Research On Network Attack Detection Technology Of Power Grid CPS Based On Situational Awareness

In recent years,with the rapid breakthrough of wide area measurement technology,communication technology and intelligent technology,the visibility and controllability of Cyber-Physical Systems(CPS)have been greatly upgraded,which immensely improves the performance of information management for power grid resources and effectively promotes the traditional power grid to digital and intelligent evolution.However,the highly networked CPS system of the power grid is inevitably threatened by the network security.In addition,the CPS network attacks are difficult to display and identify.With the continuous attack,the influence is wide-ranging,which indirectly brings serious negative impact on the stable operation of the power grid system and even social stability.The purpose of this paper is to realize the accurate identification of network attack types through power grid CPS operation situation awareness and instruct decision-makers to prevent network attacks effectively,so as to ensure the security and stability of power grid operation.Firstly,this paper summarizes and classifies network attack behaviors from four aspects,such as attack purpose,attack object,attack mode and attack strategy.Secondly,combined with the vulnerable spot in the process of power grid CPS operation,the potential network attack types of power grid are analyzed.In addition,this paper establishes the CPS network physical attack model in the basis of DC power flow,then use the CPS situation awareness method based on improved particle swarm optimization(PSO)algorithm to solve the model,which is verified by the actual operation of the power grid.Finally,the power distribution energy management system is taken as an example to evaluate the vulnerability of load-side false information injection attack in the Interactive Demand Response(IDR)mode.Then the IDR attack model is constructed based on the idea of bi-level optimization model.Meanwhile,the long short-term memory(LSTM)is introduced to form a new network attack detection method for model solving.The actual power grid example results show that the network security situation awareness method proposed in this paper can identify the types of network attacks more accurately and efficiently.The sensitivity of network attack detection is as high as 99.67%,which is nearly 10% higher than traditional detection methods.Therefore,it is valuable in engineering practice.