Soft computing applications in high impedance fault detection in distribution systems

Two methods, one based on genetic algorithm (GA) and one based on neural networks (NN), are proposed for high impedance fault (HIF) detection in distribution systems. These methods are used to discriminate HIFs from isolator leakage current (ILC) and transients such as capacitor switching, load switching (high/low voltage), ground fault, inrush current and no load line switching. Wavelet transform is used for the decomposition of signals and feature extraction in both methods. In one method, GA is used for feature vector reduction and Bayes for classification. In the other method, principal component analysis (PCA) is applied for feature vector reduction and NN for classification. HIF and ILC data was acquired by experimental tests and the data for other faults was obtained by simulating a real network using EMTP. Results show that either of the proposed procedures can be used to identify HIF from other events efficiently.