Book Chapter
Wide-area monitoring and analysis of inter-area oscillations using the Hilbert-Huang transform
Fecha
2008Autor
Messina, A.R.
Andrade, M.A.
Barocio, E.
Institución
Resumen
Many transient processes in power systems involve phenomena that vary in time andspace in complicated ways. Comprehensive monitoring of large-scale power systems bymeans of properly placed time-synchronized phasor measurement units (PMUs) providesthe opportunity to analyze and characterize complex inter-area swing dynamics involvingall or most of the power system.Wide-area real-time monitoring may prove invaluable in power system dynamicstudies by giving a quick assessment of the damping and frequency content of dominantsystem modes after critical contingencies. Measured data, however, may exhibit quitedifferent dynamics at each system location or exhibit abrupt changes, dynamicirregularities, or be complicated by nonlinear trends or noise. Traditional Fourier andProny methods for system identification are unable to resolve the localized nature ofthese processes and hence provide little useful information concerning the nature ofnoisy, time-varying oscillatory processes.In this Chapter, a new method for analyzing the temporal dynamics of nonlinear andnon-stationary inter-area oscillations using a local empirical mode decomposition (EMD)method and the Hilbert transform is presented. Two novel algorithms are developed toaddress nonlinear and non-stationary issues. The first method is a local implementation ofthe empirical mode decomposition technique. The second is an algorithm to compute theHilbert transform using finite impulse response (FIR) filters. By combining theseapproaches, the method can be used to analyze complex signals for which theconventional assumptions of linearity and stationarity may not apply and can beimplemented for on-line estimation of modal damping and frequency using synchronizedwide-area measurement systems. The physical mechanism underlying nonlinear time-varying inter-area oscillations isinvestigated and methods to characterize the observed oscillatory phenomena in terms ofphysically meaningful modal components are proposed. Emphasis is placed onidentifying modal content in the presence of noise and nonlinear trends. Issuesconcerning the implementation of the method and numerical considerations are alsodiscussed.As specific applications, data obtained from PMU measurements from a real event inthe northern systems of the Mexican interconnected system are used to examine thepotential usefulness of nonlinear time series analysis techniques to characterize thespatio-temporal characteristics of the observed oscillations and to determine the natureand propagation of the system disturbance. The efficiency and accuracy of the method isdemonstrated by comparison to other approaches. � 2008 Nova Science Publishers, Inc. All rights reserved.