Effects of parameter estimation techniques and uncertainty on the selection of fatigue crack growth model. (June 2019)

• Record Type: Journal Article
• Authors: Chowdhury, S.; Deeb, M.; Zabel, V.
• Is Part Of: Structures. Volume 19(2019)
• Page Start: 128
• Page End: 142
• Publication Date: 2019-06
• Subjects: Fatigue -- Crack growth model -- Parameter estimation -- Uncertainty -- Life cycle assessment; Structural engineering -- Periodicals; 624.1
• DOI: 10.1016/j.istruc.2018.11.018
• Languages: English
• ISSNs: 2352-0124

Abstract

Abstract: This paper presents a stochastic framework for conducting lifetime fatigue performance of a structure by integrating several fatigue crack growth models (CGM). Several parameter estimation techniques have been considered to estimate the parameters associated with each model for the comparison and selection of the most suitable technique for any particular fatigue CGM. The estimated parameters are used to predict the loading exposure of the structure from a given damage state. Influence of different levels of uncertainty has been taken into account to portray the effects on model's prediction quality. The test data under constant amplitude loading is obtained from the NASGRO manual to verify the approach whereas lab-based study data, inspired from a real bridge damage scenario under variable amplitude loading, is used for the validation. The proposition suggests how to select the suitable fatigue CGM and its corresponding parameter estimation technique, considering the effects of uncertainty, for damage prognosis when there exists limited or no structural health monitoring. Highlights: An approach to identify appropriate crack growth model for fatigue life cycle assessment is proposed. The proposed approach finds the most suitable parameter estimation method for the considered crack growth models. The study shows that the crack growth models which follow a power law form have log-normally distributed error. The uncertainties in the data have significant influence on the reliability of life cycle assessment and therefore on the selection of crack growth model.