Distance protection of a double-circuit line based on Kohonen neural network considering different operation modes

Abstract

This paper deals with problems connected to distance protection of a double-circuit transmission lines. Because of the zero sequence mutual coupling of parallel circuits, the relay sees incorrect impedance. Behavior of distance relays, which protect multi-circuit lines, is that they underreach when the parallel circuits are in service and that they overreach when the parallel circuits are out of service and earthed at both ends. In this paper an adaptive model was developed which solves problems of single-line-to-ground fault on a single line. From the local measurements and from the transmission line theory error factor can be calculated, which is the ratio of the calculated impedance seen by relay and the desired impedance at a fault distance. The aim of the adaptive model is to improve accuracy of the distance relay (to compensate error factor), according to the local measurements using a self-organizing neural network or the so called Kohonen network.