面向智能电网的PCA近似法错误数据注入攻击  

PCA Approximation Error Data Injection Attack for Smart Grid

作　　者：褚云龙[1] 谢丽荣[2,3,4] 张小东[1] 任景 刘鹏飞[2,3,4] CHU Yunlong;XIE Lirong;ZHANG Xiaodong;REN Jing;LIU Pengfei(Northwest Branch of National Power Grid Corporation,Xi'an 710048;Nari Technology Development Limited Company,Nanjing 211106;NARI Group Corporation(State Grid Electric Power Research Institute),Nanjing 211106;State Key Laboratory of Smart Grid Protection and Control,Nanjing 211106)

机构地区：[1]国家电网公司西北分,西安710048 [2]国网南京南瑞集团公司(国网电力科学研究院),南京211106 [3]国电南瑞科技股份有限公司,南京211106 [4]智能电网保护和运行控制国家重点实验室,南京211106

出　　处：《计算机与数字工程》2021年第2期360-365,共6页Computer & Digital Engineering

摘　　要：准确的状态估计对于维持智能电网的正常运行至关重要。论文在雅可比(Jacobian)矩阵和状态变量分布假设未知的情况下,采用主成分分析近似法(PCA-Sim)研究了错误数据注入攻击问题,从线路测量的相关性中对Jacobian矩阵进行推导,验证了错误数据注入攻击近似隐形攻击,并且结合累积分布函数(cdf)分析了所提方法的性能。通过对IEEE14配电网系统在均匀状态变量、高斯状态变量和混合状态变量条件下漏检概率与决策阈值的仿真研究,结果证实了所提方法的有效性。Accurate state estimation is very important to maintain the normal operation of smart grid.In this paper,the prob⁃lem of error data injection attack is studied by PCA-SIM under the assumption of unknown Jacobian matrix and state variable distri⁃bution.The Jacobian matrix is deduced from the correlation of line measurements,which verifies the approximate invisibility attack of error data injection attack.The performance of the proposed method is analyzed by combining the cumulative distribution function(CDF).The simulation study on the probability of missed detection and decision threshold of IEEE14 distribution system under uni⁃form state variables,Gauss state variables and mixed state variables proves the effectiveness of the proposed method.

关 键 词：智能电网 坏数据检测 雅可比矩阵 主成分分析 状态估计 

分 类 号：TP353[自动化与计算机技术—计算机系统结构]