A study on the propagation of aero and wind uncertainties and their effect on the dynamic loads of a wind turbine

This work is concerned with the quantification of uncertainties associated with wind turbines. A part form the understanding of the Effects of uncertainties per se, the efficient propagation of uncertainties is necessary for the implementation of robust design optimization methods, which is one of our future goals. Among all possible sources of uncertainties, here uncertainties related to the incoming wind and to the aerodynamic characteristics of the blades are propagated throughout a high-fidelity multibody aeroservoelastic tool. Different techniques, which could all be used for propagating uncertainties other than the ones considered here, are tested and compared. These include different formulations from the family of Non-Intrusive Polynomial Chaos Expansion, as well as Ordinary and Universal Kriging. By running a reduced subset of standard design load cases, a comparison among the various methods is drawn in terms of accuracy and computational efficiency with respect to a standard Monte Carlo approach. It is concluded that, for the uncertain- ties considered here, all approaches lead to a significantly higher performance compared to Monte Carlo, with Universal Kriging slightly standing out. It is also observed that the output parameters exhibit significant variations, and this highlights the importance of a comprehensive framework for the quantification and propagation of uncertainties in wind energy systems.