Investigation of Sensitivity Analysis of Crucial Dynamic Characteristics of Offshore Wind Turbine Utilizing Polynomial Chaos Expansion

Abstract The probabilistic-based design method is gaining attention in designing offshore wind turbines (OWTs) owing to its economic design. However, the numerous uncertain design variables and the tiny target annual failure probability make it unaffordable to achieve a reliable and economic design. Conducting sensitivity analysis (SA) is a common strategy to identify nonsignificant variables, thereby reducing the uncertainty space and accelerating the design process. To this end, this study aims to identify the factors significantly influencing the dynamic characteristics of OWTs from the structural geometrical, material properties, environmental, and soil uncertainties by conducting SA. Primarily, an improved method integrating the polynomial chaos expansion (PCE) strategy with the traditional Morris screening method was proposed to accelerate the SA process, and it was further validated by comparing the natural frequency results with those of the PCE-based Sobol method. A comprehensive SA study was then carried out to explore the dominant variables highly influencing the representative dynamic responses using the proposed method. The results indicated that the pile foundation bending stiffness has a remarkable effect on the natural frequency following Young's modulus E and tower thickness t, and wind loads serve as the most prominent factor influencing support structural dynamic responses. Furthermore, according to the observed remarkable influence of the pile foundation stiffness parameter on the structural stress, it proved the necessity of introducing the coupled numerical model in the SA for OWT.