Please wait a minute...
浙江大学学报(工学版)  2022, Vol. 56 Issue (6): 1212-1219    DOI: 10.3785/j.issn.1008-973X.2022.06.020
能源与机械工程     
基于小波变换和优化CNN的风电齿轮箱故障诊断
温竹鹏1(),陈捷1,2,*(),刘连华1,焦玲玲1
1. 南京工业大学 机械与动力工程学院,江苏 南京 211816
2. 江苏省工业装备数字制造及控制技术重点实验室,江苏 南京 211816
Fault diagnosis of wind power gearbox based on wavelet transform and improved CNN
Zhu-peng WEN1(),Jie CHEN1,2,*(),Lian-hua LIU1,Ling-ling JIAO1
1. School of Mechanical and Power Engineering, Nanjing Tech University, Nanjing 211816, China
2. Jiangsu Key Laboratory of Digital Manufacturing for Industrial Equipment and Control Technology, Nanjing 211816, China
 全文: PDF(1870 KB)   HTML
摘要:

针对传统故障诊断方法过于依赖人为经验的缺陷,提出小波变换和二维密集连接扩张卷积神经网络(WT-ICNN)的风电齿轮箱智能故障诊断方法. 所提方法将一维振动信号通过连续小波变换(WT)转换成二维故障图像;再将二维故障图像输入ICNN中进行训练和测试. 通过齿轮箱开源数据和风场实测数据验证结果表明,与传统故障诊断方法相比,所提方法采用密集连接的结构自适应特征提取时频图,有效加强了故障特征的利用效率;在对风电齿轮箱的故障诊断中,所提方法具有更好的特征复用能力和更高的诊断精度.

关键词: 风电齿轮箱小波变换卷积神经网络密集连接扩张卷积    
Abstract:

An intelligent fault diagnosis method for wind turbine gearbox based on wavelet transform and two-dimensional densely connected dilated convolutional neural network(WT-ICNN) was proposed, aiming at the problem that traditional fault diagnosis method dependent on human experience too much. One dimensional vibration signal was transformed into two-dimensional fault image by continuous wavelet transform. Then the two-dimensional fault image was inputted into ICNN for training and testing. The verification of open source data of gearbox and measured data of wind field showed that compared with the traditional fault diagnosis methods, the proposed method effectively enhanced the utilization efficiency of fault features by using the densely connected structure for adaptive feature extraction of time-frequency map. And in the fault diagnosis of wind power gearbox, the proposed method had better feature reuse ability and higher diagnosis accuracy.

Key words: wind power gearbox    wavelet transform    convolutional neural network    densely connect    dilated convolution
收稿日期: 2021-06-25 出版日期: 2022-06-30
CLC:  TH 132  
基金资助: 国家重点研发计划资助项目(2019YFB2005005)
通讯作者: 陈捷     E-mail: 1393532653@qq.com;njtechchenjie@163.com
作者简介: 温竹鹏(1996—),男,硕士生,从事齿轮箱故障诊断研究. orcid.org/0000-0003-0160-3113. E-mail: 1393532653@qq.com
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
作者相关文章  
温竹鹏
陈捷
刘连华
焦玲玲

引用本文:

温竹鹏,陈捷,刘连华,焦玲玲. 基于小波变换和优化CNN的风电齿轮箱故障诊断[J]. 浙江大学学报(工学版), 2022, 56(6): 1212-1219.

Zhu-peng WEN,Jie CHEN,Lian-hua LIU,Ling-ling JIAO. Fault diagnosis of wind power gearbox based on wavelet transform and improved CNN. Journal of ZheJiang University (Engineering Science), 2022, 56(6): 1212-1219.

