改进密集连接卷积网络的滚动轴承故障诊断方法

牛锐祥1,2,丁华1,2,施瑞1,2,孟祥龙1,2

振动与冲击 ›› 2022, Vol. 41 ›› Issue (11) : 252-258.

PDF(1353 KB)
PDF(1353 KB)
振动与冲击 ›› 2022, Vol. 41 ›› Issue (11) : 252-258.
论文

改进密集连接卷积网络的滚动轴承故障诊断方法

  • 牛锐祥1,2,丁华1,2,施瑞1,2,孟祥龙1,2
作者信息 +

Rolling bearing fault diagnosis method based on improved densely connected convolution network

  • NIU Ruixiang1,2, DING Hua1,2, SHI Rui1,2, MENG Xianglong1,2
Author information +
文章历史 +

摘要

针对滚动轴承工作环境噪声干扰较大、模型泛化能力不足、变工况诊断较难的问题,提出了一种改进密集连接卷积网络的故障诊断方法。将采集到的滚动轴承的原始时域信号作为模型输入,不需要任何数据处理,实现端到端的特征提取和分类任务。改进密集连接卷积网络在密集块中强调信息流动,增强特征复用,通过多尺度卷积层提取特征,利用注意力机制对多尺度特征通道加权。在堆叠的密集块和池化层完成主要特征提取后,采用多分类函数实现故障诊断。选用凯斯西储大学轴承数据集验证改进密集连接卷积网络的诊断能力,结果表明,改进密集连接卷积网络在理想实验下的识别率为99.8%、在抗噪实验下的识别率为98.22%,在泛化实验下的识别率为97.19%,识别率明显高于其他深度学习模型,证明了其在滚动轴承故障诊断方面的优越性。

Abstract

Aiming at the problems of high noise interference in the working environment of rolling bearing,insufficient generalization ability of model,and difficulty diagnosis of variable operating conditions, a fault diagnosis method based on improved densely connected convolution network is proposed.The collected original time domain signal of the rolling bearings are used as model input, without any data processing,and end-to-end feature extraction and classification task are realized.The improved densely connected convolution network emphasizes information flow in dense blocks, enhances feature reuse,through multi-scale convolution layer in dense blocks extracts feature and then uses the attention mechanism to weight the multi-scale feature channels.After the stacked dense blocks and pooling layers complete main feature extraction,through multi classification function realizs fault diagnosis.Use Case Western Reserve University bearing dataset to verify the diagnostic ability of model.The results show that the recognition rate of improved densely connected convolution network under the ideal experiment is 99.8%,the recognition rate under the anti-noise experiment is 98.22%,and the recognition rate under the generalization experiment is 97.19%,which is significantly higher than other deep learning models and proves its superiority.

关键词

滚动轴承 / 故障诊断 / 密集连接卷积网络 / 多尺度 / 注意力机制

Key words

rolling bearing / fault diagnosis / densely connected convolution network / multi-scale / attention mechanism

引用本文

导出引用
牛锐祥1,2,丁华1,2,施瑞1,2,孟祥龙1,2. 改进密集连接卷积网络的滚动轴承故障诊断方法[J]. 振动与冲击, 2022, 41(11): 252-258
NIU Ruixiang1,2, DING Hua1,2, SHI Rui1,2, MENG Xianglong1,2. Rolling bearing fault diagnosis method based on improved densely connected convolution network[J]. Journal of Vibration and Shock, 2022, 41(11): 252-258

