Interval Observer-based Cyber Attack Detection In Smart Grids

With the growing momentum of computer and communication technology,the traditional power system has gradually developed into a smart grid system through a high degree of integration with modern sensing technology,communication technology,and control methods.However,because the smart grid system is increasingly dependent on the communication network,its openness and complexity have also increased,so it is easy to be regarded as a target of attack by hackers.Among the many attack methods,the false data injection attack has attracted much attention.It can bypass the traditional attack detection mechanism without being detected.By injecting fake data into the system,the performance of the power grid system is degraded or even collapsed.Aiming at the information security of smart grid systems,this paper uses interval observers to detect fake data injection attacks in smart grid systems.The main research contents of this paper are as follows:First of all,for the attack detection problem under the false data injection attack on the sensor side of the smart grid system,a detection algorithm of fake data injection attack based on the interval unknown input observer is proposed.For the large-scale multi-regional distributed power system,considering the dynamic model of motor operation and the interconnectivity between regions,the unknown interconnection items are regarded as unknown input items,through decoupling to obtain a brand new system without unknown interconnection items.At the same time,in order to avoid the complex calculation of the threshold value,the boundary of the system disturbance is established The interval unknown input observer.Then use whether the generated interval residual always contains zero value as the detection criterion of the attack.Finally,a simulation example verifies the effectiveness of the proposed detection algorithm against false data injection attacks in the power grid.Secondly,considering the limited communication resources of the wireless communication network of the smart grid system,for network attacks,an attack detection method based on event-triggered interval observers is designed.First,a nonlinear discrete power grid model is established,and an improved time-varying threshold event trigger mechanism is designed for the power grid network system.Secondly,the disturbance,event error,and nonlinear terms are all taken into account,and the event-triggered interval observer is designed.Subsequently,sufficient conditions are proposed and the Lyapunov stability theory is used to prove the boundedness of the estimated upper bound and the estimated lower bound.Then use whether the generated interval residual always contains zero value as the detection criterion of the attack.Finally,simulations verify the effectiveness of the designed false data injection attack detection algorithm based on the event-triggered interval observer.