Balancing Model-Driven and Data-Driven Strategies in the SHM: Lessons Learned from a Case Study on the Dynamic Behaviour Identification of Complex Structural Scheme Bridge
摘要
Advancements in Structural Health Monitoring (SHM) have made monitoring systems a crucial asset in infrastructure management. The deployment of these systems across extensive infrastructure networks enables administrators to evaluate the remaining safety margins of bridges under their supervision and facilitates the prioritization and planning of maintenance interventions when necessary. However, the large-scale implementation of comprehensive monitoring systems often entails costs that may exceed available budgets, creating a demand for optimized solutions. In this direction different monitoring strategies, developed by Sacertis Ingegneria, are ongoing for testing, being categorized by the type of analysis used whether Model Driven, Data Driven, or a hybrid of both, and by the system’s scope, such as sensor placement and coverage on the monitored elements (Local vs Global monitoring). A real case study of a bridge with continuous double slab prestressed concrete deck interspersed with two Gerber half joints and framed with reinforced concrete piers, is presented. In this application, the monitoring strategy involves both spans treated with a model-based approach and full-sensors layout, and spans with data-driven approach and light-sensors layout. The paper aims to discuss pros and cons of the mix of monitoring approaches, when dealing with complex structural static schemes, in the interpretation of the dynamic response.