Abstract
<jats:p>Abstract. Leading-edge erosion of wind turbine blades due to hydrometeor impacts is observed across various climate zones. This study focuses on uncertainties in the assessment of end of incubation (EoI) and erosion-safe operation (ESO) efficiency using different types of meteorological observations. At the Risø site in Denmark the following data are retrieved: a 19-year-long time series from a rain gauge and a cup anemometer, up to 3-year long time series from disdrometers and a sonic, and 10-year-long satellite-based product IMERG (Integrated Multi-satellitE Retrievals for Global Precipitation Measurement) and modeled wind speed from NEWA (New European Wind Atlas). The Spearman coefficient between joint precipitation and wind speed time series consistently takes positive values across the time series. Three reference turbines are assumed to have been installed at Risø. An impingement damage model derived from a rain erosion test is used to assess EoI and ESO efficiency. Key results on the uncertainty in assessing EoI are: 1) a long series is recommended, as EoI from 1-year time series varies by ~170 % between the minimum and maximum EoI. Uncertainty in EoI due to the choice of droplet slowdown is ~24 % when comparing an advanced model and the DNV Recommended Practice. Disdrometers estimate longer EoI than rain gauge due to under catch and missing data. The intuitive perception that similar EoI will occur across all blades of a turbine or within a wind farm exposed to the same weather is rarely observed in the field. Probabilistic variations in EoI are assessed using Monte Carlo simulations of the Pareto fronts for weather and materials. It is found that the uncertainty due to the material is larger than that due to the weather. The combined effect explains why the intuitive perception does not hold for EoI. ESO efficiencies are assessed from Pareto fronts. A flat Pareto front indicates high ESO efficiency, with only a small amount of AEP lost to ensure a long EoI extension. High-frequency disdrometer data (1-minute) with good measurement resolution (0.1 mm h−1) show flatter Pareto fronts than IMERG data (30-minute, 0.1 mm h−1). Resampling disdrometer data from 1-minute to 30-minute intervals results in steeper Pareto fronts, as expected, like those of IMERG. Rain gauge observations are inadequate due to limited measurement resolution (1.2 mm h−1). ESO efficiency is affected by the choice of droplet slowdown model, with the advanced model yielding flatter Pareto fronts. A recommendation is to assess the relevance of EoI from long-term time series and, in synergy with high-frequency observations, assess ESO efficiency for operational decision making.</jats:p>