Novel probabilistic approach to assessing barge-bridge collision damage based on vibration measurements through transitional Markov chain Monte Carlo sampling

Barge-bridge collision has been considered a major contributor to bridge damage in the United States. Most barge-bridge collisions usually cause partial damage of bridges that may be invisible but threaten the safe operation of the bridge. After each collision accident, the bridge and the navigation waterway are usually closed for inspection and assessment of the impact of the collision on the bridge structural integrity. This can lead to significant economic losses due to substantial traffic delay or detour. Quick and reliable assessment of bridge post-collision condition can minimize those economic losses. This paper presents a novel perspective on the bridge post-collision condition assessment based on Bayesian probabilistic framework, which is aimed to promptly identify collision damage and rigorously address associated uncertainties using real-time vibration measurements. The presented approach is the first attempt to incorporate the bridge finite element model into an advanced statistical sampling algorithm of the transitional Markov chain Monte Carlo to draw samples, whose statistical distributions can approximate the updated probability distributions of extents and locations of the barge-bridge collision damage for decision making. The applicability and effectiveness of the proposed approach are illustrated using a simulation example of a prototype bridge. Simulation results indicate that the proposed approach has potential capacity for determining the extent and location of barge-bridge collision damage and their probabilistic characteristics. Finally, the limitations of this study and future research need for practical application of the proposed probabilistic framework are discussed.