Enhancing automatic generation control (AGC) for frequency stability in renewable-dominated power grids: Challenges, gaps, and future directions
Mkhutazi Mditshwa, Komla A. Folly, David T. O. Oyedokun
Abstract
This paper addresses the critical challenges posed by the high penetration of Variable Renewable Energy (VRE) in modern power systems, particularly concerning frequency stability and Automatic Generation Control (AGC). As renewable energy sources, such as solar and wind, continue to grow, traditional power systems encounter significant issues due to reduced system inertia and increased frequency fluctuations. Traditional AGC methods, which rely heavily on accurate system models, often fall short in coping with the variability of VRE, underscoring the need for adaptable, real-time control strategies. To address these issues, this paper reviews the performance and potential of model-free AGC strategies, including fuzzy logic, artificial neural networks (ANNs), deep reinforcement learning (DRL), and hybrid methods. These advanced control approaches do not depend on precise system models, allowing them to dynamically adapt to fluctuating grid conditions, making them particularly suitable for renewable-dominated power systems. Through a comparative analysis based on robustness, computational demands, and adaptability, the paper assesses the strengths and limitations of each approach: fuzzy logic is shown to be robust and cost-effective but may struggle with complex scenarios; ANNs and DRL offer high adaptability but require considerable computational resources. The review identifies key research gaps, including the need to optimize AGC systems specifically for high VRE integration, enhance the participation of renewable resources in frequency regulation, and develop hybrid and adaptive models that integrate both model-free and model-based methods. Hybrid approaches, which combine the adaptability of model-free methods with the structure of model-based techniques, hold significant promise for advancing AGC in renewable-rich power systems. Future research directions are recommended to enhance computational efficiency and robustness in these systems, ultimately fostering resilient and sustainable power grids capable of supporting a renewable-powered future.