Laser-Induced Ultrasonic Guided Waves Based Corrosion Diagnosis of Rail Foot

Abstract

Rail foot corrosion is an important issue for electrified railway system. It usually initiates at the bottom of rail foot edges covered by a rail clip and is hardly accessible to conventional online inspection systems. To address this problem, this study combines air-coupled ultrasonic transducer (AUT) and laser-induced ultrasonic guided wave to detect rail foot corrosion. The corrosion sensitivity of the guided wave below 500 kHz is examined considering the impact of rail clip. The results have shown that A <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> and S <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> are the dominant wave modes near the edge of rail foot. The preferred A <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> wave is insensitive to allowable rail foot corrosion below 100 kHz. It is neither sensitive above 270 kHz due to wave mode conversion and reduced wave penetration. The optimal frequency is about 200 kHz where the attenuation of A <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">0</sub> wave can reflect the impact of all levels of corrosion even in the presence of a rail clip. Accordingly, an end-to-end deep learning model is employed for corrosion detection based on 200-kHz air-coupled ultrasonic signals. It can automatically extract signal features for corrosion diagnosis as well as differentiate unknown anomalous signals by environmental interference. It is tested with 8122 labeled signal samples collected from field trial and achieves a detection accuracy of 99.5%. All the anomalous signals are also correctly identified.