Probabilistic fatigue life prediction of metallic components based on continuum damage mechanics

Abstract

Based on continuum damage mechanics, a probabilistic method of predicting high-cycle fatigue life for metallic components is proposed. First, macro-meso two-scale stress-strain equations are established by combining Eshelby-Kroner localization law and classic elastic-plastic constitutive equation. Second, some parameters in Lemaitre's fatigue damage evolution model are randomized. Probabilistic properties of which are obtained by inverse-analysis and optimization. Then, the probabilistic method of predicting high-cycle fatigue life is established by coupling the model with macro-meso two-scale equations. And the proposed algorithm is coded with Fortran language, which is non-intrusive post-program of FEM analysis. Next, constant amplitude fatigue test of Al 2024-T3 coupon is performed to determine probabilistic properties of model parameters. Finally, as the object of study, wallboard is analyzed by FEM. Taking skin-strip component as research object, the fatigue life of hot spot is predicted by proposed method. The corresponding fatigue test of the component is performed, which verifies the effectiveness of proposed method to predict probabilistic life of metallic components.