Home > Research > Publications & Outputs > Multi-scale deep intra-class transfer learning ...

Associated organisational unit

Electronic data

  • 1-s2.0-S0951832020305512-main

    Rights statement: This is the author’s version of a work that was accepted for publication in Reliability Engineering and System Safety. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Reliability Engineering and System Safety, 202, 2020 DOI: 10.1016/j.ress.2020.107050

    Accepted author manuscript, 4.3 MB, PDF document

    Available under license: CC BY-NC-ND

Links

Text available via DOI:

View graph of relations

Multi-scale deep intra-class transfer learning for bearing fault diagnosis

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Published

Standard

Multi-scale deep intra-class transfer learning for bearing fault diagnosis. / Wang, X.; Shen, C.; Xia, M. et al.

In: Reliability Engineering and System Safety, Vol. 202, 107050, 01.10.2020.

Research output: Contribution to Journal/MagazineJournal articlepeer-review

Harvard

Wang, X, Shen, C, Xia, M, Wang, D, Zhu, J & Zhu, Z 2020, 'Multi-scale deep intra-class transfer learning for bearing fault diagnosis', Reliability Engineering and System Safety, vol. 202, 107050. https://doi.org/10.1016/j.ress.2020.107050

APA

Wang, X., Shen, C., Xia, M., Wang, D., Zhu, J., & Zhu, Z. (2020). Multi-scale deep intra-class transfer learning for bearing fault diagnosis. Reliability Engineering and System Safety, 202, [107050]. https://doi.org/10.1016/j.ress.2020.107050

Vancouver

Wang X, Shen C, Xia M, Wang D, Zhu J, Zhu Z. Multi-scale deep intra-class transfer learning for bearing fault diagnosis. Reliability Engineering and System Safety. 2020 Oct 1;202:107050. Epub 2020 Jun 3. doi: 10.1016/j.ress.2020.107050

Author

Wang, X. ; Shen, C. ; Xia, M. et al. / Multi-scale deep intra-class transfer learning for bearing fault diagnosis. In: Reliability Engineering and System Safety. 2020 ; Vol. 202.

Bibtex

@article{2ef5cf2fb219448a90f68f5c18c1fd6d,
title = "Multi-scale deep intra-class transfer learning for bearing fault diagnosis",
abstract = "The tremendous success of deep learning in machine fault diagnosis is dependent on the hypothesis that training and test datasets are subordinated to the same distribution. This subordination is difficult to meet in practical scenarios of industrial applications. On the one hand, the working conditions of rotating machinery can change easily. On the other hand, vibration data and labels are difficult to obtain to train a specific model for each working condition. In this study, we solve these problems by constructing a novel deep transfer learning model called multi-scale deep intra-class adaptation network, which first uses the modified ResNet-50 to extract low-level features and then constructs a multiple scale feature learner to analyze these low-level features at multiple scales and obtain high-level features as input for the classifier. Pseudo labels are then computed to shorten the conditional distribution distance of vibration data collected under different working loads for intra-class adaptation. The proposed method is validated using two datasets to recognize the bearing normal state, the inner race, the ball and outer race faults, and their fault degrees under four different working loads. The high-precision diagnosis results of 24 transfer learning experiments reveal the reliability and generalizability of the constructed model. ",
keywords = "Deep transfer learning, Fault diagnosis, Intra-class adaptation, Multi-scale feature learner, Bearings (machine parts), Failure analysis, Fault detection, Learning systems, Transfer learning, Bearing fault diagnosis, Conditional distribution, High-level features, High-precision, Low-level features, Machine fault diagnosis, Multiple scale, Vibration data, Deep learning",
author = "X. Wang and C. Shen and M. Xia and D. Wang and J. Zhu and Z. Zhu",
note = "This is the author{\textquoteright}s version of a work that was accepted for publication in Reliability Engineering and System Safety. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Reliability Engineering and System Safety, 202, 2020 DOI: 10.1016/j.ress.2020.107050",
year = "2020",
month = oct,
day = "1",
doi = "10.1016/j.ress.2020.107050",
language = "English",
volume = "202",
journal = "Reliability Engineering and System Safety",
issn = "0951-8320",
publisher = "Elsevier Limited",

}

RIS

TY - JOUR

T1 - Multi-scale deep intra-class transfer learning for bearing fault diagnosis

AU - Wang, X.

AU - Shen, C.

AU - Xia, M.

AU - Wang, D.

AU - Zhu, J.

AU - Zhu, Z.

N1 - This is the author’s version of a work that was accepted for publication in Reliability Engineering and System Safety. Changes resulting from the publishing process, such as peer review, editing, corrections, structural formatting, and other quality control mechanisms may not be reflected in this document. Changes may have been made to this work since it was submitted for publication. A definitive version was subsequently published in Reliability Engineering and System Safety, 202, 2020 DOI: 10.1016/j.ress.2020.107050

PY - 2020/10/1

Y1 - 2020/10/1

N2 - The tremendous success of deep learning in machine fault diagnosis is dependent on the hypothesis that training and test datasets are subordinated to the same distribution. This subordination is difficult to meet in practical scenarios of industrial applications. On the one hand, the working conditions of rotating machinery can change easily. On the other hand, vibration data and labels are difficult to obtain to train a specific model for each working condition. In this study, we solve these problems by constructing a novel deep transfer learning model called multi-scale deep intra-class adaptation network, which first uses the modified ResNet-50 to extract low-level features and then constructs a multiple scale feature learner to analyze these low-level features at multiple scales and obtain high-level features as input for the classifier. Pseudo labels are then computed to shorten the conditional distribution distance of vibration data collected under different working loads for intra-class adaptation. The proposed method is validated using two datasets to recognize the bearing normal state, the inner race, the ball and outer race faults, and their fault degrees under four different working loads. The high-precision diagnosis results of 24 transfer learning experiments reveal the reliability and generalizability of the constructed model.

AB - The tremendous success of deep learning in machine fault diagnosis is dependent on the hypothesis that training and test datasets are subordinated to the same distribution. This subordination is difficult to meet in practical scenarios of industrial applications. On the one hand, the working conditions of rotating machinery can change easily. On the other hand, vibration data and labels are difficult to obtain to train a specific model for each working condition. In this study, we solve these problems by constructing a novel deep transfer learning model called multi-scale deep intra-class adaptation network, which first uses the modified ResNet-50 to extract low-level features and then constructs a multiple scale feature learner to analyze these low-level features at multiple scales and obtain high-level features as input for the classifier. Pseudo labels are then computed to shorten the conditional distribution distance of vibration data collected under different working loads for intra-class adaptation. The proposed method is validated using two datasets to recognize the bearing normal state, the inner race, the ball and outer race faults, and their fault degrees under four different working loads. The high-precision diagnosis results of 24 transfer learning experiments reveal the reliability and generalizability of the constructed model.

KW - Deep transfer learning

KW - Fault diagnosis

KW - Intra-class adaptation

KW - Multi-scale feature learner

KW - Bearings (machine parts)

KW - Failure analysis

KW - Fault detection

KW - Learning systems

KW - Transfer learning

KW - Bearing fault diagnosis

KW - Conditional distribution

KW - High-level features

KW - High-precision

KW - Low-level features

KW - Machine fault diagnosis

KW - Multiple scale

KW - Vibration data

KW - Deep learning

U2 - 10.1016/j.ress.2020.107050

DO - 10.1016/j.ress.2020.107050

M3 - Journal article

VL - 202

JO - Reliability Engineering and System Safety

JF - Reliability Engineering and System Safety

SN - 0951-8320

M1 - 107050

ER -