A New Approach to Diagnose Rolling Bearing Faults Based on AFD

  • Yu Liang
  • Li min Jia
  • Guo qiang Cai
  • Jin zhao Liu
Conference paper
Part of the Lecture Notes in Electrical Engineering book series (LNEE, volume 288)


A new fault diagnosis method of rolling bearing based on adaptive Fourier decomposition (AFD) is proposed. The new approach extracts the meaningful bearing vibration signal based on AFD algorithm instead of traditional band-pass filter; AFD decomposes the original bearing vibration signal into a series of mono-components, the kurtosis of each mono-component is calculated and clustered into two classes by fuzzy C-mean clustering (FCM). The mean of the two cluster centers is taken as threshold and the mono-components with large kurtosis is summed as bearing fault carrier signal because the bearing fault is sensitive to kurtosis; Then demodulated resonance technique is used to diagnose and locate the fault. The new approach can diagnose all kinds of rolling bearings’ fault. Finally, the proposed approach is used to analysis the outer ring fault in case of N205EM type rolling bearing; the experiments indicates that the effectiveness and accuracy are significantly approved.


Fault diagnosis Adaptive Fourier decomposition Mono-component Resonant demodulation 



This work is supported by the National High Technology Research and Development Program of China (863 Program) (2011AA110501).


  1. 1.
    Howard I (1994) A review of rolling element bearing vibration “detection, diagnosis and prognosis”. Melbourne, Victoria 3001 Australia. DSTO aeronautical and maritime research laboratoryGoogle Scholar
  2. 2.
    Cheng JS, Yu DJ, Yang Y (2005) Time–energy density analysis based on wavelet transform. NDT E Int 38:569–572CrossRefGoogle Scholar
  3. 3.
    Peng ZK, Peter TW, Chu FL (2005) An improved Hilbert–Huang transform and its application in vibration signal analysis. J Sound Vib 286:187–205Google Scholar
  4. 4.
    Huang NE, Shen Z, Long SR et al (1998) The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis. Proc R Soc Lond 454:903–995MathSciNetCrossRefMATHGoogle Scholar
  5. 5.
    Gai GH (2006) The processing of rotor startup signals based on empirical mode decomposition. Mech Syst Sig Proc 20:222–235CrossRefGoogle Scholar
  6. 6.
    Wang WY (2001) Early detection of gear tooth cracking using the resonance demodulation technique. Mech Syst Sig Process 15(5):887–903CrossRefGoogle Scholar
  7. 7.
    Tang DY (2006) Generalized resonance. resonant demodulation fault diagnosis and safety engineering-for railway. China railway pressGoogle Scholar
  8. 8.
    Qian T, Wang YB (2011) Adaptive decomposition into basic signals of non-negative instantaneous frequencies—a variation and realization of greedy algorithm. Adv Comput Math 34(3):279–293Google Scholar
  9. 9.
    Qian T, Zhang LM, Li ZX (2011) Algorithm of adaptive fourier decomposition. IEEE Trans Sig Process 59(12):5899–5906MathSciNetCrossRefGoogle Scholar
  10. 10.
    Qian T, Wang YB, Dang P Adaptive decomposition into mono-components. Adv Adapt Data Anal. doi:  10.1142/S1793536909000278
  11. 11.
    Bozchalooi IS, Liang M (2008) A joint resonance frequency estimation and in-band noise reduction method for enhancing the detectability of bearing fault signals. Mech Syst Sig Process 22:915–933CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • Yu Liang
    • 1
  • Li min Jia
    • 1
  • Guo qiang Cai
    • 1
  • Jin zhao Liu
    • 2
  1. 1.Beijing Jiaotong UniversityBeijingChina
  2. 2.China Academy of Railway SciencesBeijingChina

Personalised recommendations