Fault Diagnosis Method of Bearing based on LCD Cross Approximate Entropy and Relevance Vector Machine

Abstract

Aiming at the fault diagnosis problem of rolling bearing,a fault diagnosis method of rolling bearing based on local characteristic-scale decomposition（LCD） cross approximate entropy（CAE） and relevance vector machine（RVM） is proposed. Firstly,the bearing vibration signals is decomposed into several intrinsic scale components（ISC） which with different frequency components. Secondly,some ISCs that contain main fault information are shifted out by the energy analysis criterion and CAE values are calculated as fault feature vectors that could represent the operating conditions of bearings. Finally,the fault feature are put into RVM to identify different faults. The effective of the proposed method is verified by the different fault type and different fault degree of rolling bearing experiment.

Keywords

• Local characteristic-scale decomposition;Cross approximate entropy;Relevance vector machine;Fault diagnosis;Rolling bearing