A New Prognosis Model for Nonlinear Stochastic Deteriorating Devices with Multiple Degrading Modes

Remaining useful life (RUL) prediction is one of the crucial issues in prognostics and health management (PHM). By guiding the arrangement of effective maintenance activities before failure, accurate RUL prediction can reduce the probability of catastrophic failure, improving the reliability and saving operating costs for industrial devices. However, most existing prognosis models in the literature are only capable of a specific degradation model, impeding the generalization in devices with multiple degradation mode. In view of the lack of generalized stochastic modelling method which can simultaneously characterize multiple possible degradation modes for nonlinear stochastic degrading devices, a new model has been proposed in this paper to model the degradation and predict the RUL for devices with multiple degrading modes. Firstly, based on nonlinear Wiener process, a new generalized degradation model is constructed to depict the multi-mode stochastic degradation process of the concerned devices. Secondly, the probability density function (PDF)s of the RULs defined under the concept of first hitting time (FHT) and the maximum likelihood estimation (MLE) of the unknown parameters within the proposed model are derived to achieve RUL prediction. Finally, the proposed method is illustrated and verified by a numerical study. The results show that the established generalized degradation model is accurate and effective in RUL prediction.