Prescribed Performance-Based Distributed Event-Triggered Data-Driven Frequency Control for Hybrid Power Systems Under Jamming Attacks

The reliance on open communication networks poses significant challenges to power systems. This paper studies a prescribed performance-based scalable distributed event-triggered model-free adaptive control (DETMFAC) scheme for multi-area load frequency control (LFC) systems that are integrated with wind-energy, analyzing both scenarios without and with jamming attacks. First, a transformed-error-based dynamic linearization model (DLM) for distributed outputs is developed to handle the unknown system dynamics under performance constraints. By directly incorporating the DLM and the prescribed performance function into the criterion function, we derive a novel prescribed performance-based DETMFAC scheme. This algorithm improves the control performance while overcoming the problems caused by global topology information and model dependence as in the traditional distributed LFC design approaches by introducing a parameter adaptation mechanism. Additionally, a novel jamming attack model based on the signal-to-interference-plus-noise ratio (SINR) is introduced to illustrate the interaction between the power system and the malicious attacker in a Stackelberg game framework. Then a secure scheme with attack compensation is proposed to mitigate the negative attack impact by reconstructing the LFC commands via linear estimation. Experimental results for a four-area power system demonstrate the designed scheme. Note to Practitioners—LFC is a fundamental strategy for frequency regulation in networked power systems. However, traditional LFC systems are vulnerable to cyber-attacks, tracking performance constraints, and bandwidth limitations due to the interconnection between different subareas. This paper proposes a prescribed performance-based DETMFAC scheme and a secure control scheme with a resilience mechanism to effectively address these challenges in a coordinated manner, integrating an event-trigger, parameter identifier, attack compensator, and input update law. The control input is only updated when the triggering condition is met, based on the distributed output and identified adaptive parameters. In the case of an attack, the event-trigger ceases operations and the attack compensator generates an estimated input signal. Importantly, the proposed LFC scheme relies solely on local input and output data (i.e., model-free) and can be implemented in a fully distributed architecture (i.e., scalable). Thus, practitioners can seamlessly integrate the designed strategies with existing LFC systems to enhance performance, security, and efficiency. The results provide a valuable reference for scalable, cost-effective and secure distributed data-driven control design of multi-area LFC systems and promote related application research.