基于平移不变CNN的机械故障诊断研究

被引：64

作者：

朱会杰 ^{[1
,2
,3
]}

王新晴 ^{[4
]}

芮挺 ^{[4
]}

张欲保 ^{[2
]}

李艳峰 ^{[5
]}

机构：

[1] 近地面探测技术重点实验室

[2] 无锡科研一所

[3] 上海交通大学自动化系

[4] 陆军工程大学野战工程学院

[5] 陆军军事交通学院汽车士官学校

来源：

振动与冲击 | 2019年 / 38卷 / 05期

关键词：

故障诊断; 卷积神经网络; 深度学习;

D O I：

10.13465/j.cnki.jvs.2019.05.007

中图分类号：

TP183 [人工神经网络与计算]; TH17 [机械运行与维修];

学科分类号：

120111 [工业工程]; 140502 [人工智能];

摘要：

常规机械故障诊断方法需要信号预处理、特征提取、特征选择、模式识别等多个步骤,过程复杂,通用性差。卷积神经网络(Convolutional Neural Network, CNN)是一种自学习性能好、抗干扰能力强的深度神经网络。为了简化步骤、提高效率,将CNN引入到机械故障诊断,直接使用传感器测得的原始数据进行故障识别。由于机械振动信号的特征具有典型的时移性,CNN需要大量数据才能自我学习到这种特性。结合故障信号的冲击特点和CNN的不足,提出了权值求和和大尺度最大值池化策略,有效解决了特征的平移不变性,增强了小样本时的泛化能力。通过对单点和多点故障的轴承进行诊断,验证了平移不变CNN的有效性。与常规故障诊断方法和其他机器学习算法对比显示,平移不变CNN不仅准确率高,而且使用简单,为故障诊断提供了一种新的途径。

引用

页码：45 / 52

页数：8

共 9 条

[1]

Sparse representation based on adaptive multiscale features for robust machinery fault diagnosis [J].

Zhu, Huijie ;

Wang, Xinqing ;

Zhao, Yang ;

Li, Yanfeng ;

Wang, Wenfu ;

Li, Liping .

PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART C-JOURNAL OF MECHANICAL ENGINEERING SCIENCE, 2015, 229 (12) :2303-2313

[2]

Wavelet leaders multifractal features based fault diagnosis of rotating mechanism [J].

Du, Wenliao ;

Tao, Jianfeng ;

Li, Yanming ;

Liu, Chengliang .

MECHANICAL SYSTEMS AND SIGNAL PROCESSING, 2014, 43 (1-2) :57-75

[3]

Matching pursuit of an adaptive impulse dictionary for bearing fault diagnosis [J].

Cui, Lingli ;

Wang, Jing ;

Lee, Seungchul .

JOURNAL OF SOUND AND VIBRATION, 2014, 333 (10) :2840-2862

[4]

An approach to fault diagnosis of reciprocating compressor valves using Teager-Kaiser energy operator and deep belief networks [J].

Van Tung Tran ;

AlThobiani, Faisal ;

Ball, Andrew .

EXPERT SYSTEMS WITH APPLICATIONS, 2014, 41 (09) :4113-4122

[5]

Compressed sensing based on dictionary learning for extracting impulse components.[J].Xuefeng Chen;Zhaohui Du;Jimeng Li;Xiang Li;Han Zhang.Signal Processing.2013,

[6]

Early fault diagnosis of rotating machinery based on wavelet packets—Empirical mode decomposition feature extraction and neural network.[J].G.F. Bin;J.J. Gao;X.J. Li;B.S. Dhillon.Mechanical Systems and Signal Processing.2011,

[7]

Fault diagnosis of ball bearings using machine learning methods.[J].P.K. Kankar;Satish C. Sharma;S.P. Harsha.Expert Systems With Applications.2010, 3

[8]

Adaptive feature extraction using sparse coding for machinery fault diagnosis.[J].Haining Liu;Chengliang Liu;Yixiang Huang.Mechanical Systems and Signal Processing.2010, 2

[9]

Fault diagnosis of rotating machinery based on improved wavelet package transform and SVMs ensemble.[J].Qiao Hu;Zhengjia He;Zhousuo Zhang;Yanyang Zi.Mechanical Systems and Signal Processing.2006, 2

← 1 →