EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Dynamic MAML with Efficient Multi-Scale Attention for Cross-Load Few-Shot Bearing Fault Diagnosis

Qinglei Zhang, Yifan Zhang, Jiyun Qin, Jianguo Duan, Ying Zhou

Entropy · 14.10.2025 · 27 (10) · 1063

DOI: 10.3390/e27101063 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Entropy, online erschienen am 14.10.2025. Verfasst von Qinglei Zhang und 4 weiteren Autor:innen. Beteiligte Einrichtung: China Institute of FTZ Supply Chain, Shanghai Maritime University, Shanghai 201306, China. Die Arbeit wurde bisher 14-mal zitiert (Crossref) und nennt 43 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Accurate bearing fault diagnosis under various operational conditions presents significant challenges, mainly due to the limited availability of labeled data and the domain mismatches across different operating environments. In this study, an adaptive meta-learning framework (AdaMETA) is proposed, which combines dynamic task-aware model-independent meta-learning (DT-MAML) with efficient multi-scale attention (EMA) modules to enhance the model’s ability to generalize and improve diagnostic performance in small-sample bearing fault diagnosis across different load scenarios. Specifically, a hierarchical encoder equipped with C-EMA is introduced to effectively capture multi-scale fault features from vibration signals, greatly improving feature extraction under constrained data conditions. Furthermore, DT-MAML dynamically adjusts the inner-loop learning rate based on task complexity, promoting efficient adaptation to diverse tasks and mitigating domain bias. Comprehensive experimental evaluations on the CWRU bearing dataset, conducted under carefully designed cross-domain scenarios, demonstrate that AdaMETA achieves superior accuracy (up to 99.26%) and robustness compared to traditional meta-learning and classical diagnostic methods. Additional ablation studies and noise interference experiments further validate the substantial contribution of the EMA module and the dynamic learning rate components.

Schlagwörter

Stichwörter der Autor:innen

adaptive meta-learningcross-domain fault diagnosisefficient multi-scale attentionfew-shot learning

Autor:innen und Institutionen

  • China Institute of FTZ Supply Chain, Shanghai Maritime University, Shanghai 201306, China.

Zitieren

APA

Zhang, Q., Zhang, Y., Qin, J., Duan, J., & Zhou, Y. (2025). Dynamic MAML with Efficient Multi-Scale Attention for Cross-Load Few-Shot Bearing Fault Diagnosis. Entropy, 27(10), 1063. https://doi.org/10.3390/e27101063

Vancouver

Zhang Q, Zhang Y, Qin J, Duan J, Zhou Y. Dynamic MAML with Efficient Multi-Scale Attention for Cross-Load Few-Shot Bearing Fault Diagnosis. Entropy. 2025;27(10):1063. doi:10.3390/e27101063

BibTeX
@article{Zhang2025,
  author = {Zhang, Qinglei and Zhang, Yifan and Qin, Jiyun and Duan, Jianguo and Zhou, Ying},
  title = {Dynamic MAML with Efficient Multi-Scale Attention for Cross-Load Few-Shot Bearing Fault Diagnosis},
  journal = {Entropy},
  year = {2025},
  volume = {27},
  number = {10},
  pages = {1063},
  doi = {10.3390/e27101063}
}
Kostenlos starten →