基于拉格朗日乘子法的虚假数据攻击策略

发布时间：2018-04-24 09:50

本文选题：虚假数据攻击 + 信息物理融合系统　；参考：《电力系统自动化》2017年11期

【摘要】：虚假数据攻击面临掌握的电气参数存在误差,甚至不完整及量测数据中存在不良数据的问题,提出一种基于拉格朗日乘子法的虚假数据攻击策略。首先通过拉格朗日乘子法和增广状态估计法辨识不良数据和估计未知支路电抗,然后在凸松弛技术框架内,将传统的攻击单个量测点的次优虚假数据攻击向量模型转化为基追踪(BP)模型,最后采用交替方向乘子法(ADMM)快速求解次优攻击向量。以典型的IEEE节点测试系统为例进行仿真测试,仿真结果表明:与传统的线性规划算法相比,将攻击单个量测点的次优攻击向量模型转化为BP模型后,采用ADMM求解次优攻击向量具有更高的计算效率;电抗未知支路数量较少时,攻击成功率较高,但是状态变量的误差向量的标准差较小时,电抗未知支路数量对攻击成功率影响减弱;该方法不会显著增加攻击成本。
[Abstract]:The false data attack is faced with the error of the electrical parameters in control, even incomplete and the bad data in the measurement data. A false data attack strategy based on Lagrange multiplier method is proposed. At first, the Lagrange multiplier method and augmented state estimation method are used to identify bad data and estimate unknown branch reactance, and then within the framework of convex relaxation technique, The traditional suboptimal false data attack vector model which attacks a single measurement point is transformed into the base tracking BP- model, and the suboptimal attack vector is solved quickly by alternating direction multiplier method (ADMMM). A typical IEEE node test system is taken as an example. The simulation results show that compared with the traditional linear programming algorithm, the sub-optimal attack vector model which attacks a single measurement point is transformed into BP model. When the number of reactance unknown branches is small, the success rate of attack is higher, but the standard deviation of error vector of state variable is smaller. The effect of the number of reactance unknown branches on the attack success rate is weakened, and the attack cost is not significantly increased by this method.
【作者单位】：武汉大学电子信息学院;武汉大学动力与机械学院;
【基金】：国家自然科学基金资助项目(50677047) 湖北省自然科学基金资助项目(2015CFB563) 中央高校基本科研业务费专项资金资助项目(2042017kf0037)~~
【分类号】：TM76;TP309

【相似文献】