Remaining Life Prediction for High-speed Rail Bearing Considering Hybrid Data-model-driven Approach

Bearings are one of the most important rotating machinery components in high-speed railways, and prediction of their remaining useful life (RUL) is an important basis for ensuring the safe and reliable operation of equipment. The data-driven methods of remaining useful life prediction don't rely on physical or statistical models, but have the characteristics of large data demand and weak mechanism correlation; model-driven methods often have problems such as poor robustness and insensitive to changes in operating conditions. Based on the extraction of health features, this paper divides the bearing operation stages, uses the Wiener model and the BP neural network to predict the remaining useful life separately by focusing on the rapid degradation period of the bearing, and utilizes benchmark Fusion, which is put forward in this paper, to realize prediction. The prediction results show that the effect after the fusion has been greatly improved compared with that before the fusion, which proves the advanced nature and feasibility of the idea of hybrid data-model-driven. The feasibility of the benchmark fusion method proposed in this paper is demonstrated by comparing several common weight distribution methods.