基于加权域适应卷积神经网络的滚动轴承故障诊断

doi:10.16182/j.issn1004731x.joss.22-0616

系统仿真学报 ›› 2023, Vol. 35 ›› Issue (11): 2445-2453.doi: 10.16182/j.issn1004731x.joss.22-0616

基于加权域适应卷积神经网络的滚动轴承故障诊断

张文锋(), 祝志超, 吴定会()

江南大学物联网应用技术教育部工程研究中心，江苏无锡 214122

收稿日期:2022-06-06 修回日期:2022-08-10 出版日期:2023-11-25 发布日期:2023-11-23
通讯作者: 吴定会 E-mail:6211924137@stu.jiangnan.edu.cn;wdh123@jiangnan.edu.cn
第一作者简介:张文锋(1998-)，男，硕士生，研究方向为智能故障诊断。E-mail：6211924137@stu.jiangnan.edu.cn
基金资助:
国家重点研发计划(2020YFB1711102)

Rolling Bearing Fault Diagnosis Based on Weighted Domain Adaptive Convolutional Neural Network

Zhang Wenfeng(), Zhu Zhichao, Wu Dinghui()

Engineering Research Center of Internet of Things Technology Appliations Ministry of Education, Jiangnan University, Wuxi 214122, China

Received:2022-06-06 Revised:2022-08-10 Online:2023-11-25 Published:2023-11-23
Contact: Wu Dinghui E-mail:6211924137@stu.jiangnan.edu.cn;wdh123@jiangnan.edu.cn

摘要/Abstract

摘要：

针对工作负载变化引起滚动轴承振动信号的数据分布变化，导致故障诊断算法的泛化性变差的问题，提出一种基于加权域适应卷积神经网络（weighted domain adaptive convolutional neural network， WDACNN）的滚动轴承故障诊断方法。在卷积神经网络中嵌入域适应算法，使基于源域的分类器在目标域上取得良好的泛化性，并引入权重系数对源域样本进行加权，降低类权重偏差的影响。采用6组迁移任务验证所提方法的有效性，结果表明：平均故障诊断精度达到96.6%，证明了本文方法在不同工作负载条件下的有效性。

关键词: 轴承, 故障诊断, 迁移学习, 域适应, 权重系数

Abstract:

A rolling bearing fault diagnosis method based on a weighted domain adaptive convolutional neural network (WDACNN) is proposed to solve the problem that the data distribution of vibration signals of rolling bearings changes due to workload changes, which leads to poor generalization of fault diagnosis algorithm. In this method, the domain adaptation algorithm is embedded in the convolutional neural network to make the classifier based on the source domain achieve excellent generalization in the target domain, and the weight coefficient is introduced to weight the samples in the source domain to reduce the influence of the class weight deviation. In the simulation experiment, six migration tasks are used to verify the effectiveness of the proposed method. The average fault diagnosis accuracy of the proposed method reaches 96.6%, which proves the effectiveness of the proposed method under different workload conditions.

Key words: bearing, fault diagnosis, transfer learning, domain adaptation, weight coefficient

中图分类号:

TP391.9

张文锋,祝志超,吴定会 . 基于加权域适应卷积神经网络的滚动轴承故障诊断[J]. 系统仿真学报, 2023, 35(11): 2445-2453.

Zhang Wenfeng,Zhu Zhichao,Wu Dinghui . Rolling Bearing Fault Diagnosis Based on Weighted Domain Adaptive Convolutional Neural Network[J]. Journal of System Simulation, 2023, 35(11): 2445-2453.

