EUVIMEDEUVIMED entdecken
Crossref

Improved deep-learning rotor fault diagnosis based on multi vibration sensors and recurrence plots

Aroui Tarek, Marmouch Sameh

Journal of Vibration and Control · 2025-05 · 31 (9-10) · 1874-1883

DOI: 10.1177/10775463241250367 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Vibration and Control, online erschienen am 09.05.2024. Verfasst von Aroui Tarek und 1 weiteren Autor:innen. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 24 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Deep learning techniques are increasingly applied to time series data, offering promising results in various fields. Deep learning techniques can handle data from multiple sensors to detect anomalies in an industrial environment. This paper proposes a new method of anomaly detection based on a multilayer image representation of different vibration sensors’ recurrence plots. Each sensor’s recurrence plot forms a layer. The performance and reliability of our method were assessed using an experimental database collected under different load conditions and with different types of rotor anomalies. Experiment results demonstrate the effectiveness of GoogLeNet using individual and multi-layered recurrence plots to find rotor faults in an induction motors.

Zitieren

APA

Tarek, A., & Sameh, M. (2025). Improved deep-learning rotor fault diagnosis based on multi vibration sensors and recurrence plots. Journal of Vibration and Control, 31(9-10), 1874-1883. https://doi.org/10.1177/10775463241250367

Vancouver

Tarek A, Sameh M. Improved deep-learning rotor fault diagnosis based on multi vibration sensors and recurrence plots. Journal of Vibration and Control. 2025;31(9-10):1874-1883. doi:10.1177/10775463241250367

BibTeX
@article{Tarek2025,
  author = {Tarek, Aroui and Sameh, Marmouch},
  title = {Improved deep-learning rotor fault diagnosis based on multi vibration sensors and recurrence plots},
  journal = {Journal of Vibration and Control},
  year = {2025},
  volume = {31},
  number = {9-10},
  pages = {1874--1883},
  doi = {10.1177/10775463241250367}
}
Kostenlos starten →