基于WAMS与改进拉普拉斯特征映射的同调机群在线识别.pdfVIP

第37 卷 第8 期 电 网 技 术 Vol. 37 No. 8 2013 年8 月 Power System Technology Aug. 2013 文章编号：1000-3673 （2013 ）08-2157-08 中图分类号：TM 714 文献标志码：A 学科代码：470·40 基于 WAMS 和改进拉普拉斯特征映射的 同调机群在线识别 宋洪磊，吴俊勇，郝亮亮，冀鲁豫 （北京交通大学电气工程学院，北京市 海淀区100044 ） On-Line Identification of Coherent Generator Based on WAMS and Improved Laplacian Eigenmap Algorithm SONG Honglei, WU Junyong, HAO Liangliang, JI Luyu (School of Electrical Engineering, Beijing Jiao Tong University, Haidian District, Beijing 100044, China) ABSTRACT: When power system suffers from cascading 系统的时变特征。首先根据WAMS 量测可得到故障后发电 failures that would lead to system asynchronous operation, it is 机组的实时响应功角轨迹信息，利用基于类别信息和核空间 necessary to identify coherent generator groups accurately, 的改进拉普拉斯特征映射算法提取特征信息，进而识别出各 which supplies basis for the study on controlled islanding 发电机的运行特性；再利用k-way 余弦相似度因子分群算法 strategy. Because dynamic trajectory of generator measured by 对发电机组进行自主识别分群。最后通过新英格兰39 节点 WAMS tends to have nonlinear and non-stationary 系统仿真，验证了所提方法的有效性，并且适用于系统不同 characteristic, a novel approach on identification of coherent 运行方式，能在线准确识别同调机群。 generator groups is presented. This method can fully consider 关键词：广域量测系统；拉普拉斯特征映射；同调识别；特 dynamic behavior of different fault modes and time varying 征提取 characteristic of nonlinear system. Firstly, the real-time rotor angle trajectory information can be measured by WAMS 0 引言 directly. Based on categorization information and improved