A Review: Prediction Method for the Remaining Useful Life of the Mechanical System

Remaining useful life (RUL) refers to the remaining service life of a mechanical system after it runs for a period. Predicting the remaining service life of the system accurately can greatly reduce the loss caused by system downtime and improve the reliability of system operation. Various RUL prediction methods are discussed in this paper. According to the analysis and discussion of various articles, the RUL prediction methods commonly used in the literature can be divided into four categories: physics model-based approaches, statistical model-based approaches, AI approaches, and hybrid approaches. Then the definition and common methods of these classifications are systematically introduced. Finally, the advantages and disadvantages of each method are analyzed and summarized.