Online Identification Of Low Frequency Oscillation In Power System Based On WAMS

Online oscillation characteristic analysis algorithms are used to monitor online low frequency oscillations for control of wide area damping. In all kinds of algorithms, Prony method has unique advantage. However, traditional Prony analysis is very strict with the input signal, and sensitive to noise of data. With respect to the shortage of Prony method, the work in this paper is as follows:(1) A kind of digital filter based on improved EEMD(ensemble empirical mode decomposition) method is proposed. With respect to the shortage of EMD(empirical mode decomposition) filtering, signal with noise is smoothly processed with median filtering method before decomposed by EEMD. The improved method not only combine advantages of median filtering which can filter impulse noise and EEMD method which can filter random noise and high-frequency continuous noise, but also does not exit mode mixing.(2) By means of digital simulation, filtering effect of improved method is researched and compared with that of wavelet method. The simulation shows that: the method can effectively suppressed various noises encountered during power system signal acquisition, and the filtering effect is better than that of wavelet method which has difficulty of choosing basis function.(3) Proposes a method for low frequency oscillations analysis which combined improved ensemble empirical mode decomposition filtering and Prony analysis. In this method, improved ensemble empirical mode decomposition is used to adaptively filter the noise of the input signals before Prony analysis is carried out. The simulation shows that the proposed method can be accurate to abstract model parameters of low frequency oscillations in power system even in a heavy noise environment.