A comprehensive training for wavelet-based RBF classifier for power quality disturbances

In this paper we demonstrate that the dominant frequencies and Lipschitz exponents in nonstationary and transitory power quality disturbances efficiently extracted from their wavelet transform modulus maxima (WTMM) in the time-scale domain can serve as powerful discriminating features for wavelet-based classification of these disturbances. We also propose a comprehensive "knowledge-based" algorithm for the training of the radial basis function (RBF) network classifier so that at its convergence, the network gives both the optimal feature weight vector as well as the cluster centres and scaling widths.