Oshnoei, SPeykarporsan, RHeidari, JMahboubi-Moghaddam, ELie, TTKhooban, MH2026-01-142026-01-142025-12-28IET Generation, Transmission & Distribution, ISSN: 1751-8687 (Print); 1751-8695 (Online), Wiley, 20(1). doi: 10.1049/gtd2.702181751-86871751-8695http://hdl.handle.net/10292/20499The load frequency control (LFC) scheme, as a vital application in power systems' stability, makes the power system susceptible to cyber-attacks due to its dependence on information technologies and communication networks. This paper studies the LFC performance of Kundur's 4-unit-12-bus power system under false data injection (FDI) attacks. The available defence schemes are either based on the system's model or data-driven. The effectiveness of these schemes depends on the precise mathematical modelling or the extensive historical data of the power system. Therefore, it is necessary to design a defence strategy without depending on the mathematical model and the historical data of the system. To this end, this paper proposes a model-free resilient defence strategy, comprising a model-free detection scheme and an event-triggered mechanism. The presented detection scheme accomplishes the manipulated signal estimation using the measurement and control signals and compares the difference between the estimated and observed signals with a predefined threshold value. When the difference exceeds the threshold value, the detection scheme announces that an attack has occurred on the system. After detecting an attack, the event-triggered mechanism is activated to mitigate the attack's effect on the system frequency response. Accordingly, the event-triggered mechanism blocks the falsified signal and submits the estimated signal to the LFC controller. The presented scheme is independent of the system's mathematical model and historical data and can be employed in any cyber-physical power system. The design process of this strategy is simple and independent of the size and complexity of the power system. A deep reinforcement learning algorithm is also employed to tune the adjustable parameters of the proposed method. The real-time results obtained by the OPAL-RT simulator show that the developed scheme can timely identify FDI attacks and completely mitigate the attack's effect on the system's dynamic performance.This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made. © 2025 The Author(s). IET Generation, Transmission & Distribution published by John Wiley & Sons Ltd on behalf of The Institution of Engineering and Technology.46 Information and Computing Sciences4007 Control Engineering, Mechatronics and Robotics40 Engineering4010 Engineering Practice and EducationBioengineering0906 Electrical and Electronic EngineeringEnergy4008 Electrical engineering4009 Electronics, sensors and digital hardwareJournal ArticleOpenAccess10.1049/gtd2.70218