Cyber-Physical Coordinated Modeling for Synchronization-Stability of Grid-Forming Heterogeneous Renewables
Aisikaer Aisikaer, Jing Ling, Zhiqian Yang, Liangnian Lv, Lei Wu, Shuang Xi Jing · Advances in transdisciplinary engineering · 2025
The increasing integration of renewable energy sources into modern power grids has introduced significant synchronization-stability challenges, particularly in grid-forming heterogeneous systems. These challenges arise from the diverse dynamic behaviors of renewable devices, stochastic communication delays, and the complex interplay of cyber-physical interactions, which heighten the risk of transient instability. To address these issues, this study presents a novel interdisciplinary framework that combines advanced control theories, cyber-physical modeling, and machine learning techniques. Specifically, a multi-agent reinforcement learning (MARL)-based approach is proposed to enhance synchronization stability in grid-forming renewable energy systems. The framework integrates Lyapunov-based stability constraints into decentralized control policies, ensuring robust performance under transient disturbances, stochastic delays, and cyber-attacks. A high-fidelity 72-node digital twin testbed was developed to validate the proposed method against traditional control strategies such as droop control and model predictive control (MPC). Results demonstrate that the MARL framework achieves a 38% reduction in synchronization recovery time and significantly lower steady-state errors compared to conventional methods. It also exhibits superior delay tolerance, maintaining stable operation under communication delays of up to 150 ms, far exceeding the limits of droop and MPC. By effectively bridging the fields of machine learning, control systems, and power engineering, this interdisciplinary framework offers a robust, scalable, and adaptive solution to the critical challenges facing modern renewable energy grids.