EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Cross-Domain Open Set Fault Diagnosis Based on Weighted Domain Adaptation with Double Classifiers

Huaqing Wang, Zhitao Xu, Xingwei Tong, Liuyang Song iD

Sensors · 14.02.2023 · 23 (4) · 2137

DOI: 10.3390/s23042137 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Sensors, online erschienen am 14.02.2023. Verfasst von Huaqing Wang und 3 weiteren Autor:innen. Beteiligte Einrichtung: College of Mechanical Electrical Engineering, Beijing University of Chemical Technology, Beijing 100029, China. Die Arbeit wurde bisher 21-mal zitiert (Crossref) und nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: the National Key R&D Program of China, the National Natural Science Foundation of China, the Beijing Natural Science Foundation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The application of transfer learning in fault diagnosis has been developed in recent years. It can use existing data to solve the problem of fault recognition under different working conditions. Due to the complexity of the equipment and the openness of the working environment in industrial production, the status of the equipment is changeable, and the collected signals can have new fault classes. Therefore, the open set recognition ability of the transfer learning method is an urgent research direction. The existing transfer learning model can have a severe negative transfer problem when solving the open set problem, resulting in the aliasing of samples in the feature space and the inability to separate the unknown classes. To solve this problem, we propose a Weighted Domain Adaptation with Double Classifiers (WDADC) method. Specifically, WDADC designs the weighting module based on Jensen–Shannon divergence, which can evaluate the similarity between each sample in the target domain and each class in the source domain. Based on this similarity, a weighted loss is constructed to promote the positive transfer between shared classes in the two domains to realize the recognition of shared classes and the separation of unknown classes. In addition, the structure of double classifiers in WDADC can mitigate the overfitting of the model by maximizing the discrepancy, which helps extract the domain-invariant and class-separable features of the samples when the discrepancy between the two domains is large. The model’s performance is verified in several fault datasets of rotating machinery. The results show that the method is effective in open set fault diagnosis and superior to the common domain adaptation methods.

Schlagwörter

Stichwörter der Autor:innen

deep learningfault diagnosisopen set domain adaptationrotating machinerytransfer learning

Autor:innen und Institutionen

  • College of Mechanical Electrical Engineering, Beijing University of Chemical Technology, Beijing 100029, China.
  • Key Laboratory of Health Monitoring and Self-recovery for High-end Mechanical Equipment, Beijing University of Chemical Technology, Beijing 100029, China.

ORCID-Profile: Liuyang Song

Förderung

  • the National Key R&D Program of China
  • the National Natural Science Foundation of China
  • the Beijing Natural Science Foundation
  • the Joint Project of BRC-BC
  • Beijing Natural Science Foundation
  • Joint Project of BRC-BC
  • Key Laboratory Hunan Province of Health Maintenance Mechanical Equipment

Zitieren

APA

Wang, H., Xu, Z., Tong, X., & Song, L. (2023). Cross-Domain Open Set Fault Diagnosis Based on Weighted Domain Adaptation with Double Classifiers. Sensors, 23(4), 2137. https://doi.org/10.3390/s23042137

Vancouver

Wang H, Xu Z, Tong X, Song L. Cross-Domain Open Set Fault Diagnosis Based on Weighted Domain Adaptation with Double Classifiers. Sensors. 2023;23(4):2137. doi:10.3390/s23042137

BibTeX
@article{Wang2023,
  author = {Wang, Huaqing and Xu, Zhitao and Tong, Xingwei and Song, Liuyang},
  title = {Cross-Domain Open Set Fault Diagnosis Based on Weighted Domain Adaptation with Double Classifiers},
  journal = {Sensors},
  year = {2023},
  volume = {23},
  number = {4},
  pages = {2137},
  doi = {10.3390/s23042137}
}
Kostenlos starten →