基于线性判别分析与主成分分析的骨干网架生存性评估方法.pdfVIP

38 卷 2 期 电 网 技 术 Vol. 38 No. 2 20 14 年2 月 Power System Technology Feb. 20 14 文章编号：1000-3673 （20 14）02-0388-07 中图分类号：TM 72 文献标志码：A 学科代码：470 4051 基于线性判别分析与主成分分析的 骨干网架生存性评估方法 赵一婕，刘涤尘，吴军，董飞飞，宋春丽，潘旭东，王浩磊 （武汉大学 电气工程学院，湖北省 武汉市 430072 ） Survivability Evaluation of Backbone Network Based on Linear Discriminant Analysis and Principal Component Analysis ZHAO Yijie, LIU Dichen, WU Jun, Dong Feifei, SONG Chunli, PAN Xudong, WANG Haolei (School of Electrical Engineering, Wuhan University, Wuhan 430072, Hubei Province, China) ABSTRACT: From the perspective of resistibility, 决准则函数的最大值，以其所对应的最大投影方向确定三级 recoverability, security and connectivity of backbone network 指标权重，再对二级指标矩阵进行主成分分析，根据累积方 an indicator system for survivability evaluation of 差贡献率来确定二级指标权重，最终实现骨干网架综合生存 differentiated planning is established, and the analysis on 性评估。IEEE 118 节点算例的仿真结果验证了该评估方法的 benefit and cost attributes of survivability indicators and the 有效性和可信性。 standardization of them are performed. On this basis, a linear 关键词：差异化规划；骨干网架；生存性；线性判别分析； discriminant analysis (LDA) model for the second-class 主成分分析 indicators and a principal component analysis (PCA) model for DOI 10.13335/j.1000-3673.pst.2014.02.018 the first-class indicator are built. Firstly, the maximum value of LDA judgment criterion function for the second-class decision 0 引言 matrix is solved, and according to the largest proj ection 近年来，频繁发生的极端自然灾害对电力系统 direction corresponding to this maximum value the weights of [1-