基于扩展卡尔曼滤波的智能电网虚假数据检测

Detection of False Data in Smart Grid Based on Extended Kalman Filter

虚假数据注入攻击(FDIA)是一种以破坏电力数据完整性与可用性为目的的隐蔽攻击,特点是由攻击者精心设计的攻击向量可以绕过电力系统常规数据检测,因此其检测方法对电力系统的稳定运行具有重要意义。提出了一种基于残差的方法来检测系统是否遭受攻击,保留了系统原有的加权最小二乘估计(WLS),引入了一个额外的扩展卡尔曼滤波(EKF)。由于扩展卡尔曼滤波器的递归特点,在达到新的稳定值前会有一个较长的过程,利用这个过程中两估计器的残差超越所给定的阈值来判断系统是否遭受攻击。在IEEE14节点的仿真证明了检测方法对FDIA的有效性。

False data injection attack(FDIA)is a stealthy attack aimed at destroying the integrity and availability of power grid data.An attack vector carefully designed by the attacker can bypass the traditional data detection of power system.The detection method of FDIA is of great significance to the stable operation of the power system.The paper propose a residual-based detection method to detect whether the system is under attack,introducing an additional extended Kalman filter(EKF),while retaining the original weighted least squares estimator.Due to the recursive feature of EKF,there will be a long process before reaching a new stable value after an attack occurs.In this process,the deviation of the two estimators is used to judge whether the system is under attack.The simulation on IEEE14-bus system proves the effectiveness of the detection method for FDIA.