It is difficult to identify the early fault of the rolling bearings. A diagnosis method based on variational mode decomposition (VMD) and Mahalanobis distance support vector machine (SVM) is proposed. Firstly

the original vibration signal is de-noised by wavelet threshold method to obtain effective vibration signal. Secondly

according to the center frequency of each mode after VMD decomposition

the final number of decomposed layers is determined. At the same time

the energy characteristics are extracted from the decomposed variational modal components. Finally

in order to measure distance between samples more accurately

Mahalanobis distance is introduced into the calculation of the Gaussian kernel function of the SVM

and a Gaussian function kernel based on Mahalanobis distance is established

which is used to support the vector machine classifier. Improved SVM is employed to identify the running state of the bearing

the experimental results show that the proposed method has high accuracy in identifying the normal state

the inner ring

the outer ring and the ball body fault of bearings.