A novel clustering based curvature method with wavelet transform for detecting progressive damage of simply supported ultra-high performance fiber-reinforced concrete beams using laser scanner vibrometer

Monitoring the structural performance of Ultra-High-Performance Fiber-Reinforced Concrete (UHPFRC) beams is essential for understanding damage progression and improving long-term durability in advanced structural elements. This study presents a novel reference-free clustering-based algorithm that leverages high spatial resolution and higher vibrational modes obtained using a laser scanning vibrometer (LSV) during a progressive damage test on UHPFRC. In the experimental phase, the vibrational behavior of a 2-meter one-span simply-supported UHPFRC beam was measured by laser scanning vibrometry under progressively increasing damage levels. Two different reference-free damage detection methods are developed and compared. In the first method, the curvature is calculated using the central difference approximation, and a clustering-based algorithm combining LoOP outlier detection with locally weighted nonparametric regression fitting using a second-order polynomial (LOESS) is applied to construct an intact baseline directly from the damaged data. In the second method, a wavelet-based curvature formulation is introduced to overcome the noise sensitivity and boundary effects inherent in central difference schemes. Both approaches are evaluated for their ability to identify damage zone characteristics, such as position and severity, and are validated through finite element simulations where strain energy variation is used as a numerical severity index.