Probabilistic Performance Assessment of Highway Bridges Using Operational Monitoring Data

A new probabilistic performance assessment procedure using operational monitoring data is proposed. In the procedure, multiple finite element models are identified from the weighted aggregation formulation for multi-objective optimization problem, each of which can represent uncertainty condition of operational monitoring data. By applying principal component analysis andK-means clustering on numerous candidate models, they are grouped according to their similarity and contributions in performance assessment. And then bridge’s performance can be assessed in probabilistic approach using the FE models in the classified groups, at which specific types of operational monitoring data uncertainty is represented. Yeondae Bridge, a steel-box girder highway bridge in Korea, is taken as an illustrative example. Load rating factor is evaluated as performance index of the bridge, and compared to rating factors by sophisticated model and baseline mode to verify effectiveness of proposed method.