基于ICA和WVD的电力系统低频振荡参数辨识  

Identification of Low-frequency Oscillation Parameters of Power System Based on Independent Component Analysis and Wigner-Ville Distribution

作　　者：王雨虹[1,2] 郭天驰 付华 徐耀松[1] WANG Yuhong;GUO Tianchi;FU Hua;XU Yaosong(School of Electrical and Control Engineering,Liaoning Technical University,Huludao 125105,China;School of Safety Science and Engineering,Liaoning Technical University,Fuxin 123000,China)

机构地区：[1]辽宁工程技术大学电气与控制工程学院,葫芦岛125105 [2]辽宁工程技术大学安全科学与工程学院,阜新123000

出　　处：《电力系统及其自动化学报》2019年第2期74-79,共6页Proceedings of the CSU-EPSA

基　　金：辽宁省教育厅重点实验室资助项目(LJZS003)

摘　　要：为了提高电力系统中的低频振荡参数辨识的精度,提出一种基于独立分量分析ICA(independent compo?nent analysis)和Wigner-Ville分布WVD(Wigner-Ville distribution)相结合的在线辨识方法。以广域测量系统WAMS(wide area measurement system)监测到的数据为原始输入信号,采用ICA算法对信号进行降噪处理,再应用Wigner-Ville分布研究信号的频率、幅值及能量分布特点。仿真分析和应用实例研究表明,该方法相较于传统Prony辨识算法而言,具有较强的抗噪能力和较好的辨识结果,可以更好地反映电力系统中非平稳信号的局部特性,提高了电力系统低频振荡参数在线辨识的准确性。To improve the identification accuracy of low-frequency oscillation parameters of power system , an on-line identification method based on independent component analysis ( ICA ) and Wigner-Ville distribution ( WVD ) is proposed in the paper. The monitor data from the wide area measurement system ( WAMS ) are taken as the original input signal. After de-noising treatment by ICA , WVD is applied to study the frequency , amplitude and energy distribution characteristics of the signal. The simulation analysis and the study on an application example show that compared with the traditional Prony identification algorithm , the proposed method has a stronger de-noising capability and better iden? tification result , thus it can better reflect the local characteristics of the non-stationary signal of power system , and im? prove the online identification accuracy of low-frequency oscillation parameters.

关 键 词：低频振荡 独立分量分析 Wigner-Ville 分布 广域测量系统 

分 类 号：TM712[电气工程—电力系统及自动化]