Method for Bearing Fault Quantitative Diagnosis Based on MTF and Improved Residual Network

doi:10.12068/j.issn.1005-3026.2024.05.012

Abstract

Abstract:

Different from the current rolling bearing fault diagnosis which mainly focuses on the qualitative analysis stage， a quantitative fault diagnosis method for rolling bearings based on image classification is proposed. The overlapping sampling method is used to enhance the one?dimensional time series data， and then the Markov transition field （MTF） method is used to convert the one?dimensional time series data into two?dimensional images， which provide two?dimensional image samples for inputting into the neural network model and retain the time?domain information. The ResNeXt and ResNeSt modified residual networks with fine?tuning processing based on transfer learning are built and trained to classify fault images and realize fault diagnosis. The confusion matrix method and t?distributed stochastic neighbor embedding（t?SNE） visualization method are used to carry out experiments. The results show that the proposed method for rolling bearing fault diagnosis can realize the quantitative diagnosis of multi?working condition rolling bearing fault， and has higher diagnosis accuracy and faster training speed.

Key words: bearing fault, Markov transition field （MTF）, residual network, transfer learning, quantitative diagnosis

CLC Number:

TP 277

Ling-xuan LI, Zhen-wei MA, Ze-jun YU, Zhuang XING. Method for Bearing Fault Quantitative Diagnosis Based on MTF and Improved Residual Network[J]. Journal of Northeastern University(Natural Science), 2024, 45(5): 697-706.

Figures/Tables 18

Fig.1 Method flow chart

Fig.2 Schematic diagram of overlapping sampling

Fig.3 MTF transition process

Fig.4 Inception block

Fig.5 General residual connection structure

Fig.6 ResNeXt residual block structure

Table 1 ResNeXt network structure

模块	ResNeXt-50(32×4d)	输出
conv1	$7 × 7$ ,64, 步长2	112 $×$ 112
conv1	$3 × 3$ ,最大池化层, 步长2	112 $×$ 112
conv2	$1 × 1, 128 3 × 3, 128 C = 32 1 × 1, 256 × 3$	56 $×$ 56
conv3	$1 × 1, 256 3 × 3, 256 C = 32 1 × 1, 512 × 4$	28 $×$ 28
conv4	$1 × 1, 512 3 × 3, 512 C = 32 1 × 1, 1 ? 024 × 6$	14 $×$ 14
conv5	$1 × 1, 1 ? 024 3 × 3, 1 ? 024 C = 32 1 × 1, 2 ? 048 × 3$	7 $×$ 7
分类器	全局平均池化层，9维全连接层，Softmax	1 $×$ 1

Table 1 ResNeXt network structure

模块	ResNeXt-50(32×4d)	输出
conv1	$7 × 7$ ,64, 步长2	112 $×$ 112
conv1	$3 × 3$ ,最大池化层, 步长2	112 $×$ 112
conv2	$1 × 1, 128 3 × 3, 128 C = 32 1 × 1, 256 × 3$	56 $×$ 56
conv3	$1 × 1, 256 3 × 3, 256 C = 32 1 × 1, 512 × 4$	28 $×$ 28
conv4	$1 × 1, 512 3 × 3, 512 C = 32 1 × 1, 1 ? 024 × 6$	14 $×$ 14
conv5	$1 × 1, 1 ? 024 3 × 3, 1 ? 024 C = 32 1 × 1, 2 ? 048 × 3$	7 $×$ 7
分类器	全局平均池化层，9维全连接层，Softmax	1 $×$ 1

Fig.7 ResNeSt residual block structure

Fig.8 Split attention layer structure

Fig.9 Schematic diagram of fine‐tuning strategy

Table 2 Three working conditions corresponding

电动机负载/kW	电动机转速/（r·min^-1）
0.735	1 772
1.470	1 750
2.206	1 730

Table 3 Division of bearing fault data set

状态	故障程度/mm	训练集	验证集	类别标签
正常	—	400	50	0
滚动体	0.177 8	400	50	1
	0.355 6	400	50	2
	0.533 4	400	50	3
内圈	0.177 8	400	50	4
	0.355 6	400	50	5
	0.533 4	400	50	6
外圈	0.177 8	400	50	7
外圈	0.355 6	400	50	8

