Wavelets for detection and analysis of power system transients

Abstract

In this paper we present the wavelet analysis technique, a powerful tool for the detection and analysis of power system transients. Traditional Fourier analysis techniques are suitable for periodic signals whereas aperiodic signals such as transients on a power system would require a more robust analysis technique such as the wavelet analysis technique. The basic idea of this technique is to decompose the transient signal into a number of levels, which basically contain the high and low-frequency components of the signal. By analyzing the decomposed transient signal at different levels, not only do we get information regarding the frequency content of the transient but also the time at which each of these transients occurred, which helps in distinguishing transients occurring due to different reasons.