Vehicle and Truss Bridge Interaction Adopting a Simplified 2D Model

Abstract

Indirect structural health monitoring of bridges using vehicle-mounted sensors offers a promising approach for the continuous assessment of structures. Full-scale modelling of complex structures, however, is computationally intensive. In order to optimize the required computational expenses, a two-dimensional simplified numerical model whose response resembles the real structure is developed. Euler-Bernoulli frame elements are adopted for finite element structural modelling. The transitional Markov Chain Monte Carlo technique is then applied for Bayesian finite element model updating. The proposed approach proved to reach an accurate model for a real-world truss bridge in Japan. Vehicle bridge interaction elements are then developed based on the updated simplified 2D model. The generalized-α time integration scheme is used to overcome inherent numerical instabilities. Time-response of the vehicle is evaluated using the developed Vehicle-Bridge Interaction (VBI) framework to study the dynamic properties of the bridge. The proposed approach significantly reduces the computational efforts in extracting time history responses of the moving dynamic system over the bridge without sacrificing the prediction accuracy. Time history responses can be processed subsequently to identify damage(s) in the structure, if any.