Distributed Remote Secure ADP-Based Control for Interconnected Power Systems Under Cyber-Attacks

The interconnected power systems (IPSs) generally with complex communication topology are sensitive to cyber-attacks. In this article, the problem of cyber-attacks on the IPSs is considered, and distributed remote secure adaptive dynamic programming (DRSADP) is proposed. Considering the cyber-attacks on the nodes of the sensor and actuator networks of the IPSs, the distributed remote estimators (REs) are first constructed based on the unscented Kalman filter (UKF), and the cyber-attacks models for the sensor and actuator networks are constructed. Then, to compensate for the influence of cyber-attacks and optimize the system state, the estimator-based distributed optimal control strategy is proposed. By constructing the control objective with the control performance index and estimation performance index, the novel optimal attack compensation approach is proposed. Furthermore, the distributed RE-based adaptive dynamic programming with the online finite-time actor-critic algorithm is implemented to solve the secure control strategy effectively. To illustrate the stability of the IPSs with the proposed control strategy, the property of the proposed control scheme is analyzed theoretically, and the finite-time convergence performance with the finite-time gradient learning algorithm is illustrated, which is different from the asymptotic convergence for the existing work. Finally, the results analysis shows that the proposed secure control strategy for IPSs with 9-machines has superiority compared with existing work and the RTLAB-dSPACE platform is applied to further implement the effect of the proposed method.