Development of a fast numerical tool for wind flow predictions around offshore wind turbines

Purpose This study aims to develop a fast numerical tool for detailed wind flow predictions around offshore wind turbines. The tool is designed to assist in optimizing Unmanned Aerial Vehicle (UAV) flight paths for maintenance operations by providing high-resolution wind and turbulence data in real time. Design/methodology/approach A Computational Fluid Dynamics (CFD)-generated database is built using a validated Reynolds-Averaged Navier–Stokes model under different wind conditions. Proper Orthogonal Decomposition, namely High Order Singular Value Decomposition, is applied to create a surrogate model. When coupled with site-specific meteorological forecasts, it produces high-resolution wind flow and turbulence maps (spatial resolution of less than 1 m) in less than 1.5 s. These predictions are used within a UAV simulator to assess flight behaviour under realistic turbulence conditions. Findings This numerical tool speeds up wind predictions by a factor of 2,400 compared to direct CFD calculations while maintaining mean and maximum relative deviations for velocity and turbulence kinetic energy under 2 and 10%, respectively. Its integration with UAV flight assessment tools helps to identify critical regions that may compromise UAV stability, improving operational safety. Originality/value This tool enables real-time wind predictions (using meteorological data as inputs) and UAV flight analysis, improving UAV-based maintenance operations in offshore wind farms. Its computational efficiency allows real-time use on any computer, supporting pre-flight risk assessment and safe UAV trajectory planning. The integration of the tool into UAV simulators provides a novel approach to enhance the reliability of UAV flights in extreme marine conditions.