ANALYSIS OF POWER QUALITY EVENTS USING WAVELETS

Wavelets are prominently used for Power Quality (PQ) signal analysis, the features that are computed from wavelet sub bands are informative for detection and classification. Energy levels of non-stationary events that occur in PQ signal computed considering wavelet sub bands suffer from shift variant property and hence use of dual tree complex wavelets that supports shift invariance property is used for PQ event analysis. In this paper, PQ event algorithm is developed considering dual tree wavelets and the results are compared with wavelets. Various PQ signals with non-stationary events are analyzed and the shift invariant property of dual tree wavelets is demonstrated to be advantageous in terms of event classification. Dual Tree Complex wavelet Transform (DTCWT) energy levels are capable of differentiating between multiple events as well as different types of sags, swells, harmonics, interrupts and flicker. The classification accuracy using DTCWT energy bands is improved by more than 90%. DTCWT filters selected in this paper are suitable for PQ event detection as well as classification.