Application of neural network combined with improved algorithm in distorted waveform analysis

The issue of power quality has attracted considerable attention from utilities and customers due to fast development of power system industry. It is vital to ensure power quality by means of advanced technology which can monitor, locate and classify disturbances with measurement approaches and instruments. A novel approach for the power quality disturbances detection and investigation using wavelet transform and neural network is proposed. The wavelet transform technology provides an effective means for analysis of non-stationarity and transient signal in terms of shifted and scaled versions. The detection and localization processes are a series of convolution and decimation processes at each corresponding scale, which provide feature vectors as input variables for neural network designed for disturbance pattern recognition. The wavelet network combines advantages of wavelet transformation for purposes of feature extraction and with the characteristic decision capabilities of neural network. In process of training phase, the evolutionary algorithm is used to complete the network parameters adjustment. The processing phase performs waveform recognition and the output of the processing phase is the type of the disturbances. The simulation results show that the proposed method has the ability of recognizing and classifying different power disturbance types efficiently.