Quantitative Uncertainty Analysis for a Mechanistic Multistage Fatigue Model

Xue, Y., Solanki, K.N., Steele, G., & Horstemeyer, M. (2007). Quantitative Uncertainty Analysis for a Mechanistic Multistage Fatigue Model . Proceedings of the 48th AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics, and Materials Conference, AIAA-2007-1982 . Honolulu, Hawaii.

Abstract

A quantitative approach is presented to evaluate the uncertainty of a microstructurebased multistage fatigue (MSF) model for fatigue life predictions and fatigue damage estimations. The microstructurally-based MSF model manifests the varying mechanisms of fatigue damage incubation, microstructurally and physically small crack growth, and long crack growth. The MSF model also incorporates the effects of microstructural features in fatigue damage progression: 1) the size and morphology of discontinuities, such as intermetallic particles and casting pores, 2) primary and secondary dendrites in cast alloys, and 3) grain size and texture in wrought alloys. The MSF model lays a foundation to evaluate the stochastic symbiosis of fatigue cracks with microstructural features that induce large scatter in the fatigue life. The quantitative uncertainty analysis demonstrates analytically the sensitivity of the MSF model parameters, random microstructural features to the fatigue life and uncertainty propagation of each parameter with respect to the fatigue life. The uncertainty magnification factor and uncertainty percentage contribution of each model parameter are quantified to provide understanding of each parameter’s contribution to the overall fatigue behavior. This paper demonstrates that a mechanistic multistage fatigue model, with multiple parameters related directly to microstructural features, induces less uncertainty in the results. The MSF model is a true high fidelity model that possesses great potential in structural prognosis in various applications.