留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

小波变换和深度残差收缩网络在齿轮箱故障诊断中的应用

翁敏超 王海瑞 朱贵富

翁敏超,王海瑞,朱贵富. 小波变换和深度残差收缩网络在齿轮箱故障诊断中的应用[J]. 机械科学与技术,2024,43(5):790-797 doi: 10.13433/j.cnki.1003-8728.20230054
引用本文: 翁敏超,王海瑞,朱贵富. 小波变换和深度残差收缩网络在齿轮箱故障诊断中的应用[J]. 机械科学与技术,2024,43(5):790-797 doi: 10.13433/j.cnki.1003-8728.20230054
WENG Minchao, WANG Hairui, ZHU Guifu. Application of Wavelet Transform and Deep Residual Shrinkage Network in Gearbox Fault Diagnosis[J]. Mechanical Science and Technology for Aerospace Engineering, 2024, 43(5): 790-797. doi: 10.13433/j.cnki.1003-8728.20230054
Citation: WENG Minchao, WANG Hairui, ZHU Guifu. Application of Wavelet Transform and Deep Residual Shrinkage Network in Gearbox Fault Diagnosis[J]. Mechanical Science and Technology for Aerospace Engineering, 2024, 43(5): 790-797. doi: 10.13433/j.cnki.1003-8728.20230054

小波变换和深度残差收缩网络在齿轮箱故障诊断中的应用

doi: 10.13433/j.cnki.1003-8728.20230054
基金项目: 国家自然科学基金项目(61863016,61263023)
详细信息
    作者简介:

    翁敏超,硕士研究生,wengmc@foxmail.com

    通讯作者:

    朱贵富,工程师,zhuguifu@kust.edu.cn

  • 中图分类号: TH17

Application of Wavelet Transform and Deep Residual Shrinkage Network in Gearbox Fault Diagnosis

  • 摘要: 齿轮的精确故障诊断是确保旋转机械设备稳定可靠运行的有效手段,针对强噪声环境下齿轮箱中齿轮故障分类问题,提出了一种基于连续小波变换和深度残差收缩网络的故障诊断模型。首先,采用小波变换对一维时间序列的振动数据进行时频分析,将其转化为二维时频图作为深度残差收缩网络(DRSN)的输入;其次,在多层卷积神经网络的基础上加入残差结构中的跨层恒等连接解决了梯度消失和爆炸的问题,同时利用自适应阈值子网络实现软阈值化降噪;最后,将故障样本的时频图作为诊断模型的输入进行故障分类。实验结果证明了与其他模型相比,本文采用的故障诊断方法更容易识别故障特征,分类准确率达到了99.15%。
  • 图  1  常用的母小波

    Figure  1.  Common mother wavelet

    图  2  池化原理

    Figure  2.  Pooling principle

    图  3  典型CNN架构

    Figure  3.  Model of CNN

    图  4  DRSN的基本结构

    Figure  4.  Basic structure model of DRSN

    图  5  深度残差收缩网络的整体结构

    Figure  5.  DRSN backbone

    图  6  齿轮箱故障诊断模型流程图

    Figure  6.  Gearbox fault diagnosis flowchart

    图  7  Drivetrain Dynamic Simulator实验平台

    Figure  7.  Drivetrain Dynamic Simulator's experimental platform

    图  8  5种类型的时频图

    Figure  8.  Five types of time frequency

    图  9  30 Hz-2 V工况下的测试集准确率

    Figure  9.  30 Hz-2 V classification accuracy

    图  10  20 Hz-0 V工况下的测试集准确率

    Figure  10.  20 Hz-0 V classification accuracy

    图  11  30 Hz-2 V工况下的平均Loss

    Figure  11.  30 Hz-2 V average Loss

    图  12  20 Hz-0 V工况下的平均Loss

    Figure  12.  20 Hz-0 V average Loss

    图  13  原始数据的降维图

    Figure  13.  Dimension reduction of raw data

    图  14  各结构层的降维图

    Figure  14.  Dimension reduction of various structural layers

    图  15  故障分类混淆矩阵

    Figure  15.  Confusion matrix for fault classification

    表  1  DRSN网络结构参数

    Table  1.   Structural parameters of DRSN

    结构层 层参数
    卷积核 核数量 步长 Output
    Conv 1 3 × 3 × 3 64 1 (64, H, W
    Conv 2 64 × 3 × 3 64 1 (64, H, W
    Conv 3 64 × 3 × 3 128 2 (128, H/2, W/2)
    Conv 4 128 × 3 × 3 256 2 (256, H/4, W/4)
    Conv 5 256 × 3 × 3 512 2 (512, H/8, W/8)
    GAP (512, 1, 1)
    Flatten 512
    FC 512
    下载: 导出CSV

    表  2  故障类型

    Table  2.   Fault type

    类型描述
    Chipped齿轮出现裂纹甚至断裂
    Miss齿轮缺损
    Root齿轮根部出现裂纹
    Surface齿轮表面磨损
    下载: 导出CSV

    表  3  齿轮故障分类结果

    Table  3.   Gear fault classification results

    方法 分类准确率/%
    20 Hz-0 V 30 Hz-2 V
    SVM 61.28 62.65
    CNN 95.16 92.41
    VGG 96.35 95.40
    ResNet 95.70 98.65
    DRSN 98.04 99.15
    下载: 导出CSV
  • [1] 丁承君, 张良, 冯玉伯, 等. VMD和t-SNE相结合的滚动轴承故障诊断[J]. 机械科学与技术, 2020, 39(5): 758-764.

