EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Rail corrugation diagnosis from axle box vibration signals using optimized empirical wavelet transform and CNN–LSTM

Yanqing Gong, Cheng Zhu, Jiankun Xie, Tingzhou Hu, Jie Hu

Journal of Vibroengineering · 30.10.2026

DOI: 10.21595/jve.2026.25856 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Vibroengineering, online erschienen am 30.10.2026. Verfasst von Yanqing Gong und 4 weiteren Autor:innen. Die Arbeit nennt 10 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Rail corrugation is a common short-wave defect of metro rails that can excite strong wheel-rail vibration and compromise running safety and ride quality. Its efficient identification from axle box vibration signals remains challenging because weak corrugation components are easily masked by wheel-rail noise, and conventional fixed-threshold decomposition methods cannot adapt to varying operating conditions. To fill this gap, this paper optimizes the empirical wavelet transform (EWT) using the Otsu maximum between-class variance method (OSTU) to extract corrugation features adaptively from axle box signals, and proposes a CNN-LSTM intelligent diagnosis model for rail corrugation. A wheel–rail coupled dynamic model is established and a simulation method of rail corrugation is given; simulation validation shows that the decomposition algorithm can effectively separate corrugation components and reduce the influence of the preset threshold. The model is then experimentally validated on actual line data measured on an operating metro line, achieving a diagnostic accuracy of 89.5 % on the test set and demonstrating good performance across all corrugation grades.

Zitieren

APA

Gong, Y., Zhu, C., Xie, J., Hu, T., & Hu, J. (2026). Rail corrugation diagnosis from axle box vibration signals using optimized empirical wavelet transform and CNN–LSTM. Journal of Vibroengineering. https://doi.org/10.21595/jve.2026.25856

Vancouver

Gong Y, Zhu C, Xie J, Hu T, Hu J. Rail corrugation diagnosis from axle box vibration signals using optimized empirical wavelet transform and CNN–LSTM. Journal of Vibroengineering. 2026. doi:10.21595/jve.2026.25856

BibTeX
@article{Gong2026,
  author = {Gong, Yanqing and Zhu, Cheng and Xie, Jiankun and Hu, Tingzhou and Hu, Jie},
  title = {Rail corrugation diagnosis from axle box vibration signals using optimized empirical wavelet transform and CNN–LSTM},
  journal = {Journal of Vibroengineering},
  year = {2026},
  doi = {10.21595/jve.2026.25856}
}
Kostenlos starten →