轴承智能故障诊断

吴冬梅; 王福齐; 李贤功; 唐润; 张新建

doi:10.13272/j.issn.1671-251x.17986

轴承智能故障诊断

doi: 10.13272/j.issn.1671-251x.17986

1.
永城职业学院电子信息工程系, 河南永城　476600
2.
中国矿业大学矿业工程学院, 江苏徐州　221116
3.
南京财经大学管理科学与工程学院, 江苏南京　210023
4.
河南能源化工集团永煤公司陈四楼煤矿, 河南永城　476600

基金项目: 国家重点研发计划项目(2017YFC0804408)。

详细信息

作者简介:
吴冬梅（1980—）女，河南永城人，讲师，工程硕士，主要从事安全技术管理工作，E-mail：wdm0011@126.com

通讯作者:
李贤功（1981—），男，江苏连云港人，讲师，硕士研究生导师，博士，主要从事系统工程、煤矿安全管理方面的研究工作，E-mail：cumtlxg@163.com

中图分类号: TD67
计量
- 文章访问数: 271
- HTML全文浏览量: 53
- PDF下载量: 45
- 被引次数: 0
出版历程
- 收稿日期: 2022-07-20
- 修回日期: 2022-09-07
- 网络出版日期: 2022-09-15

Bearing intelligent fault diagnosis

1.
Department of Electronic Information Engineering, Yongcheng Vocational College, Yongcheng 476600, China
2.
School of Mines, China University of Mining and Technology, Xuzhou 221116, China
3.
School of Management Science and Engineering, Nanjing University of Finance & Economics, Nanjing 210023, China
4.
Chensilou Coal Mine, Henan Energy Chemical Group Yongmei Company, Yongcheng 476600, China

摘要

摘要: 轴承振动信号作为一种时间序列数据，其时间维度特征在分类中起着关键作用，单独使用卷积神经网络（CNN）进行轴承故障诊断会造成时间维度信息的丢失，导致诊断精度下降。针对上述问题，提出了一种结合一维CNN、双向门控循环单元(Bi GRU)和注意力机制的轴承故障诊断模型。首先利用CNN自适应提取一维振动信号局部空间特征；然后将该特征信息作为Bi GRU的输入，利用Bi GRU将提取的特征信息进行时间维度的融合，并引入注意力机制对多个时刻的特征信息进行加权，提取出更关键的故障特征；最后将故障特征输入全连接层得到分类结果，实现轴承智能故障诊断。实验结果表明：① 在测试集混淆矩阵上，轴承运行状态基本分类正确，只有部分标记类型没有被完全分类正确，但是召回率达到了95%以上，总的故障识别准确率为99.3%。② 利用t−SNE技术对降维处理后的数据进行可视化，轴承各运行状态的数据很好地聚集在各自的空间内，只有少量数据被混杂到其他区域，说明该模型具有较强的特征提取能力。③ 在恒定负载情况下，该模型故障诊断准确率较一维CNN、Bi GRU和注意力CNN等模型的平均准确率分别提高了0.8%、0.6%和0.3%。④ 在变负载情况下，与SVM(支持向量机)、一维CNN、Bi GRU和注意力CNN等模型相比，该模型具有更好的稳定性，当负载为2.25 kW时，准确率达85%以上。该模型既具有一维CNN局部特征提取能力，又具有Bi GRU时间依赖信息的建模能力，能够在获取轴承信号局部复杂特征后进一步融入特征之间的时间维度信息，同时注意力机制能进一步关注与故障更相关特征，因此具有较好的精度。
- 煤矿机械 /
- 轴承 /
- 智能故障诊断 /
- 注意力机制 /
- 卷积神经网络 /
- 双向门控循环单元 /
- 时间维度
Abstract: Bearing vibration signal is a kind of time series data, and its time dimension characteristic plays a key role in classification. Using convolutional neural network (CNN) alone to diagnose bearing fault will cause the loss of time dimension information. This results in the decline of diagnosis accuracy. To solve the above problems, a bearing fault diagnosis model combining one-dimensional CNN, bidirectional gated recurrent unit (Bi GRU) and attention mechanism is proposed. Firstly, CNN is used to adaptively extract the local space characteristic of one-dimensional vibration signals. Secondly, the characteristic information is taken as the input of the Bi GRU. Bi GRU is used to perform time dimension fusion on the extracted characteristic information. The attention mechanism is introduced to weigh the characteristic information of a plurality of moments so as to extract a more critical fault characteristic. Finally, the fault characteristic is input into a full connection layer to obtain a classification result, so as to realize intelligent fault diagnosis of the bearing. The experimental result shows the following points. ① On the confusion matrix of the test set, the classification of the bear running state is basically correct. Only some mark types are not completely classified correctly. But the recall rate is more than 95%, and the total fault recognition accuracy rate is 99.3%. ② The t-SNE technology is used to visualize the data after dimensionality reduction processing. The data of each running state of the bearing are well gathered in their own space. Only a small amount of data are mixed into other areas, which shows that the model has strong characteristic extraction capability. ③ Under the condition of constant load, the average accuracy of fault diagnosis of this model is 0.8%, 0.6% and 0.3% higher than that of one-dimensional CNN, Bi GRU and attention CNN models respectively. ④ Under the condition of variable load, this model has better stability than SVM, one-dimensional CNN, Bi GRU, attention CNN and other models. When the load is 2.25 kW, the accuracy rate is more than 85%. The model has the capability to extract one-dimensional CNN local characteristics and the capability to model Bi GRU time-dependent information. The model can further fuse time dimension information among the characteristics after acquiring the bear signal local complex characteristics. And the attention mechanism can further pay attention to the characteristics more relevant to faults. Therefore, the model has better precision.
- coal mine machinery /
- bearing /
- intelligent fault diagnosis /
- attention mechanism /
- convolutional neural network /
- bidirectional gated recurrent unit /
- time dimension