    DING C J, ZHANG L, FENG Y B, et al. Fault diagnosis method of rolling bearing combining VMD with t-SNE[J]. Mechanical Science and Technology for Aerospace Engineering, 2020, 39(5): 758-764. (in Chinese)
    [2] 李滨, 曾辉. 改进的深度置信网络在电主轴故障诊断中的应用[J]. 机械科学与技术, 2021, 40(7): 1051-1057.

    LI B, ZENG H. Application of improved deep belief network in electric spindle fault diagnosis[J]. Mechanical Science and Technology for Aerospace Engineering, 2021, 40(7): 1051-1057. (in Chinese)
    [3] WANG M J, CHEN Y F, ZHANG X N, et al. Roller bearing fault diagnosis based on integrated fault feature and SVM[J]. Journal of Vibration Engineering & Technologies, 2022, 10(3): 853-862.
    [4] XIA X F, RAO X J, SU Y, et al. Mechanical fault diagnosis of high voltage circuit breaker based on improved GSO-SVM algorithm[J]. Journal of Physics: Conference Series, 2021, 2087: 012033. doi: 10.1088/1742-6596/2087/1/012033
    [5] WU J D, CHEN J C. Continuous wavelet transform technique for fault signal diagnosis of internal combustion engines[J]. NDT & e International, 2006, 39(4): 304-311.
    [6] 刘型志, 田娟, 李松浓, 等. 基于连续小波变换的电表绕组故障检测方法研究[J]. 自动化仪表, 2022, 43(2): 65-68.

    LIU X Z, TIAN J, LI S N, et al. Research on fault detection method for electric meter winding based on continuous wavelet transform[J]. Process Automation Instrumentation, 2022, 43(2): 65-68. (in Chinese)
    [7] 何晓霞, 沈玉娣, 张西宁. 连续小波变换在滚动轴承故障诊断中的应用[J]. 机械科学与技术, 2001, 20(4): 571-572. doi: 10.3321/j.issn:1003-8728.2001.04.038

    HE X X, SHEN Y D, ZHANG X N. An application of continuous wavelet transform to fault diagnosis of rolling element bearings[J]. Mechanical Science and Technology for Aerospace Engineering, 2001, 20(4): 571-572. (in Chinese) doi: 10.3321/j.issn:1003-8728.2001.04.038
    [8] ALBAWI S, MOHAMMED T A, AL-ZAWI S. Understanding of a convolutional neural network[C]// 2017 International Conference on Engineering and Technology (ICET). Antalya: IEEE, 2017: 1-6.
    [9] ZHAO M H, ZHONG S S, FU X Y, et al. Deep residual shrinkage networks for fault diagnosis[J]. IEEE Transactions on Industrial Informatics, 2020, 16(7): 4681-4690. doi: 10.1109/TII.2019.2943898
    [10] 李名莉, 焦欣欣. 一种新的小波阈值去噪算法在工程中的应用[J]. 计算机与数字工程, 2019, 47(7): 1627-1630. doi: 10.3969/j.issn.1672-9722.2019.07.014

    LI M L, JIAO X X. Application of a new wavelet threshold denoising algorithm in engineering[J]. Computer and Digital Engineering, 2019, 47(7): 1627-1630. (in Chinese) doi: 10.3969/j.issn.1672-9722.2019.07.014
    [11] SHAO S Y, MCALEER S, YAN R Q, et al. Highly accurate machine fault diagnosis using deep transfer learning[J]. IEEE Transactions on Industrial Informatics, 2019, 15(4): 2446-2455. doi: 10.1109/TII.2018.2864759
    [12] YIN Z Y, HOU J. Recent advances on SVM based fault diagnosis and process monitoring in complicated industrial processes[J]. Neurocomputing, 2016, 174: 643-650. doi: 10.1016/j.neucom.2015.09.081
    [13] ZHANG K, TANG B P, DENG L, et al. A hybrid attention improved ResNet based fault diagnosis method of wind turbines gearbox[J]. Measurement, 2021, 179: 109491. doi: 10.1016/j.measurement.2021.109491
    [14] 贾京龙, 余涛, 吴子杰, 等. 基于卷积神经网络的变压器故障诊断方法[J]. 电测与仪表, 2017, 54(13): 62-67. doi: 10.3969/j.issn.1001-1390.2017.13.011

    JIA J L, YU T, WU Z J, et al. Fault diagnosis method of transformer based on convolutional neural network[J]. Electrical Measurement & Instrumentation, 2017, 54(13): 62-67. (in Chinese) doi: 10.3969/j.issn.1001-1390.2017.13.011
    [15] WEN L, LI X, LI X Y, et al. A new transfer learning based on VGG-19 network for fault diagnosis[C]// 2019 IEEE 23rd International Conference on Computer Supported Cooperative Work in Design (CSCWD). Porto: IEEE, 2019: 205-209.
    [16] KOBAK D, BERENS P. The art of using t-SNE for single-cell transcriptomics[J]. Nature Communications, 2019, 10(1): 5416. doi: 10.1038/s41467-019-13056-x
    [17] TOWNSEND J T. Theoretical analysis of an alphabetic confusion matrix[J]. Perception & Psychophysics, 1971, 9(1): 40-50.
  • 加载中
图(15) / 表(3)
计量
  • 文章访问数:  56
  • HTML全文浏览量:  17
  • PDF下载量:  11
  • 被引次数: 0
出版历程
  • 收稿日期:  2022-05-22
  • 刊出日期:  2024-05-31

目录

    /

    返回文章
    返回