The Research On The Detection And Analysis Of Power Quality Transient Based On The Wavelet Transform

With the wide application of large-scale sensitive electronic devices such as such as the computer and adjustable speed driver in the power system, Power quality transient problems have already become the focus that numerous fields have paid close attention to. The power quality transient is detected and analyzed accurately, As to realize the fault localization, characteristic extraction, and identification . Which is the essential presupposition and basis of monitoring and managing to power quality transient.This paper puts forward various kinds of electromagnetic transient phenomena according to IEEE. Five common power quality transients, namely voltage sag, voltage swell, voltage interruption, surge and impulse are described carefully. The following respect is the research on the detection and analyse of power quality transient.This paper adopts a novel digital signal processing method- the wavelet transform. According to the non-steady characteristic of power quality transient, The good time-frequency localization makes the sigularity of signal can be signified as the wavelet transform modulus maxima . Using Mallat algorithm, the wavelet transform modulus maxima of sigularity is extracted through multiresolution decomposition. By which. The accurate localization of fault signal can be realized, The duration and amplitude of which can also be obtained.To the question of denoising and compression in actual signal detection ,The paper proposed a non-linear threshold method based on wavelet package transform . Due that the time-frequency resolution of wavelet package transform is superior to wavelet transform and the wavelet package coefficients of noise and signal are irrelevant. The noise and unnecessary information can be deleted combining the threshold method.In addition. As to the decomposed characteristics in wavelet package plane of power quality transient. The most eigenvector-the energy of wavelet transform coefficients which reflects transient property is extracted by that time-shift of signal energy is unchangeable. Which is input to the artificial neural network, The power quality transient can be discerned and classified correctly. The simulation results have verified the validity of this method.