[1] Wu Jiachen, Hu Jianzhong, Xu Yadong,. Fault diagnosis method of rolling bearing based onthreshold denoising synchrosqueezing transform and CNN [J]. Journal of Southeast University (English Edition), 2020, 36 (1): 32-40. [doi:10.3969/j.issn.1003-7985.2020.01.005]

Copy

Fault diagnosis method of rolling bearing based onthreshold denoising synchrosqueezing transform and CNN()

基于阈值降噪同步压缩变换和CNN的滚动轴承故障诊断方法

Share：

Journal of Southeast University (English Edition)[ISSN:1003-7985/CN:32-1325/N]

Volumn:

36

Issue:

2020 1

Page:

32-40

Research Field:

Mechanical Engineering

Publishing date:

2020-03-20

Info

Title:

Fault diagnosis method of rolling bearing based onthreshold denoising synchrosqueezing transform and CNN

基于阈值降噪同步压缩变换和CNN的滚动轴承故障诊断方法

Author(s):

Wu Jiachen;  Hu Jianzhong;  Xu Yadong

School of Mechanical Engineering, Southeast University, Nanjing 211189, China

吴佳晨;  胡建中;  徐亚东

东南大学机械工程学院, 南京 211189

Keywords:

threshold denoising;  synchrosqueezing transform;  convolutional neural network;  rolling bearing

阈值降噪;  同步压缩变换;  卷积神经网络;  滚动轴承

PACS:

TH133.3;TP18

DOI:

10.3969/j.issn.1003-7985.2020.01.005

Abstract:

The rolling bearing vibration signal is non-stationary and is easily disturbed by background noise, so it is difficult to accurately diagnose bearing faults. A fault diagnosis method of rolling bearing based on the time-frequency threshold denoising synchrosqueezing transform(TDSST)and convolutional neural network(CNN)is proposed. Since the traditional methods of wavelet threshold denoising and wavelet adjacent coefficient denoising are greatly affected by the estimation accuracy of noise variance, a time-frequency denoising method based on the STFT spectral correlation coefficient threshold optimization is adopted, which is combined with a synchrosqueezing transform. The ability of the TDSST to reduce noise and improve time-frequency resolution was verified by simulated impact fault signals of rolling bearings. Finally, the CNN is utilized to diagnose the time-frequency diagrams obtained by the TDSST. The diagnostic results of the rolling bearing experimental data show that the proposed method can effectively improve the accuracy of diagnosis. When the SNR of the bearing signal is larger than 0 dB, the accuracy is over 95%, even when the SNR reduces to -4 dB, the accuracy is still around 80%. Moreover, the standard deviation of multiple test results is small, which means that the method has good robustness.

针对滚动轴承振动信号的非平稳性和易被背景噪声干扰导致故障难以被准确诊断的问题, 提出了一种基于时频阈值降噪同步压缩变换(TDSST)和卷积神经网络(CNN)的滚动轴承故障诊断方法.由于传统的小波阈值降噪及小波相邻系数降噪方法受信号噪声方差估计精度影响大, 因此采用了基于STFT谱相关系数阈值寻优的时频降噪方法, 将其与同步压缩变换结合, 并用滚动轴承模拟冲击故障信号验证了TDSST方法降噪及提高时频分辨率的能力.最后, 利用CNN对TDSST方法处理得到时频图进行诊断, 滚动轴承实验数据诊断结果表明了所提方法能够有效地提高诊断准确率, 当轴承信号信噪比大于0 dB时, 诊断准确率都达到了95%以上, 即使信噪比降到-4 dB时, 诊断准确率也维持在80%左右, 并且多次测试结果的标准差较小, 表明方法具有良好的鲁棒性.

References:

