A framework to perform quantification and reduction of uncertainties in a wind turbine numerical model using a global sensitivity analysis and a recursive Bayesian inference method is developed in this article. We explain how a prior probability distribution on the model parameters is transformed into a posterior probability distribution, by incorporating a physical model and real field noisy observations. Nevertheless, these approaches suffer from the so‐called curse of dimensionality. In order to reduce the dimension, Sobol' indices approach for global sensitivity analysis, in the context of wind turbine modeling, is presented. A major issue arising for such inverse problems is identifiability, that is, whether the observations are sufficient to unambiguously determine the input parameters that generated the observations. Global sensitivity analysis is also used in the context of identifiability.

In this paper, we propose a procedure for quantifying and reducing uncertainties impacting numerical simulations involved in the estimation of the fatigue of a wind turbine structure. The present study generalizes a previous work carried out by the authors proposing to quantify and to reduce uncertainties affecting the properties of a wind turbine model by combining a global sensitivity analysis and a recursive Bayesian filtering approach. We extend the procedure to include the uncertainties involved in the modeling of a synthetic wind field. Unlike the model properties having a static or slow time-variant behavior, the parameters related to the external sollicitation have a non-explicit dynamic behavior which must be taken into account during the recursive inference. A non-parametric data-driven approach to approximate the non-explicit dynamic of the inflow related parameters is used. More precisely, we focus on data assimilation methods combining a nearest neighbor or analog sampler with a stochastic filtering method such as the ensemble Kalman filter. This so-called data-driven data assimilation approach is evaluated on an industrial case of a wind turbine in operation using in situ measurements from an operating structure. The measured data are used by the method to recursively reduce the uncertainties that affect the parameters related to both model properties and wind field.