Remaining Useful Life Prediction for Complex Systems by Fusing Multi-Level Information

Remaining useful life (RUL) prediction holds great importance reliability and is regarded as a crucial step in prognostics and health management (PHM). Traditional methods often demand extensive testing durations and large sample sizes, or they might pose considerable risks in estimation. Consequently, the utilization of multi-level information is considered to overcome these limitations and a novel method for estimating the residual life of complex systems is presented. Firstly, the fusion of multisource information is employed to determine both the prior and posterior distributions of components, subsequently the multiorder moments of reliability under the posterior distributions are derived. On this basis, the moments of the whole system are calculated by these moments of components. Finally, the compound Simpson formula is applied to obtain the estimations of the system's residual life. An illustrative example is provided and by comparing with the Monte Carlo simulation results, the validity of the proposed method is proved.