A probabilistic model for low-cycle fatigue crack initiation under variable load cycles

Abstract

A probabilistic low-cycle fatigue (LCF) model, ProbLCF-VA, is proposed in this work to describe the scatter of the LCF crack initiation life for homogeneous materials subjected to variable or random load cycles. This model can be implemented either to estimate the probabilistic cumulative damage for a given load history consisting of variable cyclic loads, or to predict the probabilistic LCF crack initiation if distributional estimates of a random load spectrum are available. More importantly, the proposed model is capable of describing certain prevailing effects such as load interaction and small cycle contributions, which are very commonly observed in fatigue tests with variable amplitude cyclic loads but can hardly be described by other models with physical and mathematical foundations. Model predictions are illustrated with multiple examples and compared to those from classical cumulative damage rules. The result shows a good qualitative match between the model prediction and test data in the literature on load interaction effects.