A monitoring-based classification system for risk management of bridge scour

摘要

Flood-induced scour is the principal cause of bridge failure worldwide. Nevertheless, bridge scour risk assessment isstill based on visual inspections, which may be affected by human errors and cannot be performed during floodpeaks. This problem, together with the simplifications in scour estimation, might cause misclassification of the bridgescour risk, unnecessary bridge closures or recourse to avoidable scour mitigation measures. Structural-healthmonitoring(SHM) systems allow overcoming these issues, providing bridge managers with more accurateinformation about scour, thus supporting them in taking optimal management decisions. This paper illustrates thedevelopment of an SHM- and event-based classification system for bridge scour management, which extends andcomplements current risk-rating procedures by incorporating the various sources of uncertainty characterising thescour estimation and information from different sensors. The proposed system is based on a probabilistic frameworkfor scour risk estimation and can be used to provide transport agencies with real-time scour risk classification ofbridges under a heavy-flood event. The system is applied to a bridge network located in south-west Scotland in aheavy-flood scenario, and information from heterogeneous sources is considered for updating the knowledge ofscour. It is shown that integrating scour-monitoring data leads to an overall uncertainty reduction that is reflected inmore accurate scour risk classification, thus helping transport agencies in prioritising bridge inspections and riskmitigation actions.