Adaptive Extend Kalman Filter-based Cyber Attack Detection And Isolation In Smart Grid

With the rapid development of science and technology,the traditional power system is deeply integrated with modern sensing technology,control method and communication technology,forming a new type of smart grid system.However,the smart grid system has many new features,such as complexity and openness,making it a target for many attackers.Compared with traditional attacks,most of the attacks are stealthy in smart grid.This type of attack can bypass the traditional detection mechanism and inject false data into the system without being detected,thereby destroying the power grid system normal operation.Aiming at the information security problem of smart grid,this paper uses Kalman filter to detect and locate false data injection attacks and replay attacks in smart grid.The main work is as follows:Firstly,in order to detect false data injection attack in smart grid,a cosine matching attack detection method based on adaptive extended Kalman filter is proposed.Considering the impact of input and output nonlinearity,a nonlinear state space model of smart grid is established.Based on this model,an attack model of the false data injection attack is given and the stealthy characteristic of the attack is analyzed.Considering the time-varying noise which may exist in the system,an adaptive extended Kalman filter is designed to estimate the internal state of the system.Then,the attack can be detected by comparing the changes in the internal state of the system.Secondly,in order to locate false data injection attack in smart grid,a location method based on dimensionality reduction filter bank is proposed.At first,the power grid model is expanded to become a power grid model containing N sensor nodes,and the large power grid is divided into regions.Then,the performance of the detection method is analyzed,and an adaptive threshold based on residual is designed to shorten the time of attack detection and location successfully.Then,the dimensionality reduction filter bank and attack detection method are used to locate a single attack in the system.Finally,for the detection of hybrid attacks in smart grid,a hybrid attack detection method based on adaptive extended Kalman filter is proposed.Considering the impact of input nonlinearity,an adaptive extended Kalman filter based on model optimization is designed to estimate the state.Considering the attack characteristics of replay attack and false data injection attack under the optimized model,the concealment of the two attacks is analyzed.An adaptive extended Kalman filter based on model optimization is designed to detect hybrid attacks in smart grid.