@article{Liu2026, 
author = {Kan Liu and Zherui Zhang and Liran Li and Leiting Zhao and Yijie Zhou},
title = {Research and application of diagnosis methods and devices for urban rail traction systems},
year = {2026},
journal = {Railway Sciences},
volume = {5},
number = {1},
pages = {88-99},
keywords = {VVVF converters, High-frequency data acquisition, Edge storage, Fault pattern mining},
url = {https://www.sciopen.com/article/10.1108/RS-10-2025-0048},
doi = {10.1108/RS-10-2025-0048},
abstract = {PurposeWith the deepening integration of rail transit systems–encompassing urban rail, regional railways, trunk lines and medium-low capacity transportation–the four-network integration imposes higher demands on operation and maintenance systems regarding cross-modal coordination, full-element interconnectivity and dynamic responsiveness.Design/methodology/approachThis paper, based on policy directives and engineering practices, analyzes the operational maintenance characteristics of urban rail traction systems from perspectives including device interconnectivity and fault data mining. A non-intrusive high-frequency diagnostic device independent of vehicle control is proposed, informed by practical onboard operation experience. This innovation significantly enhances diagnostic accuracy for components requiring high sampling frequency, while integrating “Flash” storage with far greater capacity than conventional control chips.FindingsThis article will systematically introduces the key points and diagnostic methods for typical faults in urban rail traction systems. Through rational diagnostic algorithms combined with high-precision, high-storage diagnostic instrumentation, the overall safety and reliability of urban rail traction systems have been improved. The proposed non-intrusive high-frequency diagnostic solution has been validated across multiple rail lines.Originality/valueThis paper introduces an innovative non-intrusive diagnostic device with a dual-channel design for multi-system compatibility and a high-speed acquisition architecture enabling 400 kHz sampling. Its originality stems from the independent, high-fidelity capture of microsecond-level transient faults like IGBT shoot-through and pantograph arcing; Validated in operational environments, this approach provides a significant leap in diagnostic precision, directly enhancing traction system availability and operational safety by enabling precise fault localization and intelligent, adaptive protection strategies.}
}