Dynamic characterization of a curved nine-spans pre-stressed concrete box girder bridge with half-joints

Abstract

In recent years, the failure of many bridges has pressed governments all over the world to increase investments and to enact new monitoring regulations. In the Italian context, the Ministry of Infrastructure and Transport answered to that challenge by enacting the new Guidelines on Risk Classification and Management, Safety Assessment and Monitoring of Existing Bridges, in 2022. In this new regulation, Structural Health Monitoring (SHM) plays a fundamental role, complementing traditional periodic inspections by carrying out low intrusive non-destructive analyses. In this field, Operational Modal Analysis (OMA) is a vibration-based method consisting in the processing of the ambient accelerations recorded on the structure in order to extract and track its modal features. The final goal of continuous OMA is the timely identification of possible damage-induced persistent variations in the modal features of the monitored structure. This is why one of the most recent advances in OMA algorithms is related to the development of continuous Automated OMA embedded in high-potential software, to enable the infrastructure management authorities to realize large-scale monitoring of bridge networks. Many road bridges are however long multi-span bridges which require numerous sensors for their monitoring, thus making automated OMA especially challenging from the computational point of view. The aim of the work is to contribute to a more computationally affordable automated OMA for long bridges, by proposing to subdivide the OMA task considering suitable selections of subgroups of sensors. The proposed approach is illustrated in a case study consisting of a curved nine-spans pre-stressed concrete box girder bridge with 10 half-joints and equipped with more than 50 sensors.