Skip to Main Content
This paper deals with analysis of power signals using complex wavelet transform. In the first step power signals containing sag, swell, harmonic, sag-harmonic, swell harmonic, transient and spike were generated using Matlab. Various features like energy, kurtosis, entropy, skewness etc. were extracted using `db4' and complex wavelet decomposition up to 11 levels. Next, an extensive database of these features was created. A neural network based on these parameters was trained and tested. It has been shown that the classification accuracy achieved by using complex wavelet is higher than obtained by the use of `db4' wavelet.