Wavelet and kernel principal component analysis based fuzzy-neuro technique to detect and classify power transmission system faults

摘要

In this paper, identification and classification of transmission line faults are analysed by wavelet and kernel principal component analysis based fuzzy-neuro technique. When a fault occurrs, the transient behaviour of the fault is superimposed on transmission line signals, which can be seen as distorted waveforms. These distorted line xoaveforms are composed of different frequency components and need to be represented in time-frequency domain for fault analysis. For this representation of the line signals, discrete wavelet transform is used. It extracts the fault features and forwards them to a hybrid fuzzy-neuro classifier for classifying the type of fault that has occurred in the transmission system. To reduce the complexity of fault classification by fuzzy-neuro technique, kernel principal component analysis is performed on the extracted features. The proposed hybrid technique for fault identification and classification was tested on 500 kV power transmission lines using MATLAB. The performance of the proposed work is validated using statistical measures such as accuracy, sensitivity and specificity, and was also compared with other techniques. The comparison showed that the proposed technique is better than other techniques in based on these statistical measures.