Developing an atlas of rain-induced leading edge erosion for wind turbine blades in the Dutch North Sea

To support the ongoing development of offshore wind energy in the Netherlands and to maintain current assets, it is essential to provide wind farm operators with accurate estimates of wind turbine blade erosion. Unfortunately, there is currently a shortage of information on wind turbine erosion risk, especially in offshore regions. In this work, we developed an atlas detailing rain-induced leading edge erosion for wind turbine blades in the Dutch North Sea, using weather simulations spanning a decade to capture long-term climate patterns. These simulations, based on a meso-scale model, were incorporated into a fatigue-based damage model, linking weather conditions to blades’ leading edge erosion. The results reveal that the erosive impact of rainfall on wind turbine blades varies across the Dutch North Sea. The estimated average incubation period, which indicates the leading edge protection system's lifespan, ranges from 8 to 9 years in the southwestern region, decreasing to 6 to 7 years in the northeastern area. This is due to both the higher average wind speeds and greater rainfall amounts occurring in the northeastern locations compared to the southwestern ones. This paper emphasizes that the northeastern regions of the Dutch North Sea, which are being examined for potential wind farm developments post-2030, will encounter higher erosion risks compared to those currently operating in southern locations, possibly requiring enhanced mitigation strategies. Additionally, a year-long comparison of meso-scale simulations, high-resolution large eddy simulations (LESs), and measurements revealed that meso-scale simulations estimate 7 %–20 % less damage than LES, which captures more extreme events. Nonetheless, meso-scale simulations and LESs reveal alignment in the spatial patterns of erosion-related parameters, confirming that meso-scale simulations produce satisfactory atlases where regional differences are consistently captured with LES. Through the comparison between LES and measurements, it was found that LES estimates 23 %–66 % less damage than actual weather data, due to underestimating larger droplets and recording fewer extreme events.