Multi-Source Information Fault Diagnosis Method for Rolling Bearing Based on Multi-View Clustering
Journal of Engineering Mechanics and Machinery · 2026 · 11 (1)
DOI: 10.23977/jemm.2026.110117 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Engineering Mechanics and Machinery, 2026.
Abstract (Originalsprache)
: This paper addresses the difficulty that decision-level fusion in unsupervised fault diagnosis of rolling bearing multi-source data fails to effectively exploit the complementary information among multi-source data. A rolling bearing fault diagnosis method based on multi-view clustering (MVC) is proposed. First, a variational autoencoder (VAE) is employed to extract latent features from multi-source signals. On this basis, a joint matrix that integrates multi-source data features and a high-confidence matrix containing consistency information are constructed. A low-rank approximation method is then used to mine complementary information within multi-source information and enhance structural consistency. Finally, accurate identification of unsupervised fault diagnosis is achieved. Experimental results show that the proposed method can effectively utilize the complementary information among multiple views, overcome dependence on a single data source, and exhibit excellent robustness and stability.
Zitieren
APA
(2026). Multi-Source Information Fault Diagnosis Method for Rolling Bearing Based on Multi-View Clustering. Journal of Engineering Mechanics and Machinery, 11(1). https://doi.org/10.23977/jemm.2026.110117
Vancouver
Multi-Source Information Fault Diagnosis Method for Rolling Bearing Based on Multi-View Clustering. Journal of Engineering Mechanics and Machinery. 2026;11(1). doi:10.23977/jemm.2026.110117
BibTeX
@article{euvimed2026,
title = {Multi-Source Information Fault Diagnosis Method for Rolling Bearing Based on Multi-View Clustering},
journal = {Journal of Engineering Mechanics and Machinery},
year = {2026},
volume = {11},
number = {1},
doi = {10.23977/jemm.2026.110117}
}
EUVIMED