基于慢同调理论和希尔伯特-黄变换的发电机在线同调识别

摘要

提出一种在线同调机群识别方法.根据慢同调分群算法确定线性化系统的最优慢同调聚合分群数,根据分群矩阵的相关系数、振荡模式的参与因子,以及广域测量系统提供的功角信息,利用希尔伯特-黄变换提取发电机的振荡模式能量谱,实现扰动后的发电机群的同调识别.通过WSCC系统和新英格兰10机39节点系统的仿真实验验证所提方法的适用性和有效性,该方法能全面反映系统在受扰后的动态特性和系统的非线性时变特性.%A method of online identification of coherent electric generators is proposed,which adopts the slow coherency grouping algorithm to determine the group number of optimal slow coherency clustering and applies the Hilbert-Huang transform,based on the correlation coefficients of group matrix,the participation factors of oscillation model and the power angles measured by the wide area measurement system,to abstract the modal energy spectrum of generator oscillation for identifying the coherent generator group after disturbance.Simulative experiments for WSCC system and New England 10-machine 39-bus system show that,the dynamic performances and non-linear time-varying characteristics of power system after disturbance can be completely demonstrated by the proposed method reflects,verifying its applicability and effectiveness.