Study On Power Grid Low Frequency Oscillation Comprehensive Self-adaption Identification Methods

As with the scale of power grid and new-energy is getting bigger, grid structure and operation function are more changeable. With various kinds of automatic control devices brought in, the modern power grid’s dynamic characteristics are becoming more sophisticated. The study on the self-adaption identification methods of low-frequency oscillation under practical condition has its practical importance to the analysis, operation, monitoring and control of power grid’s dynamic security and stability.First, studying the Prony algorithm and its methods of selecting parameters can improve the identification’s accuracy. For the algorithm parameters has direct crucial influence to its accuracy, study on the dynamic selection of sampling frequency, time window length and model order is conducted, including the self-adaption identification method of selecting model order based on absolute value of residues and DVR index.Second, a comprehensive self-adaption identification method of low-frequency oscillation is proposed. The main contents are concluding three sides. The first point is to excavate different characteristics of tie-line power oscillation and fluctuation, and the extending Prony method based on ITAE index and mode energy is proposed, identifying the two main characteristics of oscillation energy-level and spontaneity respectively. The time window data that has oscillation characteristics is first identified, thus the defect of traditional Prony algorithm considering the presence of steady-state random noise under realistic condition can be omitted. The second point is to promote a self-adaptive damping ratio amending algorithm of low-frequency oscillation identification based on WAMS, thus improving the integral accuracy and adaptability of the method to power grid’s operation mode change. The third point is to propose an index of power grid structure coefficient based on short-circuit current to identify the physical characteristics of low-frequency oscillation triggered by grid structure weakness.At last, the comprehensive self-adaption identification flow chart is given in detail. The feasibility and effectiveness of this algorithm and method are verified by the simulation calculation of realistic power grid.