Fault diagnosis for distribution networks using enhanced support vector machine classifier with classical multidimensional scaling

Abstract

In this paper, a new fault diagnosis techniques based on time domain reflectometry (TDR) method with pseudo-random binary sequence (PRBS) stimulus and support vector machine (SVM) classifier has been investigated to recognize the different types of fault in the radial distribution feeders. This novel technique has considered the amplitude of reflected signals and the peaks of cross-correlation (CCR) between the reflected and incident wave for generating fault current dataset for SVM. Furthermore, this multi-layer enhanced SVM classifier is combined with classical multidimensional scaling (CMDS) feature extraction algorithm and kernel parameter optimization to increase training speed and improve overall classification accuracy. The proposed technique has been tested on a radial distribution feeder to identify ten different types of fault considering 12 input features generated by using Simulink software and MATLAB Toolbox. The success rate of SVM classifier is over 95% which demonstrates the effectiveness and the high accuracy of proposed method.