[1] Li J M, Yao X F, Wang H, et al. Periodic impulses extraction based on improved adaptive VMD and sparse code shrinkage denoising and its application in rotating machinery fault diagnosis[J].Mechanical Systems and Signal Processing, 2019, 126: 568-589. DOI:10.1016/j.ymssp.2019.02.056.
[2] Guo J C, Zhen D, Li H Y, et al. Fault feature extraction for rolling element bearing diagnosis based on a multi-stage noise reduction method[J].Measurement, 2019, 139: 226-235. DOI:10.1016/j.measurement.2019.02.072.
[3] Yan X A, Liu Y, Jia M P. A feature selection framework-based multiscale morphological analysis algorithm for fault diagnosis of rolling element bearing[J].IEEE Access, 2019, 7: 123436-123452. DOI:10.1109/access.2019.2937751.
[4] Yan X A, Liu Y, Jia M P, et al. A multi-stage hybrid fault diagnosis approach for rolling element bearing under various working conditions[J].IEEE Access, 2019, 7: 138426-138441. DOI:10.1109/access.2019.2937828.
[5] Li H K, Yang R, Ren Y J, et al. Rolling element bearing diagnosis using particle filter and kurtogram[J]. Journal of Mechanical Engineering, 2017, 53(3): 63-72. DOI:10.3901/JME.2017.03.063. (in Chinese)
[6] Zhou F Y, Jin L P, Dong J. Review of convolutional neural network[J]. Chinese Journal of Computers, 2017, 40(6): 1229-1251.(in Chinese)
[7] Eren L, Ince T, Kiranyaz S. A generic intelligent bearing fault diagnosis system using compact adaptive 1D CNN classifier[J].Journal of Signal Processing Systems, 2019, 91(2): 179-189. DOI:10.1007/s11265-018-1378-3.
[8] Liu X C, Zhou Q C, Zhao J, et al. Fault diagnosis of rotating machinery under noisy environment conditions based on a 1-D convolutional autoencoder and 1-D convolutional neural network[J].Sensors, 2019, 19(4): 972-1-972-20. DOI:10.3390/s19040972.
[9] Li H, Zhang Q, Qin X R, 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. DOI:10.13465/j.cnki.jvs.2018.19.020. (in Chinese)
[10] Oberlin T, Meignen S, Perrier V. The Fourier-based synchrosqueezing transform[C]//2014 IEEE International Conference on Acoustics, Speech and Signal Processing(ICASSP). Florence, Italy, 2014: 315-319. DOI:10.1109/icassp.2014.6853609.
[11] Stankovic L, Djurovic I, Stankovic S, et al. Instantaneous frequency in time-frequency analysis: Enhanced concepts and performance of estimation algorithms[J].Digital Signal Processing, 2014, 35: 1-13. DOI:10.1016/j.dsp.2014.09.008.
[12] Yu G, Wang Z H, Zhao P. Multisynchrosqueezing transform[J].IEEE Transactions on Industrial Electronics, 2019, 66(7): 5441-5455. DOI:10.1109/tie.2018.2868296.
[13] Su W S, Wang F T, Zhu H, et al. Rolling element bearing faults diagnosis based on optimal Morlet wavelet filter and autocorrelation enhancement[J].Mechanical Systems and Signal Processing, 2010, 24(5): 1458-1472. DOI:10.1016/j.ymssp.2009.11.011.
[14] Cai T T, Silverman B W. Incorporating information on neighboring coefficients into wavelet estimation[J]. Sankhya, 2001, 63(2): 127-148.
[15] Wang S B, Chen X F, Cai G G, et al. Matching demodulation transform and synchrosqueezing in time-frequency analysis[J]. IEEE Transactions on Signal Processing, 2014, 62(1): 69-84. DOI:10.1109/tsp.2013.2276393.
[16] Wang Y H, Xu C, Xu C, et al. Packing convolutional neural networks in the frequency domain[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 41(10): 2495-2510. DOI:10.1109/tpami.2018.2857824.
[17] Sun J, Cao W, Xu Z, et al. Learning a convolutional neural network for non-uniform motion blur removal[C]//2015 IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Boston, MA, USA, 2015: 769-777. DOI:10.1109/cvpr.2015.7298677.
[18] Yuan J H, Han T, Tang J, et al. Intelligent fault diagnosis method for rolling bearings based on wavelet time-frequency diagram and CNN[J]. Machine Design and Research, 2017, 33(2): 93-97.(in Chinese)

Memo

Memo:

Biographies: Wu Jiachen(1995—), male, graduate; Hu Jianzhong(corresponding author), male, doctor, associate professor, hjz@seu.edu.cn.
Citation: Wu Jiachen, Hu Jianzhong, Xu Yadong. Fault diagnosis method of rolling bearing based on threshold denoising synchrosqueezing transform and CNN[J].Journal of Southeast University(English Edition), 2020, 36(1):32-40.DOI:10.3969/j.issn.1003-7985.2020.01.005.

Common functions