Identification and classification of high impedance fault on a medium voltage distribution line using radial base function neural network (RBFNN)

Abstract

High impedance fault (HIF) can be described as a fault which does not produce enough fault current to be detectable by conventional protection schemes which present high risks to lives and electrical equipment’s. Several methods have been employed to identification and classification of HIF but come with their limitations. In this research, RBFNN has been used in which it has been trained from the data extracted from HIF current signal, using wavelet decomposition at level 8 and Daubechies4 wavelet transform (db4) so as to have a unique feature or finger print of this fault, Radial base function neural network was used and the training performance of this model is 0.0176164 and accuracy of 94.45% has been obtained