Paradox of Shorter Residual Fatigue Life due to Omission of Low-Amplitude Cycles and Its Significance for Testing

The work investigates nonvalidity of the common presumption that the nondamaging cycles do not influence residual fatigue life. Paradoxically, application of the full loading spectrum (more cycles) resulted in approximately 2.3 times longer life of the fatigue crack growth specimens than application of the spectrum with 33% of the smallest amplitudes omitted. Unlike in humid air (controlled relative humidity of 50% at 23°C), the effect disappeared in a dry-air chamber (relative humidity <10% at 23°C), where both fatigue lives were short. The mechanism responsible for these effects was identified as the oxide-induced crack closure, an extrinsic mechanism unrelated to material damage. Oxide debris developed on fracture surfaces was observed by light and scanning electron microscopy, whereas crack closure was measured during the experiments. The presented counterintuitive behavior in humid air may result in wrong assessment or prediction of components residual fatigue lives, which can be nonconservative in some scenarios.