Anomaly Detection Through Long-Term SHM: Some Interesting Cases on Bridges

The growing demand for health assessment of civil infrastructure has allowed for a dense control, among others, of many bridges. The new industrial design of monitoring systems, based on cheaper hardware, at the same time granting satisfactory performances, is making available huge data flows, allowing researchers not just to create a wide database but also to create logbooks of possible anomalies and defects, which can be both related to the infrastructure and the monitoring system. When prompt alarms are to be sent to the bridge manager, to avoid failures, there is the risk to produce false positives, due to specific data features or ambiguous recorded structural behaviors that may appear as damage, therefore requiring a deeper knowledge level of the expected structural response, to be fully understood and classified as not dangerous in terms of risk for the human life (structural reliability).Only the combination of a continuous monitoring with the availability of an updated digital twin and a meticulous data interpretation of an expert engineer in the monitoring field can lead to a thorough understanding of the structural behavior and a proper asset management. In the present paper, an apparent damage identified by monitoring the dynamic response of a reinforced concrete bridge, under operational conditions, has proven to be a change of the static scheme of the structure, with a consequent bi-stable periodic variation of the structural behavior, explained, thanks to a fit-to-the-purpose numerical modeling.