Modeling fatigue small-crack growth with confidence – A multistage approach

Abstract

Measurement of fatigue small-crack growth on the surface of smooth or notch specimens has evolved into a mature experimental technique. However, post data processing and model correlation for fatigue small-crack growth has not progressed in a parallel manner to the measurement technique. A new physics-based, mathematically precise fatigue small-crack growth modeling method is proposed for the specifics of surface small-crack growth in heterogeneous metals and rigorous statistical analysis. A multistage fatigue modeling philosophy was implemented and the fatigue damage incubation, small-crack growth, and long-crack growth regimes were identified. The uncertainty in measurement errors and the effects of random variation of microstructural features on the small-crack growth were explored using Monte Carlo (MC) simulations. The reliability of life prediction is directly correlated to the uncertainty-sources of measurements and the random nature of material microstructures through uncertainty propagation in the model correlation process. Fatigue damage incubation life is proven to be an essential quantity, both mathematically and physically, and it is introduced into the modeling process when correlating multiple experiments under the same loading condition. The proposed model is demonstrated using fatigue surface crack growth measurements of a dual-phase Ti-alloy.