基于对比学习的滚动轴承早期故障在线检测方法

PDF(1699 KB)

振动与冲击 ›› 2023, Vol. 42 ›› Issue (14) : 229-236.

论文

王岩红1，温笑欢1 ，揭永琴1，王少伟2

作者信息 +

Online detection method for bearing incipient faults based on contrastive learning

WANG Yanhong1,WEN Xiaohuan1,JIE Yongqin1,WANG Shaowei2

Author information +

文章历史 +

摘要

早期故障检测不仅是故障诊断与设备健康管理的难点，也是工业生产实践中亟待攻坚的技术重点。本文在工业大数据驱动视域下，提出了一种基于对比学习的滚动轴承异常及早期故障在线检测方法。首先，建立一个基于深度可分离卷积及残差连接的深度编码器，在能有效提取信号特征的同时进一步降低模型参数量和计算量；其次，设置特定代理任务以实现基于对比学习方法的无监督编码器训练，使编码器胜任不同采样点信号之间差异的学习任务；最后，通过训练后的编码器对信号进行特征提取，并设计一种在线检测算法，该算法能够识别并区分滚动轴承的异常及早期故障。本文引入XJTU-SY数据集对上述方法进行验证，结果表明，与现有无监督故障检测方法相比，本方法准确性高，时效性强，丰富了不同工况下轴承全生命周期的健康管理方法。

Abstract

Incipient fault detection is not only an important role in prognostics and health management, but also has an increasing practical value in industry. This paper proposes a novel online detection method for bearing incipient fault based on contrastive learning from the perspective of industrial big data. Firstly, a novel deep encoder is proposed based on depth-wise separable convolution and residual connections, which is able to extract features from signal effectively. Then, contrastive learning method is utilized to train the proposed encoder by setting up a specific proxy task. The trained encoder is able to learn differences between signal samples. Finally, a novel online detection algorithm is proposed. The algorithm is able to detect abnormal condition and incipient fault effectively. Experiments are carried out on the XJTU-SY dataset. The results demonstrate our method outperforms existing ones and contributes to enriching bearing health management method under different working conditions.

导出引用

王岩红1，温笑欢1,揭永琴1，王少伟2. 基于对比学习的滚动轴承早期故障在线检测方法[J]. 振动与冲击, 2023, 42(14): 229-236

WANG Yanhong1,WEN Xiaohuan1,JIE Yongqin1,WANG Shaowei2. Online detection method for bearing incipient faults based on contrastive learning[J]. Journal of Vibration and Shock, 2023, 42(14): 229-236

参考文献

[1] Soualhi A, Razik H, Clerc G, et al. Prognosis of bearing failures using hidden Markov models and the adaptive neuro-fuzzy inference system[J]. IEEE Transactions on Industrial Electronics, 2013, 61(6): 2864-2874.
[2] 胡若晖, 张敏, 许文鑫. 基于DCGAN和DANN网络的滚动轴承跨域故障诊断[J]. 振动与冲击, 2022, 41(06): 21-29. DOI: 10.13465/j.cnki.jvs. 2022.06.004.
HU Ruo-hui, ZHANG Min, XU Wen-xin. Cross-domain fault diagnosis of rolling element bearings using DCGAN and DANN [J]. Journal of Vibration and Shock, 2022, 41(06): 21-29. DOI: 10.13465/j.cnki.jvs. 2022.06.004.
[3] Li P, Zhang Z, Grosu R, et al. An end-to-end neural network framework for state-of-health estimation and remaining useful life prediction of electric vehicle lithium batteries[J]. Renewable and Sustainable Energy Reviews, 2022, 156: 111843.
[4] Scheffer C, Girdhar P. Practical machinery vibration analysis and predictive maintenance[M]. Elsevier, 2004.
[5] 王玉静, 李少鹏, 康守强, 谢金宝, MIKULOVICH V I. 结合CNN和LSTM的滚动轴承剩余使用寿命预测方法[J]. 振动.测试与诊断, 2021, 41(03): 439-446+617.
WANG Yu-jing, LI Shao-peng, KANG Shou-qiang, et al. Method of Predicting Remaining Useful Life of Rolling Bearing Combining CNN and LSTM[J]. Journal of Vibration，Measurement & Diagnosis, 2021, 41(3): 439-446.
[6] Chen H, Jiang B, Lu N, et al. Deep PCA based real-time incipient fault detection and diagnosis methodology for electrical drive in high-speed trains[J]. IEEE Transactions on Vehicular Technology, 2018, 67(6): 4819-4830.
[7] 赵志宏, 李乐豪, 杨绍普, 李晴. 一种无监督的轴承健康指标及早期故障检测方法[J/OL]. 中国机械工程: 1-11 [2022-04-20]. http://kns.cnki.net/kcms/detail/42.1294.TH.20210915.1328.002.html
ZHAO Zhi-hong, LEE Le-hao, YANG Shao-pu, et al. An Unsupervised Bearing Health Indicator and Early Fault Detection Method[J/OL]. China Mechanical Engineering, 2022, 33(10): 1234-1243.
[8] Ding Y, Zhuang J, Ding P, et al. Self-supervised pretraining via contrast learning for intelligent incipient fault detection of bearings[J]. Reliability Engineering & System Safety, 2022, 218: 108126.
[9] 李巍华,何琛,陈祝云,黄如意,晋刚.基于对称式对比学习的齿轮箱无监督故障诊断方法[J].仪器仪表学报, 2022, 43(03): 121-131.
LI Wei-hua, He Chen, Chen Zhu-yun, et al. Unsupervised Fault Diagnosis of Gearbox Based on Symmetrical Contrast Learning. [J]. Chinese Journal of Scientific Instrument, 2022, 43(03): 121-131.
[10] Chen X, He K. Exploring simple siamese representation learning[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2021: 15750-15758.
[11] Sifre L, Mallat P S. Rigid-motion scattering for image classification author[J]. Ecole Polytechnique, CMAP PhD thesis, 2014.
[12] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 770-778.
[13] Wu Z, Xiong Y, Yu S X, et al. Unsupervised feature learning via non-parametric instance discrimination[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 3733-3742.
[14] Wang B, Lei Y, Li N, et al. A hybrid prognostics approach for estimating remaining useful life of rolling element bearings[J]. IEEE Transactions on Reliability, 2018, 69(1): 401-412.
[15] Yosinski J, Clune J, Bengio Y, et al. How transferable are features in deep neural networks?[J]. Advances in neural information processing systems, 2014, 27.
[16] Wang R, Chen H, Guan C, et al. Research on the fault monitoring method of marine diesel engines based on the manifold learning and isolation forest[J]. Applied Ocean Research, 2021, 112: 102681.
[17] Fernández-Francos D, Martínez-Rego D, Fontenla-Romero O, et al. Automatic bearing fault diagnosis based on one-class ν-SVM[J]. Computers \& Industrial Engineering, 2013, 64(1): 357-365.
[18] Qiu H, Lee J, Lin J, et al. Wavelet filter-based weak signature detection method and its application on rolling element bearing prognostics[J]. Journal of sound and vibration, 2006, 289(4-5): 1066-1090.