Bearing residual life prediction using hybrid model based on CNN and BiGRU

With the enhancement of the level of industrial equipment intelligence, the health condition monitoring and residual life prediction (RUL) of bearings are particularly important for improving equipment reliability and reducing maintenance costs. In this paper, a comprehensive model utilizing failure prediction time (FPT), variational modal decomposition (VMD), and the hybrid model based on convolutional neural network (CNN) and bi-directional gated recurrent unit (BiGRU) is proposed, firstly, the bearing life is classified into healthy and declining periods by FPT, and then VMD extracts the intrinsic mode functions, and next, hybrid model is constructed to learn the mapping relationship between multiple modes and the residual life, which is used to realize the accurate prediction of the residual life of bearings. In this paper, the XJTU-SY bearing dataset is used for experimental validation. The results show that the proposed hybrid model significantly improves the accuracy of the residual life prediction of bearings compared with the traditional methods.