Vibration-based SHM of a centenary bridge: a comparative study between two different automated OMA techniques

Since Autumn 2011 a permanent dynamic monitoring system was installed by Politecnico di Milano in the San Michele bridge, a centenary iron-arch bridge which crosses the Adda river about 50 km far from Milan. During the first 13 months of monitoring, the automated identification of modal parameters from continuously recorded acceleration data has been performed using two different procedures: the first one is based on the Frequency Domain Decomposition technique, whereas the second one involves the Covariance-based Stochastic Subspace Identification algorithm. The paper summarizes the results obtained in the first year of monitoring, also with the aim of comparing the performance of the developed automatic procedures within the framework of vibration-based Structural Health Monitoring. Differences in terms of modal identification and frequency tracking are discussed and the correlation between natural frequencies and environmental/operational conditions was explored in order to distinguish between the normal response of the structure to changes in its environmental/operational conditions and possible damage or structural anomaly. Furthermore, a methodology of structural anomaly detection based on mode shape variation is discussed.