Research On Method Of Online Identification For Low Frequency Oscillation Based On WAMS

As the development of power grid inter-connection, power oscillation in transmission lines and between generators caused by low frequency oscillation have become a major obstacle for the expand of power grid and . Accurate and timely identification for low frequency oscillation is a main part of the self-healing function in smart grid, and is of great importance for the stability and safety in power system.Low frequency oscillation belong to small signal stability, and weakly damped or negative damped modes are the major cause of small signal unstability. As the WAMS developed, taking this system as a fast, effective platform for online identification for LFO has become a research focus at home and abroad.This paper at first introduces system identification theory based on the ARMA model, including: definition, structure, and physical meaning of such a special case, and then introduced model order determination problems and the estimation of ARMA model parameters, the frequency domain characteristics of ARMA model is also discussed.A brief introduction of the current wide area measurement system and application status, describing current study on low-frequency oscillation including both offline and online methods in order to compare and validate the proposed method. A method based on ARMA model and measured noise like signal is also proposed, which includes: data preprocessing, adaptive order determination for ARMA model, the identification for domain modes of low frequency oscillation and so on. On this basis, an online identification framework for LFO combined with Prony algorithm is proposed based on WAMS, simulation and analysis shows that, the framework proposed can achieve online identification for LFO.Based on MATLAB and the Power System Toolbox Software, white noise is modulated to the load in benchmark system, identification of parameters in different circumstances, such as machine angle and active power flow on transmission line in time domain simulation, obtained some relevant conclusions; For wide area measured signals, an active power signal on a transmission line in China Southern Power Grid is identified to verify the correctness of the proposed method.On the previous content, to research the nonlinear interaction between different oscillation modes, the improved Hilbert-Huang transform, through the analysis of various examples, show that the proposed two improvements are effective for the analysis for nonlinear characteristics of low-frequency oscillation: the improved HHT can inhibite the terminal effect and the distortion of frequency caused by cubic spline, such improved HHT method provides a new way for the research of LFO .