Research On Cyber Attack Detection Strategy In Smart Grid Based On Robust Observer

With the continuous evolution and development of the power grid,a significant breakthrough is the emergence of the smart grid,which is a power system that integrates information technology and automation technology to improve the efficiency and reliability of the power grid.However,due to its high integration with communication technology,the smart grid has more vulnerabilities,making it more susceptible to malicious cyber attacks.Among them,false data injection attacks can bypass traditional attack detection strategies,affecting the normal operation of the smart grid and even causing it to collapse.This article focuses on the security issues of the information layer of the smart grid and conducts research on observer-based detection strategies for possible attacks in the grid.The main work of this paper is as follows.Firstly,this section proposes an attack detection strategy based on an RBF neural network observer for false data injection attacks in smart grids with unknown parameters.Considering that the unknown parameters in actual power grids can affect state observation and threshold design,thereby reducing the accuracy of attack detection,an RBF neural network is first introduced to approximate the unknown parameters in the model.Based on this,a neural network observer and a verification threshold are designed to ensure the stability of the observer,and the effectiveness of the proposed detection strategy is verified through simulation examples.Secondly,in this section,a robust on interval sliding-mode observer is designed for the threshold design problem faced by threshold-based power grid attack detection,and an attack detection algorithm is proposed based on this observer.Considering the difficulty of threshold design encountered by most detection strategies,an interval observer is introduced to eliminate the problem of threshold design,and the robustness of the observer is enhanced by adding sliding-mode control.Meanwhile,the introduction of the observer performance reduces the impact of unknown disturbances on interval residuals,improves the efficiency of the detection algorithm,and finally,the effectiveness of the algorithm is verified through simulation examples.Thirdly,this section addresses the issue of insufficient timeliness in observer-based attack detection strategies by proposing a hierarchical observer-based attack detection strategy to highlight the attacked parts.First,the system is decomposed based on the parts that may be attacked.On the basis of the interval observer in Chapter 2,a hierarchical strategy and sliding mode control are added.By progressively deriving the observer needed for attack detection based on the three parts of the system’s state variables,the area where attacks may occur is highlighted,thus achieving a rapid response of the detection strategy.Finally,the effectiveness of the strategy is verified through simulation examples.