A Generalized False Data Injection Attacks Against Power System Nonlinear State Estimator and Countermeasures
A Generalized False Data Injection Attacks Against Power System Nonlinear State Estimator and Countermeasures
复制标题
DOI：
10.1109/tpwrs.2018.2794468
复制
发表时间：
2018-01
影响因子：
6.6
通讯作者：
Junbo Zhao;L. Mili;Meng Wang
中科院分区：
文献类型：
--
作者：
Junbo Zhao;L. Mili;Meng Wang
本文开发了一个通用框架，使我们能够从操作员的角度研究电力系统非线性状态估计器对虚假数据注入攻击（FDIA）的脆弱性，并采取一些对策。与大多数现有的 FDIA 方法不同，这些方法假设黑客完全了解系统测量和拓扑，我们推导并分析了成功启动 FDIA 的不确定性及其上限。为了有效防御 FDIA，我们提出了一种强大的检测器，使用安全 PMU 测量的子集来检查测量统计一致性。我们首先证明，如果这些安全 PMU 测量没有不良数据，同时使系统可观察，则 FDIA 是可检测的。然后，我们表明，如果放松这些条件，同时使用来自短期节点同步相量预测的替代冗余测量以及鲁棒的 Huber M 估计器，也可以确保可检测性。 IEEE 30 总线和 118 总线系统上的数值仿真结果证明了该方法的有效性和鲁棒性，即使安全测量包含噪声和不良数据。
This paper develops a generalized framework that allows us to investigate the vulnerability of the power system nonlinear state estimator to false data injection attacks (FDIAs) from the operator's perspective and to initiate some countermeasures. Unlike most existing FDIA methods, which assume a perfect knowledge of the system measurements and topology by a hacker, we derive and analyze the uncertainties for launching successful FDIAs along with their upper bounds. To effectively defend against an FDIA, we propose a robust detector that checks the measurement statistical consistency using a subset of secure PMU measurements. We first show that if these secure PMU measurements are free of bad data while making the system observable, the FDIA is detectable. We then show that detectability is also ensured if these conditions are relaxed while using alternative redundant measurements from short-term nodal synchrophasor predictions together with the robust Huber M-estimator. Numerical simulation results on the IEEE 30-bus and 118-bus systems demonstrate the effectiveness and robustness of the proposed method even the secure measurements contain noise and bad data.