Probabilistic considerations for growth and detection of cracks from rivet holes in a lap joint

In this paper, probabilistic considerations are introduced in a model for fatigue life prediction for a riveted lap joint using Paris' law. Initial crack sizes are distributed according to a truncated lognormal distribution, which is chosen to avoid known complications due to Paris' law. The stress intensity factor for a single rivet hole is calculated, and is generalized to a lap joint. The probability of the existence of a crack in two domains of interest are evaluated, and the effect of a single inspection, modeled using the Probability of Detection, is studied. Additionally, the probability of detection concept is extended by linking it to applied stress and number of elapsed cycles using Bayes' theorem and ramifications are explored.