Fault localization on the transmission lines by wavelet technique combined radial basis function neural network

Abstract

The rapid growth of the electricity system in the of a country's economic development has led to an increase in the number of power transmission lines operating at different voltage levels and the total length of power transmission lines. Thus, Faults on the transmission line are unavoidable. In this paper, Wavelet transforms for the recognition and localization of short circuits faults on the power transmission lines. In that, the voltage waves and current waves on the lines are simulated by Simulink - Matlab. The Wavelet transform is used with the RBF Neural network to find out where the short circuit faults occurred. The proposed method is applied for a power system which is a power transmission lines, one load and one generator. The results are achieved and demonstrated the potential for on-line identification capabilities in Vietnam Power System. Third-level DWT with mother wavelet db4 is utilized to calculate detailed discrete wavelet energies of fault current, the values of these features are fed to the RBF neural network to estimate fault location on an overhead power line. The algorithm is fast as only half cycle. The accuracy of fault localization is very high at different locations on the transmission line.