参考文献

[1] 刘艳芳,刘尚旺.MED和分层模糊熵在滚动轴承故障诊断中的应用[J].机械设计与制造,2018(11):49-52.
LIU Yanfang,LIU Shangwang.Application of MED and Hierarchical Fuzzy Entropy to Rolling Bearing Fault Diagnosis[J].Machinery Design & Manufacture,2018(11):49-52.
[2] 刘韬,陈进,董广明.基于频带熵的滚动轴承故障诊断研究[J].振动与冲击,2014,33(1):77-80.
LIU Tao,CHEN Jin,DONG Guangming.Rolling element bearing fault diagnosis based on frequency band entropy[J].Journal of Vibration and Shock 2014,33 (1):77-80.
[3] CHEN J,LI Z,PAN J,et al.Wavelet transform based on inner product in fault diagnosis of rotating machinery: A review[J]. Mechanical Systems and Signal Processing,2016,70-71,1-35.
[4] ZHANG W,PENG G,LI C,et al. A new deep learning model for fault diagnosis with good anti-noise and domain adaptation ability on raw vibration signals[J]. Sensors,2017, 17(2):425.
[5] LI Y,WANG N,SHI J,et al.Revisiting batch normalization for practical domain adaptation[J].arXiv preprint arXiv:1603.04779,2016.
[6] ZHENBO Y,MINPING J.GA-1DLCNN method and its application in bearing fault diagnosis[J].东南大学学报 (英文版),2019 (1):6.
[7] HE K,ZHANG X,REN S,et al.Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2016:770-778.
[8] ZHAN M,KANG M,TANG B,et al.Deep residual networks with dynamically weighted wavelet coefficients for fault diagnosis of planetary gearboxes[J].IEEE Transactions on Industrial Electronics,2017,65(5):4290-4300.
[9] ZHAO M,ZHONG S,FU X,et al.Deep Residual Networks with Adaptively Parametric Rectifier Linear Units for Fault Diagnosis[J].IEEE Transactions on Industrial Electronics, 2020.
[10] ZHANG W,LI X,DING Q.Deep residual learning-based fault diagnosis method for rotating machinery[J].ISA Transactions,2019,95:295-305.
[11] 汪久根,柯梁亮.基于残差网络的 RV 减速器故障诊断[J]. 机械工程学报,2019 (3):73-80.
WANG Jiugen,KE Liangliang.Fault Diagnosis for RV Reducer Based on Residual Network[J].Journal of Mechanical Engineering,2019,55(03):73-80.
[12] HUANG G,LIU Z,VAN Der Maaten L,et al.Densely connected convolutional networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2017:4700-4708.
[13] 熊鹏,汤宝平,邓蕾,等.基于动态加权密集连接卷积网络的变转速行星齿轮箱故障诊断[J].机械工程学报,2019,55(07):52-57.
XIONG Peng,TANG Baoping,DENG Lei,et al. Fault Diagnosis for Planetary Gearbox by Dynamically Weighted Densely Connected Convolutional Networks [J] . Journal of Mechanical Engineering,2019,55(07):52-57.
[14] LIU R,WANG F,YANG B,et al.Multiscale Kernel Based Residual Convolutional Neural Network for Motor Fault Diagnosis Under Nonstationary Conditions[J]. IEEE Transactions on Industrial Informatics,2019,16(6):3797-3806.
[15] 孔子迁,邓蕾,汤宝平等.基于时频融合和注意力机制的深度学习行星齿轮箱故障诊断方法[J].仪器仪表学报,2019,40(6):221-227.
KONG Ziqiang,Deng Lei,Tang Baoping.Fault diagnosis of planetary gearbox based on deep learning with time-frequency fusion and attention mechanism[J].Chinese Journal of Scientific Instrument,2019,40(6):221-227.
[16] 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 Eenjie,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.
[17] IOFFE S,SZEGEDY C.Batch normalization: Accelerating deep network training by reducing internal covariate shift[J]. arXiv preprint arXiv:1502.03167,2015.
[18] KRIZHEVSKY A,HINTON G.Convolutional deep belief networks on cifar-10[J].Unpublished Manuscript,2010,40(7): 1-9.
[19] HOWARD A,SANDLER M,CHU G,et al.Searching for mobilenetv3[C]//Proceedings of the IEEE International Conference on Computer Vision.2019:1314-1324.
[20] MNIH V,HEESS N,GRAVES A.Recurrent models of visual attention[C]//Advances in Neural Information Processing Systems.2014:2204-2212.
[21] HU J,SHEN L,SUN G.Squeeze-and-excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.2018:7132-7141.
[22] WANG Q,WU B,ZHU P,et al.ECA-net: Efficient channel attention for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.2020:11534-11542.
[23] 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.
[24] HE K,ZHANG X,REN S,et al.Identity mappings in deep residual networks[C]//European Conference on Computer Vision. Springer,Cham,2016:630-645.
[25] VAN Der Maaten L,HINTON G.Visualizing data using t-SNE[J]. Journal of Machine Learning Research,2008,9(Nov):2579-2605.

PDF(1353 KB)

361

Accesses

0

Citation

Detail

段落导航
相关文章

/