SUMMARY

Traditional fatigue analysis schemes used for geometries with stress gradient fields (such as notches) have required many experiments to determine the probability of fatigue failure. Due to the extensive amount of scatter in high cycle fatigue (HCF) analysis, typically fatigue life data is fitted from a significant amount of experiments into an assumed distribution, such as a Weibull or lognormal distribution. The resulting notch effect on fatigue life is characterized via a notch root fatigue strength reduction factor, often alternatively called the fatigue notch factor, Kf. The experimental results are beneficial for life prediction of a given geometry and microstructure, but do not offer predictive insight into the underlying physical mechanisms that explain scatter, size effects, and gradient effects on fatigue damage. Moreover, if the material is changed, the fatigue notch factor changes for a given geometry. Computational crystal plasticity models that characterize sensitivity to microstructure variability, notch size and gradient effects, and extrinsic defects (inclusions, FOD) can be used to assist in characterizing these fatigue mechanisms and to provide insight into materials design and selection for a given application. This research develops approaches that combine computational crystal plasticity with nonlocal notch root plasticity and damage approaches for small crack formation in HCF, LCF, and mixed conditions, with applications to aircraft gas turbine engine materials. These approaches combine elements of crystal plasticity with new probabilistic methods for notch sensitivity based on slip distributions in the microstructure at the notch root.