Abstract & Overview
Millions of catenary droppers are critical connection components in the electrified railway, ensuring the safety and reliability to train operations. However, the operation performance of the train will be affected by the fatigue state of droppers. Therefore, it is crucial to quickly and accurately assess dropper fatigue for electrified railways. In this study, we firstly propose a novel detection method based on laser-induced dual-modal signals, which combining Laser-Induced Breakdown Spectroscopy (LIBS) and Laser-Induced Acoustic Spectrum (LIAS). LIBS and LIAS are respectively used to capture chemical properties and evaluate mechanical properties, and machine learning techniques are utilized to classify the fatigue levels of the dropper. In order to improve model interpretability, the Shapley Additive Explanations (SHAP) is applied to deeply explain the LIBS and LIAS signals contributions of fatigue classification results. Experiment results indicate that the warning stage of fatigue degradation is mainly driven by the degradation of mechanical properties, while over time the degradation of chemical composition becomes the dominant factor. And the dual-modal signals from LIBS and LIAS effectively distinguish between different fatigue levels, achieving a classification accuracy of 99.17%. The proposed method in this study has potential to be extended to broader applications in the detection of metal aging and degradation in other engineering application. © 2012 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of Global Science and Technology Forum Pte Ltd