Crack growth based probability modeling of S-N response for high strength steel

Fatigue life prediction has historically been a challenging problem analytically as well as experimentally. Current analyses are predominantly statistical and do not adequately reflect long-term operating conditions. Recent observations of fatigue in the giga cycle regime have suggested that that there might be two distinct mechanisms for the nucleation and early growth of fatigue cracks; one associated with surface damage and the other with internal inclusions. Herein, a simple crack growth based probability model is used to examine other plausible contributors to the observed response, e.g., those associated with manufacturing and material properties, which are readily identifiable. A connection between the crack growth model and S-N response, and the impact of residual stresses induced by specimen preparation, and the distribution between external and internal nucleation sites are examined. These likely connections are demonstrated through analysis of an extensive set of S-N data on SUJ2 steel.