Time-Frequency Multi-Domain 1D Convolutional Neural Network with Channel-Spatial Attention for Noise-Robust Bearing Fault Diagnosis
Abstract
:1. Introduction
- It proposes a time-frequency multi-domain feature extraction and fusion model for accurate bearing fault diagnosis.
- The proposed model has a simple but effective 1D CNN architecture for low overhead. Furthermore, it uses minimum preprocessing of bearing-physics-informed envelop extraction and fast Fourier transform.
- The proposed model applied channel-wise and spatial-wise attention modules that enhanced the overall noise-robustness to be implemented under a strong noise and disturbance environment.
2. Background
2.1. Frequencies of Bearing Fault Signals
2.2. Envelope Extraction
2.3. Convolutional Block Attention Module (CBAM)
3. Proposed Network Model
3.1. Proposed Architecture Overview
3.2. Data Preprocessing
3.2.1. Dataset Augmentation
3.2.2. Envelope Extraction and Fast Fourier Transform
3.3. Multi-Domain Feature Extraction
3.3.1. Time Domain Feature Extraction Network
3.3.2. Frequency Domain Feature Extraction Network
3.4. Multi-Domain Feature Fusion
3.5. Attention Module for Noise Robustness
4. Experiment and Analysis
4.1. Network Parameters
4.2. Case Study 1: Case Western Reserve University (CWRU) Dataset
4.2.1. Dataset Description
4.2.2. Comparison Model
4.2.3. Experiment Results under No Noise
4.2.4. Experiment Results under Random Noise
t–SNE Analysis
4.3. Case Study: Paderborn University (PU) Dataset
4.3.1. Dataset Description
4.3.2. Comparison Study
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- Nandi, S.; Toliyat, H.A.; Li, X. Condition monitoring and fault diagnosis of electrical motors—A review. IEEE Trans. Energy Convers. 2005, 20, 719–729. [Google Scholar] [CrossRef]
- Randall, R.B.; Antoni, J. Rolling element bearing diagnostics—A tutorial. Mech. Syst. Signal Process. 2011, 25, 485–520. [Google Scholar] [CrossRef]
- Lebold, M.; McClintic, K.; Campbell, R.; Byington, C.; Maynard, K. Review of vibration analysis methods for gearbox diagnostics and prognostics. In Proceedings of the 54th Meeting of the Society for Machinery Failure Prevention Technology, Virginia Beach, VA, USA, 1–4 May 2000; Volume 634, p. 16. [Google Scholar]
- Konar, P.; Chattopadhyay, P. Bearing fault detection of induction motor using wavelet and Support Vector Machines (SVMs). Appl. Soft Comput. 2011, 11, 4203–4211. [Google Scholar] [CrossRef]
- Samanta, B.; Al-Balushi, K.; Al-Araimi, S. Artificial neural networks and support vector machines with genetic algorithm for bearing fault detection. Eng. Appl. Artif. Intell. 2003, 16, 657–665. [Google Scholar] [CrossRef]
- Safizadeh, M.; Latifi, S. Using multi-sensor data fusion for vibration fault diagnosis of rolling element bearings by accelerometer and load cell. Inf. Fusion 2014, 18, 1–8. [Google Scholar] [CrossRef]
- Wang, Z.; Zhang, Q.; Xiong, J.; Xiao, M.; Sun, G.; He, J. Fault diagnosis of a rolling bearing using wavelet packet denoising and random forests. IEEE Sensors J. 2017, 17, 5581–5588. [Google Scholar] [CrossRef]
- Zhang, J.; Wang, C.; Gui, P.; Wang, M.; Zou, T. State assessment of rolling bearings based on the multiscale bubble entropy. In Proceedings of the 2021 International Conference on Electronics, Circuits and Information Engineering (ECIE), Zhengzhou, China, 22–24 January 2021; pp. 179–182. [Google Scholar]
- Qin, X.; Li, Q.; Dong, X.; Lv, S. The fault diagnosis of rolling bearing based on ensemble empirical mode decomposition and random forest. Shock Vib. 2017, 2017, 2623081. [Google Scholar] [CrossRef]
- He, G.; Li, J.; Ding, K.; Zhang, Z. Feature extraction of gear and bearing compound faults based on vibration signal sparse decomposition. Appl. Acoust. 2022, 189, 108604. [Google Scholar] [CrossRef]
