Bayesian Damage Detection of a Steel Plate Girder Bridge

摘要

This study proposed a novel Bayesian damage detectionmethod. The proposed Bayes factor quantifies changes inthe damage sensitive features, which is relevant to themodal properties. A damage experiment was conductedwith artificial damages imitating fatigue cracks at thegirder end. Anomaly due to the artificial fatigue crackswere successfully detected by the proposed method. Theseverity of the damage was also quantified through theBayes factors.