|
|
Intelligent fault diagnosis of rolling bearing based on EMDPWVD time-frequency images and improved ViT network |
FAN Hongwei1,2, MA Ningge1, MA Jiateng1, CHEN Buran1, CAO Xiangang1,2, ZHANG Xuhui1,2 |
1. School of Mechanical Engineering, Xi’an University of Science and Technology, Xi’an 710054, China;
2. Shaanxi Provincial Key Lab of Mine Electromechanical Equipment Intelligent Detection and Control, Xi’an 710054, China |
|
|
Abstract Rolling bearing is one of the key components of mechanical equipment, and its fault diagnosis is crucial for the safe and stable operation of equipment. For rolling bearings with non-stationary vibration signal, a new intelligent fault diagnosis method of EMDPWVD time-frequency images combined with improved Vision Transformer(ViT) network model is proposed. For the actual signals, three time-frequency analysis methods as Short-time Fourier Transform(STFT), Continuous Wavelet Transform(CWT), and Empirical Mode Decomposition & Pseudo-Wigner-ville Distribution(EMDPWVD) are firstly studied. Considering that STFT and CWT cannot simultaneously achieve the high time and frequency resolution, EMDPWVD is selected as the time-frequency image preparation method used for the intelligent fault diagnosis network. Secondly, a typical ViT is used as the basic fault diagnosis model, which divides time-frequency images into the blocks with predetermined size and then linearly maps them into input sequences. Meanwhile, the global information of the images is integrated by a self attention mechanism, and the network transmission is completed using a stacked Transformer encoder to finally achieve the fault diagnosis. To further improve the accuracy of fault diagnosis, the pooling layer is used as the preprocessing network of ViT to obtain an improved Pooling Vision Transformer(PiT) model, which extends the spatial features of time-frequency images and enhances the sensitivity of the model to the input images. The results show that the proposed method has the high diagnosis accuracy for different fault types of rolling bearings, and the accuracy of PiT is 4.40% higher than that of ViT, which proves that adding pooling layers to ViT can improve the effect of rolling bearing fault diagnosis.
|
Received: 17 May 2023
Published: 15 June 2024
|
|
|
|
[1] 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 Qin, QIN Xianrong, 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.
[2] 何岭松,李巍华. 用Morlet小波进行包络检波分析[J]. 振动工程学报,2002,(01): 123-126.
HE Lingsong, LI Weihua. Morlet Wavelet and its application in enveloping[J]. Journal of Vibration Engineering, 2002, (01): 123-126.
[3] HUANG N E, SHEN Z, LONG S R. The empirical mode decom-position and the Hilbert spectrum for nonlinear and nonsta-tionary time series analysis[J]. Proc.R.Soc, 1998, 454: 903-905.
[4] CHEN G, CHEN J, DONG G M. Chirplet Wigner–Ville distribution for time-frequency representation and its application[J]. Mechanical Systems and Signal Processing, 2013, 41(1): 1-13.
[5] LIU X, JIA Y X, HE Z W, et al. Application of EMD-WVD and particle filter for gearbox fault feature extraction and remaining useful life prediction[J]. Journal of Vibroengineering, 2017, 19(3).
[6] 牟伟杰,石林锁,蔡艳平,等. 基于EMD-WVD与LNMF的内燃机故障诊断[J]. 振动与冲击,2016,35(23): 191-202.
MU Weijie, SHI Linsuo, CAI Yanping, et al. IC engine fault diagnosis method based EMD-WVD and LNMF[J]. Journal of Vibration and Shock, 2016, 35(23): 191-202.
[7] FAN H W, SHAO S J, ZHANG X H, et al. Intelligent fault diagnosis of rolling bearing using FCM clustering of EMD-PWVD vibration images[J]. IEEE Access, 2020(8):145194-145206.
[8] ZHAO Z, XU Q, JIA M. Improved shuffled frog leaping algorithm-based BP neural network and its application in bearing early fault diagnosis[J]. Neural Computing and Applications, 2016, 27(2): 375-385.
[9] 张超,陈建军,郭迅. 基于EMD能量熵和支持向量机的齿轮故障诊断方法[J]. 振动与冲击,2010,29(10): 216-220+261.
ZHANG Chao, CHEN Jianjun, GUO Xun. A gear fault diagnosis method based on EMD energy entropy and SVM[J]. Journal of Vibration and Shock, 2010, 29(10): 216-220+261.
[10] DI J, WANG L L. Application of Improved Deep Auto-Encoder Network in Rolling Bearing Fault Diagnosis[J]. Journal of Computer and Communications, 2018, 6(7): 41-53.
[11] 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(02): 340-347.
LI Weihgua, SHAN Waiping, ZENG Xueqiong. Bearing fault identification based on deep belief network[J]. Journal of Vibration Engineering, 2016, 29(02): 340-347.
[12] 刘星辰,周奇才,赵炯,等. 一维卷积神经网络实时抗噪故障诊断算法[J]. 哈尔滨工业大学学报,2019,51(07): 89-95.
LIU Xingchen, ZHOU Qicai, ZHAO Jiong, et al. Real-time and anti-noise fault diagnosis algorithm based on1-D convolutional neural network[J]. Journal of Harbin Institute of Technology, 2019, 51(07): 89-95.
[13] FAN H W, XUE C Y, ZHANG X H, et al. Vibration Images-Driven Fault Diagnosis Based on CNN and Transfer Learning of Rolling Bearing under Strong Noise[J]. Shock and Vibration, 2021.
[14] LECUN Y., BOSER B., DENKER J. S., et al. Backpropagation applied to handwritten zip code recognition[J]. Neural Computation, 1989, 541-551.
[15] ASHISH V., NOAM S., NIKI P., et al. Attention is all you need[J]. Neural Inf Process Syst, 2017, 30: 5998-6008.
[16] LIU H, ZHOU J Z, ZHENG Y, et al. Fault diagnosis of rolling bearings with recurrent neural network basedautoencoders[J]. ISA Transactions, 2018, 77167-178.
[17] 陈伟,陈锦雄,江永全,等. 基于RS-LSTM的滚动轴承故障识别[J]. 中国科技论文,2018, 13(10): 1134-1141.
CHEN Wei, CHEN Jinxiong, JIANG Yongquan, et al. Fault identification of rolling bearing based on RS-LSTM[J]. China Sciencepaper, 2018, 13(10): 1134-1141.
[18] DOSOVITSKIY A., BEYER L., KOLESNIKOV A., et al. An image is worth 16×16 words: Transformers for image recognition at scale. arXiv 2020, arXiv: 2010.11929.
[19] 郝欢,王华力,魏勤. 经验模态分解理论及其应用[J]. 高技术通讯, 2016, 26(01): 67-80.
HAO Huan, WANG HuaLi, WEI Qin, et al. Empirical mode decomposition theory and its application[J]. High Technology Letters, 2016, 26(01): 67-80. |
|
|
|