Gaussian process regression for fatigue reliability analysis of offshore wind turbines

Impact of climate-change scenarios on offshore wind turbine structural performance

“…In other words, surrogate models mimic the behaviour of the simulation model as closely as possible while remaining computationally cheaper to evaluate. Various surrogate modelling approaches have been recently proposed in the literature for OWTs on monopile foundations, including Gaussian Process (GP) regression (8)(9)(10)(11) and for onshore wind turbines using Polynomial Chaos Expansion (12,13). These models significantly improve the computational efficiency of FLS assessment, and therefore also enable structural reliability assessment and the development of probabilistic risk models for OWTs to predict the potential financial impact of OWT failures.…”

“…The limit state depends on two random variables defined by the probability distributions shown in Table 1, although it should be noted that a larger set of random variables can be considered to capture other modelling uncertainties e.g., (11). Failure of the structure/structural component is assumed to occur when the limit state equation assumes negative values, and the probability of failure can be computed as Pr (𝐺 ≤ 0).…”

“…In fact, OWT structures are normally designed to obtain very low probabilities of failure, which implies that the limit state equation must be solved several thousand times, resulting in a prohibitively large computational effort. Full details on the proposed simulation-based structural reliability assessment can found in (11). An accurate technique for problems with a low number of random variables is importance sampling (35) and is used in this study, as further described in Section 5.3.…”

“…( 4), lifetime fatigue damage prediction is a problem of mean estimation: a random sampling scheme is expected to perform well in this context (36). This study follows the recommendations in Wilkie and Galasso (11) and uses 500 samples and 20 seeds.…”

Reliability Modeling Using an Adaptive Neuro-Fuzzy Inference System: Gas Turbine Application

“…This helps to underline the importance of integrating uncertainty and ensuring constant levels of reliability leading to cost reductions. And Wilkie and Galasso [5] have proposed a fatigue reliability analysis of offshore wind turbines, using Gaussian process regression, for the assessment of fatigue damage over the expected design lifetime of this system. This improves the sensitivity of various goodness-of-fit measures of the model and further reduces the computational effort required to perform regressions / predictions.…”