Reliability analysis of offshore wind turbine support structures using Kriging models

This works presents an efficient framework for the reliability analysis of offshore wind turbine support structures based on a computational interface between an aero-hydro-servo-elastic simulator FAST and a finite element analysis software Abaqus (R). This framework is used to evaluate the safety levels achieved by following the largely used IEC 61400-3 standard. The existence of uncertainties such as geometric imperfections, material properties or environmental loads as well as the presence of nonlinearities introduced by the control systems may have a remarkable influence on the dynamic response of the structure, hence this computational link uses stochastic fully coupled simulations aiming to capture both the dynamic behavior of the structure and inherent uncertainties. However, when time-domain simulations and finite element model analyses are employed, each evaluation of the limit state function may imply a large computational cost making traditional methods such as Monte Carlo simulations infeasible. Therefore, the Kriging method is proposed to approximate the computational interface. First, several design points are generated through a random sampling technique, then using these points a model is constructed to replace the original computationally intensive interface. Finally, Monte Carlo simulations are used to perform the reliability analysis and to derive the reliability index. The analysis is applied to the NREL 5 MW prototype turbine model, mounted on a monopile with a rigid foundation.