Crack growth evaluation based on the extended finite element and particle filter combined method

Abstract

In the issue of significant prediction deviations in fatigue crack predictions, the parameter uncertainties are usually neglected. To deal with this drawback, this paper proposes a crack growth evaluation method which takes parameter uncertainties into account to predict crack fatigue life. Firstly, the life model of crack growth is established through the combination of extended finite element method (XFEM) and Paris law. Then, the particle filter (PF) is used to reduce the uncertainty of material parameters through the actual monitoring data, and the material parameters in Paris law are estimated. Finally, the updated material parameters are put into the crack growth model for calculation, enabling a more accurate prediction of fatigue crack life. The feasibility of the proposed method is verified by two numerical examples, and the analysis results show that the proposed method can predict the fatigue crack growth life accurately.