This article aims to give a complete review of previous and current research on numerous types of out-of-roundness (OOR) failures in train wheels, as well as diagnostic approaches based on machine learning and vibration data. The study provides a comprehensive overview of the current state of research by categorizing reviews into three primary domains: (1) types of OOR failures in train wheels, (2) fault diagnosis methodologies, and (3) the use of machine learning and vibration data to diagnose train wheel OOR failures. Initially, the study investigates the characteristics, causes, and consequences of railway wheel OOR failures, including their impact on vibrations. It then dives further into diagnostic methods, comparing the effectiveness of statistical methods to machine learning-based methods for diagnosing failures. Furthermore, the study addresses current advances in machine learning and vibration-based diagnostic methods to diagnose train wheel OOR failures, providing information on their applications and results. This article highlights that by utilizing machine learning methods with vibration data offers a promising way for accurately diagnosing OOR faults in train wheels and predicting their potential failure and remaining useful life, resulting to enhanced maintenance efficiency and less downtime.
Copyrights © 2024