Fig.10 Legends for MTF transition

Fig.11 MTF‐ResNeXt model loss and model accuracy

Fig.12 MTF‐ResNeSt model loss and model accuracy

Fig.13 Confusion matrix

Fig.14 T‐SNE visualization of MTF‐ResNeXt

Fig.15 T‐SNE visualization of MTF‐ResNeSt

References 20

1	Cocconcelli M， Zimroz R， Rubini R，et al.STFT based approach for ball bearing fault detection in a varying speed motor［C］//2nd International Conference on Condition Monitoring of Machinery in Non‑stationary Operations.Hammamet：Springer，2012：41-50.
2	Sun W， Yang G A， Chen Q，et al.Fault diagnosis of rolling bearing based on wavelet transform and envelope spectrum correlation［J］.Journal of Vibration and Control，2013，19（6）：924-941.
3	Huang N E， Shen Z， Long S R，et al.The empirical mode decomposition and the Hilbert spectrum for nonlinear and non‐stationary time series analysis ［J］.Proceedings of the Royal Society of London.Series A：Mathematical，Physical and Engineering Sciences，1998，454（1971）：903-995.
4	Yu D J， Cheng J S， Yang Y.Application of EMD method and Hilbert spectrum to the fault diagnosis of roller bearings［J］.Mechanical Systems and Signal Processing，2005，19（2）：259-270.
5	甄冬，田少宁，郭俊超，等.改进型EEMD和MSB解调方法及其在轴承故障特征提取中的应用［J］.振动工程学报，2023，36（5）：1447-1456.
	Zhen Dong， Tian Shao‐ning， Guo Jun‐chao，et al.An improved decomposition method using EEMD and MSB and its application in rolling bearing fault feature extraction［J］.Journal of Vibration Engineering，2023，36（5）：1447-1456.
6	Wu Z H， Huang N E.Ensemble empirical mode decomposition：a noise‐assisted data analysis method［J］.Advances in Adaptive Data Analysis，2009，1（1）：1-41.
7	崔玲丽，刘银行，王鑫.基于改进奇异值分解的滚动轴承微弱故障特征提取方法［J］.机械工程学报，2022，58（17）：156-169.
	Cui Ling‐li， Liu Yin‐hang， Wang Xin.Feature extraction of weak fault for rolling bearing based on improved singular value decomposition［J］.Journal of Mechanical Engineering，2022，58（17）：156-169.
8	Gu R， Chen J， Hong R J，et al.Incipient fault diagnosis of rolling bearings based on adaptive variational mode decomposition and Teager energy operator［J］.Measurement，2020，149：106941.
9	Shenfield A， Howarth M.A novel deep learning model for the detection and identification of rolling element‐bearing fault［J］.Sensors，2020，20（18）：5112.
10	Fu W L， Zhou J Z， Zhang Y C.Fault diagnosis for rolling element bearings with VMD time‑frequency analysis and SVM［C］//2015 Fifth International Conference on Instrumentation and Measurement，Computer，Communication and Control （IMCCC）.Qinhuangdao，2015：69-72.
11	路敦利，宁芊，杨晓敏.KNN-朴素贝叶斯算法的滚动轴承故障诊断［J］.计算机测量与控制，2018，26（6）：21-23，27.
	Lu Dun‐li， Ning Qian， Yang Xiao‐min.Fault diagnosis of rolling bearing based on KNN‐naive Bayesian algorithm［J］.Computer Measurement and Control，2018，26（6）：21-23，27.
12	Eren L， Ince T， Kiranyaz S.A generic intelligent bearing fault diagnosis system using compact adaptive 1D CNN classifier［J］.Journal of Signal Processing Systems，2019，91（2）：179-189.
13	Liu C Y， Gryllias K.Simulation‐driven domain adaptation for rolling element bearing fault diagnosis［J］.IEEE Transactions on Industrial Informatics，2022，18（9）：5760-5770.
14	王太勇，王廷虎，王鹏，等.基于注意力机制BiLSTM的设备智能故障诊断方法［J］.天津大学学报（自然科学与工程技术版），2020，53（6）：601-608.
	Wang Tai‐yong， Wang Ting‐hu， Wang Peng，et al.An intelligent fault diagnosis method based on attention‐based bidirectional LSTM network［J］.Journal of Tianjin University（Science and Technology），2020，53（6）：601-608
15	张伟.基于卷积神经网络的轴承故障诊断算法研究［D］.哈尔滨：哈尔滨工业大学，2017.
	Zhang Wei.Study on bearing fault diagnosis algorithm based on convolutional neural network［D］.Harbin：Harbin Institute of Technology，2017.
16	Wang Z G， Oates T.Encoding time series as images for visual inspection and classification using tiled convolutional neural networks［C］//Workshops at the Twenty‐Ninth AAAI Conference on Artificial Intelligence.Austin，2015：275970614.
17	Xie S N， Girshick R， Dollár P，et al.Aggregated residual transformations for deep neural networks［C］//2017 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）.Honolulu：IEEE，2017：5987-5995.
18	Zhang H， Wu C R， Zhang Z Y，et al.ResNeSt：split‐attention networks［C］// 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshop（CVPR）.New Orleans：IEEE，2022：2735-2745.
19	Yosinski J， Clune J， Bengio Y，et al.How transferable are features in deep neural networks［C］//28th Conference on Neural Information Processing Systems （NIPS）.Montreal，2014：8-13.
20	Case Western Reserve University Bearing Data Center.Data file［EB/OL］.（2018-06-10）［2021-03-10］..

