Thesis title: Piezoelectric Bearings (PiBs) for Smart and Sustainable Structural Health Monitoring of Bridges
The growing emphasis on sustainability in civil infrastructure highlights the need to reduce the energy demand and maintenance burden of Structural Health Monitoring (SHM) systems while extending the service life of ageing bridge networks. This thesis introduces Piezoelectric Bearings (PiBs), a multifunctional class of elastomeric bridge bearings in which one or more flexible unimorph piezoelectric laminas are integrated within an internal cavity, enabling the simultaneous conversion of traffic-induced bearing deformations into electrical energy and information-rich electrical signatures of vehicle passages. A multi-level, multiphysics framework is developed to model the device and its interaction with real bridges: at the macrolevel, the bridge is represented through a detailed finite-element model in which PiBs are introduced via an equivalent vertical rheological representation to reproduce realistic moving-load distribution; at the mesolevel, a three-dimensional bearing model resolves the local compression response and converts support reactions into vertical closure time histories; and at the microlevel, a coupled electromechanical transient model predicts voltage and power output under imposed settlement histories, including geometric imperfections, electrode idealization, and resistive-load optimization.To ensure predictive capability, a full-scale laboratory campaign is conducted by installing a commercial piezoelectric lamina inside a sectioned elastomeric bearing and applying displacement-controlled cyclic compression at traffic-representative amplitudes and frequencies, thereby extracting power–resistance curves used to validate and calibrate the microlevel model. The validated framework is then extended to standard and custom transducer geometries, including substrate effects, and applied to representative bridge scenarios. Bridge-scale results show harvested energy levels in the order of tens of mJ per heavy-vehicle passage, supporting the feasibility of self-powered, duty-cycled monitoring concepts. In parallel, physically guided traffic-identification algorithms are developed to exploit the PiB electrical response as a sensing signal, extracting axle-related and kinematic information including axle count, axle spacing, vehicle category, and crossing speed. The methodology is progressively extended from a simplified single-lane formulation to a more general two-sensor framework capable of reconstructing complex traffic scenarios without imposing lane-specific vehicle sequencing. System-level feasibility is assessed through two complementary architectures: (X) event-driven traffic logging via PiB voltage features, and (Y) PiB-powered external MEMS sensing cycles for intermittent dynamic identification and anomaly screening, both implementable with shared low-power electronics and adaptive buffer-aware scheduling. These results position PiBs not only as energy-harvesting devices, but as bridge-integrated multifunctional sensing components with particular potential for traffic characterization and future load-monitoring applications. Finally, to bridge the gap between numerical predictions and deployment, a dedicated PiB prototype is designed for testing at the JRC laboratories in Ispra, where a small-span experimental bridge and real vehicles will be used to evaluate (i) the effective harvested energy versus model predictions and (ii) the performance of the identification algorithm on measured electrical power signals under realistic traffic passages.