Operating condition domain generalization fault diagnosis method for bearings based on a multiscale convolutional neural network and the two-stage attention mechanism

PDF(3868 KB)

Journal of Vibration and Shock ›› 2025, Vol. 44 ›› Issue (2) : 267-278.

QIAO Huihui1, ZHAO Erxian1, HAO Rujiang*1, LIU Jie2, LIU Shuai1, WANG Yongchao1

Author information +

History +

Abstract

Under variable operating conditions, the training set and testing set of the deep learning-based train wheelset bearing fault diagnosis model usually come from different operating conditions, and the domain shift problem caused by the distribution difference of vibration signals under different operating conditions leads to the diagnosis accuracy reduction. The domain adaptation-based bearing fault diagnosis method under variable operating condition needs to obtain sample of the target domain to participate in the training process, which is infeasible in engineering practice. Thus the domain adaptation-based method is not suitable for bearing fault diagnosis of unseen conditions. Aiming at the above problems, an operating condition domain generalization fault diagnosis method based on multi-scale CNN and two-stage attention mechanism network model (TSAMCNN) is proposed in this paper. First, the multi-scale feature extraction module can extract richer fault information in the time-domain vibration signals from multiple scales. Then, the two-stage attention module can adaptively enhance the fault-sensitive features and suppress the condition-sensitive and useless features from both the channel and space dimensions. The model is ultimately capable of extracting condition-invariant fault features, which is the key to achieving the operating condition domain generalization fault diagnosis. Through the variable speed and the variable load bearing fault diagnosis experiments, it is proved that the proposed TSAMCNN model can improve the accuracy of bearing fault diagnosis, noise immunity performance and operating condition domain generalization ability. In addition, the weight vectors of the two-stage attention mechanism and the features extracted by each module of the model are visualised and analysed to improve the model interpretability.

Key words

train wheelset bearings / operating condition domain generalization fault diagnosis / convolutional neural network / multiscale feature extraction / attention mechanism

Cite this article

EndNote

Ris (Procite)

Bibtex

Download Citations

QIAO Huihui1, ZHAO Erxian1, HAO Rujiang1, LIU Jie2, LIU Shuai1, WANG Yongchao1. Operating condition domain generalization fault diagnosis method for bearings based on a multiscale convolutional neural network and the two-stage attention mechanism[J]. Journal of Vibration and Shock, 2025, 44(2): 267-278