[1]	ZHAO Hai， WANG Xiang， SHI Han， CHEN Jia-wei. A Transfer Learning Framework for EEG Emotion Recognition [J]. Journal of Northeastern University(Natural Science), 2023, 44(7): 913-921.
[2]	ZHAO Hai， CHEN Jia-wei， SHI Han， WANG Xiang. A Transfer Learning Algorithm Applied to Human Activity Recognition [J]. Journal of Northeastern University(Natural Science), 2022, 43(6): 776-782.
[3]	HOU Dong-xiao， MU Jin-tao， FANG Cheng， SHI Pei-ming. Fault Diagnosis of Variable Speed Bearings Based on GADF and ResNet34 Introduced Transfer Learning [J]. Journal of Northeastern University(Natural Science), 2022, 43(3): 383-389.
[4]	WANG Xin-gang， HAN Kai-zhong， WANG Chao， LI Lin. Bearing Remaining Useful Life Prediction Method Based on Transfer Learning [J]. Journal of Northeastern University(Natural Science), 2021, 42(5): 665-672.
[5]	WANG Shu， GUAN Zhan-xu， WANG Jing， SUN Xiao-hui. Bayesian Network Parameter Learning Method Based on Transfer Learning [J]. Journal of Northeastern University(Natural Science), 2021, 42(4): 509-515.
[6]	FANG Liang， ZHOU Yun， TANG Zhi-quan. Image Classification of Corroded Steel Reinforcement Based on Optimized Residual Network [J]. Journal of Northeastern University(Natural Science), 2021, 42(11): 1625-1633.
[7]	CHANG Yu-qing， ZHAO Wei-wei， LIU Le-yuan， KANG Xiao-yun. Modeling of Coal Mill Process Monitoring Based on Instance-based Transfer Learning [J]. Journal of Northeastern University(Natural Science), 2021, 42(10): 1369-1375.
[8]	WANG Xin， WANG Cui-rong， WANG Cong， YUAN Ying. Dual-channel Multi-perception Convolutional Network for Image Super-Resolution [J]. Journal of Northeastern University Natural Science, 2020, 41(11): 1564-1570.
[9]	XU Li-sheng， ZHANG Wen-xu， PANG Yu-xuan， WU Cheng-yang. Driver Drowsiness Detection Algorithm Using Short-Time ECG Signals [J]. Journal of Northeastern University Natural Science, 2019, 40(7): 937-941.
[10]	QI Lin， LYU Xu-yang， YANG Ben-qiang， XU Li-sheng. Segmentation of Left Ventricle Endocardium Based on Transfer Learning of Fully Convolutional Networks [J]. Journal of Northeastern University Natural Science, 2018, 39(11): 1577-1582.