Decoupled Cyber--Physical Defense for Resilient Control of AC Microgrids Under Hybrid Threats

Abstract

This article proposes a decoupled cyber–physical defense strategy for AC microgrids (MGs) under hybrid threats involving internal actuator faults and external false data injection (FDI) attacks. Under such conditions, traditional “estimate-compensate” schemes may misinterpret FDI-corrupted data as actuator faults, whereas “assess-isolate” methods may mistakenly identify nonmalicious devices affected by actuator faults as attackers. To address this issue, a data-source-independent decoupled defense architecture is established. In particular, a self-triggered extended state observer (ST-ESO) is employed at the physical layer to estimate actuator faults and compensate for their effects using only local measurements. At the cyber layer, a composite trust mechanism is integrated with a logic-enhanced finite-state machine (FSM) and a soft-recovery protocol to detect FDI attacks, isolate compromised links, and facilitate gradual link recovery. On this basis, a resilient controller is constructed by integrating local fault compensation with trust-based secure cooperation. Rigorous analysis establishes the global boundedness of the closed-loop error signals, the asymptotic convergence of the tracking and estimation errors, and the absence of Zeno behavior under the stated nondegeneracy condition. Finally, the effectiveness of the proposed strategy is validated through real-time OPAL-RT experiments.