用于轴承故障诊断的两步迁移学习法

doi:10.3778/j.issn.1002-8331.2009-0207

摘要/Abstract

摘要： 由于轴承故障数据存在数据量少和分布不均衡的问题，将迁移学习引入故障诊断领域，同时由于轴承故障数据的分布与源数据集分布差异巨大，直接采用迁移学习的方法会产生负迁移效应，即由于源数据集与目标数据集间分布差异过大而导致无法学习到源数据集的知识，提出一种对迁移学习进行改进的诊断新方法：即两步迁移学习法，使用DCGAN来制作辅助数据集，在辅助数据集上进行迁移学习，再将网络放在目标数据集上再次进行迁移学习训练，根据与普通迁移学习和不使用迁移学习的对比实验，新方法相较于目前已有的方法具有更快的速度与更高的准确率。

关键词: 迁移学习, 轴承故障诊断, 深度卷积对抗生成网络

Abstract: Due to the problem of small amount of data and uneven distribution of bearing fault data, transfer learning is introduced into the field of fault diagnosis. At the same time, because the distribution of bearing fault data is very different from the distribution of the source data set, the direct use of transfer learning methods will produce negative transfer effects. That is, because the distribution difference between the source data set and the target data set is too large, it is impossible to learn the knowledge of the source data set. A new diagnosis method for improving transfer learning is proposed：the two-step transfer learning method, using DCGAN to make the auxiliary for the data set. It first performs transfer learning on the auxiliary data set, and then puts the network on the target data set for transfer learning training again. According to the comparison experiment with ordinary transfer learning and without transfer learning, the new method has faster speed and higher accuracy compared with the current methods.

Key words: transfer learning, bearing fault diagnosis, deep convolutional generative adversarial network

陶启生, 彭成, 满君丰, 刘翊. 用于轴承故障诊断的两步迁移学习法[J]. 计算机工程与应用, 2022, 58(2): 303-312.

TAO Qisheng, PENG Cheng, MAN Junfeng, LIU Yi. Two-Step Transfer Learning Method for Bearing Fault Diagnosis[J]. Computer Engineering and Applications, 2022, 58(2): 303-312.

参考文献

[1] HOANG D T，KANG H J.Rolling element bearing fault diagnosis using convolutional neural network and vibration image[J].Cognitive Systems Research，2019，53：42-50.
[2] 李巍华，单外平，曾雪琼.基于深度信念网络的轴承故障分类识别[J].振动工程学报，2016，29（2）：340-347.
LI W H，SHAN W P，ZENG X Q.Bearing fault identification based on deep belief network[J].Journal of Vibration Engineering，2016，29（2）：340-347.
[3] 赵海生，刘政，张伟，等.突变磁场对W6Mo5Cr4V2高速钢刀具硬度的影响[J].钢铁研究学报，2018，30（5）：368-372.
ZHAO H S，LIU Z，ZHANG W.Effects of alternating magnetic field on hardness of W6Mo5Cr4V2 high-speed steel cutting tool[J].Journal of Iron and Steel Research，2018，30（5）：368-372.
[4] TAN B，ZHANG Y，PAN S J，et al.Distant domain transfer learning[C]//Proceedings of AAAI，2017：2604-2610.
[5] 莫代一，崔玲丽，王婧.基于双重Q因子的稀疏分解放在滚动轴承早期故障诊断中的引用[J].机械工程学报，2013，24（1）：35-41.
MO D Y，CUI L L，WANG J.Sparse singal decomposition method based on the daul Q-factor and its application to rolling bearing early fault diagnosis[J].Journal of Mechanical Engineering，2013，24（1）：35-41.
[6] 程军圣，于德介，杨宇.基于内禀模态奇异值分解与支持向量机的故障诊断方法[J].自动化学报，2006，32（3）：476-480.
CHENG J S，YU D J，YANG Y.Fault diagnosis approach based on intrinsic mode singular value decomposition and support vector machines[J].Acta Automatica Sinica，2006，32（3）：476-480.
[7] LIU B，RIEMENSCHNEIDER S，XU Y.Gearbox fault diagnosis using empirical mode decomposition and Hilbert spectrum[J].Mechanical Systems & Signal Processing，2006，20（3）：718-734.
[8] 郑慧峰，喻桑桑，王月兵，等.基于经验模态分解和奇异值分解的振动声调制信号分析方法研究[J].计量学报，2016（37）：398-401.
ZHENG H F，YU S S，WANG Y B.Research on the analysis method of vibro-acoustic modulation signal based on EMD and SVD[J].Acta Metrologica Sinica，2016（37）：398-401.
[9] ZOSSO D，DRAGOMIRETSKIY K，BERTOZZI A L，et al.Two-dimensional compact variational mode decomposition[J].Journal of Mathematical Imaging & Vision，2017，58（2）：294-320.
[10] MCDONALD G L，ZHAO Q，ZUO M J.Maximum correlated kurtosis deconvolution and application on gear tooth chip fault detection[J].Mechanical Systems and Signal Processing，2012，33：237-255.
[11] GAO Z W，CECATI C，DING S X.A survey of fault diagnosis and fault-tolerant techniques—part I：fault diagnosis with model-based and signal-based approaches[J].IEEE Transactions on Industrial Electronics，2015，62（6）：3757-3767.
[12] 刘正平，胡浚，张龙.基于堆栈降噪自编码的轴承故障诊断方法[J].机床与液压，2018，46（15）：177-181.
LIU Z P，HU J，ZHANG L.Fault diagnosis method of bearing based on stacked denoising autoencoder[J].Machine Tool & Hydraulics，2018，46（15）：177-181.
[13] YANG B，LEI Y，JIA F，et al.An intelligent fault diagnosis approach based on transfer learning from laboratory bearings to locomotive bearings[J].Mechanical Systems and Signal Processing，2019，122：692-706.
[14] 李林杰.基于二维数据训练的深度学习滚动轴承故障诊断方法研究[D].成都：电子科技大学，2020.
LI L J.Research on deep learning rolling bearing fault diagnosis method based on two-dimensional data training[D].Chengdu：University of Electronic Science and Technology of China，2020.
[15] BENGIO Y，COURVILLE A.Deep learning of representations[C]//Handbook on neural information processing.[S.l.]：Springer，2013：1-28.
[16] PAN S J，YANG Q.A survey on transfer learning[J].IEEE Transactions on Knowledge and Data Engineering，2010，22（10）：1345-1359.
[17] GOODFELLOW I J，POUGET-ABADIE J，MIRZA M，et al.Generative adversarial nets[C]//Proceedings of the 27th International Conference on Neural Information Processing Systems，Canada，December 13，2014.Montreal：MIT Press，2014：2672-2680.
[18] RADFORD A，METZ L，CHINTALA S.Unsupervised representation learning with deep convolutional generative adversarial networks[J].arXiv：1511.06434，2015.
[19] HE K M，ZHANG X Y，REN S Q，et al.Deep residual learning for image recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition，2016：770-778.
[20] XU B，WANG N，CHEN T，et al.Empirical evaluation of rectified activations in convolutional network[J].arXiv：1505.00853，2015.
[21] KINGMA D P，BA J.Adam：a method for stochastic optimization[J].arXiv：1412.6980，2014.