Neural networks based algorithm for detecting high impedance faults on power distribution lines

This paper investigates a new technique for accurate high impedance fault detection on power distribution lines using artificial neural networks (ANN). The need for ANN techniques in such applications is described and the implementation for power distribution lines is described. The backpropagation learning algorithm is used for adjusting the weights in a multilayer neural network to minimize the prediction error with respect to the connection weights in the network. The paper shows the ability of the new protection scheme to identify high impedance faults for improved protection discrimination.