Time-Dependent Reliability Analysis of Bridge Piers for Cross-Sea Bridges Based on Dynamic Bayesian Networks

Abstract

To accurately assess the time-dependent reliability of bridge piers in marine environments, this paper proposes a time-dependent reliability evaluation method for bridge piers based on a Dynamic Bayesian Network (DBN). By establishing a resistance degradation model under the combined effects of reinforcement corrosion and concrete deterioration and embedding it into the DBN framework, dynamic prediction and updating of pier time-dependent reliability are achieved. A case study of a twin-column pier was conducted for verification, showing that the computational results of the proposed DBN model align well with the first-order reliability method (FORM), validating its accuracy and feasibility in time-dependent reliability prediction. Further, using the most severe service condition, that is, the tidal-spray zone as an example, the DBN prediction results were updated with inspection data to achieve dynamic assessment of the actual pier lifespan. Additionally, a comparative analysis of environmental zones revealed that the tidal-spray zone exhibits the fastest reliability degradation, followed by the atmospheric zone, while the submerged zone shows the slowest.