图/表 8

图1

图2

图3

表1

图4

表2

表3

图5

参考文献 21

1	Wang Huaqing, Li Shi, Song Liuyang, et al. A Novel Convolutional Neural Network Based Fault Recognition Method via Image Fusion of Multi-vibration-signals[J]. Computers in Industry, 2019, 105: 182-190.
2	陈仁祥, 杨星, 胡小林, 等. 深度置信网络迁移学习的行星齿轮箱故障诊断方法[J]. 振动与冲击, 2021, 40(1): 127-133, 150.
	Chen Renxiang, Yang Xing, Hu Xiaolin, et al. Planetary Gearbox Fault Diagnosis Method Based on Deep Belief Network Transfer Learning[J]. Journal of Vibration and Shock, 2021, 40(1): 127-133, 150.
3	张群, 唐振浩, 王恭, 等. 基于长短时记忆网络的超短期风功率预测模型[J]. 太阳能学报, 2021, 42(10): 275-281.
	Zhang Qun, Tang Zhenhao, Wang Gong, et al. Ultra-short-term Wind Power Prediction Model Based on Long and Short Term Memory Network[J]. Acta Energiae Solaris Sinica, 2021, 42(10): 275-281.
4	Liu Han, Zhou Jianzhong, Zheng Yang, et al. Fault Diagnosis of Rolling Bearings with Recurrent Neural Network-based Autoencoders[J]. ISA Transactions, 2018, 77: 167-178.
5	Ben-David S, Blitzer J, Crammer K, et al. A Theory of Learning From Different Domains[J]. Machine Learning, 2010, 79(1): 151-175.
6	Guo Liang, Lei Yaguo, Xing Saibo, et al. Deep Convolutional Transfer Learning Network: A New Method for Intelligent Fault Diagnosis of Machines with Unlabeled Data[J]. IEEE Transactions on Industrial Electronics, 2019, 66(9): 7316-7325.
7	Jin Tongtong, Yan Chuliang, Chen Chuanhai, et al. New Domain Adaptation Method in Shallow and Deep Layers of the CNN for Bearing Fault Diagnosis Under Different Working Conditions[J]. The International Journal of Advanced Manufacturing Technology, 2023, 124(11): 3701-3712.
8	Li Xiang, Zhang Wei. Deep Learning-based Partial Domain Adaptation Method on Intelligent Machinery Fault Diagnostics[J]. IEEE Transactions on Industrial Electronics, 2021, 68(5): 4351-4361.
9	孙琦钰, 赵超强, 唐漾, 等. 基于无监督域自适应的计算机视觉任务研究进展[J]. 中国科学(技术科学), 2022, 52(1): 26-54.
	Sun Qiyu, Zhao Chaoqiang, Tang Yang, et al. A Survey on Unsupervised Domain Adaptation in Computer Vision Tasks[J]. Scientia Sinica(Technologica), 2022, 52(1): 26-54.
10	许亚雲, 严华. 无监督域适应的表示学习算法[J]. 哈尔滨工业大学学报, 2021, 53(2): 40-46.
	Xu Yayun, Yan Hua. Representation Learning for Unsupervised Domain Adaptation[J]. Journal of Harbin Institute of Technology, 2021, 53(2): 40-46.
11	Zhao Minghang, Kang M, Tang Baoping, et al. Deep Residual Networks with Dynamically Weighted Wavelet Coefficients for Fault Diagnosis of Planetary Gearboxes[J]. IEEE Transactions on Industrial Electronics, 2018, 65(5): 4290-4300.
12	Liang Mingxuan, Cao Pei, Tang J. Rolling Bearing Fault Diagnosis Based on Feature Fusion with Parallel Convolutional Neural Network[J]. The International Journal of Advanced Manufacturing Technology, 2021, 112(3): 819-831.
13	崔石玉, 朱志宇. 基于参数迁移和一维卷积神经网络的海水泵故障诊断[J]. 振动与冲击, 2021, 40(24): 180-189.
	Cui Shiyu, Zhu Zhiyu. Seawater Pump Fault Diagnosis Based on Parameter Transfer and One-dimensional Convolutional Neural Network[J]. Journal of Vibration and Shock, 2021, 40(24): 180-189.
14	Kang Guoliang, Jiang Lu, Yang Yi, et al. Contrastive Adaptation Network for Unsupervised Domain Adaptation[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ, USA: IEEE, 2019: 4888-4897.
15	赵小强, 张亚洲. 利用改进卷积神经网络的滚动轴承变工况故障诊断方法[J]. 西安交通大学学报, 2021, 55(12): 108-118.
	Zhao Xiaoqiang, Zhang Yazhou. Improved CNN-based Fault Diagnosis Method for Rolling Bearings Under Variable Working Conditions[J]. Journal of Xi'an Jiaotong University, 2021, 55(12): 108-118.
16	谢银成, 黎曦, 李天, 等. 基于改进ResNet和损失函数的表情识别[J]. 自动化与仪表, 2022, 37(4): 64-69.
	Xie Yincheng, Li Xi, Li Tian, et al. Expression Recognition Based on Improved ResNet and Loss Function[J]. Automation & Instrumentation, 2022, 37(4): 64-69.
17	刘俊锋, 董宝营, 俞翔, 等. 基于FSC-MPE与BP神经网络的滚动轴承故障诊断方法[J]. 中国舰船研究, 2021, 16(6): 183-190.
	Liu Junfeng, Dong Baoying, Yu Xiang, et al. Rolling Bearing Fault Diagnosis Method Based on FSC-MPE and BP Neural Network[J]. Chinese Journal of Ship Research, 2021, 16(6): 183-190.
18	Long Mingsheng, Cao Yue, Cao Zhangjie, et al. Transferable Representation Learning with Deep Adaptation Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 41(12): 3071-3085.
19	Jin Tongtong, Yan Chuliang, Chen Chuanhai, et al. New Domain Adaptation Method in Shallow and Deep Layers of the CNN for Bearing Fault Diagnosis Under Different Working Conditions[J]. The International Journal of Advanced Manufacturing Technology, 2023, 124(11): 3701-3712.
20	Ganin Y, Lempitsky V. Unsupervised Domain Adaptation by Backpropagation[C]//Proceedings of the 32nd International Conference on Machine Learning. Chia Laguna Resort, Sardinia, Italy: PMLR, 2015: 1180-1189.
21	Laurens van der Maaten, Hinton G. Visualizing Data Using t-SNE[J]. Journal of Machine Learning Research, 2008, 9(86): 2579-2605.

