Enhanced Resilient Model Predictive Control Electrolyzers for Frequency Regulations Under Severe Denial-of- Service Attacks

Abstract

Proton exchange membrane electrolyzers (PEMEL) installed with renewable energy resources can be used for power system ancillary services such as frequency regulation and virtual inertia emulation. However, the performance of PEMEL-based frequency ancillary services is threatened by cyberattacks that compromise the communication control. Considering denial-of-service (DoS) attacks on PEMEL communication, this paper proposes an enhanced resilient model predictive control (ER-MPC) for a PEMEL controller. The proposed ER-MPC consists of two procedures. First, the combination of autoregressive (AR) model-based prediction method and hold signal method is used to reconstruct attacked signals during severe DoS attacks. Then a model predictive control (MPC) is used to compute the control signal for PEMEL stack. The objective of ER-MPC-based controlling of PEMEL is to regulate the frequency deviation during contingency and normal operation under severe DoS attack. The effectiveness of the proposed ER-MPC was compared with that of AR-based resilient MPC and resilient MPC methods. The simulation results revealed that the proposed ER-MPC successfully improved microgrid frequency regulations under severe DoS attacks. In addition, the proposed ER-MPC-based PEMEL has a performance effect over other techniques in terms of the reduction in frequency deviation and the rate of change of frequency during severe DoS attacks, disconnection, and successful connection of wind turbine generation.