Offshore wind parks are central to the global energy transition, yet efficiently utilizing resources remains challenging. This review synthesizes data-driven approaches for improving offshore wind system performance, covering natural resources (e.g., wind and meteorological conditions) and artificial resources (e.g., turbines and supporting infrastructure). Three recurring data challenges—data imbalance, distribution shift, and data sparsity—are identified as key constraints in wind power forecasting, turbine layout optimization, and maintenance planning. For natural resource management, machine learning and deep learning models enhance wind and power forecasting under highly variable offshore conditions, while surrogate modeling and spatial optimization reduce wake losses and layout costs. Across representative studies, hourly and daily capacity-factor prediction errors are typically reduced by around 10% and 15–20%, respectively. For artificial resources, data-driven lifecycle management and preventive maintenance increasingly rely on SCADA data, sensor networks, and digital twins to assess equipment health and mitigate operational risks, although limited failure data remains a major bottleneck. Existing evidence suggests maintenance cost reductions of approximately 20–30%. Beyond operational efficiency, this review examines the integration of sustainability and circular-economy principles into offshore wind development, including material recycling, component lifetime extension, and wind-to-hydrogen integration. Overall, effective data-driven optimization depends on advanced algorithms, high-quality data, and system integration. By linking data challenges with circular-economy objectives, this study proposes a unifying framework to support intelligent and sustainable offshore wind systems.
Lyu et al. (2026) studied this question.