On 2D UGW-based probabilistic diagnostic imaging of web structure as a complex waveguide

The ultrasonic-guided waves (UGWs)-based sensor network, together with the probabilistic diagnostic imaging algorithm, has met extensive applications in structural health monitoring for thin-walled structures, which are not restricted to plates, but also include oil pipes, tubes, and cables. While UGWs are effective for flat or curved surfaces with simple, uniform waveguides, damage imaging on complex waveguides with varying thicknesses has been rarely investigated. This study presents the development of a UGW-based probabilistic damage imaging algorithm for a complex waveguide, denoted as RailPDI, using the rail web as a demonstration and verification example. The developed algorithm involves mode analysis using a semi-analytical finite element approach for damage-sensitive index calculation and considers the influence of thickness variation on weight function, as well as the diffusion mechanism of ultrasonic-guided waves in rail. Moreover, a piecewise normalization process is introduced to enhance the damage-distinguishing capability of slim and slender sensing pairs, consequently delivering a high-precision damage inference model for single- and multiple-damage scenarios. Verification experiments on rail web structure demonstrate the superior effectiveness and performance of RailPDI compared to other baseline methods. The proposed approach is also expected to be applicable for structural health monitoring of other critical components with complex waveguides.