Power Quality Disturbances Detection And Classification Based On Lifting Wavelet

The proliferation of power electronic devices and nonlinear loads has triggered a rowing concern with power quality.Traditional power quality analysis is based on ffective value theory. As a main analysis tool, Fourier transform is not suitable for the on-stationary signals processing. So it is indispensable to establish effective detection nd analysis system to detect and classify dynamic power quality disturbances.Using lifting wavelet transform theory and ANN techniques,this paper presents a systematic study of power quality disturbances detection and classification.The main contents of which are as follows:1.It summarizes the lifting wavelet transform theory, points out the advantage of the lifting wavelet transform. Then a novel power quality analysis method based on lifting wavelet transform with adaptive performance is proposed, and successfully applied in dynamic power quality analysis for the first time.2.Because the characteristic of the wavelet transform modulus maxima related to the power quality disturbance singularity points and the noise is diferent,so it is not difficult to distinguish each other in the high scales.Based on this phenomena,a soft-threshold denoising method is introduced,which can locate the power quality disturbance start point and end point in the noisy condition exactly.3.Through analyzing the relation between butterworth filters and orthonormal wavelets,a new approach for power measurment is proposed,which can improve the accuracy of magnitude of disturbances.4.In order to classify and identify various types of power quality disturbances, a classifier based on lifting wavelet and neural network is used.Furthermore,some feature vectors based on the lifting wavelet transform are deduced which can keep unique.Then an artificial neural network integrated with lifting wavelet is used to classify various types of power quality disturbances.