Improved VMD for feature visualization to identify wheel set bearing fault of high speed locomotive

As a critical component of high-speed locomotive, wheel set bearing fault identification and prognosis based on vibration analysis has attracted an increasing attention in recent years. However, heavy background noise and adverse working conditions make it difficult to excavate the hidden weak fault feature. Variational Mode Decomposition (VMD), which can extract the Intrinsic Mode Functions from the non-stationary signal, brings a feasible method. However, the inaccurate pre-set mode number may lead to information loss or over decomposition problem. In this paper, an improved VMD method via correlation coefficient is proposed to automatically extract modes. To overcome the information loss problem, the appropriate mode number is determined by the criterion of approximate complete reconstruction. Then the similar modes are combined to solve the over decomposition problem according to the similarity of their envelops. Finally, an application to wheel set bearing fault of high speed locomotive verify the validity of the proposed method.