Abstract: S-N curve in gigacycle regime is obtained from rotary bending fatigue tests by using 35CrMo railway axle steel and the crack initiation site is observed. The results show that the specimens have conventional fatigue limit and the crack initiating from inclusion is not observed. It has been reported that inclusions cannot cause failure in small specimens made of axle steel while the full-scale components fracture from inclusions. This difference is caused by that the bending stress gradient in the small specimens is much larger than that in the full-scale railway axle; meanwhile large inclusion is more likely to occur. Ultrasonic fatigue tests are used to investigate the maximum inclusion in the specimen and the statistic maximum values method is used to forecast the maximum inclusion in a small specimen and a real axle. Based on the difference of maximum inclusion sizes between a small specimen and a real axle, the effects of inclusion on fatigue strength are analyzed. The results provide a theoretical basis for both alloy axles development and strength design.

Key words: Fatigue, Inclusion, Railway axle steel, S-N curve, Statistics of extreme values