A reliability analysis strategy for main shaft of wind turbine using importance sampling and Kriging model

Abstract

PurposeWhen dealing with simple functional functions, traditional reliability calculation methods, such as the linear second-order moment and quadratic second ordered moment, Monte Carlo simulation method, are powerful. However, when the functional function of the structure shows strong nonlinearity or even implicit, traditional methods often fail to meet the actual needs of engineering in terms of calculation accuracy or efficiency.Design/methodology/approachTo improve the reliability analysis efficiency and calculation accuracy of complex structures, the reliability analysis methods based on parametric and semi-parametric models are analyzed.FindingsThis paper proposes a reliability method that combines the Kriging model and the importance sampling method to improve the calculation efficiency of traditional reliability analysis methods.Originality/valueThis method uses an active learning function and introduces an importance sampling method to screen sample points and shift the center of gravity, thereby reducing the sample size and the amount of calculation.