Pi-bracket sensor for crack detection monitoring near stiffeners in bridge girders

Abstract

This study investigates the design, validation, and performance of a pi-bracket sensor system for structural health monitoring of fatigue cracks in steel bridge girders through fiber optic sensors (FOSs), with a focus on detecting cracks near stiffeners, areas traditionally not monitored due to bending limitations of FOSs. Combining experimental testing, Finite Element Analysis, and theoretical validation, the research demonstrates that the pi-bracket configuration overcomes the limitations of FOSs by enabling strain measurements in critical, hard-to-access regions. Laboratory experiments confirmed the system’s sensitivity to crack openings as small as 0.2mm, with measurable strain magnitudes (129 µɛ) captured by FOS on the pi-bracket. FEA simulations isolated strain variations attributable solely to crack formation near stiffeners, revealing distinct strain differences at the pi-bracket crown (-2.37 µɛ for 0.1mm cracks and -4.74 µɛ for 0.2mm cracks) and contact point (-29.52 µɛ and -59.04 µɛ, respectively). The strain subtraction methodology effectively distinguished crack-induced strains from load-related effects, enabling targeted assessments of structural integrity. Integration with Brillouin Optical Time Domain Analysis technology ensured reliable distributed strain measurements while protecting sensors from mechanical damage. The system’s crack detection ability was confirmed through FEA simulations on full-scale girder models, demonstrating its adaptability for the field applications. These findings validate the pi-bracket system as a robust addition for detecting cracks in traditionally undetectable areas near stiffeners, overcoming FOS bending limitations and enhancing infrastructure resilience through monitoring. A comprehensive case study on a steel bridge in Manitoba, Canada, further demonstrated the capability of the pi-bracket sensor system for continuous, high-resolution strain measurement, specifically in fatigue-critical stiffened regions. A conservative crack detection threshold of 37µɛ was established and validated through extensive field instrumentation. The proposed crack detection procedure, grounded in FEA simulations and enhanced by signal processing techniques to compensate for temperature and environmental noise, was successfully validated against field data. Although no active cracks were detected during monitoring, the system showed sensitivity and robustness in differentiating crack-related strain signals from structural and environmental variations. These results provide a strong foundation for future deployments aimed at early crack detection, promising improvements in structural safety, maintenance prioritization, and longevity of steel bridges.