False Data Injection Attack Detection And Partition Defense Based On Node Trust

In view of the malicious behavior of False Data Injection Attacks that can avoid bad Data detection mechanism and tamper with state estimation results,this paper proposes a detection and defense scheme of false data injection attack based on node trust.In this paper,a power system network node trust model is established to achieve the goal of FDIA detection and location,and the node trust and WLS state estimation algorithm are combined to correct the state estimation in real time.Finally,we define a new control distance by considering the node trust degree.And use the hierarchical clustering method and the mapping partition algorithm to optimize the partition of the power grid to mitigate the impact of the attack on the power system.Firstly,the failure of the residual threshold method to the covert data injection attack carefully constructed by the attacker is discussed.According to the characteristics of network node state estimation and updating,this paper proposes a node trust model and a trust updating algorithm.The detection of false data injection was carried out by measuring the degree of trust,and the validity under different attack scenarios was verified by IEEE 14-node and IEEE-118 node standard test systems.Secondly,in order to eliminate the interference of the false data injected by the attacker on the normal operation of the power grid,this paper combines the trust updating algorithm with the classical weighted least square algorithm to improve the robustness of the state estimation and alleviate the impact of the false data injection attack on the power network.In this paper,simulation experiments are carried out on 14-node and 118-node systems to test the accuracy of the optimized state estimation algorithm.Finally,this paper defines a control distance based on the trust degree of nodes,and proposes an optimized partition model based on the step-by-step partitioning method of the first reactive source node and the back-load node.First,according to the control distance between the reactive power sources based on the trust degree,the hierarchical clustering algorithm is adopted to directly partition the reactive power sources.Then,based on the principle of the shortest control distance,load mapping algorithm is adopted to partition load nodes.Thus,the coupling between the nodes in the region is strong and the coupling between the regions is weak,and the voltage control of the load nodes by the reactive power source is satisfied.In the end,the simulation results of IEEE-14 and39-node standard test systems under different scenarios are analyzed,and it is proved that the method used in this paper can effectively partition the system after being attacked,which provides theoretical basis and technical support for ensuring the stability of the system.