Abstract: In order to solve the problems that the fault feature of rolling bearing in early failure period
is difficult to extract, an incipient fault diagnosis method for rolling bearing based on variational
mode decomposition (VMD) and wavelet packet transform (WPT) was proposed. The variational
mode decomposition was firstly used to decompose the multi-component signal into a number of
intrinsic mode functions (IMF), and then the IMFS of the maximum kurtosis were selected to form
the new information based on Kurtosis Criterion. Finally, the new signal was decomposed and
reconstructed by adopting wavelet packet transform, after that, the energy of every frequency band
was calculated, and the frequency band with the maximal signal was chosen and demodulated into
energy spectrum with Teager energy operator demodulation method. In order to verify the
effectiveness of the proposed method, practical engineering experiments had been carried out and the
effect was compared with the EEMD-WPT method for rolling bearing inner fault signals. The results
show that compared with the other method, the proposed method can not only reduces the effect of
noise but also implement accurate diagnosis.

Key words: variational mode decomposition, wavelet packet transform, fault diagnosis, Teager energy
operator demodulation