Analysis of Bridge Time-Dependent Performance Based on Dynamic Bayesian Networks

Current analysis of bridge time-dependent performance is mostly based on present test data or certain single approximate computational formulas. Thus, it is usually unable to combine prior knowledge of bridge time-dependent performances with on-site detection information effectively, and the evaluation model or parameters could not be updated on time. On the basis of the bridge characteristic during the time varying process, the Dynamic Bayesian Network (DBN), which could effectively assess bridge time-dependent performance, is proposed in this study. Two bridges, the performance of which is changed by internal and external environments, are taken as research objects. DBN is used to approximate the model of time-dependent performance of two bridges. Model updating is also realized by utilizing detection information. Results validate the feasibility and effectiveness of the proposed method.