Characterization method of rolling bearing operation state based on feature information fusion

Aiming at the problem of incomplete information covered by the bearing state discrimination using a single physical quantity as the data source, a characterization method based on the feature information fusion of multi-physical quantities including acceleration RMS and stiffness that are sensitive to the bearing operating state is proposed. The accelerated life test of the bearing is carried out through the bearing experiment bench, and the effect of different feature information fusion schemes to characterize the bearing state is compared, and the effectiveness of the method is verified. This paper also analyzes the different characteristics of the state evolution law between the bearing early failure caused by specific factors and the bearing reaching normal fatigue life, which provides a new method and idea for the maintenance strategy of rolling bearing such as condition-based maintenance.