2026-10-12T23:04:19Z https://www.scidb.cn/oai

10.57760/sciencedb.j00173.00009 2026-04-02T14:57:56Z

2026-04-02 250850 EP丨A Multi-Level Resilience Control Framework for Networked Microgrids to Counter False Data Injection Attacks - Code doi:10.57760/sciencedb.j00173.00009 en In response to the vulnerability of networked microgrids to false data injection attacks (FDIA), frequency instability, and communication congestion in high penetration renewable energy and open communication environments, this paper proposes a multi-level integrated resilience control method of "event triggered communication attack observation deep reinforcement learning compensation". This method designs an adaptive event triggering mechanism based on frequency integration error to reduce redundant data transmission, and constructs a radial basis function unknown input observer (RBF-UIO) to achieve high-precision detection and state estimation of FDIA. On this basis, dual replay Q-learning (DRQL) is introduced to carry out online compensation optimization, achieving attack suppression and performance recovery. Furthermore, simulation verification was conducted on a typical 4-node distributed microgrid platform, and the results showed that the proposed method significantly reduced the number of communication events while effectively improving frequency recovery speed and steady-state accuracy, ensuring safe and stable operation under false data attacks.  microgrid; FDIA; reinforcement learning Huang Yu Cao Zhengyang Hu Songlin Yue Dong Chen Yonghua Yan Yunsong PUBLIC https://creativecommons.org/licenses/by-nc/4.0/ dataset Science Data Bank