A wavelet entropy approach for detecting lightning faults on transmission lines

Many authors proposed methods to classify transients on transmission lines. However, all of these methods fail when a lightning strike evolves to a fault. In this paper we propose a new method that can be used to detect whether a lightning strike evolves to a fault or not. The mechanism for lightning strike evolving to a fault is first reviewed. Clarke Transformation is used to transform phase currents into 0-α-β components. The α-component is then normalized. Discrete Wavelet Transform (DWT) is used afterwards to extract the coefficients of approximation_5 of the a-component. A window of one half of a cycle of post lighting data is then taken to calculate the entropy of the aerial current at that window by summing up the squares of the coefficients extracted. This energy (entropy) is compared against a preselected threshold. If the energy calculated is higher than this threshold then the strike is evolving to a fault. Numerous simulations have been done using ATP/EMTP to verify the validly of the method proposed.