As global demand for renewable energy increases, offshore wind power is gaining attention due to stable wind conditions and high output potential. Floating offshore wind turbines, suitable for deep waters such as those off Japan’s coast, offer flexibility in site selection. However, their design relies on complex numerical simulations, which are time-consuming and computationally expensive. To address this, we propose a machine learning-based surrogate model to predict turbine blade displacement using fluid forces obtained from wave and wind analysis. The surrogate model is trained on simulation data and can rapidly deliver accurate predictions, greatly reducing the computational burden. This approach enables efficient design exploration, real-time modification, and broader parameter studies. Additionally, the model may be applied beyond the design phase for condition monitoring and failure prediction. By incorporating data-driven methods, this study contributes to the practical deployment and maintenance optimization of floating offshore wind turbines in complex marine environments.
Endo et al. (Wed,) studied this question.