Classification of Single Line to Ground Faults on Double Circuit Transmission Line using ANN

Abstract

Abstract— This paper discusses the potential application of ANN techniques for detection of single line to ground faults and fault type classification on double circuit transmission lines with remote end infeed. Distance protection of double circuit transmission lines has been a very challenging task. The problems arise principally as a result of the mutual coupling between the two circuits under different fault conditions. An accurate algorithm for fault detection and classification of single line-to-ground faults (A1N, A2N, B1N, B2N, C1N & C2N) in double circuit transmission line considering the effects of mutual coupling, high fault resistance, varying fault location, fault inception angle and remote source infeed is presented using feed forward neural network (FFNN) algorithm. The algorithm employs the fundamental components of voltage and current signals. This technique neither requires communication link to retrieve the remote end data nor zero sequence current compensation for healthy phases are required. This is a major advantage of the proposed algorithm for protection of double circuit line fed from both the ends. Results of study on a 220 kV transmission line are presented as an illustration. Simulation results indicate that algorithm is immune to the effect of mutual coupling, fault type, fault inception angle, fault resistance, fault location and remote end infeed. Index Terms— Artificial neural network, Double circuit transmission line, Fault detection & classification, High impedance fault, Single line-to-ground fault.