·1490 工程科学学报.第43卷第11期 2018,99:1 for health state identification of planetary gearboxes. [7]Wang B,Lei Y G,Li N P,et al.Deep separable convolutional Measurement,2019,146:268 network for remaining useful life prediction of machinery [17]Zhang W,Li C H,Peng G L,et al.A deep convolutional neural MechSystSignal Process,2019,134:106330 network with new training methods for bearing fault diagnosis [8]Li X,Zhang W,Ding Q.Deep leaming-based remaining useful life under noisy environment and different working load.Mech Syst estimation of bearings using multi-scale feature extraction. Signal Process,2018,100:439 ReliabEngSystSaf,2019,182:208 [18]Peng DD,Liu Z L,Wang H,et al.A novel deeper one- [9] Dong X F,Lian JJ,Wang HJ.Vibration source identification of dimensional CNN with residual learning for fault diagnosis of offshore wind turbine structure based on optimized spectral wheelset bearings in high-speed trains.IEEE Access,2019,7: kurtosis and ensemble empirical mode decomposition.Ocean Eng, 10278 2019,172:199 [19]Hoang DT,Kang H J.A survey on Deep Learning based bearing [10]Zou YY,de Zhang Y,Mao HC.Fault diagnosis on the bearing of fault diagnosis.Neurocomputing.2019,335:327 traction motor in high-speed trains based on deep learning.Alex [20]KrizhevskyA,Sutskeverl,HintonGE.ImageNet classification with EgJ,2021,60(1):1209 deep convolutional neural networks.Commun 4CM,2017,60(6): [11]Shao H D,Jiang H K,Wang F A,et al.Rolling bearing fault 84 diagnosis using adaptive deep belief network with dual-tree [21]Howard A G,Zhu M L,Chen B,et al.Mobilenets:Efficient complex wavelet packet./S4 Trans,2017,69:187 convolutional neural networks for mobile vision [12]Wang X,Qin Y,Wang Y,et al.ReLTanh:An activation function applications[J/OL].arXiv preprint online (2017-4-17)[2020-12- with vanishing gradient resistance for SAE-based DNNs and its 09].https://arxiv.org/abs/1704.04861 application to rotating machinery fault diagnosis.Neurocompuring, [22]He K M,Zhang X Y,Ren S Q,et al.Delving deep into rectifiers: 2019,363:88 Surpassing human-level performance on ImageNet classification / [13]Deng F Y.Ding H.Yang S P.et al.An improved deep residual 2015 IEEE International Conference on Computer Vision (ICCV). network with multiscale feature fusion for rotating machinery fault Santiago,2015:1026 diagnosis.Meas Sci Technol,2021,32(2):024002 [23]Zhang X Y,Zhou X Y,Lin M X,et al.ShuffleNet:an extremely [14]Wang F T,Liu X F,Deng G,et al.Remaining life prediction efficient convolutional neural network for mobile devices //2018 method for rolling bearing based on the long short-term memory IEEE/CVF Conference on Computer Vision and Pattern network.Neural Process Lett,2019,50(3):2437 Recognition.Salt Lake City,2018:6848 [15]Zhang X,Wan ST.He YL,et al.Teager energy spectral kurtosis [24]Hoang D T.Kang H J.Rolling element bearing fault diagnosis of wavelet packet transform and its application in locating the using convolutional neural network and vibration image sound source of fault bearing of belt conveyor.Measurement. CognSystRes,2019,53:42 2021,173:108367 [25]Chollet F.Xception:deep learning with depthwise separable [16]Chen H P,Hu N Q,Cheng Z,et al.A deep convolutional neural convolutions /2017 IEEE Conference on Computer Vision and network based fusion method of two-direction vibration signal data Pattern Recognition (CVPR).Honolulu,2017:18002018, 99: 1 Wang  B,  Lei  Y  G,  Li  N  P,  et  al.  