抄録
The large-scale integration of renewable energy is fundamentally challenging the operational paradigm of modern power systems by intensifying intermittency, reducing system inertia, and increasing spatiotemporal uncertainties across generation, grid, and demand sectors. Developing coordinated regulation and dynamic balancing mechanisms for heterogeneous resources across multiple temporal and spatial scales has therefore become a fundamental challenge for next-generation power systems. This review systematically reconstructs the research landscape of source–grid–load coordinated regulation through a unified framework integrating theoretical foundations, collaborative architectures, enabling technologies, optimization methodologies, dynamic balancing mechanisms, and future directions. From the complementary perspectives of source-side flexibility, grid-side active support, and demand-side interactive response, representative technologies—including renewable power forecasting and grid-forming control, coordinated energy storage, flexible transmission and distribution, wide-area monitoring and digital twins, demand response, aggregated electric vehicle control, and behind-the-meter storage—are classified and comparatively evaluated in terms of regulation capability, response characteristics, economic performance, engineering maturity, and application scenarios. Coordinated optimization strategies encompassing multi-timescale scheduling, multi-agent game-theoretic optimization, artificial intelligence-assisted decision-making, data-driven optimal control, and robust coordinated control are further synthesized to elucidate the mechanisms of resource complementarity, spatiotemporal coordination, information interaction, and value coupling underlying dynamic system balance. The review identifies key bottlenecks, including the lack of a unified theoretical framework, limited heterogeneous data fusion, insufficient interpretability and real-time adaptability of AI models, weak integration of market mechanisms with operational control, and limited resilience under extreme disturbances. Building on these findings, an integrated research roadmap is proposed emphasizing source–grid–load–storage–cloud collaborative intelligence, AI-enabled digital twins, autonomous multi-agent coordination, cross-energy-system optimization, and resilient self-healing power systems. By establishing a unified knowledge architecture and clarifying the mechanisms linking coordinated regulation with dynamic system balance, this review provides a systematic foundation for advancing theoretical innovation, intelligent operation, and engineering implementation in future carbon-neutral power systems.
