Wavelet-based detection, localization, quantification and classification of short duration power quality disturbances

This paper introduces a wavelet-based method for detecting, localizing, quantifying and classifying short duration power quality disturbances (SDPQDs). A difference signal is formed for detection and localization in real-time. Wavelet transform (WT) with a new complex wavelet is utilized to quantify SDPQDs. It can take the fast dyadic scheme not the time consuming shift-invariant scheme and just needs the detailed version at one scale not at many scales. Novel binary features are defined and extracted from time or WT domain, and then a simple binary-decimal conversion is sufficient for classification purpose. The method benefits from both speed and accuracy. Simulation results for five types of SDPQDs show that the proposed method is much better than others.