Damage Identification of Beam Structures Using Strain Mode Shape–Derived Modal Curvature: Theory, Simulation, and Field Test

Abstract Structural damage identification is essential for the serviceability assessment of beam structures. This paper presents a practical method that directly computes modal curvature from strain measurements and utilizes modal curvature difference (MCD) for damage detection. Numerical simulations established that strain has better sensitivity to structural damage than displacement, demonstrating the method’s numerical stability, effective control of truncation errors, and accurate performance under nonuniform sensor arrangements. Field validation was conducted through bolt loosening monitoring in a monopole tower. Under artificial excitation, the method identified bolt loosening, with the MCD magnitude indicating relative damage severity and its location and sign providing localization. The tests further revealed distinct variations in sensitivity across different layer flanges. Wind-induced vibration tests simulated the complete health-damage-repair monitoring process, demonstrating the method’s capability to identify bolt loosening under natural excitation while maintaining stable MCD responses. An in-service case study identifies typhoon-induced bolt loosening using sparse nonuniform sensors, confirming its applicability in real-world conditions. Further, the study discovered a distance attenuation effect in the MCD, suggesting the potential to infer the approximate circumferential position of loose bolts by comprehensively analyzing MCD changes from sensors in adjacent orientations.