Novel power quality indices based on wavelet packet transform for non-stationary sinusoidal and non-sinusoidal disturbances

High power quality (PQ) level represents one of the main objectives towards smart grid. The currently used PQIs that are a measure of the PQlevel are defined under the umbrella of the Fourier foundation that produces unrealistic results in case of non-stationary PQ disturbances. In order to accurately measure those indices, wavelet packet transform (WPT) is used in this paper to reformulate the recommended PQIs and hence benefiting from the WPT capabilities in accurately analyzing non-stationary waveforms and providing a uniform time-frequency sub-bands leading to reduced size of the data to be processed which is a necessity to facilitate the implementation of smart grid. Numerical examples' results considering non-stationary waveforms prove the suitability of the WPT for PQIs measurement: also the results indicate that Daubechies 10 could be the best candidate wavelet basis function that could provide acceptable accuracy while requiring less number of wavelet coefficients and hence reducing the data size. Moreover, a new time-frequency overall and node crest factors are introduced in this paper. The new node crest factor is able to determine the node or the sub-band that is responsible for the largest impact which could not be achieved using traditional approaches. (C) 2009 Elsevier B.V. All rights reserved.