EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Research on the fault diagnosis technology of the railway vehicle transmission system based on a convolutional neural network

Xiang Yu, Lei Han, Jili Wu, Buzhao Niu

Transportation Safety and Environment · 14.01.2025 · 7 (1)

DOI: 10.1093/tse/tdaf020 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Transportation Safety and Environment, online erschienen am 01.04.2025. Verfasst von Xiang Yu und 3 weiteren Autor:innen. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 28 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key Research and Development Program of China.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The transmission system is an important part of rail vehicles. Its operation safety is of great significance to the safe and reliable operation of rail vehicles. Therefore, it is particularly important to monitor the status and diagnose faults of the transmission system. In this paper, a parallel wide first layer depth convolutional neural network (P-WDCNN) for multi-input fusion is proposed for fault diagnosis of the rail vehicle transmission system. By analysing the fault mechanism of the transmission system and exploring relevant signal data, four time series were selected as inputs for the neural network: speed, motor current, average current of each carriage of this train and standard deviation of current of each carriage of this train. In response to the problem of imbalanced sample size caused by the introduction of a large amount of new data, this paper proposes a sample incremental learning method based on group resampling. This method utilizes group resampling to obtain a balanced number of positive and negative samples, and constructs a new training set using the balanced positive and negative samples to update the fault diagnosis model. The experimental results show that the P-WDCNN model established in this paper has good identification ability and achieves an accuracy of over 99% on the validation set. The sample incremental learning method based on group resampling adopted in this article helps to improve the problem of recall fluctuating with different proportions of faulty samples in the training set, achieving a recall rate higher than 97% on the validation set.

Förderung

  • National Key Research and Development Program of China

Zitieren

APA

Yu, X., Han, L., Wu, J., & Niu, B. (2025). Research on the fault diagnosis technology of the railway vehicle transmission system based on a convolutional neural network. Transportation Safety and Environment, 7(1). https://doi.org/10.1093/tse/tdaf020

Vancouver

Yu X, Han L, Wu J, Niu B. Research on the fault diagnosis technology of the railway vehicle transmission system based on a convolutional neural network. Transportation Safety and Environment. 2025;7(1). doi:10.1093/tse/tdaf020

BibTeX
@article{Yu2025,
  author = {Yu, Xiang and Han, Lei and Wu, Jili and Niu, Buzhao},
  title = {Research on the fault diagnosis technology of the railway vehicle transmission system based on a convolutional neural network},
  journal = {Transportation Safety and Environment},
  year = {2025},
  volume = {7},
  number = {1},
  doi = {10.1093/tse/tdaf020}
}
Kostenlos starten →