A novel fault classification method using Wavelet transform and Artificial Neural Networks

A novel fault classification method using Wavelet transform and Artificial Neural Networks is presented in this paper. First, Clarke's transformation is applied to the fault signals in order to obtain the alpha, beta, and zero components. Next, the space-vector magnitude is extracted from alpha and beta components. Then, the Wavelet transform is applied to the space-vector magnitude and zero sequence component in order to calculate the energy distribution in the Wavelet detailed levels. Finally, this Wavelet energy distribution is fed up to an Artificial Neural Network classifier in order to match the input energy pattern to its output. By applying this method, an accuracy of 100% is obtained, even with noisy data. The application of this method creates an opportunity for protecting transmission lines and providing additional information to the system operator in a short time.