Semi-random subspace with Bi-GRU: Fusing statistical and deep representation features for bearing fault diagnosis

•Statistical and deep representation features have widely used.•SRS-BG is proposed for fault diagnosis of rolling bearings.•Experiments validate that SRS-BG outperforms other methods. Statistical features and deep representation features have been widely used in bearing fault diagnosis. These two ki...

Full description

Saved in:
Bibliographic Details
Published inMeasurement : journal of the International Measurement Confederation Vol. 173; p. 108603
Main Authors Mao, XuTing, Zhang, Feng, Wang, Gang, Chu, Yan, Yuan, KaiFu
Format Journal Article
LanguageEnglish
Published London Elsevier Ltd 01.03.2021
Elsevier Science Ltd
Subjects
Bearing fault diagnosis
Bearings
Ensemble learning
Fault diagnosis
Fusion features
Gate Recurrent Unit
Learning
Representations
Signal processing
Sparsity learning
Statistical analysis
Subspace methods
Subspaces
Bearing fault diagnosis
Sparsity learning
Ensemble learning
Gate Recurrent Unit
Fusion features
Online AccessGet full text

Cover

More Information
Summary:•Statistical and deep representation features have widely used.•SRS-BG is proposed for fault diagnosis of rolling bearings.•Experiments validate that SRS-BG outperforms other methods. Statistical features and deep representation features have been widely used in bearing fault diagnosis. These two kinds of features have their superiorities, however, few studies have explored combining them and considering their heterogeneousness. Therefore, a Semi-Random Subspace method with Bidirectional Gate Recurrent Unit (Bi-GRU), i.e., SRS-BG, is proposed in this paper, to take full advantage of fusion features for bearing fault diagnosis. Firstly, the statistical features are obtained by multiple signal processing methods, and the deep representation features are obtained by Bi-GRU. Secondly, the heterogeneousness among these features are considered by proposing a novel structure sparsity learning model, which is further utilized to produce based classifiers in the proposed Semi-Random Subspace method. Finally, experiments on bearing vibration datasets derived from Case Western Reserve University validate that the proposed feature fusion strategy greatly enhances diagnostic performance, and outperforms other existing ensemble learning methods.
ISSN:0263-2241
1873-412X
DOI:10.1016/j.measurement.2020.108603