Mode Identification Of Power System Low Frequency Oscillation Based On Signal Energy Analysis

Power system is a complex dynamic system, and with the interconnection of electric network, the low frequency oscillation has pose a severe threat to operation security and stability of power system. At present, the study of low frequency oscillation concentrates on the research of oscillation mechanics,analysis methods and preventive strategies. With the continuously application of Wide-Area Measurement/Monitoring System, the research of identification methods of dominant mode has been paid more attention to.Aiming at how to abstract the dominant mode, several analysis methods are introduced firstly, then a novel method - signal energy method based on empirical mode decomposition (EMD) has been proposed. An energy-error criterion has been introduced to avoid the influence of false component in the process of EMD, and by means of the proposed method, the applied scene of signal energy method can be expanded.Cases on measured phase-angle trajectories by phasor measurement unit (PMU) and the curves simulated by PSASP are studied, in order to demonstrate effectiveness of the proposed algorithms. By comparison of proposed method with Prony algorithm and eigenvalue-analysis, the results show that the method can be used to identify the signals of off-axis-symmetry and complex modes, Otherwide, by analyzing the influence on EMD of measured signals, the applied precondition of this method is studied.In order to study the application of this method to in multi-machine power systems, a method of equivalent trajectories ananlysis based on clustering algorithms of coherent-generators is suggested. On the basis of the coherency recognitions and clustering of trajectories, equivalent curve is selected to identify the dominant modes between different groups. The results of case studies shows that it is a powerful tool for oscillation research of large power network.