A major limitation of current fatigue life prediction methods for polymer matrix composite laminates is that they rely on empirical S-N data. In contrast to fatigue life prediction methods for metals which are based on physical crack growth models, the heart of fatigue life models for composites is empirical S-N data for each specific material system and specific loading conditions. This implies that the physical nature and processes responsible for tensile fatigue are not well understood.

In this work a mechanism-based approach is used to model the damage growth and failure of uniaxial polymer matrix composites under uni-axial tension-tension fatigue loading. The model consists of three parts: an initial damage model, a damage growth model, and a tensile failure model. The damage growth portion of the model is based on fracture mechanics at the fiber/matrix level. The tensile failure model is based on a chain of bundles failure theory originally proposed for predicting the static strength of unidirectional laminates using fiber strength distributions.

The tensile fatigue life prediction model developed in this work uses static tensile strength data and basic material properties to calculate the strength degradation due to fiber-matrix damage growth caused by fatigue loading and does not use any experimental S-N data. The output of the model is the probability of failure under tensile fatigue loading for a specified peak load level. Experimental data is used to validate and refine the model and good correlation between the model and experimental data has been shown.

The principal contribution of this work is a hybrid-mechanistic model for analyzing and predicting the tension-tension fatigue life behavior of uniaxial polymer matrix composites. This model represents the very foundation to build upon a comprehensive model for fatigue. It demonstrates the validity of the ideas as they apply to uniaxial laminates that may in turn be used to apply to more complex laminates. Additionally, because the model is mechanism based it can be used for evaluation of the effects of constituent property changes such as matrix stiffness and toughness, or environmental conditions such as temperature and moisture.