Review of rolling bearings performance degradation trend prediction

Abstract

Rolling bearings are widely used in rotating machinery in modern industry, and ensuring their stability during operation is one of the prerequisites for the overall safety of the equipment. Predicting performance degradation can play a key role in preventing accidents and extending equipment life. With the development of big data and deep learning, more trend prediction methods are emerging in the field of performance degradation prediction of rolling bearings. Therefore, this paper reviews the evaluation indicators and performance degradation prediction models for rolling bearing performance degradation prediction. The advantages and disadvantages of physical degradation indicators, and virtual degradation indicators are analyzed. It is presented to utilize the powerful feature self-extraction ability and nonlinear function characterization ability of deep learning methods to construct bearing evaluation indicators. It also analyzes the research progress of traditional performance degradation prediction models and deep learning prediction models. In this review, future developments in rolling bearing performance degradation prediction are summarized in this paper as deep learning-based, digital twin correlation, high dimensionality, and adaptive, which guide researchers and practitioners to effectively identify suitable performance degradation prediction models.