ABSTRACT The design of wind turbine gearboxes presents increasing challenges in balancing efficiency, reliability, and cost as turbines trend toward larger scales. Despite their critical role in power transmission, current modular design methodologies suffer from a lack of systematic mathematical modeling and inadequate leveraging of existing product data. Consequently, development cycles are protracted—often lasting 1–3 months—due to iterative calculations and cross‐disciplinary coordination, which can lead to project delays and client dissatisfaction. While prevailing optimization efforts, such as those based on the nondominated sorting genetic algorithm II (NSGA‐II) or on MATLAB‐based volume minimization, primarily address localized improvements (e.g., in gear profiles), they fall short of a holistic solution. To address these limitations, this paper proposes a novel modular classification framework that seamlessly integrates component functionality with digital modeling. This approach establishes rigorous mathematical–geometric relationships among components and enhances product databases to enable rapid data fitting between legacy and new designs. As a result, it facilitates immediate feasibility assessments during preliminary client consultations, significantly accelerates development timelines, and maximizes modular reuse rates, thereby optimizing both design efficiency and manufacturing resource utilization.
ABSTRACT Short‐term wind power forecasting is essential for reliable grid operation, yet newly commissioned wind farms often provide only limited supervisory control and data acquisition (SCADA) histories for model adaptation. Existing cross‐farm methods face two related limitations: Data‐driven models must relearn site‐specific interturbine dependencies, while probabilistic forecasts can become poorly calibrated under layout and operating‐regime shifts. This study proposes WakeMeta, a physics‐guided graph meta‐adaptation framework. WakeMeta converts an established Bastankhah–Porté–Agel Gaussian wake relation into a directed, training‐independent graph operator determined by turbine geometry, wind direction, and fixed engineering wake parameters. The graph constrains a spatiotemporal encoder with dilated temporal convolutions and graph attention, while a probabilistic head combines Gaussian likelihood, continuous ranked probability score (CRPS), coverage calibration, and head‐level Monte Carlo (MC) dropout. A first‐order meta‐learning procedure is applied only to data‐driven parameters, and target adaptation uses validation‐gated early stopping in which the unadapted initialization is a legal candidate. Under the frozen evaluation protocol (V3‐final) with five paired nonoverlapping target windows, 4‐h forecasts, and 14 days of target data, WakeMeta achieves CRPS and calibration error (CE) values of 0.1117/0.0386 on Penmanshiel and 0.0640/0.0742 on La Haute Borne. The corresponding CRPS is 5.0% lower than seasonal autoregression on Penmanshiel and 14.1% lower than persistence on La Haute Borne, while direct source pretraining with fine‐tuning remains strongest on La Haute Borne. Robustness experiments show that backbone, predictive family, graph, and adaptation effects are target dependent, supporting a deliberately scoped interpretation centered on portable physical structure and calibrated probabilistic transfer.
Offshore wind has become a key focus for achieving carbon neutrality. However, offshore wind turbines face considerable operational and maintenance challenges due to harsh conditions and difficult access. Integrating digital twin technology helps support operation and maintenance activities. This study conducts a systematic literature review of recent wind energy digital twin research. It aims to identify the underlying modelling trend of wind energy digital twins, assess their implementation maturity and their compliance with the adopted digital twin definition, classify the functions supporting offshore wind operation and maintenance and examine the integration of emerging enabling technologies. The findings reveal a scale-dependent redistribution of modelling approaches: physics-based approaches dominant at the component level, to data-driven approaches prominent at the turbine level, and both data-driven and hybrid approaches at the wind farm level. A functional analysis further shows that predictive maintenance and real-time monitoring dominate digital twin applications at the component level, whereas wind farm-level digital twin increasingly demonstrate integrated capabilities that combine multiple functions. Meanwhile, enabling technologies, including the Internet of Things, cloud computing and machine learning, are increasingly being integrated into digital twin systems, establishing a foundation for more intelligent and autonomous operation and maintenance strategies. The findings also reveal that most existing digital twin research remain at the pre-implementation stage, and only a limited number satisfy the requirements of a full digital twin. Therefore, significant gaps remain between research and operational implementation, particularly regarding effective feedback mechanisms from virtual models to physical assets.
ABSTRACT Australia's offshore wind ambition presents a transformative opportunity to strengthen supply chains, expand port infrastructure and build capability in critical components and installation assets. This study introduces a tailored supply chain demand modelling framework, applied to the Gippsland area of the Bass Strait region, to highlight pathways for enhancing readiness across turbine components, ports, vessels and workforce. Using three development scenarios aligned with government targets and licensed projects, this paper quantifies supply chain requirements from 2029 through to 2040 and maps opportunities for strategic investment across infrastructure and components. Findings demonstrate clear potential for growth: Domestic capability in towers and foundations can be developed, port resources expanded and vessel acquisition and chartering pursued to meet demand. The workforce pipeline offers a strong foundation for targeted training and skills development. A coordinated strategy will unlock these opportunities, combining short‐term mobilisation of existing ports, targeted domestic manufacturing incentives, international vessel acquisition strategies and investment in workforce training. The modelling framework provides a replicable and transparent tool to guide policy, investment and research planning, supporting Australia's transition to a globally competitive offshore wind industry.
ABSTRACT The ability of vertical axis wind turbines (VAWTs) to self‐start is critical to their performance. There is disagreement in the literature over some of the factors affecting their self‐starting ability and there is very limited information on how the starting azimuthal angle affects self‐starting. We present a series of wind tunnel experiments in which we measure the static torque and self‐starting of a three‐blade VAWT for a range of initial azimuthal angles, which we find to have a very strong effect on the ability to self‐start. The blade pitch angle has a complex and non‐monotonic effect on the static torque and the probability of self‐starting, with self‐starting more likely when the static torque coefficient is large and uniform across all azimuthal angles. Increasing the freestream turbulence from 0.34% to 1.32% tends to induce peaks in the static torque coefficient at some angles, but leads to a relatively small improvement in self‐starting, while the addition of leading‐edge tubercles significantly improves self‐starting, ensuring the turbine self‐starts from almost any angle. Based on these insights, we introduce a quantitative criterion using the integral of the static torque curve to predict whether the turbine will self‐start for a given angle. Using an empirical threshold, this method can correctly classify 89% of the 189 cases presented in this study.
