Comparison between discrete STFT and wavelets for the analysis of power quality events Francisco Jurado a, *, Jose ´ R. Saenz b a Department of Electrical Engineering, University of Jae ´n, 23700 EUP Linares, Jae ´n, Spain b Department of Electrical Engineering, University of the Basque Country, 48013 ESI Bilbao, Spain Received 3 September 2001; received in revised form 24 January 2002; accepted 25 January 2002 Abstract This paper deals with the comparison of signal processing tools for power quality analysis. Two signal processing techniques are considered: the wavelet filters and the discrete short-time Fourier transforms. Then, examples of the two most frequent disturbances met in the power system are chosen. It is designed an adjustable speed drive with a six-pulse converter using EMTP/ATP and it is presented normal energizing of utility capacitors. The analysis is tested on a system consists of 13 buses and is representative of a medium-sized industrial plant. Finally, each kind of electrical disturbance is analyzed with example representing each tool. A qualitative comparison of results shows the advantages and drawbacks of each signal processing technique applied to power quality analysis. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Discrete short-time Fourier transforms; Power quality; Wavelet transforms 1. Introduction In signal processing the time  /frequency domain has often been exploited for analyzing signals with fast changing spectral contents. The wavelet analysis [1 /3] can be used for similar purposes, and has been exploited recently for several types of voltage and current disturbances [4  /9] and for power system protection [10]. Several possible power system applications for wavelet analysis have been proposed, among others: automated disturbance classification [7]; recognition using wavelet-based neural classifier [4]; propagation of power system transients [6]; detection of faults [10,11]; and visualization of time-varying harmonics [12]. The stated advantage of using wavelets compared to Fourier transform [13  /15] is the tradeoff between frequency and time-resolution at different frequencies. The wavelet transform gets a lot of attention in the literature, at the expense of the short-time Fourier transform (STFT). The advantage of the latter is its ease of interpretation. Power engineers are used to think in terms of sinusoidal signals and the STFT fits closest to this. Most of the results obtained by using wavelets can equally be obtained by using STFT. When using wavelet transforms it is often difficult to extract the fundamental or any other single harmonic component of the signal. For an analysis tool to become widely excepted in power system engineering, it is important that it enables analysis in terms of harmonic signals. This paper shows how the spectral contents as a function of time can be obtained by using the STFT [16,17]. Although the STFT has a fixed frequency resolution for all frequencies once the size of the window is chosen, it enables an easier interpretation in terms of harmonics. Studies of power quality phenomena have emerged as an important subject in recent years due to renewed interest in improving the quality of the electricity supply. As sensitive electronic equipment continues to prolifer- ate, the studies of power quality will be further emphasized. The identifying features are derived from well-documented theories [14,15,18  /21], power engi- neers’ heuristics gained through long years of experi- ence, and power quality data collected in recent years. * Corresponding author. E-mail addresses: fjurado@ujaen.es (F. Jurado), iepsaruj@bi.ehu.es (J.R. Saenz). Electric Power Systems Research 62 (2002) 183 /190 www.elsevier.com/locate/epsr 0378-7796/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII:S0378-7796(02)00035-4