Ensemble learning for remaining fatigue life prediction of structures with stochastic parameters: A data-driven approach

Abstract An effective approach is proposed to predict the remaining fatigue life (RFL) of structures with stochastic parameters. The extended finite element method (XFEM) was firstly used to produce a large amount of datasets associated with structural responses and RFL. Then, a RFL prediction model was developed using the ensemble learning algorithm, which employed multiple machine-learning algorithms to learn useful degradation patterns of the structures from the XFEM datasets. Several numerical examples were investigated to evaluate the performance of proposed RFL prediction approach. The analysis results demonstrate that the ensemble learning is able to effectively predict the RFL of the structures with stochastic parameters.