HTML全文

图 1 GRU单元

Figure 1. GRU unit

下载: 全尺寸图片幻灯片

图 2 Bi GRU结构

Figure 2. Bi GRU structure

下载: 全尺寸图片幻灯片

图 3 注意力机制结构

Figure 3. Attention mechanism structure

下载: 全尺寸图片幻灯片

图 4 故障诊断模型结构

Figure 4. Fault diagnosis model structure

下载: 全尺寸图片幻灯片

图 5 故障诊断流程

Figure 5. Fault diagnosis process

下载: 全尺寸图片幻灯片

图 6 准确率和损失曲线

Figure 6. Accuracy and loss curves

下载: 全尺寸图片幻灯片

图 7 混淆矩阵

Figure 7. Confusion matrix

下载: 全尺寸图片幻灯片

图 8 t−SNE降维可视化图

Figure 8. t-SNE dimension reduction visualization

下载: 全尺寸图片幻灯片

图 9 变负载准确率

Figure 9. Accuracy under variable load

下载: 全尺寸图片幻灯片

表 1 故障诊断模型结构参数

Table 1. Parameters of the fault diagnosis model structure

层数	结构参数	输出特征
M0	输入层	2048
M1	卷积层 (64×1/16/16)	128×16
M2	池化层 (2×1/2)	64×16
M3	卷积层(5×1/1/32)	64×32
M4	池化层(2×1/2)	32×32
M5	Bi GRU层（32）	32×64
M6	注意力模块	64
M7	全连接层	10
M8	Softmax层	10

下载: 导出CSV

表 2 实验样本划分

Table 2. Experimental sample division

训练样本数	测试样本数	故障种类	故障直径/mm	类型标记
400	100	滚动体	0.177 8	L0
400	100	滚动体	0.355 6	L1
400	100	滚动体	0.533 4	L2
400	100	内圈	0.177 8	L3
400	100	内圈	0.355 6	L4
400	100	内圈	0.533 4	L5
400	100	外圈	0.177 8	L6
400	100	外圈	0.355 6	L7
400	100	外圈	0.533 4	L8
400	100	正常	0	L9

下载: 导出CSV

表 3 5种故障诊断模型的准确率对比

Table 3. Accuracy comparison of five kinds of fault diagnosis models

模型	平均准确率/%
SVM	90.3
一维CNN	98.5
Bi GRU	98.7
注意力CNN	99.0
本文模型	99.3

下载: 导出CSV

参考文献(17)

