一种智能电网虚假数据攻击的检测和防御  

Detection and Defense of False Data Attacks in Smart Grid

作　　者：诸德律 翟晓萌 方向 刘毅 ZHU Delü;ZHAI Xiaomeng;FANG Xiang;LIU Yi(Economic Research Institute,State Gird Jiangsu Electric Power Co.,Ltd.,Nanjing 210008,China;State Gird Jiangsu Electric Power Co.,Ltd.,Nanjing 210008,China)

机构地区：[1]国网江苏省电力有限公司,经济技术研究院,江苏南京210008 [2]国网江苏省电力有限公司,江苏南京210008

出　　处：《微型电脑应用》2023年第2期24-26,29,共4页Microcomputer Applications

基　　金：国网江苏成本课题(B110K0207405)。

摘　　要：数据驱动的虚假数据注入攻击是智能电网的网络安全面临的主要威胁之一。现有方法是在不知道系统测量矩阵的情况下,从估计的信号子空间构造跟踪向量。估计子空间的精度对其隐身性能有很大的影响。此外,由于模型阶次选择需要全尺寸奇异值分解(SVD),计算量大。因此,在这里主要探究了一种基于截断奇异值分解(SVD)的攻击方案的原理,该方案仅使用第一主特征向量。通过探究该方法,从而达到对相关攻击鉴别的目的。In the monitoring of smart meter data, the false data injection attack driven by data is one of the main threats to the network security of smart grid. The existing method is to construct the tracking vector from the estimated signal subspace without knowing the system measurement matrix. However, the accuracy of subspace estimation has a great impact on its stealth performance. In addition, due to the need for full scale singular value decomposition(SVD) for model order selection, the computational complexity is large. Therefore, we mainly explore the principle of an attack scheme based on truncated singular value decomposition(SVD), which only uses the first principal eigenvector. Through exploring this method, the purpose of identifying relevant attacks can be achieved.

关 键 词：智能电网 注入攻击 网络安全 数据检测 虚假数据 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]