基于SVD-Prony及主成分回归的次同步振荡阻尼特性影响因素研究  

Analysis of Influencing Factors on Damping Characteristics of Subsynchronous Oscillation Based on Singular Value Decomposition-Prony and Principal Component Regression

作　　者：张骞 边晓燕 徐鑫裕 黄阮明 李灏恩 Zhang Qian;Bian Xiaoyan;Xu Xinyu;Huang Ruanming;Li Hanen(Shanghai University of Electric Power,Shanghai 200090,China;State Grid Shanghai Pudong Electric Power Supply Company,Shanghai 200120,China;State Grid Shanghai Electric Power Economic&Technological Institute,Shanghai 200223,China)

机构地区：[1]上海电力大学电气工程学院,上海200090 [2]国网上海浦东供电公司,上海200120 [3]国网上海市电力公司经济技术研究院,上海200223

出　　处：《电工技术学报》2022年第17期4364-4376,共13页Transactions of China Electrotechnical Society

基　　金：国家自然科学基金项目(51977127);上海市科学技术委员会项目(19020500800)资助。

摘　　要：风电并网系统次同步振荡阻尼特性及其影响因素的准确有效辨识,是解决实际风电场并网系统工程中次同步振荡问题的前提和关键。为避免传统次同步振荡分析方法的复杂建模,该文提出一种利用风电场参数及运行数据的次同步振荡阻尼特性及其影响因素分析方法。首先,基于奇异值增长率谱的奇异值分解法(SVD)提升Prony法的抗噪性,从振荡数据中提取模态信息;其次,选取振荡的影响因素并进行相关性分析及共线性诊断;然后,采用主成分回归(PCR)减少回归分析中的共线性问题,建立阻尼比估计模型,辨识出影响风电场次同步振荡阻尼比的主导因素;最后,通过双馈风电场经柔直并网系统仿真验证了所提方法的有效性及准确性。Accurate and effective identification of damping characteristics and its influencing factors of subsynchronous oscillation(SSO)in wind power grid connected system is the premise and key to solve the problem of SSO in actual wind power grid-connected system engineering.To avoid the complex modeling of traditional SSO analysis method,this paper proposes a method to analyze the influencing factors of SSO damping characteristics based on wind farm parameters and operation data.Firstly,singular value decomposition(SVD)based on singular value growth rate spectrum is applied to improve noise resistance performance of Prony and modal information is extracted from oscillation data.Secondly,the influencing factors of oscillation are selected and the correlation and collinearity diagnosis are carried out.Then,principal component regression(PCR)is adopted to reduce the collinearity in the regression analysis,the damping ratio estimation model is established to identify the dominant factors affecting the subsynchronous stability of the wind farm.Finally,the effectiveness and accuracy of the proposed method are verified by power system with DFIG integrated through VSC-HVDC.

关 键 词：奇异值分解 PRONY法 次同步振荡 主成分回归 

分 类 号：TM712[电气工程—电力系统及自动化] TM614