Research on bearing fault diagnosis based on multi-scale adaptive residual network combined with three-domain attention

Xiaobo Huang; Zhansi Jiang; Yulong Tan; Peixin Chen; Hui Jiang; Xiaobo Huang; Zhansi Jiang; Yulong Tan; Peixin Chen; Hui Jiang

doi:10.3934/jimo.2026110

Journal of Industrial and Management Optimization

2026, Volume 22, Issue 6: 2988-3022. doi: 10.3934/jimo.2026110

Previous Article Next Article

Research article Special Issues

Research on bearing fault diagnosis based on multi-scale adaptive residual network combined with three-domain attention

1.
School of Mechanical and Electrical Engineering, Guilin University of Electronic Technology, Guilin, 541000, China
2.
School of Future Transportation, Guangzhou Maritime University, Guangzhou, 510725, China
3.
School of Intelligent Manufacturing, Guangzhou Maritime University, Guangzhou, 510725, China

Received: 26 January 2026 Revised: 26 April 2026 Accepted: 15 May 2026 Published: 27 May 2026

To address the challenges of weak fault feature coupling, overlapping mixed faults, and low diagnosis accuracy for rolling bearings in strong noise environments, this paper presents a novel fault diagnosis method named multi-scale adaptive residual network with a three-domain attention mechanism (MSADC-TDA). The method enables robust end-to-end diagnosis under both single and mixed fault conditions. First, variational mode decomposition (VMD) is applied to adaptively denoise the vibration signal. Then, continuous wavelet transform (CWT) is employed to convert the one-dimensional signal into a two-dimensional time-frequency image for comprehensive fault feature extraction. An MSADC layer is proposed by combining multi-scale convolution and attention mechanisms. To learn the features of time-frequency images, MSADC can dynamically adjust the weights of convolutional layers at different scales. In addition, a TDA module is constructed to adaptively focus on extracting fault features. To enhance the feature learning capability of the proposed method, residual blocks are constructed using MSADC and residual connections. Multiple residual blocks are combined, and dropout layers are utilized to prevent overfitting. Experimental results on the self-built bearing dataset and Case Western Reserve University (CWRU) bearing dataset demonstrate that the proposed model maintains high diagnostic accuracy under strong noise and mixed fault interference, and possesses excellent generalization ability.
- multi-scale adaptive residual network,
- three-domain attention,
- fault diagnosis,
- deep learning,
- rolling bearings
Citation: Xiaobo Huang, Zhansi Jiang, Yulong Tan, Peixin Chen, Hui Jiang. Research on bearing fault diagnosis based on multi-scale adaptive residual network combined with three-domain attention[J]. Journal of Industrial and Management Optimization, 2026, 22(6): 2988-3022. doi: 10.3934/jimo.2026110

Related Papers:

Abstract

To address the challenges of weak fault feature coupling, overlapping mixed faults, and low diagnosis accuracy for rolling bearings in strong noise environments, this paper presents a novel fault diagnosis method named multi-scale adaptive residual network with a three-domain attention mechanism (MSADC-TDA). The method enables robust end-to-end diagnosis under both single and mixed fault conditions. First, variational mode decomposition (VMD) is applied to adaptively denoise the vibration signal. Then, continuous wavelet transform (CWT) is employed to convert the one-dimensional signal into a two-dimensional time-frequency image for comprehensive fault feature extraction. An MSADC layer is proposed by combining multi-scale convolution and attention mechanisms. To learn the features of time-frequency images, MSADC can dynamically adjust the weights of convolutional layers at different scales. In addition, a TDA module is constructed to adaptively focus on extracting fault features. To enhance the feature learning capability of the proposed method, residual blocks are constructed using MSADC and residual connections. Multiple residual blocks are combined, and dropout layers are utilized to prevent overfitting. Experimental results on the self-built bearing dataset and Case Western Reserve University (CWRU) bearing dataset demonstrate that the proposed model maintains high diagnostic accuracy under strong noise and mixed fault interference, and possesses excellent generalization ability.