- Deng, W.; Li, Z.; Li, X.; Chen, H.; Zhao, H. Compound fault diagnosis using optimized MCKD and sparse representation for rolling bearings. IEEE Trans. Instrum. Meas. 2022, 71, 1–9. [Google Scholar] [CrossRef]
- Hamadache, M.; Jung, J.H.; Park, J.; Youn, B.D. A comprehensive review of artificial intelligence-based approaches for rolling element bearing PHM: Shallow and deep learning. JMST Adv. 2019, 1, 125–151. [Google Scholar] [CrossRef]
- Ren, L.; Sun, Y.; Wang, H.; Zhang, L. Prediction of bearing remaining useful life with deep convolution neural network. IEEE Access 2018, 6, 13041–13049. [Google Scholar] [CrossRef]
- Yu, H.; Wang, K.; Li, Y.; He, M. Deep subclass reconstruction network for fault diagnosis of rotating machinery under various operating conditions. Appl. Soft Comput. 2021, 112, 107755. [Google Scholar] [CrossRef]
- Zuo, L.; Xu, F.; Zhang, C.; Xiahou, T.; Liu, Y. A multi-layer spiking neural network-based approach to bearing fault diagnosis. Reliab. Eng. Syst. Saf. 2022, 225, 108561. [Google Scholar] [CrossRef]
- Jin, Y.; Qin, C.; Huang, Y.; Liu, C. Actual bearing compound fault diagnosis based on active learning and decoupling attentional residual network. Measurement 2021, 173, 108500. [Google Scholar] [CrossRef]
- Wang, X.; Mao, D.; Li, X. Bearing fault diagnosis based on vibro-acoustic data fusion and 1D-CNN network. Measurement 2021, 173, 108518. [Google Scholar] [CrossRef]
- Eren, L.; Ince, T.; Kiranyaz, S. A generic intelligent bearing fault diagnosis system using compact adaptive 1D CNN classifier. J. Signal Process. Syst. 2019, 91, 179–189. [Google Scholar] [CrossRef]
- Chen, X.; Zhang, B.; Gao, D. Bearing fault diagnosis base on multi-scale CNN and LSTM model. J. Intell. Manuf. 2021, 32, 971–987. [Google Scholar] [CrossRef]
- Huang, W.; Cheng, J.; Yang, Y.; Guo, G. An improved deep convolutional neural network with multi-scale information for bearing fault diagnosis. Neurocomputing 2019, 359, 77–92. [Google Scholar] [CrossRef]
- Ji, M.; Peng, G.; He, J.; Liu, S.; Chen, Z.; Li, S. A two-stage, intelligent bearing-fault-diagnosis method using order-tracking and a one-dimensional convolutional neural network with variable speeds. Sensors 2021, 21, 675. [Google Scholar] [CrossRef]
- Ding, X.; Wang, H.; Cao, Z.; Liu, X.; Liu, Y.; Huang, Z. An Edge Intelligent Method for Bearing Fault Diagnosis Based on a Parameter Transplantation Convolutional Neural Network. Electronics 2023, 12, 1816. [Google Scholar] [CrossRef]
- Bertocco, M.; Fort, A.; Landi, E.; Mugnaini, M.; Parri, L.; Peruzzi, G.; Pozzebon, A. Roller bearing failures classification with low computational cost embedded machine learning. In Proceedings of the 2022 IEEE International Workshop on Metrology for Automotive (MetroAutomotive), Modena, Italy, 4–6 July 2022; pp. 12–17. [Google Scholar]
- Li, B.; Chow, M.Y.; Tipsuwan, Y.; Hung, J.C. Neural-network-based motor rolling bearing fault diagnosis. IEEE Trans. Ind. Electron. 2000, 47, 1060–1069. [Google Scholar] [CrossRef]
- Wang, J.; Mo, Z.; Zhang, H.; Miao, Q. A deep learning method for bearing fault diagnosis based on time-frequency image. IEEE Access 2019, 7, 42373–42383. [Google Scholar] [CrossRef]
- Note, S.F.A. Rolling Element Bearings; REB, Sales Technology: League City, TX, USA, 2012. [Google Scholar]
- Huang, T.; Fu, S.; Feng, H.; Kuang, J. Bearing fault diagnosis based on shallow multi-scale convolutional neural network with attention. Energies 2019, 12, 3937. [Google Scholar] [CrossRef]
- Ding, X.; He, Q. Energy-fluctuated multiscale feature learning with deep convnet for intelligent spindle bearing fault diagnosis. IEEE Trans. Instrum. Meas. 2017, 66, 1926–1935. [Google Scholar] [CrossRef]