Deep  separable  convolutional network  for  remaining  useful  life  prediction  of  machinery. MechSystSignal Process, 2019, 134: 106330 [7] Li X, Zhang W, Ding Q. Deep learning-based remaining useful life estimation  of  bearings  using  multi-scale  feature  extraction. ReliabEngSystSaf, 2019, 182: 208 [8] Dong X F, Lian J J, Wang H J. Vibration source identification of offshore  wind  turbine  structure  based  on  optimized  spectral kurtosis and ensemble empirical mode decomposition. Ocean Eng, 2019, 172: 199 [9] Zou Y Y, de Zhang Y, Mao H C. Fault diagnosis on the bearing of traction  motor  in  high-speed  trains  based  on  deep  learning. Alex Eng J, 2021, 60（1）: 1209 [10] Shao  H  D,  Jiang  H  K,  Wang  F  A,  et  al.  Rolling  bearing  fault diagnosis  using  adaptive  deep  belief  network  with  dual-tree complex wavelet packet. ISA Trans, 2017, 69: 187 [11] Wang X, Qin Y, Wang Y, et al. ReLTanh: An activation function with  vanishing  gradient  resistance  for  SAE-based  DNNs  and  its application to rotating machinery fault diagnosis. Neurocomputing, 2019, 363: 88 [12] Deng  F  Y,  Ding  H,  Yang  S  P,  et  al.  An  improved  deep  residual network with multiscale feature fusion for rotating machinery fault diagnosis. Meas Sci Technol, 2021, 32（2）: 024002 [13] Wang  F  T,  Liu  X  F,  Deng  G,  et  al.  Remaining  life  prediction method for rolling bearing based on the long short-term memory network. Neural Process Lett, 2019, 50（3）: 2437 [14] Zhang X, Wan S T, He Y L, et al. Teager energy spectral kurtosis of  wavelet  packet  transform  and  its  application  in  locating  the sound  source  of  fault  bearing  of  belt  conveyor. Measurement, 2021, 173: 108367 [15] Chen H P, Hu N Q, Cheng Z, et al. A deep convolutional neural network based fusion method of two-direction vibration signal data [16] for  health  state  identification  of  planetary  gearboxes. Measurement, 2019, 146: 268 Zhang W, Li C H, Peng G L, et al. A deep convolutional neural network  with  new  training  methods  for  bearing  fault  diagnosis under  noisy  environment  and  different  working  load. Mech Syst Signal Process, 2018, 100: 439 [17] Peng  D  D,  Liu  Z  L,  Wang  H,  et  al.  A  novel  deeper  one￾dimensional  CNN  with  residual  learning  for  fault  diagnosis  of wheelset  bearings  in  high-speed  trains. IEEE Access,  2019,  7: 10278 [18] Hoang D T, Kang H J. A survey on Deep Learning based bearing fault diagnosis. Neurocomputing, 2019, 335: 327 [19] KrizhevskyA, SutskeverI, HintonGE. ImageNet classification with deep convolutional neural networks. Commun ACM, 2017, 60（6）: 84 [20] Howard  A  G,  Zhu  M  L,  Chen  B,  et  al.  Mobilenets:  Efficient convolutional  neural  networks  for  mobile  vision applications[J/OL]. arXiv preprint online (2017-4-17)  [2020-12- 09].https://arxiv.org/abs/1704.04861 [21] He K M, Zhang X Y, Ren S Q, et al. Delving deep into rectifiers: Surpassing human-level performance on ImageNet classification // 2015 IEEE International Conference on Computer Vision (ICCV). Santiago, 2015: 1026 [22] Zhang X Y, Zhou X Y, Lin M X, et al. ShuffleNet: an extremely efficient  convolutional  neural  network  for  mobile  devices  //2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, 2018: 6848 [23] Hoang  D  T,  Kang  H  J.  Rolling  element  bearing  fault  diagnosis using  convolutional  neural  network  and  vibration  image. CognSystRes, 2019, 53: 42 [24] Chollet  F.  Xception:  deep  learning  with  depthwise  separable convolutions  //2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, 2017: 1800 [25] · 1490 · 工程科学学报，第 43 卷，第 11 期