基于FSST和D-K聚类的次同步振荡分析  

Analysis of subsynchronous oscillation based on FSST and D-K clustering

作　　者：阳育德[1,2] 莫富钧 卢建洛 覃智君 YANG Yu-de;MO Fu-jun;LU Jian-luo;QIN Zhi-jun(School of Electrical Engineering, Guangxi University, Nanning 530004, China;Guangxi Key Laboratory of Power System Optimization and Energy Technology, Guangxi University, Nanning 530004, China)

机构地区：[1]广西大学电气工程学院,广西南宁530004 [2]广西电力系统最优化与节能技术重点实验室,广西南宁530004

出　　处：《广西大学学报（自然科学版）》2021年第4期948-961,共14页Journal of Guangxi University（Natural Science Edition）

基　　金：国家自然科学基金资助项目(51767001)。

摘　　要：为了更准确地识别电力系统次同步振荡的模态数量和频率,以及时进行频率定位和重构,需要提高次同步实测数据的模态辨识的结果精度。针对DBSCAN聚类可以被应用于划分类簇但无法自动计算类簇中心,以及Kmeans聚类需提前确定类簇数量才能计算类簇中心的特点,提出了一种基于傅里叶同步挤压变换和DBSCAN-Kmeans混合聚类(以下简称D-K聚类)的电力系统次同步振荡分析法。模拟数值信号和IEEE次同步谐振(SSR)标准模型算例的仿真实验数据表明,FSST方法能被用于分离距离较近的相邻信号模态。通过算例分析,验证了结合FSST和D-K聚类的次同步振荡分析方法可以避免FSST方法无法自动获取模态的缺陷,且参数辨识结果有较高的精度。In order to identify the frequencies and quantities of the subsynchronous oscillation modes more accurately,and to locate and reconstruct the frequencies,it is necessary to enhance the accuracy of the modal identification results of the subsynchronously measured data.In view of the characteristics that density-based spatial clustering of applications with noise(DBSCAN)can be leveraged to divide the class cluster but cannot calculate the center of the class cluster automatically,and Kmeans clustering needs to determine the number of clusters in advance to calculate the cluster centers,a method of subsynchronous oscillation analysis based on Fourier-based synchrosqueezing transform(FSST)and DBSCAN-Kmeans hybrid clustering(Hereinafter referred to as D-K clustering)is proposed.Simulation experimental data of simulated digital signal and IEEE subsynchronous resonance(SSR)benchmark system model show that the FSST method can be used to separate the adjacent signal modes with close distance.Through numerical example analysis,it is verified that the subsynchronous oscillation analysis method combined with FSST and D-K clustering can avoid the defect that FSST method can not obtain the mode automatically,and its parameter identification results have a high accuracy.

关 键 词：傅里叶同步挤压变换 DBSCAN-Kmeans混合聚类 次同步振荡 模态辨识 参数辨识 

分 类 号：TM743[电气工程—电力系统及自动化]