[1]	宫涛,杨建华,单振,等. 强噪声背景与变转速工况条件下滚动轴承故障诊断研究[J]. 工矿自动化,2021,47(7):63-71. GONG Tao,YANG Jianhua,SHAN Zhen,et al. Research on rolling bearing fault diagnosis under strong noise background and variable speed working condition[J]. Industry and Mine Automation,2021,47(7):63-71.
[2]	黄重谦. 基于多隐层小波卷积极限学习神经网络的滚动轴承故障识别[J]. 工矿自动化,2021,47(7):63-71. HUANG Zhongqian. Fault identification of rolling bearing based on multi hidden layers wavelet convolution extreme learning neural network[J]. Industry and Mine Automation,2021,47(7):63-71.
[3]	吴静然,丁恩杰,崔冉,等. 采用多尺度注意力机制的旋转机械故障诊断方法[J]. 西安交通大学学报,2020,54(2):51-58. WU Jingran,DING Enjie,CUI Ran,et al. A diagnostic approach for rotating machinery using multi-scale feature attention mechanism[J]. Journal of Xi'an Jiaotong University,2020,54(2):51-58.
[4]	HE Miao,HE D. Deep learning based approach for bearing fault diagnosis[J]. IEEE Transactions on Industry Applications,2017,53(3):3057-3065. doi: 10.1109/TIA.2017.2661250
[5]	LI Xiangong,ZHANG Yuzhi,WANG Fuqi,et al. A fault diagnosis method of rolling bearing based on wavelet packet analysis and deep forest[J]. Symmetry,2022,14(2):267. doi: 10.3390/sym14020267
[6]	汪峰,周凤星,严保康. 基于特征量融合和支持向量机的滚动轴承故障诊断[J]. 科学技术与工程,2022,22(6):2351-2356. doi: 10.3969/j.issn.1671-1815.2022.06.026 WANG Feng,ZHOU Fengxing,YAN Baokang. Rolling bearing fault diagnosis based on feature fusion and support vector machine[J]. Science Technology and Engineering,2022,22(6):2351-2356. doi: 10.3969/j.issn.1671-1815.2022.06.026
[7]	雷亚国,贾峰,周昕,等. 基于深度学习理论的机械装备大数据健康监测方法[J]. 机械工程学报,2015,51(21):49-56. doi: 10.3901/JME.2015.21.049 LEI Yaguo,JIA Feng,ZHOU Xi,et al. A deep learning based method for machinery health monitoring with big data[J]. Journal of Mechanical Engineering,2015,51(21):49-56. doi: 10.3901/JME.2015.21.049
[8]	WANG Lihua,ZHAO Xiaoping,WU Jiaxin,et al. Motor fault diagnosis based on short-time fourier transform and convolutional neural network[J]. Chinese Journal of Mechanical Engineering,2017,30(6):1357-1368. doi: 10.1007/s10033-017-0190-5
[9]	LEVENT E,TURKER I,SERKAN K. A generic intelligent bearing fault diagnosis system using compact adaptive 1D CNN classifier[J]. Journal of Signal Processing Systems for Signal,Image,and Videl Technology,2019,91(2):179-189. doi: 10.1007/s11265-018-1378-3
[10]	李长文,程泽银,张小刚,等. 基于深度残差网络的采煤机摇臂齿轮故障诊断[J]. 工矿自动化,2021,47(3):71-78. LI Changwen,CHENG Zeyin,ZHANG Xiaogang,et al. Fault diagnosis of shearer rocker gear based on deep residual network[J]. Industry and Mine Automation,2021,47(3):71-78.
[11]	姜家国,郭曼利,杨思国. 基于GAF和DenseNet的滚动轴承故障诊断方法[J]. 工矿自动化,2021,47(8):84-89. JIANG Jiaguo,GUO Manli,YANG Siguo. Fault diagnosis of rolling barings based on GAF and DenseNet[J]. Industry and Mine Automation,2021,47(8):84-89.
[12]	LEI Jinhao,LIU Chao,JIANG Dongxiang. Fault diagnosis of wind turbine based on long short-term memory networks[J]. Renewable Energy,2019,133:422-432. doi: 10.1016/j.renene.2018.10.031
[13]	樊家伟,郭瑜,伍星,等. 基于LSTM神经网络和故障特征增强的行星齿轮箱故障诊断[J]. 振动与冲击,2021,40(20):271-277. doi: 10.13465/j.cnki.jvs.2021.20.034 FAN Jiawei,GUO Yu,WU Xing,et al. Fault diagnosis of planetary gearboxes based on LSTM neural network and fault feature enhancement[J]. Journal of Vibration and Shock,2021,40(20):271-277. doi: 10.13465/j.cnki.jvs.2021.20.034
[14]	张立鹏,毕凤荣,程建刚,等. 基于注意力Bi GRU的机械故障诊断方法研究[J]. 振动与冲击,2021,40(5):113-118. ZHANG Lipeng,BI Fengrong,CHENG Jiangang,et al. Mechanical fault diagnosis method based on attention Bi GRU[J]. Journal of Vibration and Shock,2021,40(5):113-118.
[15]	吉兴全,曾若海,张玉敏,等. 基于注意力机制的CNN-LSTM短期电价预测[J]. 电力系统保护玉控制,2022,50(17):125-132. JI Xingquan,ZENG Ruohai,ZHANG Yumin,et al. CNN-LSTM short-term electricity prediction based on an attention mechanism[J]. Power System Protection and Control,2022,50(17):125-132.
[16]	毛煜,尚海昆,于卓琦. 基于长短期记忆网络的电网同调机群快速辨识[J]. 电气工程学报,2022,17(2):201-207. MAO Yu,SHANG Haikun,YU Zhuoqi. A fast prediction method of coherernt generators based on long short-term memory network[J]. Journal of Electrical Engineering,2022,17(2):201-207.
[17]	MAATEN L V,HINTON G. Visualizing data using t-SNE[J]. Journal of Machine Learning Research,2008,9(11):2579-2605.