Multi-scale engineering of lead-free piezoelectric pavement sensing systems

The development of intelligent road infrastructure requires sensing systems that are reliable, durable, and compatible with long-term pavement service conditions. Piezoelectric sensing technologies are widely used in traffic monitoring because they convert mechanical loading into electrical signals without an external power supply. Most pavement sensing systems are based on lead-containing piezoelectric ceramics, especially lead zirconate titanate (PZT), due to their high intrinsic sensitivity. However, environmental considerations and sustainability requirements motivate the exploration of lead-free alternatives. Potassium sodium niobate (KNN)-based ceramics offer promising thermal stability and are suitable for elevated pavement temperatures, yet their intrinsic piezoelectric response is lower than that of PZT. Therefore, the challenge of applying KNN-based materials in pavement sensing systems lies not only in optimizing the material properties but also in understanding how material behavior interacts with pavement structures and sensing arrangements. This dissertation adopts a multi-scale engineering perspective for integrating lead-free piezoelectric materials into pavement sensing systems. The study links multiple scales, including material-level piezoelectric behavior, device-level sensor packaging, pavement-scale structural interaction, and system-level sensing performance. At the material level, KNN-based ceramics were designed and evaluated under pavement-relevant thermal conditions to assess their functional stability. The experimental results confirm their suitability for pavement environments while revealing that the relatively limited piezoelectric response requires further structural optimization at higher system levels. At the device scale, the influence of sensor packaging on sensing behavior is systematically investigated. Because signal generation in pavement sensors strongly depends on mechanical confinement and load transfer, the effects of packaging on sensitivity, linearity, and durability under cyclic loading are analyzed. The results show that mechanical strength or peak voltage alone does not determine sensing reliability. Instead, stable signal evolution and consistent response characteristics are more critical for long-term monitoring applications. At the structural scale, the sensing response of embedded sensors within pavement structures is investigated through numerical simulations. Parametric studies are conducted to examine how different traffic conditions and sensor embedding configurations influence the generated piezoelectric signals. The analyses consider factors such as load magnitude, vehicle type, sensor depth, and spatial position. The results provide insight into how pavement structural conditions affect sensor output signals and help clarify the sensitivity of the sensing response to traffic loading and installation parameters. At the system scale, a simulation-driven framework is developed to evaluate sensor arrangement strategies for pavement monitoring. The framework integrates pavement structural parameters, sensor characteristics, and traffic scenarios to assess signal distinguishability for vehicle classification. Rather than assuming dense sensor placement, the framework identifies configurations that preserve essential sensing information while reducing the number of embedded sensing units, introducing an information-efficiency perspective into pavement sensing design. The investigation combines laboratory experiments and numerical modeling with emphasis on mechanism clarification and methodological development. Full-scale field validation and long-term in-situ monitoring are beyond the scope of the present study and remain subjects for future work. This dissertation presents a multi-scale engineering pathway for translating lead-free piezoelectric materials into pavement sensing systems. It contributes a mechanics-based description of signal attenuation in layered pavement structures and proposes a rational framework for sensor arrangement. The findings demonstrate that reliable traffic classification can be achieved through coordinated system design, even when the intrinsic piezoelectric response is limited, thereby providing a foundation for subsequent field implementation and model refinement.