Presentations

Abstract

Wake interactions between offshore wind farms are a major source of energy‑production losses, and cluster wakes can persist for tens of kilometres. Although wake losses generally decline as wind speed rises, losses at wind speeds ≥ 10 m/s remain significant and model predictions diverge. High‑fidelity computational fluid dynamics (CFD) and mesoscale Weather Research and Forecasting (WRF) simulations, often yield contrasting trends, especially at wind speeds above the plateau of the thrust curve. To investigate these uncertainties, we have initiated a study of high‑wind‑speed wake behaviour using data from operational offshore wind farms in the North Sea. The datasets have been cleaned from curtailment events and filtered to isolate conditions between 10 and 15 m/s, by utilizing the average of nacelle anemometers. Numerical simulations are ongoing with a variety of models, including LongSim, the Eddy‑Viscosity model, TurboPark, WRF with the Fitch parameterisation and the Vollmer correction, a machine‑learning surrogate CFD.ML, and full CFD. Early comparisons suggest that aggregated cluster‑wake losses at 10-15 m/s from high‑fidelity CFD are about 4-6% of annual energy production, whereas WRF predicts up to 6-9% losses. A separate preliminary case using 15 MW turbines in a Dutch offshore cluster found that the turbine interaction loss factor was around 22% in CFD and 27% in WRF, with the larger contribution to this gap observed at high wind speed levels. These early findings indicate that high‑wind‑speed wakes remain significant; however, additional model runs and comparisons across the full 10-15 m/s range are planned to firm up these trends. The work will expand to more sites and refine the modelling approaches. These preliminary results, therefore, serve as a starting point for reducing uncertainty in resource assessments and informing design and operational strategies for large offshore clusters.