Search

Damage progression in unidirectional (UD) composites in tension-tension fatigue is vital to understanding fatigue in more complex laminates and loading conditions. We therefore developed the UD fatigue model that considers fiber breaks, fiber-matrix debond growth, fiber fatigue as well as interactions of these damage modes. The contribution of each mode of damage in fatigue is assessed and those that dominate fatigue damage are identified. The model accounts for Weibull statistics of fiber strength and stress redistribution around fiber breaks. The predictions for fatigue loading indicate a significantly higher density of fiber breaks and clusters compared to quasi-static loading at the same stress level. The model predicts S–N curves in pure uniaxial tension-tension fatigue loading and does not consider stress concentrations caused near the grips; the predictions therefore tend to overpredict fatigue life compared to experimental S–N curves available in the literature. The model has potential to be developed for different loading conditions.

A key aspect of the longitudinal tensile failure of composites is the stress redistribution that occurs around broken fibres. Work on this topic has focussed mainly on the stress field surrounding a single broken fibre; however, this is an important limitation as unstable failure in carbon fibre bundles occurs when a cluster of about 16 or more broken fibres is formed. Therefore, we have developed a detailed Finite Element (FE) model to investigate how stress redistribution varies with the number of broken fibres in a cluster. The results show that both the recovery length and stress concentration factor increase significantly with increasing number of broken fibres in a cluster. We have also developed an analytical model, suitable to be included in existing or new fibre bundle models, that captures how the recovery length and stress concentration factor vary with the broken cluster size, and validated its predictions against our FE simulations. Finally, we extended our FE model to predict the survival probability of fibre bundles using Monte Carlo simulations, and found that these predictions were in good agreement with experimental and analytical results on microcomposites.