Dynamic real-time reliability prediction of bridge structures based on Copula-BHDLM and measured stress data

This study presents a novel dynamic real-time reliability prediction method by considering the time-varying correlation among the data at multiple measurement points. The Bayesian Hilbert dynamic linear model (BHDLM) is developed to predict the stress response, and the dynamic real-time reliability is then predicted by using the Pair-Copula theory. To illustrate the feasibility of the proposed method, the dynamic real-time reliability of a steel-concrete composite girder bridge based on the simulated stress data due to a single vehicle with various weights and random traffic flow load is predicted. Finally, the dynamic reliability of a real bridge based on the monitoring stress data is predicted. The results show that the proposed BHDLM is effective for predicting coupled stresses, and the dynamic real-time reliability can be predicted. In addition, the predicted dynamic reliability of the real bridge demonstrates that the influence of the data correlation should be considered.