Statistical extrapolation methods for estimating extreme loads on operating wind turbine blades

Abstract

The wind turbine blade is designed against extreme load during power production with a return period of 50 years. The statistical extrapolation method is commonly used to estimate the extreme load on wind turbine blades and is recommended by the IEC standard 61400-1. However, various statistical extrapolation methods were used in previous studies and their accuracy and efficiency have not been evaluated comprehensively, which is studied in this research. The validation procedure is assessed by simulating the one-year load of the reference wind turbine. It is found the fitting before aggregation method with peak loads extracted by block maxima techniques and Lognormal fitting function has the best accuracy and robustness for estimating out-of-plane extreme load. The in-plane extreme load estimated by the peak over threshold (1.4) data extraction method combined with the Normal fitting function and the peak over threshold (2.0) data extraction method combined with the Lognormal fitting function have high precision. Forty samples for each wind speed are enough and the data below 9 m/s can be neglected for the optimal statistical extrapolation methods.