Fatigue crack-growth predictions for a railway axle under material data variability

This document reports a probabilistic approach for residual fatigue life predictions of a cracked railway axle with allowance for variability of the input data. A NASGRO formulation was used for such a purpose, with its main parameters considered as uncorrelated random variables varying in given ranges. A load spectrum was considered in order to take into account of the actual loading conditions that these components undergo. The presented approach aimed to perform residual fatigue life predictions in terms of both mean value and dispersion around it. The results were also used to drive deterministic approaches in calculating conservative crack-growth assessments.