Frequency domain analysis of power system transients using Welch and Yule–Walker AR methods Ahmet Alkan a, * , Ahmet S. Yilmaz b a Department of Computer Engineering, Yasar University, 35500 Izmir, Turkey b Department of Electrical and Electronics Engineering, Kahramanmaras Su ¨ tcu ¨ Imam University, 46050-9 Kahramanmaras, Turkey Received 15 May 2006; received in revised form 11 December 2006; accepted 24 December 2006 Available online 12 March 2007 Abstract In this study, power quality (PQ) signals are analyzed by using Welch (non-parametric) and autoregressive (parametric) spectral esti- mation methods. The parameters of the autoregressive (AR) model were estimated by using the Yule–Walker method. PQ spectra were then used to compare the applied spectral estimation methods in terms of their frequency resolution and the effects in determination of spectral components. The variations in the shape of the obtained power spectra were examined in order to detect power system tran- sients. Performance of the proposed methods was evaluated by means of power spectral densities (PSDs). Graphical results comparing the performance of the AR method with that of the Welch technique are given. The results demonstrate superior performance of the AR method over the Welch method. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Power quality; Power system transients; Spectral estimation; Welch; AR method 1. Introduction Power quality (PQ) is defined as ‘‘the concept of power- ing and grounding sensitive equipment in a matter that is suitable to the operation of that equipment’’ in IEEE Std. 1100 [1]. Another definition is also given as the ‘‘set of parameters defining the properties of the power supply as delivered to the user in normal operating conditions in terms of continuity of supply and characteristics of voltage (symmetry, frequency, magnitude, and waveform) [2]. However, PQ deals with not only voltage quality but also current quality. It includes the combination of voltage and current quality [3,4]. In practice, there are several types of PQ disturbances, such as voltage sag/swell/interruptions, switching tran- sients, flickers, harmonics, notches, etc., caused by faults, nonlinear loads and dynamic operating conditions [5]. Especially, increasing the usage of nonlinear loads, such as diode and thyristor rectifiers, lighting equipments, unin- terruptible power supplies (UPSs), arc furnaces and adjust- able speed motor drives, plays a major role in PQ validations in industrial, commercial and residential power systems. These loads disturb the current and voltage wave- form and change their magnitude and frequency. Also, they generate current harmonics and cause oscillatory and impulsive transients when they are started and stopped. Mostly, pure sinusoidal currents and voltages cannot be provided to the customers. The spectral estimation methods presented in this study have been mostly used for some power electronic applica- tions, such as flicker prediction, harmonic identification [6], switching converters [7] and resonant link inverters [8]. Spectrum estimation of discretely sampled processes is usually based on procedures employing the fast Fourier transform (FFT). This approach is computationally effi- cient and produces reasonable results for a large class of signal processes. In spite of the advantages, there are 0196-8904/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.enconman.2006.12.017 * Corresponding author. Tel.: +90 232 461 41 11x316; fax: +90 232 461 41 21. E-mail addresses: ahmet.alkan@yasar.edu.tr (A. Alkan), asyilmaz@ ksu.edu.tr (A.S. Yilmaz). www.elsevier.com/locate/enconman Energy Conversion and Management 48 (2007) 2129–2135