JZUS - Journal of Zhejiang University SCIENCE

Frontiers of Information Technology & Electronic Engineering

Accepted manuscript available online (unedited version)

Multisensor contrast neural network for prediction of remaining useful life of rolling bearing under scarce labeled data

Author(s): Binkun LIU^1, ^2, ^3, ⁴, Zhenyi XU^‡^1, ^2, ³, Yu KANG^3, ⁴, Yang CAO^3, ⁴, Yunbo ZHAO^‡^3, ⁴
Affiliation(s): ¹Jianghuai Advance Technology Center, Hefei, 230000, China ;²Key Laboratory of System Control and Information Processing, Ministry of Education, Shanghai 200240, China ;³Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, Hefei 230088, China ;⁴Department of Automation, University of Science and Technology of China, Hefei 230027, China
Corresponding email(s): xuzhenyi@mail.ustc.edu.cn, ybzhao@ustc.edu.cn
Key Words: Self-supervised; Remaining useful life prediction; Contrast Learning

Share this article to： More <<< Previous Paper \|Next Paper >>>

Binkun LIU^1,^2,^3,⁴, Zhenyi XU^‡^1,^2,³, Yu KANG^3,⁴, Yang CAO^3,⁴, Yunbo ZHAO^‡^3,⁴. Multisensor contrast neural network for prediction of remaining useful life of rolling bearing under scarce labeled data[J]. Frontiers of Information Technology & Electronic Engineering,in press.https://doi.org/10.1631/FITEE.2400753

@article{title="Multisensor contrast neural network for prediction of remaining useful life of rolling bearing under scarce labeled data",
author="Binkun LIU^1,^2,^3,⁴, Zhenyi XU^‡^1,^2,³, Yu KANG^3,⁴, Yang CAO^3,⁴, Yunbo ZHAO^‡^3,⁴",
journal="Frontiers of Information Technology & Electronic Engineering",
year="in press",
publisher="Zhejiang University Press & Springer",
doi="https://doi.org/10.1631/FITEE.2400753"
}

%0 Journal Article
%T Multisensor contrast neural network for prediction of remaining useful life of rolling bearing under scarce labeled data
%A Binkun LIU^1
%A^2
%A^3
%A⁴
%A Zhenyi XU^‡^1
%A^2
%A³
%A Yu KANG^3
%A⁴
%A Yang CAO^3
%A⁴
%A Yunbo ZHAO^‡^3
%A⁴
%J Frontiers of Information Technology & Electronic Engineering
%P
%@ 2095-9184
%D in press
%I Zhejiang University Press & Springer
doi="https://doi.org/10.1631/FITEE.2400753"

TY - JOUR
T1 - Multisensor contrast neural network for prediction of remaining useful life of rolling bearing under scarce labeled data
A1 - Binkun LIU^{1
A1 -}^{2
A1 -}^{3
A1 -}⁴
A1 - Zhenyi XU^‡^{1
A1 -}^{2
A1 -}³
A1 - Yu KANG^{3
A1 -}⁴
A1 - Yang CAO^{3
A1 -}⁴
A1 - Yunbo ZHAO^‡^{3
A1 -}⁴
J0 - Frontiers of Information Technology & Electronic Engineering
SP -
EP -
%@ 2095-9184
Y1 - in press
PB - Zhejiang University Press & Springer
ER -
doi="https://doi.org/10.1631/FITEE.2400753"

Abstract
Chinese Summary
Academic Network
Reviewer Comment

Abstract: Predicting remaining useful life (RUL) of bearing under scarce labeled data is significant for intelligent manufacturing. Current approaches typically encounter the challenge that different degradation stages have similar behaviors in multisensor scenarios. Given that cross-sensor similarity improves the discrimination of degradation features, we propose a multisensor contrast method for RUL prediction under scarce RUL-labeled data, in which we exploit cross-sensor similarity to mine multisensor similar representations that indicate machine health condition from rich unlabeled sensor data in co-occurrence space. Specifically, we use Resnet18 to span the features of different sensors into a co-occurrence space. We then obtain multisensor similar representations of abundant unlabeled data through alternate contrast based on cross-sensor similarity in co-occurrence space, which indicate the machine degradation stage. Finally, we fine-tune these similar representations with attention to achieve RUL prediction with limited labeled sensor data. The proposed method is evaluated on a publicly available bearing dataset, and the results show that the mean absolute percentage error is reduced by at least 0.058, and the score is improved by at least 0.122 compared to the state-of-the-art methods.

Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article

Reference

Open peer comments: Debate/Discuss/Question/Opinion

<1>

- Go to

Darkslateblue:Affiliate; Royal Blue:Author; Turquoise:Article

Reference