- Chen, H.Y.; Lee, C.H. Vibration signals analysis by explainable artificial intelligence (XAI) approach: Application on bearing faults diagnosis. IEEE Access 2020, 8, 134246–134256. [Google Scholar] [CrossRef]
- Wang, H.; Xu, J.; Yan, R.; Gao, R.X. A new intelligent bearing fault diagnosis method using SDP representation and SE-CNN. IEEE Trans. Instrum. Meas. 2019, 69, 2377–2389. [Google Scholar] [CrossRef]
- Chen, Z.; Mauricio, A.; Li, W.; Gryllias, K. A deep learning method for bearing fault diagnosis based on cyclic spectral coherence and convolutional neural networks. Mech. Syst. Signal Process. 2020, 140, 106683. [Google Scholar] [CrossRef]
- Gao, D.; Zhu, Y.; Wang, X.; Yan, K.; Hong, J. A fault diagnosis method of rolling bearing based on complex morlet CWT and CNN. In Proceedings of the 2018 prognostics and system health management conference (PHM-Chongqing), Chongqing, China, 26–28 October 2018; pp. 1101–1105. [Google Scholar]
- Du, J.; Li, X.; Gao, Y.; Gao, L. Integrated gradient-based continuous wavelet transform for bearing fault diagnosis. Sensors 2022, 22, 8760. [Google Scholar] [CrossRef]
- Ruan, D.; Wang, J.; Yan, J.; Gühmann, C. CNN parameter design based on fault signal analysis and its application in bearing fault diagnosis. Adv. Eng. Inform. 2023, 55, 101877. [Google Scholar] [CrossRef]
- Sun, J.; Wen, J.; Yuan, C.; Liu, Z.; Xiao, Q. Bearing fault diagnosis based on multiple transformation domain fusion and improved residual dense networks. IEEE Sensors J. 2021, 22, 1541–1551. [Google Scholar] [CrossRef]
- Liu, X.; Sun, W.; Li, H.; Hussain, Z.; Liu, A. The Method of Rolling Bearing Fault Diagnosis Based on Multi-Domain Supervised Learning of Convolution Neural Network. Energies 2022, 15, 4614. [Google Scholar] [CrossRef]
- Dong, K.; Lotfipoor, A. Intelligent Bearing Fault Diagnosis Based on Feature Fusion of One-Dimensional Dilated CNN and Multi-Domain Signal Processing. Sensors 2023, 23, 5607. [Google Scholar] [CrossRef]
- Wang, H.; Xu, J.; Yan, R.; Sun, C.; Chen, X. Intelligent bearing fault diagnosis using multi-head attention-based CNN. Procedia Manuf. 2020, 49, 112–118. [Google Scholar] [CrossRef]
- Huang, Y.J.; Liao, A.H.; Hu, D.Y.; Shi, W.; Zheng, S.B. Multi-scale convolutional network with channel attention mechanism for rolling bearing fault diagnosis. Measurement 2022, 203, 111935. [Google Scholar] [CrossRef]
- Hao, Y.; Wang, H.; Liu, Z.; Han, H. Multi-scale CNN based on attention mechanism for rolling bearing fault diagnosis. In Proceedings of the 2020 Asia-Pacific International Symposium on Advanced Reliability and Maintenance Modeling (APARM), Vancouver, BC, Canada, 20–23 August 2020; pp. 1–5. [Google Scholar]
- Plakias, S.; Boutalis, Y.S. Fault detection and identification of rolling element bearings with Attentive Dense CNN. Neurocomputing 2020, 405, 208–217. [Google Scholar] [CrossRef]
- Woo, S.; Park, J.; Lee, J.Y.; Kweon, I.S. Cbam: Convolutional block attention module. In Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, 8–14 September 2018; pp. 3–19. [Google Scholar]
- Poddar, S.; Chandravanshi, M.L. Ball bearing fault detection using vibration parameters. Int. J. Eng. Res. Technol. 2013, 2, 1239–1244. [Google Scholar]
- Vaswani, A.; Shazeer, N.; Parmar, N.; Uszkoreit, J.; Jones, L.; Gomez, A.N.; Kaiser, Ł.; Polosukhin, I. Attention is all you need. Adv. Neural Inf. Process. Syst. 2017, 30. [Google Scholar]
- Guo, M.H.; Xu, T.X.; Liu, J.J.; Liu, Z.N.; Jiang, P.T.; Mu, T.J.; Zhang, S.H.; Martin, R.R.; Cheng, M.M.; Hu, S.M. Attention mechanisms in computer vision: A survey. Comput. Vis. Media 2022, 8, 331–368. [Google Scholar] [CrossRef]
- Case Western Reserve University Bearing Data Center. Available online: https://www.kaggle.com/datasets/brjapon/cwru-bearing-datasets (accessed on 12 September 2023).
