基于机组相关性的低频振荡多信号分类Prony分析.pdfVIP

第 35 卷 第 6 期 电 网 技 术 Vol. 35 No. 6 2011 年 6 月 Power System Technology Jun. 2011 文章编号：1000-3673 （2011 ）06-0128-06 中图分类号：TM 744 文献标志码：A 学科代码：470·4054 基于机组相关性的低频振荡多信号分类 Prony 分析 王辉，苏小林 （山西大学工程学院，山西省 太原市 030013 ） Generation Unit Correlativity-Based Prony Analysis on Multi-Signal Classification of Low-Frequency Oscillation WANG Hui, SU Xiaolin (College of Engineering, Shanxi University, Taiyuan 030013, Shanxi Province, China) ABSTRACT: Power system signals from wide area 进算法。文献[12]提出了用于低频振荡分析的改进 measurement system (WAMS) can be divided into two 多信号 Prony 算法，首先利用求和方式将所有信号 categories, namely the similar signals such as the angles of all 叠加在一起，然后采用单信号 Prony 算法进行分析， generation units and non-similar signals such as angle, angular 该算法并未考虑不同信号之间的差异、信号过多相 velocity and active power and so on of one and the same 加容易产生样本函数矩阵计算误差等问题，而且系 generation unit. During the analysis on low-frequency oscillation, in order to avoid both numerical value submerging 统规模较大时并非所有信号均能反映系统的振荡 and cumulative error of sample function matrix, a multi signal 特性。为此，本文在改进多信号 Prony 算法的基础 classification-based Prony algorithm is proposed. For 上，将广域测量系统获得的低频振荡信号进行分 large-scale power system, according to correlation factors of 类，同时对大规模电力系统根据机组相关性对机组 generation units the correlation matrix is built and according to 进行分群，进而对相关机群的低频振荡信号进行分 the built matrix the generation units are divided into groups by the α decomposition method