A machine learning-based approach for product maintenance prediction with reliability information conversion

Abstract Predictive maintenance (PdM) cannot only avoid economic losses caused by improper but also maximize the operation reliability of product. It has become core management. As an important issue in PdM, time between failures (TBF) prediction can realize early detection and products. The information is main basis for TBF prediction. Therefore, purpose this paper to establish intelligent model complex mechanical conversion method used solve problems collection difficulty, high cost small data samples process based on product fully mined enriched obtain more reliable accurate results. Firstly, Fisher algorithm employed convert expand sample, compatibility test discussed. Secondly, BP neural network final TBF, PSO optimize initial weight threshold falling into local extreme value improve convergence speed. Thirdly, mean-absolute-percentage-error Coefficient determination are selected evaluate performance proposed method. Finally, a case study remanufactured CNC milling machine tool (XK6032-01) studied paper, results show that feasibility superiority