A predictive maintenance scheduling framework utilizing residual life prediction information

摘要

This article proposes a component-level predictive maintenance scheduling framework in which system residual life prediction information is utilized. In the proposed framework, predictive maintenance scheduling is only triggered when a system is close to failure. A lifetime margin is used to infer that a system is close to failure but still safely operational when predictive maintenance scheduling is triggered. Some technical elements are detailed to facilitate the implementation of the proposed framework. A simulation study based on an example predictive maintenance scheduling framework (framework B) is provided to illustrate the implementation of the proposed framework, and compare framework B with the corresponding counterpart only consisting of predictive maintenance scheduling models (framework A). Given a relatively low noise level of system degradation processes, compared with the classical prognostics framework only consisting of predictive maintenance scheduling models, the proposed framework is similarly effective in failure prevention and more economic in that several premature preventive maintenance acts are saved.