Journal of Systems Engineering and Electronics ›› 2023, Vol. 34 ›› Issue (4): 1074-1084.doi: 10.23919/JSEE.2023.000109

• RELIABILITY • Previous Articles    

Rolling bearing fault diagnostics based on improved data augmentation and ConvNet

Delanyo Kwame Bensah KULEVOME1,2(), Hong WANG1,2,*(), Xuegang WANG1()   

  1. 1 School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu 611731, China
    2 Yangtze Delta Region Institute (Huzhou), University of Electronic Science and Technology of China, Huzhou 313001, China
  • Received:2022-05-12 Online:2023-08-18 Published:2023-08-28
  • Contact: Hong WANG;;
  • About author:
    KULEVOME Delanyo Kwame Bensah was born in 1983. He received his M.E. degree in electronic science and engineering in 2019 from the University of Electronic Science and Technology of China, Chengdu, China, where he is currently pursuing his Ph.D. degree in information and communication engineering. His research interests include prognostics and health management of systems, fault diagnostics, signal processing, and deep learning. E-mail:

    WANG Hong was born in 1974. He received his B.S., M.S. and Ph.D. degrees from Northwestern Polytechnical University, Chongqing University, and University of Electronic Science and Technology of China (UESTC), respectively. He is a faculty member with UESTC since 2003. From 2007 to 2009 he was engaged in doctoral research with the Second Research Institute of Civil Aviation Administration. From 2009 to 2010, he was a research scholar with Polytechnic Institute of New York University and research assistant with New Jersey Institute of Technology, USA. His research interests include radar signal processing, avionics, aeronautical telecommunication, and surveillance technologies in air traffic control. E-mail:

    WANG Xuegang was born in 1962. He received his Ph.D. degree from Xidian University in 1992. He is now a professor and Ph.D. supervisor with University of Electronic Science and Technology of China. His research interests include radar signal processing, and millimeter wave radar. E-mail:
  • Supported by:
    This work was supported by the National Natural Science Foundation of China (42027805), and the National Aeronautical Fund (ASFC-20172080005)


Convolutional neural networks (CNNs) are well suited to bearing fault classification due to their ability to learn discriminative spectro-temporal patterns. However, gathering sufficient cases of faulty conditions in real-world engineering scenarios to train an intelligent diagnosis system is challenging. This paper proposes a fault diagnosis method combining several augmentation schemes to alleviate the problem of limited fault data. We begin by identifying relevant parameters that influence the construction of a spectrogram. We leverage the uncertainty principle in processing time-frequency domain signals, making it impossible to simultaneously achieve good time and frequency resolutions. A key determinant of this phenomenon is the window function’s choice and length used in implementing the short-time Fourier transform. The Gaussian, Kaiser, and rectangular windows are selected in the experimentation due to their diverse characteristics. The overlap parameter ’s size also influences the outcome and resolution of the spectrogram. A 50% overlap is used in the original data transformation, and ±25% is used in implementing an effective augmentation policy to which two-stage regular CNN can be applied to achieve improved performance. The best model reaches an accuracy of 99.98% and a cross-domain accuracy of 92.54%. When combined with data augmentation, the proposed model yields cutting-edge results.

Key words: bearing failure, short-time Fourier transform, prognostics and health management, data augmentation, fault diagnosis