EUVIMEDEUVIMED entdecken
Crossref

Multisensor Fusion With Feature Separation and Dual Alignment for Induction Motor Fault Diagnosis Under Variable-Speed Conditions

Yicong Tang, He-Sheng Zhang, Hengzhi Liu, Jianfeng Hong

IEEE Transactions on Instrumentation and Measurement · 2026 · 75 · 3519516-3519516

DOI: 10.1109/tim.2026.3712935 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Transactions on Instrumentation and Measurement, 2026. Verfasst von Yicong Tang und 3 weiteren Autor:innen. Die Arbeit nennt 40 Literaturangaben. Gefördert von: Beijing Natural Science Foundation, Systemic Major Project of China State Railway Group Company Ltd., Key Project of China State Railway Group Company Ltd..

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The widespread deployment of multiple sensors in motor condition monitoring has propelled the application of multisensor fusion technology in motor fault diagnosis. However, existing fusion methods suffer from insufficient extraction of shared and private features across sensors, inadequate capture of sensor fault semantic correlations, and lack of both local and global sensor feature alignment under variable-speed conditions, hindering diagnosis accuracy while limiting method applicability. Thus, we propose a multisensor fusion method based on feature separation and dual alignment (FSDA). This method projects each sensor’s data into two distinct feature spaces to separate shared information and private information. In the cross-sensor alignment stage, fault category-guided contrastive learning (FCGCL) is proposed to achieve intersensor and intrasensor fault semantic alignment. In the cross-domain alignment stage, sensor-shared and sensor-private features are aligned separately. Moreover, sensor-shared alignment reduces domain discrepancy at both local and global sensor levels. For 18 transfer tasks in four induction motor operation scenarios, the average diagnostic accuracy reaches 90.46 %, outperforming other methods. The results verify the effectiveness of the proposed method.

Förderung

  • Beijing Natural Science Foundation
  • Systemic Major Project of China State Railway Group Company Ltd.
  • Key Project of China State Railway Group Company Ltd.

Zitieren

APA

Tang, Y., Zhang, H., Liu, H., & Hong, J. (2026). Multisensor Fusion With Feature Separation and Dual Alignment for Induction Motor Fault Diagnosis Under Variable-Speed Conditions. IEEE Transactions on Instrumentation and Measurement, 75, 3519516-3519516. https://doi.org/10.1109/tim.2026.3712935

Vancouver

Tang Y, Zhang H, Liu H, Hong J. Multisensor Fusion With Feature Separation and Dual Alignment for Induction Motor Fault Diagnosis Under Variable-Speed Conditions. IEEE Transactions on Instrumentation and Measurement. 2026;75:3519516-3519516. doi:10.1109/tim.2026.3712935

BibTeX
@article{Tang2026,
  author = {Tang, Yicong and Zhang, He-Sheng and Liu, Hengzhi and Hong, Jianfeng},
  title = {Multisensor Fusion With Feature Separation and Dual Alignment for Induction Motor Fault Diagnosis Under Variable-Speed Conditions},
  journal = {IEEE Transactions on Instrumentation and Measurement},
  year = {2026},
  volume = {75},
  pages = {3519516--3519516},
  doi = {10.1109/tim.2026.3712935}
}
Kostenlos starten →