DIGITAL RELAYING OF HIGH VOLTAGE TRANSMISSION LINES BY ARTIFICIAL NEURAL NETWORKS

Abstract

Power system protection using digital relays has an active area of research and development for last years of all computer-relaying topics. In this thesis we are interested in studying and testing digital relaying system using artificial neural network approach for purposes of protection of a transmission line. Electro magnetic program (EMTP) is used for simulation the power system for models under study to obtain the voltages and current signals that are used to feed the digital relay.

This thesis present usmg of artificial neural network for digital protection purposes of high voltage transmission line, fault detector, fault classifier and fault locator. The first model at which the transmission line is fed by two generators one at each end. The second model simulation is three generators feed the transmission line two at each end and the position of the third generators is changed to show the effect of infeed on the relay performance and time of response. The input to artificial neural network (ANN) of the two models is the samples of voltages and current signals or the magnitude of impedance or reactance of the transmission line. For each input to the relaying system the suggested ANN is proposed, learned and tested, moreover fault detector, fault classifier and fault locator for each input are learned and tested. The time of relaying of each suggested relaying system is compared. The accuracy of fault locator of each input is• calculated and compared. This thesis is formatted as follows: