Probabilistic fatigue resistance model for steel welded details under variable amplitude loading – Inference and uncertainty estimation

• Published in: International journal of fatigue 2020-06, Vol.135, p.105515-18, Article 105515
• Main Authors: Leonetti, Davide, Maljaars, Johan, Snijder, H.H.
• Format: Article
• Language: English
• Subjects: Bridge loading; Damage; Elastic limit; Fatigue life; Fatigue limit; Fatigue strength; Fracture mechanics; Life prediction; Likelihood; Linear elastic fracture mechanics; Materials fatigue; Metal fatigue; Prediction models; Probabilistic inference; Probability theory; Reliability analysis; Reliability engineering; S N diagrams; Statistical analysis; Structural reliability; Uncertainty; Variable amplitude loading; Welded details
• ISSN: 0142-1123; 1879-3452
• DOI: 10.1016/j.ijfatigue.2020.105515

•A novel model to infer VA fatigue data of welded joints is proposed.•The damage limit concept has been formalized and justified using fracture mechanics.•Test data of four welded structural details are inferred and uncertainty are determined. This paper presents a probabilistic fatigue life prediction model for welded details subjected to variable amplitude random loading, using the nominal stress approach. The proposed model makes use of CA S-N curves in combination with the relative damage rule which allows to globally account for load interaction effects, and the damage limit concept that permits the estimation of the damage for the stress ranges lower than the fatigue limit and is based on linear elastic fracture mechanics. Epistemic (sampling) and aleatory (statistical) uncertainties are estimated in a statistical framework, making the model applicable for structural reliability analyses.