ABSTRACT
The prestressed concrete (PC) bridge is one of the most widely-used bridges. In recent years, their maintenance has become a significant problem in civil engineering. This study aims to investigate the applicability of a probabilistic-based model updating method to the structural health monitoring (SHM) of a PC box girder. The static loading test and vibration measurement were conducted on ta PC box girder. A finite element (FE) model was established to reproduce the experiment process. Although the tendency between simulation results and measured data was similar, the discrepancies between them were also observed. To this end, a probabilistic-based model updating method using structural modal properties was used to consider the uncertainties in the system to improve the performance of the simulation. The transitional Markov Chain Monte Carlo (TMCMC) algorithm is taken as a sampler to estimate the distribution of model parameters. The results showed that the discrepancy in the modal properties decreased after updating. The updated FE model was also applied in the simulation of the static loading test, and better fitness in the linear state was observed. However, responses in a nonlinear state did not match well.
