The role of parameter uncertainties in the damage prediction of the alongwind-induced fatigue

Literature dealing with the calculation of fatigue damage under wind loads is rich in procedures in the time domain or in the frequency probabilistic domain, whose parameters are usually assigned according to deterministic values. Starting from an advanced model of alongwind-induced fatigue damage and using a probabilistic assessment of the parameters involved, this study discusses and propagates the uncertainties inherent in the input parameters over the fatigue damage prediction. The use of Taylor series is particularly simple and expressive. It shows that errors in parameter estimates give rise to very large scatter in the fatigue damage assessment, pointing out a set of quantities whose role is crucial for a reliable estimate. Because of large uncertainties and strong non-linearity of the problem, the actual distribution of the fatigue damage is then obtained by Monte Carlo simulations. The results are discussed and compared with conventional verifications, based on deterministic evaluations with standard safety coefficients.