References

[1]	M. Hakim, A. A. B. Omran, A N. Ahmed, M. Al-Waily, A. Abdellatif, A systematic review of rolling bearing fault diagnoses based on deep learning and transfer learning: Taxonomy, overview, application, open challenges, weaknesses and recommendations, Ain Shams Eng. J., 14 (2023), 101945. https://doi.org/10.1016/j.asej.2022.101945 doi: 10.1016/j.asej.2022.101945
[2]	Y. Liu, H. Xiang, Z, S. Jiang, J. W. Xiang, Second-order transient-extracting S transform for fault feature extraction in rolling bearings, Reliab. Eng. Syst. Safe., 230 (2023), 108955. https://doi.org/10.1016/j.ress.2022.108955 doi: 10.1016/j.ress.2022.108955
[3]	Y. Liu, H. Xiang, Z, S. Jiang, J. W. Xiang, Iterative synchrosqueezing-based general linear chirplet transform for time-frequency feature extraction, IEEE Trans. Instrum. Meas., 72 (2023), 1–11. https://doi.org/10.1109/TIM.2022.3232090 doi: 10.1109/TIM.2022.3232090
[4]	J. X. Gao, R. Pan, Y. P. Yuan, J. X. Zhou, H. Y. Ding, W. Y. Kong, A novel multiaxial fatigue life prediction method based on multi-scale physical neural network, Reliab. Eng. Syst. Saf., 269 (2026), 112049. https://doi.org/10.1016/j.ress.2025.112049 doi: 10.1016/j.ress.2025.112049
[5]	R. H. Wang, S. B. Sun, P. C. Zhao, X. G. Yang, X. J. Wei, C. Y. Hu, Multi-scale fusion network using time-division fourier transform for rolling bearing fault diagnosis, Comput. Mater. Con., 84 (2025), 3519–3539. https://doi.org/10.32604/cmc.2025.066212 doi: 10.32604/cmc.2025.066212
[6]	P. W. Tse, Y. A. Peng, R. Yam, Wavelet analysis and envelope detection for rolling element bearing fault diagnosis—their effectiveness and flexibilities, J. Vib. Acoust., 123 (2001), 303–310. https://doi.org/10.1115/1.1379745 doi: 10.1115/1.1379745
[7]	J. Gu, Y. X. Peng, H. Lu, X. D. Chang, G. A. Chen, A novel fault diagnosis method of rotating machinery via VMD, CWT and improved CNN, Measurement, 200 (2022), 111635. https://doi.org/10.1016/j.measurement.2022.111635 doi: 10.1016/j.measurement.2022.111635
[8]	Y. G. Lei, J. Lin, Z. G. He, M. J. Zuo, A review on empirical mode decomposition in fault diagnosis of rotating machinery, Mech. Syst. Signal Process., 35 (2013), 108–126. https://doi.org/10.1016/j.ymssp.2012.09.015 doi: 10.1016/j.ymssp.2012.09.015
[9]	M. R. Sethi, A. B. Subba, M. Faisal, S. Sahoo, D. K. Raju, Fault diagnosis of wind turbine blades with continuous wavelet transform based deep learning model using vibration signal, Eng. Appl. Artif. Intell., 138 (2024), 109372. https://doi.org/10.1016/j.engappai.2024.109372 doi: 10.1016/j.engappai.2024.109372
[10]	X. Yu, Z. T. Liang, Y. J. Wang, H. S. Yin, X. W. Liu, W. L. Yu, et al., A wavelet packet transform-based deep feature transfer learningmethod for bearing fault diagnosis under different working conditions, Measurement, 201 (2022), 111597. https://doi.org/10.1016/j.measurement.2022.111597 doi: 10.1016/j.measurement.2022.111597
[11]	J. L. Chen, Y. Y. Zi, Z. J. He, J. Yuan, Compound faults detection of rotating machinery using improved adaptive redundant lifting multiwavelet, Mech. Syst. Signal Process., 38 (2013), 36–54. https://doi.org/10.1016/j.ymssp.2012.06.025 doi: 10.1016/j.ymssp.2012.06.025
[12]	J. Y. Jiao, M. Zhao, J. Lin, K. Liang, A comprehensive review on convolutional neural network in machine fault diagnosis, Neurocomputing, 417 (2020), 36–63. https://doi.org/10.1016/j.neucom.2020.07.088 doi: 10.1016/j.neucom.2020.07.088