健康状态	标签	工作负载	测试样本
滚动体故障	0	A	320
		B	300
		C	280
内圈故障	1	A	300
		B	280
		C	260
外圈故障	2	A	300
		B	280
		C	320
正常	3	A	260
		B	320
		C	300

层数	类型	核数量	核尺寸	步长	激活函数
1	卷积层	1	3×3	2	ReLU
2	池化层	—	2×2	2	—
3	卷积层	8	3×3	1	ReLU
4	池化层	—	2×2	2	—
5	卷积层	16	3×3	1	ReLU
6	池化层	—	2×2	2	—
7	全连接层	1	800	—	Sigmoid
8	全连接层	1	100	—	Sigmoid
9	全连接层	1	20	—	Sigmoid
10	输出层	1	4	—	Softmax

诊断任务	BP	LSTM	CNN	DAN	MACNN	WDACNN
AVG	79.8	83.3	88.5	92.1	94.6	96.6
AB	85.8	87.3	92.9	96.2	97.8	99.1
AC	81.5	82.3	88.2	91.3	93.9	96.5
BA	79.6	86.1	91.3	94.1	96.2	97.1
BC	77.9	81.8	86.1	92.1	95.4	98.9
CA	77.5	79.6	84.6	88.3	91.5	93.2
CB	76.7	82.6	88.1	89.9	92.9	95.3

[1]	袁标, 黄友锐, 徐善永, 荣雪. 光纤二次套塑车间数字孪生系统的构建与应用[J]. 系统仿真学报, 2023, 35(9): 2011-2022.
[2]	马娜, 温廷新, 贾旭, 李晓会. 基于辅助分类网络的跨领域文本情感分类[J]. 系统仿真学报, 2023, 35(4): 721-733.
[3]	刘家瑞, 杨国田, 王孝伟. 基于孪生深度神经网络的风电机组故障诊断方法[J]. 系统仿真学报, 2022, 34(11): 2348-2358.
[4]	祝志超, 吴定会, 岳远昌. 基于监督核熵成分分析的发动机磨损故障诊断[J]. 系统仿真学报, 2022, 34(1): 45-52.
[5]	李元, 耿泽伟. 基于LLE与K均值聚类算法的工业过程故障诊断[J]. 系统仿真学报, 2021, 33(9): 2066-2073.
[6]	王印松, 孙天舒. 基于二分类NMF网络的执行器故障状态评估[J]. 系统仿真学报, 2021, 33(8): 1882-1891.
[7]	尹诗, 侯国莲, 迟岩, 弓林娟, 胡晓东. 风电机组发电机前轴承健康度预测方法及实现[J]. 系统仿真学报, 2021, 33(6): 1323-1333.
[8]	位晶晶, 刘勤明, 叶春明, 李冠林. 基于GA-SVR的小样本数据缺失下的设备故障诊断[J]. 系统仿真学报, 2021, 33(6): 1342-1349.
[9]	吴定会, 祝志超, 韩欣宏. 基于BO-SDAE多源信号的风电机组轴承故障诊断[J]. 系统仿真学报, 2021, 33(5): 1148-1156.
[10]	胡小曼, 王艳, 纪志成. 模糊信息粒化与改进RVM的滚动轴承寿命预测[J]. 系统仿真学报, 2021, 33(11): 2561-2571.
[11]	王素琴, 张峰, 高宇豆, 石敏. 基于图像序列的学习表情识别[J]. 系统仿真学报, 2020, 32(7): 1322-1330.
[12]	李炜, 周丙相, 蒋栋年. 基于深度学习序贯检验的电源车故障诊断方法[J]. 系统仿真学报, 2020, 32(4): 638-648.
[13]	樊名鲁, 王艳, 纪志成. 基于特征生成的轴承不均衡数据故障诊断[J]. 系统仿真学报, 2020, 32(12): 2438-2448.
[14]	张圣东, 杨秀英. 计入多因素作用的船舶轴承负荷研究[J]. 系统仿真学报, 2019, 31(6): 1158-1164.
[15]	李炜, 葛振福, 申富媛, 龚建兴. 基于FDRE的3D威亚编码器故障检测与容错控制[J]. 系统仿真学报, 2019, 31(4): 740-749.

基于加权域适应卷积神经网络的滚动轴承故障诊断

Rolling Bearing Fault Diagnosis Based on Weighted Domain Adaptive Convolutional Neural Network

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 21

相关文章 15

编辑推荐

Metrics

本文评价