Data equalization processing method for high-speed train wheelset bearings based on CYCLEGAN-FAM

LIU Suyan1,2, WANG Haoning2, MA Zengqiang1,2,3, YUAN Zonghao3,4

Journal of Vibration and Shock ›› 2024, Vol. 43 ›› Issue (15) : 32-43.

PDF(4072 KB)
PDF(4072 KB)
Journal of Vibration and Shock ›› 2024, Vol. 43 ›› Issue (15) : 32-43.

Data equalization processing method for high-speed train wheelset bearings based on CYCLEGAN-FAM

  • LIU Suyan1,2, WANG Haoning2, MA Zengqiang1,2,3, YUAN Zonghao3,4
Author information +
History +

Abstract

Once the rolling bearings of high-speed trains fail, they will be stopped for maintenance, resulting in extremely unbalanced sample data. The unbalance of data sets will have an important impact on the accuracy and stability of fault diagnosis results. To solve this problem, a Cyclegan-FAM method for processing unbalanced bearing data based on feature attention matching (FAM) and CYCLEGAN was proposed. In this method, a feature attention matching module was added to the discriminator of CYCLEGAN to align the features extracted from real images and generated images, thereby improving the quality of generated samples. Experiments show that this method can generate generated samples that are highly similar to real samples. With unbalanced data sets being gradually balanced, the accuracy of fault diagnosis can reach 99.8% and 99.2% respectively on CWRU Class 4 and Class 10 data sets, and 99.4% and 99.6% respectively on QPZZ-II Class 4 and Class 10 data sets.

Key words

Generative adversarial network / Feature attention matching / Unbalanced data sets / Fault diagnosis

Cite this article

Download Citations
LIU Suyan1,2, WANG Haoning2, MA Zengqiang1,2,3, YUAN Zonghao3,4. Data equalization processing method for high-speed train wheelset bearings based on CYCLEGAN-FAM[J]. Journal of Vibration and Shock, 2024, 43(15): 32-43

