FFT are given also.