2互信息理论-电测与仪表.doc

基于独立分量分析和互信息的多谐波源定位 苗长新1，冯学俊1，汤家升1，李璇1，刘鹏1，高月奎2 （1.中国矿业大学信息与电气工程学院，江苏 徐州 221008；2.山东能源枣庄矿业有限责任公司，山东 枣庄 277000） 摘要：针对独立分量分析算法在谐波源定位中估计得到的各次谐波电流与谐波源电压不存在固定的对应关系，提出了一种基于独立分量分析（ICA）（MI）PSCAD环境中，对有多个谐波源的IEEE-14母线网络进行了仿真。结果表明，这种方法在对电网参数未知的情况下，能够准确判断出多个谐波源的位置。 关键词：独立分量分析；；；PSCAD 中图分类号： 文献标识码： 文章编号：1001-1390（201）00-0000-00Localization of Multi-harmonic Sources Based on Independent Component Analyses and Mutual Information MIAO Chang-xin1,FENG Xue-jun1,TANG Jia-sheng1,LI Xuan1,LIU Peng1,GAO Yue-kui2 (1. College of Information and Electrical Engineering, China University of Mining and Technology,Xuzhou 221008,Jiangsu,China. 2.Shangdong Energy Zaozhuang Mining Co. Ltd.,Zaozhuang 277000,Shandong,China ) Abstract: In view of the non-fixed relationship between the subharmonic current estimated by ICA and the harmonic source voltage, a multi-harmonic source localization method was proposed, which was based on the independent component analysis (ICA) algorithm and the mutual information (MI) theory. The ICA algorithm was applied to estimate the harmonic currents generated by the harmonic source. The MI theory could accurately locate multi-harmonic sources according to the relevant information between the estimated harmonic current and bus voltage. To verify the feasibility of this method, PSCAD simulations were made to locate multi-harmonic sources in the IEEE-14 bus system. The results showed that this approach could accurately locate multi-harmonic sources without knowledge of the network parameters. Key words:independent component analysis, mutual information, harmonic source localization, multi-harmonic sources, PSCAD 0 引 言 近年来，由于非线性负载的广泛运用导致电网谐波污染日趋严重。为了区分各谐波源谐波责任、制定谐波奖惩方案和综合治理谐波问题，研究谐波源定位技术已经刻不容缓[1-2]。 文献[3]最先系统研究了基于独立分量分析算法的盲源分离问题，通过定义目标函数，当各输出信号相互独立时，目标函数趋近于最小或最大，为研究盲源分离问题指明了方向。文献[4]将电力系统谐波辨识的问题抽象为盲源分离问题，从而将独立分量分析算法运用于谐波辨识。但是在ICA算法中，估计得到的信号与源信号不存在固定的对应关系，主要表现在估计得到的信号矢量的排列位置、幅值和初始相位会发生变化[5]。这种分离存在不确定性。 因此，本文针对独立分量分析算法估计得到的各次谐波电流与谐波源电流不存在固定的对应关系，引入了互信息