References

1.石佳,黄宇峰,王锋.基于ACMD与改进MOMEDA的滚动轴承故障诊断[J].振动与冲击,2023,42(16) :218-226+261. Shi Jia,HUANG Yufeng,Wang Feng. Fault Diagnosis of Rolling Bearing Based on ACMD and Improved MOMEDA [J]. Journal of Vibration and Shock,2023,42(16) :218-226+261. 2.李俊,苏凯,张皓光等.基于多注意力机制的端到端滚动轴承故障诊断方法[J].空军工程大学学报,2023,24(04) :28-34. Li Jun,Su Kai,Zhang Haoguang,et al. End-to-end Rolling Bearing Fault diagnosis Method based on Multi-attention Mechanism [J]. Journal of Air Force Engineering University,2023,24(04) :28-34. 3.BAO S Y,FENG J,XU X B,et al. Multi-input parallel graph neural network for semi-supervised rolling bearing fault diagnosis[J]. Measurement Science and Technology,2023,34(5). 4.WANG H S,WEI J,LI P J.Research on Fault Diagnosis Technology Based on Deep Learning[J].Journal of Physics: Conference Series,2022,2187(1). 5.TANG S J,WANG C G,ZHOU F N,et al. Multi-Scale Recursive Semi-Supervised Deep Learning Fault Diagnosis Method with Attention Gate[J]. Machines,2023,11(2) :153-153. 6.金江涛,许子非,李春等.基于深度学习与混沌特征融合的滚动轴承故障诊断[J].控制理论与应用,2022,39(01) :109-116. Jin Jiangtao,Xu Zifei,Li Chun,et al. Fault Diagnosis of Rolling Bearing based on Deep Learning and Chaotic Feature Fusion [J]. Control Theory and Applications,202,39(01) :109-116. 7.曹洁,何智栋,余萍等.数据不平衡分布下轴承故障诊断方法[J].吉林大学学报(工学版),2022,52(11) :2523-2531. Cao Jie,HE Zhidong,Yu Ping,et al. Bearing Fault Diagnosis Method with Unbalanced Data Distribution [J]. Journal of Jilin University (Engineering and Technology Edition), 2022,52 (11) :2523-2531. 8.李梦男,李琨,吴聪.基于IWAE的不平衡数据集下轴承故障诊断研究[J].机械强度,2023,45(03):569-575. Li Mengnan,LI Kun,WU Cong. Research on Bearing Fault Diagnosis Based on IWAE Unbalanced Data Set [J]. Journal of Mechanical Strength,2023,45 (03) :569-575. 9.陶新民,李晨曦,李青等.不均衡最大软间隔SVDD轴承故障检测模型[J].振动工程学报,2019,32(04) :718-729. Tao Xinmin,LI Chenxi,Li Qing,et al. Fault Detection Model of SVDD Bearing with Unbalanced Maximum Soft Interval [J]. Journal of Vibration Engineering, 2019,32 (04) :718-729. 10.ZONG X,YANG R,WANG H S,et al.Semi-Supervised Transfer Learning Method for Bearing Fault Diagnosis with Imbalanced Data[J].Machines,2022,10(7) :515-515. 11.李梦男,李琨,吴聪.基于IWAE的不平衡数据集下轴承故障诊断研究[J].机械强度,2023,45(03) :569-575. Li Mengnan,LI Kun,WU Cong. Research on Bearing Fault Diagnosis Based on IWAE Unbalanced Data Set [J]. Journal of Mechanical Strength,2023,45 (03) :569-575. 12.Goodfellow I,Mirza M,XU B,et al. Generative Adversarial Nets[J]. Advances in Neural Information Processing Systems,2014,1406.2661. 13.Radford A , Metz L , Chintala S .Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks[J].Computerence,2015.1511.06434. 14.HU Z Y,HAN T R,BIAN J,et al.A deep feature extraction approach for bearing fault diagnosis based on multi-scale convolutional autoencoder and generative adversarial networks[J]. Measurement Science and Technology,2022,33(6). 15.LI D P,YANG Y C,SHEN S K,et al. Research on Fault Diagnosis based on Improved Generative Adversarial Network under Small Samples[J]. IAENG International Journal of Computer Science,2023,50(1). 16.ZHU F Z,WANG C,ZHU B,et al. An improved generative adversarial networks for remote sensing image super-resolution reconstruction via multi-scale residual block[J]. The Egyptian Journal of Remote Sensing and Space Sciences,2023,26(1):151-160. 17.TONG Q B,LU F Y,FENG Z W,et al. A Novel Method for Fault Diagnosis of Bearings with Small and Imbalanced Data Based on Generative Adversarial Networks[J]. Applied Sciences,2022,12(14):7346-7346. 18.ZHU J Y,PQRK T,ISOLA P,et al. Unpaired image-to-image translation using cycle-consistent adversarial networks[C]//Proceedings of the IEEE international conference on computer vision. 2017:2223-2232. 19.李泽东,李志农,陶俊勇等. 基于特征融合的注意力增强卷积神经网络的航空发动机滚动轴承故障诊断方法[J].兵工学报,2022,43(12):3228-3239. Li Zedong,Li Zhinong,Tao Junyong,et al. An Attention-enhanced Convolutional Neural network based on Feature Fusion for aero-engine rolling bearing fault diagnosis [J]. Journal of Ordnance Engineering,2002,43(12) :3228-3239. 20.李静娇,陈恩利,刘永强.BSS与Morlet小波变换在轴承声学故障诊断中的研究[J].机械强度,2018,40(03):528-533. Li Jingjiao,Chen Enli,Liu Yongqiang.Research of BSS and Morlet wavelet transform in bearing acoustic fault diagnosis[J].mechanical intensity,2018,40(03):528-533. 21.刘兆宸,谢庆,王春鑫等.基于CycleGAN与深度残差网络的局放数据增强与模式识别方法[J].高压电器,2022,58(11):106-113. Liu Zhaochen,Xie Qing,Wang Chunxin,et al. Local release data enhancement and pattern recognition method based on CycleGAN and deep residual network [J]. High Voltage Electrical Apparatus, 2019,58(11) :106-113.
PDF(4072 KB)

Accesses

Citation

Detail

Sections
Recommended

/