Wavelet-based algorithm for power quality analysis

This paper presents a computational algorithm for identifying integer harmonics and non-integer harmonics by using wavelet-based transform. The Continuous Wavelet Transform (CWT) based on the modified Complex Morlet Wavelet (CMW) is adopted to estimate harmonics present in a power waveform. A frequency detection technique is developed from the wavelet scalogram and ridges. The necessary conditions for the discrimination of adjacent frequencies, the minimum sampling frequency without aliasing and the minimum waveform length for accurate estimation of harmonics are established. Experimental works have been conducted to verify the accuracy of the proposed algorithm in estimating harmonics of power system waveforms. The proposed algorithm is comparable to Discrete Fourier Transform (DFT) in estimating stationary waveforms however the result is much better than that derived with DFTin estimating non-stationary waveforms. Copyright © 2009 John Wiley & Sons, Ltd.