Dynamic performance investigation of a long-span suspension bridge using a Bayesian approach

Abstract

In the past few decades, many large-scale bridges have been built to meet the different needs of styles and characteristics. Large bridges usually carry heavy traffic, and some bridges work in harsh environments. All these make the condition assessment of bridges very significant. On-site vibration testing is a general way to obtain the real response of the structure. Among a variety of measurements, the ambient vibration test has become popular due to its convenient usage and better economics. Based on the vibration data, modal identification can be used to identify the modal parameters that represent the dynamic characteristics of a structure. In this study, before being opened to traffic, a short-term non-destructive field vibration test was conducted on a newly-built long-span suspension bridge. Due to the limited number of sensors, multiple setups were designed to complete the entire measurement. For the measurement locations on the bridge deck, three-dimensional coordinates were determined using an electronic level. Analysis was performed based on the fast Bayesian FFT method, which can provide theoretical estimates for the posterior uncertainty and the most probable values of modal parameters. Nine modes of the bridge deck and four modes of the bridge tower were determined. The analysis results are discussed, and the posterior coefficient of variations (COVs) and sample COVs in different setups are compared. The analysis results in a healthy state can be used to verify the finite element model built in the design stage. It can also provide a benchmark for the future SHM of the bridge and improve the decision-making during usage and maintenance.