S-transform based spectral energy feature space for fault location approximation

Abstract

A methodology for fault location calculation of transmission lines using S-transform based spectral energy feature space and multi-layer perceptron is presented in this paper. S-transform being capable of providing the time domain and frequency domain information at the same time, gives remarkable insight in to the protection issues of a transmission line. The energy distributions of the faulted current waveforms are used to compute the fault location using the pattern recognition approach. This paper proposes the fault detection and classification through S-transform energy by using synchronized measurement of the differential and average energy of current signals. The S-transform energy calculations are used for fault location using a neural network.