多能系统中协调能源管理的技术经济评估：地下储能与基于人工智能的海上调度的成本有效集成

Abstract Integrated energy systems (IES) hold significant potential for enhancing energy flexibility and decarbonization. However, current implementations often suffer from siloed operation and limited cross-sector coordination. In particular, traditional underground hydrogen storage (UHS) systems typically operate independently, reducing their effectiveness within interconnected energy networks, especially in coastal regions where geological storage and maritime infrastructure could be synergistically utilized. To address these limitations, this study proposes a hydrogen-centric IES framework that integrates multi-vector energy modeling, a hybrid UHS configuration, and an AI-driven maritime coordination strategy. These innovations enhance dynamic operation, energy conversion efficiency, and system-wide flexibility, especially in coastal regions where geological storage and port infrastructure converge. The framework introduces three core innovations: (1) a unified multi-vector modeling approach that integrates electrical, thermal, gas, and water networks with hybrid UHS for comprehensive sector-coupled analysis; (2) coordinated operation of electrolyzers, fuel cells, and geological storage within a hybrid UHS system to optimize hydrogen production and utilization; and (3) a maritime energy strategy in which hydrogen-powered vessels serve as mobile energy assets during port stays, contributing to local grid services via an AI-driven scheduling system based on XGBoost, which dynamically aligns vessel operations with fluctuations in renewable generation and energy demand. The framework is formulated as an MILP optimization model, with the AI-driven scheduling simultaneously optimizing IES-wide objectives while respecting maritime operational constraints. Simulation results demonstrate notable improvements, including an 11.81% increase in hydrogen system efficiency and a 67.9% rise in port-based revenue, while ensuring a continuous supply–demand balance across all integrated energy vectors. Comparative analysis validates the superiority of the AI-driven approach, and sensitivity analysis across key cross-sectoral parameters confirms the framework’s robustness. These findings highlight the framework’s potential as a scalable solution for achieving flexible, efficient, and economically optimized IES. It is particularly advantageous for coastal regions, capitalizing on the convergence of subsurface storage capabilities and maritime infrastructure to enable enhanced cross-sector integration and coordinated energy system operation.