基于多尺度深度卷积神经网络的故障诊断方法

摘要
图/表
参考文献(22)
相关文章 (15)

全文: PDF (1735 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要针对机电装备故障诊断需要大量专家经验、故障特征识别困难的问题，在一维深度卷积神经网络基础上进行改进，构建多尺度一维深度卷积神经网络(M1DCNN)，提出基于多尺度一维深度卷积神经网络的故障诊断方法:首先在网络输入层构建多个含有不同尺寸卷积核通道的特征提取层，对一维时序信号中故障特征进行多尺度特征提取，丰富智能体诊断信息，将所提取特征通过输入到包含多尺寸卷积核以及多样池化层中进行特征处理，最后合并多通道所处理的特征，使网络完成自我学习，从而实现故障诊断。将该方法应用到西储大学轴承故障数据及行星齿轮箱的故障数据诊断实验，结果表明该方法具有诊断精度高、鲁棒性强的特点，相较于一维卷积神经网络准确率提高1.25%，与反向传播神经网络、循环神经网络相比准确率平均提高3%以上，对网络特征提取效果进行可视化分析，结果表明该方法特征提取效果与诊断精度优于一维卷积神经网络。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	卞景艺
	刘秀丽
	徐小力
	吴国新

关键词 ：深度卷积神经网络(DCNN), 多尺度特征提取, 特征可视化, 故障诊断

Abstract：Aiming at the fault diagnosis of mechanical and electrical equipments that requires a lot of expert experience and usually has difficulty in fault identification, a method based on a multi-scale one-dimensional deep convolutional neural network (M1DCNN) was proposed to improve the original network algorithm by introducing the multi-scale processing.Firstly, several feature extraction layers with different scale convolution kernel channels were constructed in the network input layer, and the fault features in one-dimensional time series signals were extracted by multi-scale feature extraction to enrich the diagnostic information.Then, the extracted features were input into the multi-scale convolution kernel and multiple pooling layers for feature processing.Finally, the features processed by multi-channels were combined to enable the network to complete self-learning to achieve fault diagnosis.The method has been applied to the bearing fault data and planetary gearbox fault data at Case Western Reserve University.The results show that the method has the characteristics of high diagnostic accuracy and strong robustness.Compared with the original one-dimensional convolutional neural network, the accuracy rate is improved by 1.25%, and the accuracy rate is increased by more than 3% on average compared with the BP neural network and recurrent neural network.A visual analysis on the effect of network feature extraction was carried out and the results show that the model feature extraction effect and the diagnostic accuracy of the proposed method are better than the conventional one-dimensional convolutional neural network.

Key words： deep convolutional neural network(DCNN) multi-scale feature extraction feature visualization fault diagnos

收稿日期: 2020-04-29 出版日期: 2021-09-28

引用本文:

卞景艺，刘秀丽，徐小力，吴国新. 基于多尺度深度卷积神经网络的故障诊断方法[J]. 振动与冲击, 2021, 40(18): 204-211.
BIAN Jingyi,LIU Xiuli,XU Xiaoli,WU Guoxin. Fault diagnosis method based on a multi-scale deep convolutional neural network. JOURNAL OF VIBRATION AND SHOCK, 2021, 40(18): 204-211.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2021/V40/I18/204

［1］辛玉,李舜酩,王金瑞,等.基于迭代经验小波变换的齿轮故障诊断方法［J］.仪器仪表学报,2018,39(11):79-86.
XIN Yu, LI Shunming, WANG Jinrui, et al.Gear fault diagnosis method based on iterative empirical wavelet transform［J］.Chinese Journal of Scientific Instrument,2018,39(11):79-86

［2］龙英,何怡刚,张镇,等.基于小波变换和ICA特征提取的开关电流电路故障诊断［J］.仪器仪表学报,2015,36(10):2389-2400.
LONG Ying, HE Yigang, ZHANG Zhen, et al.Switched current circuit fault diagnosis based on wavelet transform and ICA feature extraction［J］.Chinese Journal of Scientific Instrument,2015,36(10):2389-2400.

［3］朱喜华,李颖晖,周飞帆,等.基于改进EMD算法的永磁同步电机故障特征提取［J］.微电机,2011,44(2): 65-69.
ZHU Xihua, LI Yinghui, ZHOU Feifan, et al.Feature extraction for PMSM based on ameliorated EMD arithmetic［J］.Micromotors,2011,44(2):65-69.

［4］高冠琪,黄伟国,李宁,等.基于时频挤压和阶比分析的变转速轴承故障检测方法［J］.振动与冲击,2020,39(3):205-210.
GAO Guanqi, HUANG Weiguo, LI Ning, et al.Fault detection method for varying rotating speed bearings based on time-frequency squeeze and order analysis［J］.Journal of Vibration and Shock,2020,39(3):205-210.

［5］李志农,朱明,褚福磊,等.基于经验小波变换的机械故障诊断方法研究［J］.仪器仪表学报,2014,35(11):2423-2432.
LI Zhinong, ZHU Ming, CHU Fulei, et al.Mechanical fault diagnosis method based on empirical wavelet transform［J］.Chinese Journal of Scientific Instrument, 2014,35(11):2423-2432.