链接本文:

https://www.zjujournals.com/eng/CN/10.3785/j.issn.1008-973X.2022.06.020        https://www.zjujournals.com/eng/CN/Y2022/V56/I6/1212

图 1  密集连接的结构
图 2  扩张卷积结构
图 3  密集连接扩张卷积神经网络的诊断模型
图 4  动力传动实验模拟平台 [ 14]
图 5  案例1中5种状态的小波时频图
图 6  故障状态训练结果
图 7  案例1分类结果的混淆矩阵
图 8  案例2风电齿轮箱的4类故障
图 9  案例2中5种状态的小波时频图
图 10  案例2分类结果的混淆矩阵
图 11  案例2中5种状态的输入层特征可视化
图 12  案例2中5种状态的Softmax层特征可视化
实验 Acc N Acc Y 实验 Acc N Acc Y
%
1 94.6 99.8 4 93.7 99.6
2 95.7 99.5 5 92.8 99.5
3 94.4 99.3 平均 94.2 99.5
表 1  密集连接对模型分类精度的影响
实验 Acc/% t'/s
N Y N Y
1 99.3 100 261 224
2 98.9 99.8 251 227
3 99.1 99.6 253 228
4 99.1 99.8 247 225
5 99.6 98.2 246 226
平均 99.2 99.5 252 226
表 2  扩张卷积对模型训练效果的影响
模型 $ \eta $ Acc/% 模型 $ \eta $ Acc/%
1DCNN 10 96.0 LSSVM 10 91.6
2DCNN 10 96.7 WT-ICNN 10 99.3
KELM 10 95.7
表 3  不同模型的诊断精度对比
1 DAI J C, YANG X, WEN L Development of wind power industry in China: a comprehensive assessment[J]. Renewable and Sustainable Energy Reviews, 2018, 97: 156- 164
doi: 10.1016/j.rser.2018.08.044
2 LEI Y G, YANG B, JIANG X W, et al Applications of machine learning to machine fault diagnosis: a review and roadmap[J]. Mechanical Systems and Signal Processing, 2020, 138: 106587
doi: 10.1016/j.ymssp.2019.106587
3 张西宁, 郭清林, 刘书语. 深度学习技术及其故障诊断应用分析与展望[J]. 西安交通大学学报, 2020, 54(12): 1-13.
ZHANG Xi-ning, GUO Qing-lin, LIU Shu-yu. Analysis and prospect of deep learning technology and its fault diagnosis application [J]. Journal of Xi'an Jiaotong University. 2020, 54 (12): 1-13.
4 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 Xian-rong, 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
5 陈仁祥, 黄鑫, 杨黎霞, 等 基于卷积神经网络和离散小波变换的滚动轴承故障诊断[J]. 振动工程学报, 2018, 31 (5): 883- 891
CHEN Ren-xiang, HUANG Xin, YANG Li-xia, et al Rolling bearing fault diagnosis based on convolution neural network and discrete wavelet transform[J]. Journal of Vibration Engineering, 2018, 31 (5): 883- 891
6 ZHAO D Z, WANG T Y, CHU F L Deep convolutional neural network based planet bearing fault classification[J]. Computers in Industry, 2019, 107: 59- 66
doi: 10.1016/j.compind.2019.02.001
7 LIANG P F, DENG C, WU J, et al Intelligent fault diagnosis of rotating machinery via wavelet transform, generative adversarial net and convolutional neural network[J]. Measurement, 2020, 159: 107768
doi: 10.1016/j.measurement.2020.107768
8 JIAO J Y, ZHAO M, LIN J, et al A multivariate encoder information based convolutional neural network for intelligent fault diagnosis of planetary gearboxes[J]. Knowledge-Based Systems, 2018, 160: 237- 250