- Springenberg, J.T.; Dosovitskiy, A.; Brox, T.; Riedmiller, M. Striving for simplicity: The all convolutional net. arXiv 2014, arXiv:1412.6806. [Google Scholar]
- Lin, M.; Chen, Q.; Yan, S. Network in network. arXiv 2013, arXiv:1312.4400. [Google Scholar]
- Zhang, W.; Peng, G.; Li, C.; Chen, Y.; Zhang, Z. A new deep learning model for fault diagnosis with good anti-noise and domain adaptation ability on raw vibration signals. Sensors 2017, 17, 425. [Google Scholar] [CrossRef]
- Paderborn University Bearing Data Center. Available online: https://mb.uni-paderborn.de/en/kat/main-research/datacenter/bearing-datacenter (accessed on 12 September 2023).
Conv. Layer | SNR [dB] | |||||||
---|---|---|---|---|---|---|---|---|
None | 6 | 4 | 2 | 0 | −2 | −4 | −6 | |
1st layer | 100.00 | 99.83 | 99.42 | 98.37 | 95.99 | 89.71 | 79.36 | 64.77 |
2nd layer | 100.00 | 99.65 | 99.13 | 97.79 | 95.06 | 89.13 | 76.80 | 61.16 |
3rd layer | 100.00 | 99.53 | 99.07 | 97.91 | 95.35 | 90.87 | 78.84 | 62.85 |
4th layer | 100.00 | 99.77 | 99.36 | 98.26 | 95.93 | 90.06 | 77.73 | 60.17 |
5th layer | 100.00 | 99.77 | 99.36 | 98.31 | 95.41 | 90.23 | 79.24 | 62.56 |
TD-CNN | FD-CNN | ||||||||
---|---|---|---|---|---|---|---|---|---|
Layer |
Kernel Size
/Stride |
No. of
Kernels |
Output
Shape | Padding | Layer |
Kernel Size
/Stride |
No. of
Kernels |
Output
Shape | Padding |
Conv1 | 256/5 | 4 | 154 × 4 | - | Conv1 | 256/5 | 4 | 58 × 4 | 2 |
Conv2 | 15/3 | 8 | 47 × 8 | - | CBAM | 64/1 | 4 | 58 × 4 | - |
Conv3 | 7/2 | 16 | 21 × 16 | - | Conv2 | 5/2 | 8 | 29 × 8 | 2 |
Conv4 | 7/2 | 32 | 8 × 32 | - | Conv3 | 5/2 | 16 | 15 × 16 | 2 |
Conv5 | 3/1 | 64 | 6 × 64 | - | Conv4 | 5/2 | 32 | 8 × 32 | 2 |
GAP | - | - | 64 × 1 | - | Conv5 | 5/2 | 64 | 4 × 64 | 2 |
- | GAP | - | - | 64 × 1 | - | ||||
DFMLP | |||||||||
Layer | Input shape | Output shape | |||||||
FC1: Domain fusion | 128 × 1 | 64 × 1 | |||||||
FC2 | 64 × 1 | 32 × 1 | |||||||
FC3 | 32 × 1 | 10 × 1 |
Inner Race | Outer Race | Ball | |
---|---|---|---|
Defect freq. [Hz] |
Class No. | Fault Location | Fault Size [mils] | Working Loads [HP] | Train Samples | Valid Samples | Test Samples | |||
---|---|---|---|---|---|---|---|---|---|
C0 | Normal | - | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C1 | Inner race fault | 7 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C2 | Inner race fault | 14 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C3 | Inner race fault | 21 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C4 | Outer race fault | 7 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C5 | Outer race fault | 14 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C6 | Outer race fault | 21 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C7 | Ball fault | 7 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C8 | Ball fault | 14 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 | ||||||
C9 | Ball fault | 21 | 0 | 129 | 516 | 43 | 172 | 43 | 172 |
1 | 129 | 43 | 43 | ||||||
2 | 129 | 43 | 43 | ||||||
3 | 129 | 43 | 43 |
Model | Attention Module | No. of Params | Accuracy [%] |
---|---|---|---|
TF-MD+NA | None | 38,162 | 99.94 |
TF-MDA (SA) | Spatial attention only | 38,293 | 100 |