[13]	C. Liang, X. L. Mu, X. G. Zhang, F. F. Lu, C. C. Wang, Y. B. Shao, Enhanced fault diagnosis of rolling bearings using attention-augmented separable residual networks, Eng. Sci. Technol. Int. J., 61 (2025), 101930. https://doi.org/10.1016/j.jestch.2024.101930 doi: 10.1016/j.jestch.2024.101930
[14]	Y. N. Yan, Z. K. Liu, J. W. Xu, H. Zhang, N. Guo, L. Zhao, et al., A temperature-decoupled impedance-based mass sensing using CBAM-CNN and adaptive weighted average preprocessing with high accuracy, Mech. Syst. Signal Process., 213 (2024), 111347. https://doi.org/10.1016/j.ymssp.2024.111347 doi: 10.1016/j.ymssp.2024.111347
[15]	S. Y. Han, H. J. Choi, S. K. Choi, J. S. Oh, Fault diagnosis of planetary gear carrier packs: a class imbalance and multiclass classification problem, Int. J. Precis. Eng. Manuf., 20 (2019), 167–179. https://doi.org/10.1007/s12541-019-00082-4 doi: 10.1007/s12541-019-00082-4
[16]	J. Q. Cui, J. W. Gao, R. X. Xing, W. K. Wu, M. Y. Li, Y. P.Yang, Fault Diagnosis Method for Rolling Bearing Based on Attention Mechanism and BiTCN model, Digit. Signal Process., 167 (2025), 105454. https://doi.org/10.1016/j.dsp.2025.105454 doi: 10.1016/j.dsp.2025.105454
[17]	D. W. Hong, B. Kim, 1D convolutional neural network-based adaptive algorithm structure with system fault diagnosis and signal feature extraction for noise and vibration enhancement in mechanical systems, Mech. Syst. Signal Process., 197 (2023), 110395. https://doi.org/10.1016/j.ymssp.2023.110395 doi: 10.1016/j.ymssp.2023.110395
[18]	X. A. Yan, W. J. Yan, Y. D. Xu, K. V. Yuen, Machinery multi-sensor fault diagnosis based on adaptive multivariate feature mode decomposition and multi-attention fusion residual convolutional neural network, Mech. Syst. Signal Process., 202 (2023), 110664. https://doi.org/10.1016/j.ymssp.2023.110664 doi: 10.1016/j.ymssp.2023.110664
[19]	Q. Z. Wei, X. C. Tian, L. Cui, F. Q. Zheng, L. D. Liu, WSAFormer-DFFN: A model for rotating machinery fault diagnosis using 1D window-based multi-head self-attention and deep feature fusion network, Eng. Appl. artif. intell., 124 (2023), 106633. https://doi.org/10.1016/j.engappai.2023.106633 doi: 10.1016/j.engappai.2023.106633
[20]	Y. D. Xu, X. A. Yan, K. Feng, X. Sheng, B. B. Sun, Z. Liu, Attention-based multiscale denoising residual convolutional neural networks for fault diagnosis of rotating machinery, Reliab. Eng. Syst. Saf., 226 (2022), 108714. https://doi.org/10.1016/j.ress.2022.108714 doi: 10.1016/j.ress.2022.108714
[21]	L. J. Wang, Z. J. Wang, Z. X. Wu, X. F. Yang, Y. X. Xie, Rolling bearing fault diagnosis method based on improved multi-scale residual networks and light gradient boosting machine, J. Braz. Soc. Mech. Sci. Eng., 48 (2026), 104. https://doi.org/10.1007/s40430-025-06071-1 doi: 10.1007/s40430-025-06071-1
[22]	W. Huang, J. Cheng, Y. Yang, G. Guo, An improved deep convolutional neural network with multi-scale information for bearing fault diagnosis, Neurocomputing, 359 (2019), 77–92. https://doi.org/10.1016/j.neucom.2019.05.052 doi: 10.1016/j.neucom.2019.05.052
[23]	Q. X. Zhu, Y. S. Qian, N. Zhang, Y. L. He, Y. Xu, Multi-scale Transformer-CNN domain adaptation network for complex processes fault diagnosis, J. Process Control., 130 (2023), 103069. https://doi.org/10.1016/j.jprocont.2023.103069 doi: 10.1016/j.jprocont.2023.103069
[24]	Q. Song, P. Jiang, A multi-scale convolutional neural network based fault diagnosis model for complex chemical processes, Process Saf. Environ. Prot., 159 (2022), 575–584. https://doi.org/10.1016/j.psep.2021.11.020 doi: 10.1016/j.psep.2021.11.020
