基于CEEMDAN与改进一维多尺度卷积神经网络结合的滚动轴承故障诊断

兰州理工大学学报 ›› 2025, Vol. 51 ›› Issue (1): 45-54.

基于CEEMDAN与改进一维多尺度卷积神经网络结合的滚动轴承故障诊断

马宁¹, 赵荣珍^*1,2, 郑玉巧¹

1.兰州理工大学机电工程学院, 甘肃兰州 730050;
2.广州理工学院智能制造与电气工程学院, 广东广州 510540

收稿日期:2022-06-23 出版日期:2025-02-28 发布日期:2025-03-03
通讯作者: 赵荣珍(1960-),女,山东枣庄人,博士,教授,博导.Email:zhaorongzhen@lut.edu.cn
基金资助:
国家自然科学基金(52465012,52475134)

Fault diagnosis of rolling bearing based on CEEMDAN combined with improved one-dimensional multi-scale convolutional neural network

MA Ning¹, ZHAO Rong-zhen^1,2, ZHENG Yu-qiao¹

1. School of Mechanical and Electronical Engineering, Lanzhou University of Technology, Lanzhou 730050, China;
2. College of Intelligent Manufacturing and Electrical Engineering, Guangzhou Institute of Science and Technology, Guangzhou 510540, China

Received:2022-06-23 Online:2025-02-28 Published:2025-03-03

摘要/Abstract

摘要： 针对滚动轴承信号微弱故障特征提取困难、故障诊断依靠大量专家经验和故障识别率低等问题,提出了融合自适应噪声完备集合经验模态分解与改进一维多尺度卷积神经网络的滚动轴承故障诊断方法.首先,采用自适应噪声完备集合经验模态分解对轴承信号进行消噪处理,并利用皮尔逊相关系数法对所得IMF分量进行信号重构;其次,在网络首层将大尺寸卷积核与空洞卷积结合,并引入金字塔场景解析网络提出改进的一维多尺度卷积神经网络,对故障特征信息进行提取,采用PSO算法对卷积核进行参数寻优;最后,融合多尺度特征信息完成网络学习,并输入Sofmax分类器,实现滚动轴承故障诊断.采用西储大学轴承数据集和HZXT-DS-001型双跨综合故障模拟实验台的滚动轴承故障数据进行了验证.结果表明,相比传统故障诊断方法该方法可以得到良好的诊断结果.

关键词: 自适应噪声完备集合经验模态分解, 一维卷积神经网络, 多尺度特征提取, 特征可视化, 故障诊断

Abstract: To solve the weak feature extraction of the fault signals in rolling bearing and the low fault identification accuracy relying on expert's experience, a fault Identification method combining complete ensemble empirical mode decomposition with adaptive denoise and improved one-dimensional convolutional neural network was proposed. First, the complete ensemble empirical mode decomposition is used to denoise the bearing signal, followed by signal reconstruction using Person correlation coefficient method on the obtained intrinsic mode functions (IMFs). Next, the first layer of the network adopts a method of combining large-scale kernels and atrous convolutions, the pyramid scene parsing network is introduced and leading to a proposed one-dimensional multi-scale convolutional network for extracting the fault features and the scale of convolution kernel is optimized by PSO algorithm. Finally, the multi-scale feature information is integrated to complete the network learning. And it is input into the Softmax classifier to achieve the fault category. The method is verified by the bearing data set of Western Reserve University and the rolling bearing faults data collected at the simulating faults on HZXT-DS-001 rig. The results show that to compare with the traditional method, the method proposed in this paper has good diagnosis results.

Key words: adaptive noise complete ensemble empirical mode decomposition, 1D convolutional neural network, multiscale feature extraction, feature visualization, fault diagnosis

中图分类号:

TH165
TP18

马宁, 赵荣珍, 郑玉巧. 基于CEEMDAN与改进一维多尺度卷积神经网络结合的滚动轴承故障诊断[J]. 兰州理工大学学报, 2025, 51(1): 45-54.

MA Ning, ZHAO Rong-zhen, ZHENG Yu-qiao. Fault diagnosis of rolling bearing based on CEEMDAN combined with improved one-dimensional multi-scale convolutional neural network[J]. Journal of Lanzhou University of Technology, 2025, 51(1): 45-54.

