基于约束EKF的低频振荡模态参数辨识  

Identification for Modal Parameters of Low Frequency Oscillation Based on Constrained EKF

作　　者：刘亚南[1] 王义[2] 钟永洁 孙永辉[2] LIU Yanan;WANG Yi;ZHONG Yongjie;SUN Yonghui(Jiangsu Frontier Electric Technology Co.,Ltd,Nanjing,Jiangsu 211102,China;College of Energy and Electrical Engineering,Hohai University,Nanjiang,Jiangsu 210098,China)

机构地区：[1]江苏方天电力技术有限公司,江苏南京211102 [2]河海大学能源与电气学院,江苏南京210098

出　　处：《广东电力》2018年第7期77-83,共7页Guangdong Electric Power

基　　金：国家自然科学基金项目(61673161);江苏省自然科学基金项目(BK20161510);中央高校基本科研业务资助项目(2017B13914)

摘　　要：低频振荡已成为限制电力系统区域间功率传输能力的突出问题,严重影响了电力系统的安全稳定运行。为了准确有效地提取低频振荡信号所包含的特征信息,分析低频振荡信号模态构成的特点,在扩展卡尔曼滤波(extended Kalman filter,EKF)算法的基础上,结合信号自身的物理约束,提出一种低频振荡模态参数辨识方法,实现了约束条件下的低频振荡模态参数实时在线辨识。所提方法能有效避免运用EKF算法进行低频振荡模态参数辨识时收敛性差和参数越界的问题,提高低频振荡模态参数辨识的精度。最后,对不同的低频振荡信号进行仿真测试分析,结果表明该约束EKF方法不但能够实现低频振荡模态参数的约束辨识,而且较EKF算法具有更好的收敛性和更高的辨识精度。Low frequency oscillation has become a prominent problem of restricting inter-regional power transmission ability of the power system,which has seriously affect safe and stable operation of the power system.In order to accurately and effectively extract feature information of low frequency oscillation signals and analyze the modal characteristics,this paper combines physical constraint of signals and presents a kind of identification method for modal parameters of low frequency oscillation based on the extended Kalman filter（EKF）algorithm that has realized real-time online identification for modal parameters of low frequency oscillation under constrained conditions.It proves the proposed method can effectively avoid problems of poor convergence and parameter transition in the process of identifying modal parameters by using the EKF algorithm.Finally,it makes simulation and testing analysis for different low frequency oscillation signals and the results indicate this constrained EKF method can not only realize constraint identification for modal parameters,but also has better convergence and higher precision compared with the EKF algorithm.

关 键 词：卡尔曼滤波 约束方法 低频振荡 粒子群算法 参数辨识 

分 类 号：TM71[电气工程—电力系统及自动化]