Propagação de incertezas e análise de sensibilidade aplicadas à operação não ideal de aerogeradores

Abstract

One of the trends in the wind industry is the use of comprehensive and validated models for simulating wind turbines, which makes up one of the bases for the construction of digital twins. Starting from a consolidated and open source model for wind turbine simulation (OpenFAST), the effects of operation in non-ideal conditions on power and dynamic responses of a theoretical 5 MW wind turbine were evaluated, aiming to help in the definition of priorities for preventive action and risk quantification. Parameters representative of non-ideal operational situations were identified (rotor misalignment in relation to the wind, deviations in the pitch and twist angles of the blades and error in the blades’ masses) and probabilistic models were proposed for the associated uncertainties. Afterward, an analysis of the propagation of uncertainties for the responses of interest was carried out, and finally, through a global sensitivity analysis, it was identified which situations should be primarily treated. The results indicated that the average generation considering all non-idealities was up to 9% lower than the reference value, with the possibility of losses greater than 30% in extreme cases, depending on the frequency distribution of the wind. For the damage equivalent loads, the torque and bending at tower top, in addition to the bending moments at tower base, stood out as the most critical loads, with higher risks of failure over the wind turbines’ service life. Finally, the sensitivity analysis indicated that the variables with the greatest influence on generation and life of the wind turbines are the rotor misalignment with the wind and deviation in the pitch angle of the blades.