Using limited roving sensors to monitor bridge subjected to random traffic load

The development of operational modal analysis (OMA) techniques has enabled the monitoring of large civil engineering structures, such as long-span bridges and high-rise buildings. However, implementing these techniques in real-world testing sites with limited resources remains a challenge. This study introduces a novel approach to obtain the modal characteristics of a box girder bridge in a more cost-effective and simplified manner, with experiments performed in a gap of more than 4 years (Years: 2018 and 2023). The study proposes two output-only frequency domain system identification techniques, namely roving reference normalized power spectrum and roving reference frequency domain decomposition, as part of the OMA-based approach. These techniques involve measuring responses from a pair of roving accelerometers. The collected vibration data records the concrete bridge's response to various external factors, including vehicular traffic on the carriageway, pedestrian movement, river flow, and wind. Despite weak environmental excitations and sensor noise, the findings suggest that accurate modal properties can still be extracted. The two proposed OMA approaches yielded five mode shapes and modal frequencies with closely matching results. When these experimental findings are compared with the numerical results, they exhibit a notable level of consistency. The paper identifies the dynamic characteristics of bridge structures from ambient vibration responses using only two accelerometers with two-point roving technique. This is especially important in real-world testing sites where data are inevitably noisy, unlike in laboratory environments. Additionally, it significantly reduces the cost of the bridge health monitoring.