Wavelet Transform Application In Harmonic Detection Of Power System

In the recent year, as a result of the wide use of new style electronically controlled equipments, harmonic which their electricity characteristic of nonlinear, impact and unbalance leads to, has seriously polluted power-supply system. It has become a curse to the quality of electric energy. Detecting harmonic effectively has become very important to improve the quality of electric energy. But traditional Fourier transform can not analyze both in time and frequent domain. Wavelet analysis bas become a useful signal time-frequent tool after Fourier transform, it also has time-frequent localizing and multi-resolution analysis ability.Proceeding from the harmonic of power system, this paper emphatically discusses wavelet transform and analyzes the different time-frequent character and applicative rang between wavelet and Fourier transform, according to comparing with recent harmonic detection. It introduces two main parts round wavelet transform application in harmonic detection. The one part shows that wavelet transform can detect steady harmonic of power system instead of Fourier transform. The other part includes using extreme value of amplitude based on wavelet transform to analyze the singularity of signals. It also includes according difference of noise and singular points to extract singular points in noising signals. It uses the conventional soft and hard threshold wavelet de-noising method and elicits that choosing right threshold is key of de-noising. So it puts forwards a selecting threshold algorithm based on adaptive learning algorithm of gradient according to simulation. Against the exiting questions about measurement of changing harmonic, it uses a method on the merits of wavelet transform combining with Fourier transform to detect them. It also figures out that this complex method has validity with lots of simulation.