Adaptive observer-based attack location and defense strategy in smart grid

Abstract

This article mainly focuses on the attack location and defense of cyber-physical power systems under dynamic load altering attacks. Taking the smart grid as an example, first, the cyber-physical power system model under attack is established to provide conditions for attack detection. On this basis, an adaptive observer with [Formula: see text] performance is designed to realize detection and location through residual signal and attack estimation, respectively. Furthermore, a power redistribution strategy based on optimal economic scheduling is presented to realize attack defense by means of the observer data. Finally, theoretical analysis proves the feasibility of the algorithm, and simulation cases prove the effectiveness of attack location and defense strategies.

Event-triggered dynamic output feedback control of switched cyber-physical systems with stochastic cyber attacks

“…With the rapid development of computing, communication, and control (3C) technology, cyber-physical systems (CPSs) have attracted extensive attention in the past few years due to their wide range of applications including smart grids, 1 electromechanical system, 2 electric vehicle, 3 and so on. As the next generation intelligent system integrating 3C, CPSs can make physical devices have the functions of calculation, communication, precise control, remote coordination, and self-management, and realize the deep integration of the virtual network world and the real physical world.…”

Guaranteed-performance resilient consensus for multi-agent systems subject to actuator attacks

“…10 Furthermore, the quality of control signal input is the key to control performance, then the security control problems under actuator attack has received increasing attention. 11 For the study of cyber attacks, fruitful theoretical results have been produced, which can be mainly classified in three categories include attack detection, 12,13 attack location, 14 secure state estimation 15,16 and resilient control. 17,18 For the first class, authors in article 19 propose a fault matrix method based on state and output for attack detection.…”

“…Different M f indicates different attack distribution. The gain matrix R of the adaptive observer is obtained by solving inequality(14), that is Meanwhile, we get the matrix M = ½0:8711, 2:3103, 0.3240, 0:1244: Let b = 2. By Theorem 2, we get the matrix H and the controller gain K as…”