［6］薛瑞,赵荣珍.EWT-MFE与t-SNE结合的旋转机械故障诊断方法［J］.机械设计与研究,2019,35(4):53-57.
XUE Rui, ZHAO Rongzhen.Fault diagnosis method forrotating machinery based on EWT-MFE and t-SNE［J］.Machine Design & Research, 2019,35(4):53-57.

［7］唐立力.基于粗糙遗传BP神经网络的滚动轴承故障诊断［J］.机械工程与自动化,2018(3):138-140.
TANG Lili.Fault diagnosis on rolling bearing based on rough genetic BP neural network［J］.Mechanical Engineering & Automation,2018(3): 138-140.

［8］吴伟,郑娟.基于BP神经网络的齿轮故障诊断［J］.机械研究与应用,2015,28(1): 88-90.
WU Wei, ZHENG Juan.Gear fault diagnosis based on BP neural network［J］.Mechanical Research & Application,2015,28(1):88-90.

［9］张建,李艳军,曹愈远,等.免疫支持向量机用于航空发动机磨损故障诊断［J］.北京航空航天大学学报,2017,43(7):1419-1425.
ZHANG Jian, LI Yanjun, CAO Yuyuan, et al.Immune SVM used in wear fault diagnosis of aircraft engine［J］.Journal of Beijing University of Aeronautics and Astronautics,2017,43(7): 1419-1425.

［10］贺立敏,王岘昕,韩冰.基于随机森林和支持向量机的船舶柴油机故障诊断［J］.中国航海,2017,40(2): 29-33.
HE Limin, WANG Xianxin, HAN Bing.Fault diagnosis of marine diesel engine based on random forest and support vector machine［J］.Navigation of China,2017,40(2): 29-33.

［11］CHENG J, WANG P S, LI G, et al.Recent advances in efficient computation of deep convolutional neural networks［J］.Frontiers of Information Technology & Electronic Engineering,2018,19(1): 64-77.

［12］GU J X, WANG Z H, KUEN J, et al.Recent advances in convolutional neural networks［J］.Pattern Recognition,2018, 77(5): 354-377.

［13］KRIZHEVSKY A, SUTSKEVER I, HINTON G.ImageNet classification with deep convolutional neural networks［J］.Advances in Neural Information Processing Systems, 2012, 25(2):1097-1105.

［14］曲建岭,余路,袁涛,等.基于一维卷积神经网络的滚动轴承自适应故障诊断算法［J］.仪器仪表学报,2018,39(7):134-143.
QU Jianling, YU Lu, YUAN Tao, et al.Adaptive fault diagnosis algorithm for rolling bearings based on one-dimensional convolutional neural networik［J］.Chinese Journal of Scientific Instrument,2018,39(7):134-143.

［15］KAREN S, ANDREW Z.Very deep convolutional networks for large-scale image recognition［J］.Computer Science, 2014, 18:1724-1734.

［16］李恒,张氢,秦仙蓉,等.基于短时傅里叶变换和卷积神经网络的轴承故障诊断方法［J］.振动与冲击,2018,37(19):124-131.
LI Heng, ZHANG Qing, QIN Xianrong, et al.Fault diagnosis method for rolling bearings based on short-time Fourier transform and convolution neural network［J］.Journal of Vibration and Shock,2018,37(19):124-131.

［17］韩涛,袁建虎,唐建,等.基于MWT和CNN的滚动轴承智能复合故障诊断方法［J］.机械传动,2016,40(12):139-143.
HAN Tao, YUAN Jianhu, TANG Jian, et al.An approach of intelligent compound fault diagnosis of rolling bearing based on MWT and CNN［J］.Journal of Mechanical Transmission,2016,40(12): 139-143.

［18］吴春志,江鹏程,冯辅周,等.基于一维卷积神经网络的齿轮箱故障诊断［J］.振动与冲击,2018,37(22): 51-56.
WU Chunzhi, JIANG Pengcheng, FENG Fuzhou, et al.Faults diagnosis method for gearboxes based on a 1D convolutional neural network［J］.Journal of Vibration and Shock,2018,37(22):51-56.

［19］李东东,王浩,杨帆,等.基于一维卷积神经网络和Soft-Max分类器的风电机组行星齿轮箱故障检测［J］.电机与控制应用,2018,45(6):80-87.
LI Dongdong, WANG Hao, YANG Fan, et al.Fault detection of wind turbine planetary gear box using 1D convolution neural networks and Soft-Max classifier［J］.Electric Machines & Control Application,2018,45(6): 80-87.

［20］SZEGEDY C, LIU W, JIA Y, et al.Going deeper with convolutions［C］∥ 2015 IEEE Conference on Computer Vision and Pattern Recognition.Boston: IEEE,2015.

［21］LIN M, CHEN Q, YAN S.Network in network［C］∥ International Conferenceon Learning Representations.Banff: ICLR, 2004.

［22］CHO K, VAN MERRIENBOER B, GULCEHRE C, et al.Learning phrase representations using RNN encoder-decoder for statistical machine translation［C］∥Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing.Doha: EMNLP, 2014.