[25]	Z. D. Hei, H. Y. Yang, W. F. Sun, M. P. Zhong, G. H. Wang, A. Kumar, et al., Multiscale conditional adversarial networks based domain-adaptive method for rotating machinery fault diagnosis under variable working conditions, ISA Trans., 154 (2024), 352–370. https://doi.org/10.1016/j.isatra.2024.08.027 doi: 10.1016/j.isatra.2024.08.027
[26]	K. Liu, N. Y. Lu, F. Wu, R. D. Zhang, F. R. Gao, Model fusion and multiscale feature learning for fault diagnosis of industrial processes, IEEE Trans. Cybern., 53 (2022), 6465–6478. https://doi.org/10.1109/TCYB.2022.3176475 doi: 10.1109/TCYB.2022.3176475
[27]	Y. W. Sun, H. F. Tao, V. Stojanovic, Autoregressive data generation method based on wavelet packet transform and cascaded stochastic quantization for bearing fault diagnosis under unbalanced samples, Eng. Appl. Artif. Intell., 138 (2024), 109402. https://doi.org/10.1016/j.engappai.2024.109402 doi: 10.1016/j.engappai.2024.109402
[28]	V. Djordjevic, L. Dubonjic, M. M. Morato, D. Prsic, V. Stojanovic. Sensor fault estimation for hydraulic servo actuator based on sliding mode observer, Math. Model. Control., 2 (2022), 34–43. https://doi.org/10.3934/MMC.2022005 doi: 10.3934/MMC.2022005
[29]	Y. W. Sun, H. F. Tao, Y. Z. Ni, V. Stojanovic, A generic single-source domain generalization framework for fault diagnosis via wavelet packet augmentation and pseudo-domain generation, IEEE Internet Things J., 12 (2025), 31629–31642. https://doi.org/10.1109/JIOT.2025.3573752 doi: 10.1109/JIOT.2025.3573752
[30]	K. M. He, X. Y. Zhang, S. Q. Ren, J. Sun, Deep residual learning for image recognition, IEEE Conf. Comput. Vis. Pattern Recognit., 2016, 770–778. https://doi.org/10.48550/arXiv.1512.03385 doi: 10.48550/arXiv.1512.03385
[31]	G. C. Nie, Z. W. Zhang, Z. H. Jiao, Y. J. Li, M. Shao, X. Dai, A novel intelligent bearing fault diagnosis method based on image enhancement and improved convolutional neural network, Measurement, 242 (2025), 116148. https://doi.org/10.1016/j.measurement.2024.116148 doi: 10.1016/j.measurement.2024.116148
[32]	W. Zhang, G. L. Peng, C. H. Li, Y. H. Chen, Z. J. Zhang, A new deep learning model for fault diagnosis with good anti-noise and domain adaptation ability on raw vibration signals, Sensors, 17 (2017), 425. https://doi.org/10.3390/s17020425 doi: 10.3390/s17020425
[33]	B. Q. Hu, J. Liu, Y. Xu, A novel multi-scale convolutional neural network incorporating multiple attention mechanisms for bearing fault diagnosis, Measurement, 242 (2025), 115927. https://doi.org/10.1016/j.measurement.2024.115927 doi: 10.1016/j.measurement.2024.115927
[34]	L. Y. Wang, W. G. Zhao, An ensemble deep learning network based on 2D convolutional neural network and 1D LSTM with self-attention for bearing fault diagnosis, Appl. Soft Comput., 172 (2025), 112889. https://doi.org/10.1016/j.asoc.2025.112889 doi: 10.1016/j.asoc.2025.112889
[35]	Q. M. Shen, Z. Q. Zhang, Fault diagnosis method for bearing based on attention mechanism and multi-scale convolutional neural network, IEEE Access., 12 (2024), 12940–12952. https://doi.org/10.1109/ACCESS.2024.3357113 doi: 10.1109/ACCESS.2024.3357113
[36]	R. S. Qin, J. S. Zhao, Adaptive multiscale convolutional neural network model for chemical process fault diagnosis, Chin. J. Chem. Eng., 50 (2022), 398–411. https://doi.org/10.1016/j.cjche.2022.10.001 doi: 10.1016/j.cjche.2022.10.001

Reader Comments

Your name:*

Email:*
© 2026 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)