Cyber attack estimation and detection for cyber-physical power systems

•The cyber-physical power systems dynamic model is established by considering the state attack and sensor attack, but unlike other excellent papers, our results provide a method that does not require linear transformation for the attack detection.•There are many methods of attack detection, but the detection method of our article takes into account the estimation of attack.•Attack estimation observers are designed for state attack and sensor attack with a prescribed H∞ performance, respectively. This paper mainly studies the problem of cyber attack estimation and detection in cyber-physical power systems (CPPS). Firstly, CPPS under cyber state attack and sensor attack are modeled through system characteristics. Secondly, to achieve attack estimation, two robust observers are designed for the CPPS under state attack and sensor attack respectively, which can guarantee the error systems are asymptotically stable with a prescribed H∞ performance. Next, to achieve attack detection, attack detection logic is introduced by comparing the threshold with the estimated attack. Finally, in order to show the feasibility of the state attack and sensor attack detection, taking power systems with three generators and six buses as an example, the simulation results are given.