Fault classification using multi-resolution analysis and discrete wavelet transforms

Abstract

The continuity of service in power systems has a vital economical and social impact on all shareholders; generation, transmission and distribution, and end users. Fault classification within transmission and distribution networks plays an important role in power restoration for guaranteed service continuity. With advances in digital signal processing in terms of speed and algorithms, the use of wavelets transform is made easy and feasible for real-time applications in power systems. In this paper we present two methods of fault classification using Discrete Wavelet Transforms (DWT). The coefficients of the wavelet decomposition of fault signals are correlated with the coefficients of signals in normal working conditions to deduce fault information. Haar wavelets and multi-resolution analysis are used for detecting the faulty phase while Daubechies wavelet is used to determine if the fault to ground or not. Both suggested methods succeeded in all types of faults simulated using Simulink.