Structural health monitoring of high-speed railways using ultrasonic guided waves

This paper presents a damage detection strategy for high-speed railways using piezoelectric active sensors. Multimodal ultrasonic guided waves generated by a piezoelectric transmitter propagate along the rail track, undergo dispersion, interact with the damage zone, and are finally picked up by the sensors. First, numerical investigations are carried out to understand the guided wave features and their interaction mechanism with typical damage scenarios in the railways. The modal analysis of a finite element scheme with Bloch-Floquet condition is conducted to obtain the dispersion characteristics and the mode shapes of the rail track guided waves. Optimum wave generation location and frequency were explored using a small-size local coupled field finite element model. Further, a Local Interaction Simulation Approach (LISA) model was developed to achieve efficient simulation of elastic wave propagation in railway structures. The LISA procedure was coded using the Compute Unified Device Architecture (CUDA), which enables the highly parallelized computing on powerful Graphics Processing Units (GPUs). This transient dynamic analysis reveals the influence of rail track features and damage signature on the sensing signals. Finally, full-scale experiments on a BS 90A rail track with embedded piezoelectric sensors are carried out to compare with the numerical investigations. This study shows that the active sensing system possess promising potential for the in-situ health monitoring of railway structures.