A rolling bearing state evaluation method based on deep learning combined with Wiener process

Abstract

As a key component of rotating parts, rolling bearings largely determine the operation safety of equipment. However, in practical applications, because the degradation trajectory of rolling bearings cannot be truly characterized, the existing model cannot accurately describe the degradation trajectory of rolling bearings, resulting in the running state of rolling bearings cannot be directly evaluated. Therefore, a method of rolling bearing state assessment based on deep learning combined with Wiener process is proposed in this paper. Firstly, a deep network model is constructed by deep learning to mine the degradation information of rolling bearings. Secondly, the mined degradation information is fused, and then the degradation indicator used to characterize the degraded trajectory of the rolling bearing is constructed. Then, based on Wiener process, the degradation model of rolling bearing is established to describe the degradation mode of rolling bearing. Finally, the constructed degradation indicator is input into the established degradation model to predict its RUL, and then the running state of the rolling bearing is evaluated.