Identification and Classification of Power System Faults using Ratio Analysis of Principal Component Distances

Abstract

Power system reliability operation has been one of the most vital topics under research. The power system network, mostly the long transmission lines is often subjected to different types of faults leading to maloperation of power flow. The idea of a reliable protection system is to most accurately and efficiently identifying the fault, classifying and the locating of fault. This paper represents the application of dynamic phasors in the form of Principal Component Analysis (PCA) to identify fault in a three phase one end fed 150 km long radial power system transmission line. In the proposed work, (1/4) cycle pre-fault and (1/2) cycle post fault line voltages have been extracted from Electromagnetic Transient Programming (EMTP) simulation. The proposed algorithm is trained using only one set of receiving end data carrying out fault only at the midpoint of the line to generate fault signatures using PCA. The eigenvectors and the score matrix thus obtained corresponding to the three phases using the above analysis have been utilized to construct the component distances, which have been analyzed using ratio analysis to extract the similar features of any particular fault individually.