The application of neural networks and Clarke-Concordia transformation in fault location on distribution power systems

Abstract

This paper presents a new approach to fault location on distribution power lines. This approach uses an artificial neural network based learning algorithm and Clarke-Concordia transformation. The /spl alpha/,/spl beta/,0 components of line currents resulting from the Clarke-Concordia transformation are used to detect all types of fault. The neural network is trained to map the nonlinear relationship existing in fault location equations. The proposed approach is able to identify and locate all different types of faults (single line to ground, double line to ground, line-to-line and three-phase short-circuit). This approach is subdivided into several main steps: Data acquisition, corresponding on three-phase current signals; Mathematical treatment by the Clarke-Concordia transformation; Fault identification, obtained by the analysis of fault and pre-fault data; Fault location artificial neural network based learning algorithm. The fault position is presented as the output of the neural network on which, as the input, it was considered the eigenvalue of matrix representing transformed line current. Results are presented which shows the effectiveness of the proposed algorithm for a correct fault location on distribution power system networks.