A data-driven remaining useful life prediction methodology: Optimization based on digital twin

Remaining useful life (RUL) estimation is one of the main objectives in prognosis. Digital twins (DTs) technology is a hot research topic in recent years, and its combination with prediction models can well solve the problem of traditional methods. This paper proposed a novel RUL prediction methodology based on DTs. Kalman filter and Bayesian networks are used to construct degradation modeling of physical entity and digital twin. Degradation model is built on the basis of Wiener process. The parameters of Wiener process are obtained by the maximum likelihood estimation method. The optimal value of RUL is obtained by supplementing the data of the digital twin to the physical entity. The degradation and RUL analysis of the subsea valves group is used to demonstrate the proposed methodology. The optimization of RUL is further analyzed. The results show that the methodology has good feasibility and applicability.