References

[1] 李翠省, 廖英英, 刘永强. 基于EEMD和参数自适应VMD的高速列车轮对轴承故障诊断[J]. 振动与冲击, 2022, 41(01): 68-77.
LI Cuisheng, LIAO Yingying, LIU Yongqiang. Fault diagnosis of wheelset bearing of high-speed train based on EEMD and parameter adaptive VMD[J]. Vibration and Shock, 2022, 41(01): 68-77.
[2] 黄晨光, 林建辉, 易彩, 等. 延伸奇异值分解包及其在高速列车轮对轴承故障诊断中的应用[J]. 振动与冲击, 2020, 39(05): 45-56.
HUANG Chenguang, LIN Jianhui, YI Cai, et al. Extended singular value decomposition package and its application in high-speed train wheel set bearing fault diagnosis[J]. Vibration and Shock, 2020, 39(05): 45-56.
[3] 黄衍, 林建辉, 刘泽潮, 等. 基于自适应VMD的高速列车轴箱轴承故障诊断[J]. 振动与冲击, 2021, 40(03): 240-245.
HUANG Yan, LIN Jianhui, LIU Zechao, et al. High-speed train axle box bearing fault diagnosis based on adaptive VMD[J]. Vibration and Shock, 2021, 40(03): 240-245.
[4] 黄洪钟, 蔡吴斌, 杨圆鉴, 等. 动车组轴承及其可靠性研究进展[J]. 电子科技大学学报, 2023, 52(03): 458-474.
HUANG Hongzhong, CAI Wubin, YANG Yuanjian, et al. Research progress on EMU bearings and their reliability[J]. Journal of University of Electronic Science and Technology of China, 2023, 52(03): 458-474.
[5] 乔卉卉. 基于深度学习的复杂条件下旋转机械健康状态识别与预测[D]. 天津：天津大学机械工程学院, 2023.
QIAO Huihui. Identification and prediction of rotating machinery health status under complex conditions based on deep learning[D]. Tianjin: School of Mechanical Engineering, Tianjin University, 2023.
[6] LI X, ZHANG W, DING Q. Cross-domain fault diagnosis of rolling element bearings using deep generative neural networks[J]. IEEE Transactions on Industrial Electronics, 2018, 66(7): 5525-5534.
[7] LIU D, CHENG W, WEN W. Rolling bearing fault diagnosis via ConceFT-based time-frequency reconfiguration order spectrum analysis[J]. IEEE Access, 2018, 6: 67131-67143.
[8] 王金福, 李富才. 机械故障诊断技术中的信号处理方法: 时频分析[J]. 噪声与振动控制, 2013, 33(03): 198-202.
WANG Jinfu, LI Fucai. Signal processing method in mechanical fault diagnosis technology: time-frequency analysis[J]. Noise and Vibration Control, 2013, 33(03): 198-202.
[9] JI M, PENG G, HE J, et al. A two-stage, intelligent bearing-fault-diagnosis method using order-tracking and a one-dimensional convolutional neural network with variable speeds[J]. Sensors, 2021, 21(3): 675.
[10] DU Y, WANG A, WANG S, et al. Fault Diagnosis under Variable Working Conditions Based on STFT and Transfer Deep Residual Network[J]. Shock and Vibration, 2020: 1274380.
[11] SHI H, CHEN J, SI J ,et al. Fault Diagnosis of Rolling Bearings Based on a Residual Dilated Pyramid Network and Full Convolutional Denoising Autoencoder[J]. Sensors, 2020, 20: 5734.
[12] ZHAN C, ZIO E, SHEN, W. Domain generalization for cross-domain fault diagnosis: An application-oriented perspective and a benchmark study[J]. Reliability Engineering and System Safety, 2024, 245: 109964.
[13] LI X, ZHANG W, DING Q, et al. Multi-layer domain adaptation method for rolling bearing fault diagnosis[J]. Signal processing, 2019, 157: 180-197.
[14] 袁壮, 董瑞, 张来斌, 等. 深度领域自适应及其在跨工况故障诊断中的应用[J]. 振动与冲击, 2020, 39(12): 281-288.
YUAN Zhuang，DONG Rui，ZHANG Laibin，DUAN Lixiang. Deep domain adaptation and its application in fault diagnosis across working conditions[J]. Journal of Vibration and Shock, 2020, 39(12): 281-288.
[15] Ben-David S, Blitzer J, Crammer K, Pereira F. Analysis of representations for domain adaptation. Adv Neural Inf Process Syst 2007:137–44.
[16] CHEN L, LI Q, SHEN C, et al. Adversarial domain-invariant generalization: A generic domain-regressive framework for bearing fault diagnosis under unseen conditions[J]. IEEE Transactions on Industrial Informatics, 2021, 18(3): 1790-1800.
[17] ZHANG Q, ZHAO Z, ZHANG X, et al. Conditional adversarial domain generalization with a single discriminator for bearing fault diagnosis[J]. IEEE Transactions on Instrumentation and Measurement, 2021,70: 3514515.
[18] YANG W, WANG K. Domain Invariant Feature Learning Based on Cluster Contrastive Learning for Intelligence Fault Diagnosis With Limited Labeled Data[J]. IEEE Signal Processing Letters, 2023, 30: 1787-1791.
[19] JING L, ZHAO M, LI P, et al. A convolutional neural network based feature learning and fault diagnosis method for the condition monitoring of gearbox[J]. Measurement, 2017, 111: 1-10.
[20] VASWANI A, SHAZEER N, PARMAR N, et al. Attention is all you need[J]. Advances in neural information processing systems, 2017, 30.
[21] HU J, SHEN L, SUN G. Squeeze-and-excitation networks[C]// Proceedings of the IEEE conference on computer vision and pattern recognition, Salt Lake City: IEEE, 2018: 7132-7141.
[22] LI C, HOU Y, WANG P, et al. Joint distance maps based action recognition with convolutional neural networks[J]. IEEE Signal Processing Letters, 2017, 24(5): 624-628.
[23] WANG H, XU J, YAN R, et al. A new intelligent bearing fault diagnosis method using SDP representation and SE-CNN[J]. IEEE Transactions on Instrumentation and Measurement, 2019, 69(5): 2377-2389.
[24] QIAO H, WANG T, WANG P, et al. An adaptive weighted multiscale convolutional neural network for rotating machinery fault diagnosis under variable operating conditions[J]. IEEE Access, 2019, 7: 118954-118964.
[25] 丁雪, 邓艾东, 李晶, 等. 基于多尺度和注意力机制的滚动轴承故障诊断[J]. 东南大学学报(自然科学版), 2022, 52(1): 172-178.
DING Xue, DENG Aidong, LI Jing, et al. Rolling bearing fault diagnosis based on multi-scale and attention mechanism[J]. Journal of Southeast University (Natural Science Edition), 2022, 52(1): 172-178.
[26] HUANG Y J, LIAO A H, HU D Y, et al. Multi-scale convolutional network with channel attention mechanism for rolling bearing fault diagnosis[J]. Measurement, 2022, 203: 111935.
[27] PENG D, WANG H, DESMET W, et al. RMA-CNN: A residual mixed-domain attention CNN for bearings fault diagnosis and its time-frequency domain interpretability[J]. Journal of Dynamics, Monitoring and Diagnostics, 2023, 2: 115-132.
[28] SMITH W A, RANDALL R B. Rolling element bearing diagnostics using the Case Western Reserve University data: A benchmark study[J]. Mechanical systems and signal processing, 2015, 64: 100-131.