A Fatigue-Life Correction Model for Finite Element Analysis of Flexible Composite Structures Under Fatigue Loading
Abstract
Fatigue-life prediction of flexible composite structures using finite element analysis can be affected by discrepancies between conventional fatigue-life models obtained from standard specimen tests and their application at the finite-element integration-point level. In this study, a fatigue-life correction model (FLCM) is proposed to reduce this discrepancy by considering the evolution of the fatigue state during cyclic loading. The FLCM incorporates damage-induced stiffness degradation, local stress redistribution, and the resulting evolution of the effective stress level into the fatigue-life formulation, thereby providing a correction to the conventional fatigue-life relationship when it is applied in finite element fatigue analysis. An artificial-neural-network-based fatigue stress–damage mapping method (FSDM-ANN) is further developed to represent the relationship between stress level, normalized fatigue life, and fatigue stiffness degradation, and is integrated with the FLCM to establish a fatigue progressive damage model for finite element analysis. The proposed approach is implemented in ABAQUS through a user-defined material subroutine and evaluated using flexible composite structures subjected to combined tensile–bending loading. The experimental program includes two static tests and two fatigue tests, with the fatigue tests performed using staged displacement amplitudes. The finite element predictions are compared with the experimental results in terms of structural load–cycle response and normalized fatigue stiffness degradation. The normalized fatigue stiffness error is 2.93% for the FLCM, compared with 7.04% for the conventional fatigue-life model. The results indicate that the proposed fatigue-life correction formulation can reduce the fatigue-life prediction discrepancy associated with the direct application of conventional fatigue-life data in finite element analysis and can support fatigue prediction under varying effective stress levels within the calibrated material and structural conditions.