ABSTRACT
This study is intended to discuss a vibration-based scour monitoring of railway bridges under inservice condition. A long-term vibration monitoring was conducted on a railway bridge to investigate feasibility of the vibration-based scour monitoring in real-time. A high-precision three-axis wireless accelerometer was installed on the pier top to measure pier vibrations. Vibrations of bridge girders adjacent the observation pier were also monitored in order to clarify influences of bridge girder vibrations to pier vibrations. Both train-induced and ambient vibrations were investigated, and frequencies relevant to the pier were identified and adopted as a scour-sensitive feature. A Bayesian operational modal analysis in frequency domain was also applied to reduce the prediction error. Observations through this study showed that the pattern of the identified frequencies varied according to the water level as well as wind speed during the monitoring.
