Multi-scale fatigue model to predict stiffness degradation in short-fiber reinforced composites

The prediction of stiffness degradation in short-fiber reinforced composites under high-cycle fatigue loading is studied. To reduce the experimental characterization effort involved, we propose a computational multi-scale method to model the stiffness degradation under fatigue loading of fiber-reinforced polymer composites. First, we introduce a damage model for the thermoplastic matrix directly resolved in time scale. In contrast to formulations in cycle space, this enables us to predict damage evolution of the material accounting for vavrious stress ratios. For industrial-scale components, fully coupled FE²-simulations are currently infeasible with conventional hardware. Thus, we employ a model order reduction for the time scale model. For the high cycle regime, the computational costs for the prediction of an engineering component are still unfeasible using the time scale model. We thus use a cycle jump technique for sinusoidal loading. The reformulation of the model yields a cycle-dependent damage evolution with a driving effective stress dependent on stress ratio and stress amplitude.