Fatigue cracks are observed to form at microstructural features and inhomogeneities. Constituent particles were the dominant fatigue crack nucleation site in thin sheet aluminum 2024-T3. The distribution of particles sizes has been used previously to predict the fatigue lives and scatter of notched specimens at a single stress level. In this work, the influence of stress level was investigated and a total life model to predict fatigue life and its variability was developed. The distributions of nucleation lives and crack nucleating particle sizes, obtained from experimental tests on notched aluminum 2024-T3 specimens, exhibited a significant dependence on stress level.

The measured distributions of crack nucleating particle sizes for the three stress levels indicated that the use of the traditional threshold produces unconservative results. Accordingly, a probability of crack nucleation (POCN) concept was developed which modeled the experimental observations more accurately than the threshold or area based approaches. The effect of stress level on the POCN function was quantified and the ability of the POCN transformation to predict the distribution of crack nucleating particles for other conditions was shown.

The measured distributions provided the foundation for the total fatigue life model, which used a probabilistic Monte Carlo method in conjunction with Newman’s FASTRAN II crack closure model. By accounting for the mechanisms of nucleation and propagation, the total life model closely predicted the cumulative distribution function (CDF) of fatigue lives for the three stress levels examined. Life predictions made using the crack nucleating particle width and length distributions resulted in better and slightly conservative predictions than those made based on the area distribution. In addition to predicting the CDF of fatigue lives, the model was used to develop S-N and probability of failure curves. The total life model and the POCN concept were also applied effectively to other materials (aluminum 7075-T6 and 7050-T7451) and types of defects (microporosity). The total life model can be a valuable tool in the design of components against fatigue and can lead to reductions in the amount of time and experimental testing required to validate components.