doi: 10.1016/j.knosys.2018.07.017
9 JING L Y, ZHAO M, LI P, et al A convolutional neural network based feature learning and fault diagnosis method for the condition monitoring of gearbox[J]. Measurement, 2017, 111: 1- 10
doi: 10.1016/j.measurement.2017.07.017
10 WANG S X, CHEN H W A novel deep learning method for the classification of power quality disturbances using deep convolutional neural network[J]. Applied Energy, 2019, 235: 1126- 1140
doi: 10.1016/j.apenergy.2018.09.160
11 于重重, 宁亚倩, 秦勇, 等 基于T-SNE样本熵和TCN的滚动轴承状态退化趋势预测[J]. 仪器仪表学报, 2019, 40 (8): 39- 46
YU Chong-chong, NING Ya-qian, QIN Yong, et al Prediction of rolling bearing state degradation trend based on T-SNE sample entropy and TCN[J]. Chinese Journal of Scientific Instrument, 2019, 40 (8): 39- 46
12 HAN Y, TANG B P, DENG L An enhanced convolutional neural network with enlarged receptive fields for fault diagnosis of planetary gearboxes[J]. Computers in Industry, 2019, 107: 50- 58
doi: 10.1016/j.compind.2019.01.012
13 SHAO S, 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
14 邵思羽. 基于深度学习的旋转机械故障诊断方法研究[D]. 南京: 东南大学, 2019.
SHAO Si-yu. Methodologies for fault diagnosis of rotary machine based on deep learning [D]. Nanjing: Southeast University, 2019.
[1] 何立,庞善民. 结合年龄监督和人脸先验的语音-人脸图像重建[J]. 浙江大学学报(工学版), 2022, 56(5): 1006-1016.
[2] 王云灏,孙铭会,辛毅,张博宣. 基于压电薄膜传感器的机器人触觉识别系统[J]. 浙江大学学报(工学版), 2022, 56(4): 702-710.
[3] 温佩芝,陈君谋,肖雁南,温雅媛,黄文明. 基于生成式对抗网络和多级小波包卷积网络的水下图像增强算法[J]. 浙江大学学报(工学版), 2022, 56(2): 213-224.
[4] 任松,朱倩雯,涂歆玥,邓超,王小书. 基于深度学习的公路隧道衬砌病害识别方法[J]. 浙江大学学报(工学版), 2022, 56(1): 92-99.
[5] 郑英杰,吴松荣,韦若禹,涂振威,廖进,刘东. 基于目标图像FCM算法的地铁定位点匹配及误报排除方法[J]. 浙江大学学报(工学版), 2021, 55(3): 586-593.
[6] 陈岳林,田文靖,蔡晓东,郑淑婷. 基于密集连接网络和多维特征融合的文本匹配模型[J]. 浙江大学学报(工学版), 2021, 55(12): 2352-2358.
[7] 刘近贞,叶方方,熊慧. 基于卷积神经网络的多类运动想象脑电信号识别[J]. 浙江大学学报(工学版), 2021, 55(11): 2054-2066.
[8] 黄靖,钟书远,文元桥,罗坤. 用于交通流预测的自适应图生成跳跃网络[J]. 浙江大学学报(工学版), 2021, 55(10): 1825-1833.
[9] 陈巧红,陈翊,李文书,贾宇波. 多尺度SE-Xception服装图像分类[J]. 浙江大学学报(工学版), 2020, 54(9): 1727-1735.
[10] 周登文,田金月,马路遥,孙秀秀. 基于多级特征并联的轻量级图像语义分割[J]. 浙江大学学报(工学版), 2020, 54(8): 1516-1524.
[11] 张彦楠,黄小红,马严,丛群. 基于深度学习的录音文本分类方法[J]. 浙江大学学报(工学版), 2020, 54(7): 1264-1271.
[12] 闫旭,范晓亮,郑传潘,臧彧,王程,程明,陈龙彪. 基于图卷积神经网络的城市交通态势预测算法[J]. 浙江大学学报(工学版), 2020, 54(6): 1147-1155.
[13] 沈宗礼,余建波. 基于迁移学习与深度森林的晶圆图缺陷识别[J]. 浙江大学学报(工学版), 2020, 54(6): 1228-1239.
[14] 汪海晋,尹宗宇,柯臻铮,郭英杰,董辉跃. 基于一维卷积神经网络的螺旋铣刀具磨损监测[J]. 浙江大学学报(工学版), 2020, 54(5): 931-939.
[15] 杨萍,王丹,康子健,李童,付利华,余悦任. 基于模式识别和集成CNN-LSTM的阵发性房颤预测模型[J]. 浙江大学学报(工学版), 2020, 54(5): 1039-1048.