An In-Depth Tutorial on BJTU-RAO Bogie Datasets for Fault Diagnosis
Yong Qin, Yiran Wang, Zhaojun Steven Li, Biao Wang, Ao Ding, Chengcheng Wang, Yuanjing Qin, Yang Wang
Abstract
The reliability and safety of trains have always been the top priority in the railway industry. As the critical subsystem of trains, the health states of bogie transmission systems directly affect the operation safety of trains. Train fault diagnosis, predictive maintenance, and algorithms evaluation become crucial to ensure operational safety. Thus, there is an urgent need for high-quality fault datasets of train transmission systems. This tutorial introduces the first publicly available fault datasets of train transmission systems, referred to as Beijing Jiaotong University - Rail Autonomous Operations (BJTU-RAO) bogie datasets. The datasets contain multi-sensor data streams of transmission systems of subway train bogies including a health state and 50 faulty states, which are acquired by fault simulation experiments and are released at the 2024 Global Reliability and Prognostics and Health Management (PHM) Conference. For each type of state, samples under nine different working conditions are collected to represent the different operating conditions of trains, and the signals are recorded at a sampling frequency of 64 kHz. The datasets are collected, organized, and made publicly available by the research team from the State Key Lab of Advanced Rail Autonomous Operation at BJTU, aiming to encourage scholars and engineers to investigate and validate their fault detection or diagnosis algorithms for key components of train transmission systems, such as driving gearboxes, axle boxes, and traction motors.