Optimization of Sensor Configurations for Cost-Efficient Monitoring of Infrastructure Systems

Cost-cutting in maintenance and inspection budgets are causing a mismatch between the current state of resourcing allocated to bridge preventive management and the real budget that would be needed to ensure safe and functioning transport routes in the long run. SHM is rarely implemented and, very often, only a sufficient spatial density of the sensor network is guaranteed. Issues related to both network topology and optimal sensor placement (OSP) are therefore pivotal to maximise the quality of SHM information that can be retrieved from a limited quantity of sensors while reducing overall SHM costs to the minimum. The present work investigates this optimisation problem by analysing different heuristic algorithms for OSP and assessing the effectiveness and reliability of the candidate sensor configurations derived therefrom in extracting accurate dynamic features and identifying damage occurrence in infrastructure systems. A real multi-span bridge subjected to progressive damage scenarios is used as a benchmark case study and a data-driven modal-based multi-objective approach is exploited to measure the utility of each configuration, allowing to determine the best possible sensor deployment.