参考文献

[1] 雷亚国,贾峰,孔德同,等.大数据下机械智能故障诊断的机遇与挑战[J].机械工程学报,2018,54(5):94-104.
[2] 雷亚国,杨彬,杜兆钧,等.大数据下机械装备故障的深度迁移诊断方法[J].机械工程学报,2019,55(7):1-8.
[3] 李军宁,罗文广,陈武阁.面向振动信号的滚动轴承故障诊断算法综述[J].西安工业大学学报,2022,42(2):105-122.
[4] HINTON G E,SALAKHUTDINOV R R.Reducing the dimensionality of data with neural networks[J].Science,2006,313(5786):504-507.
[5] CHEN S Y,MENG Y Q,TANG H C,et al.Robust deep learning-based diagnosis of mixed faults in rotating machinery[J].IEEE/ASME Transactions on Mechatronics,2020,25(5):2167-2176.
[6] WANG Y J,DING X X,ZENG Q,et al.Intelligent rolling bearing fault diagnosis via vision ConvNet[J].IEEE Sensors Journal,2021,21(5):6600-6609.
[7] WANG J X,WANG D Z,WANG S H,et al.Fault diagnosis of bearings based on multi-sensor information fusion and 2D convolutional neural network[J].IEEE Access,2021,9:23717-23725.
[8] SHI Y W,DENG A D,DENG M Q,et al.Enhanced lightweight multiscale convolutional neural network for rolling bearing fault diagnosis[J].IEEE Access,2020,8:217723-217734.
[9] MA B,CAI W D,HAN Y M,et al.A novel probability confidence CNN model and its application in mechanical fault diagnosis[J].IEEE Transactions on Instrumentation and Measurement,2021,70:3517111.
[10] LIU Z L,WANG H,LIU J J,et al.Multitask learning based on lightweight 1DCNN for fault diagnosis of wheelset bearings[J].IEEE Transactions on Instrumentation and Measurement,2021,70:3501711.
[11] ZHANG X L,HAN P,XU L,et al.Research on bearing fault diagnosis of wind turbine gearbox based on 1DCNN-PSO-SVM[J].IEEE Access,2020,8:192248-192258.
[12] CHEN J B,HUANG R Y,ZHAO K,et al.Multiscale convolutional neural network with feature alignment for bearing fault diagnosis[J].IEEE Transactions on Instrumentation and Measurement,2021,70:3517010.
[13] 郭家昕,程军圣,杨宇.改进多线性主成分分析网络及其在滚动轴承故障诊断中的应用[J].中国机械工程,2022,33(2):187-193.
[14] 胡茑庆,陈徽鹏,程哲,等.基于经验模态分解和深度卷积神经网络的行星齿轮箱故障诊断方法[J].机械工程学报,2019,55(7):9-18.
[15] LIU F Z,GAO J W,LIU H B.A fault diagnosis solution of rolling bearing based on MEEMD and QPSO-LSSVM[J].IEEE Access,2020,8:101476-101488.
[16] 许同乐,孟良,孔晓佳,等.基于EEMD的ICNN故障诊断方法[J].北京邮电大学学报,2022,45(2):110-116.
[17] ZHOU B,KHOSLA A,LAPEDRIZA A,et al.Object detectors emerge in deep scene CNNs[C]//Proceedings of the 2015 International Conference on Learning Representations.[S.l.]:ICLR,2015.
[18] ZHAO H S,SHI J P,QI X J,et al.Pyramid scene parsing network[C]//30th IEEE/CVF Conference on Computer Vision and Pattern Recognition.Piscataway,N J:IEEE,2017.
[19] 李世晓,杜锦华,龙云.基于一维卷积神经网络的机电作动器故障诊断[J].电工技术学报,2022,37(增刊1):62-73.
[20] 王琦,邓林峰,赵荣珍.基于改进一维卷积神经网络的滚动轴承故障识[J].振动与冲击,2022,41(3):216-223.
[21] ZHANG W,PENG G L,LI C H,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):425-445.