TF-MDA (CA) | Channel attention only | 38,188 | 99.94 |
TF-MDA (CSA) | Spatial–Channel Attention | 38,319 | 100 |
Model | No. of Params | Accuracy [%] |
---|---|---|
WDCNN [49] | 38,162 | 99.95 |
MCNN–LSTM [19] | 73,480 | 97.97 |
TD–CNN | 14,910 | 99.24 |
FD–CNN | 17,406 | 99.94 |
TF–MDA (CSA) | 38,319 | 100 |
Model Name | SNR [dB] | ||||||
---|---|---|---|---|---|---|---|
6 | 4 | 2 | 0 | −2 | −4 | −6 | |
TF–MD+NA | 99.88 | 99.65 | 98.78 | 96.74 | 90.81 | 81.86 | 66.40 |
TF–MDA (SA) | 99.83 | 99.59 | 98.84 | 96.05 | 90.93 | 77.44 | 64.13 |
TF–MDA (CA) | 99.88 | 99.77 | 99.13 | 96.86 | 90.99 | 80.76 | 66.63 |
TF–MDA (CSA) | 99.88 | 99.71 | 99.36 | 98.37 | 95.06 | 85.87 | 70.93 |
Model | SNR [dB] | No. of Params | ||||||
---|---|---|---|---|---|---|---|---|
6 | 4 | 2 | 0 | −2 | −4 | −6 | ||
WDCNN [49] | 99.62 | 97.77 | 92.88 | 86.74 | 79.67 | 69.40 | 59.29 | 54,510 |
MCNN–LSTM [19] | 78.18 | 68.54 | 57.97 | 50.73 | 45.05 | 41.09 | 37.29 | 73,480 |
TD–CNN | 97.62 | 95.87 | 89.07 | 81.22 | 69.24 | 57.15 | 46.98 | 14,910 |
FD–CNN | 99.71 | 99.19 | 98.14 | 94.94 | 87.67 | 77.27 | 62.97 | 17,406 |
TF–MDA (CSA) | 99.88 | 99.71 | 99.36 | 98.37 | 95.06 | 85.87 | 70.93 | 38,319 |
Description | Nominal Torque | Nominal Speed | Nominal Current | Pole Pair |
---|---|---|---|---|
Value | 1.35 [Nm] | 3000 [rpm] | 2.3 [A] | 4 |
No. | Rotational Speed [rpm] | Load Torque [Nm] | Radial Force [N] |
---|---|---|---|
OC 1 | 1500 | 0.7 | 1000 |
OC 2 | 1500 | 0.1 | 1000 |
OC 3 | 1500 | 0.7 | 400 |
Notation | Meaning | Value |
---|---|---|
D | Pitch circle diameter | 28.55 [mm] |
d | Rolling element diameter | 6.75 [mm] |
Z | No. of rolling elements | 8 |
Contact angle | 0° |
Inner Race | Outer Race | |
---|---|---|
Defect freq. [Hz] |
Class No. | Bearing Code | Fault Generation | Damage Method /Fault Type | Fault Location | Fault Level | Operating Condition |
---|---|---|---|---|---|---|
C0 | K001 | Normal | - | - | OC 1, 2, 3 | |
C1 | KI01 | Artificial | EDM | Inner race fault | 1 | |
C2 | KI07 | Artificial | electric engraver | Inner race fault | 2 | |
C3 | KI17 | Real | fatigue | Inner race fault | 1 | |
C4 | KI18 | Real | fatigue | Inner race fault | 2 | |
C5 | KA01 | Artificial | EDM | outer race fault | 1 | |
C6 | KA03 | Artificial | electric engraver | outer race fault | 2 | |
C7 | KA15 | Real | Plastic deform | outer race fault | 1 | |
C8 | KA16 | Real | fatigue | outer race fault | 2 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Kim, Y.; Kim, Y.-K. Time-Frequency Multi-Domain 1D Convolutional Neural Network with Channel-Spatial Attention for Noise-Robust Bearing Fault Diagnosis. Sensors 2023, 23, 9311. https://0-doi-org.brum.beds.ac.uk/10.3390/s23239311
Kim Y, Kim Y-K. Time-Frequency Multi-Domain 1D Convolutional Neural Network with Channel-Spatial Attention for Noise-Robust Bearing Fault Diagnosis. Sensors. 2023; 23(23):9311. https://0-doi-org.brum.beds.ac.uk/10.3390/s23239311
Chicago/Turabian StyleKim, Yejin, and Young-Keun Kim. 2023. "Time-Frequency Multi-Domain 1D Convolutional Neural Network with Channel-Spatial Attention for Noise-Robust Bearing Fault Diagnosis" Sensors 23, no. 23: 9311. https://0-doi-org.brum.beds.ac.uk/10.3390/s23239311