Prognostic-information-driven Policy for Joint Spare Parts Ordering and Postponed Replacement Optimization

This paper proposes a joint replacement and spare ordering policy, which utilizes prognostic information to update the adaptive decision of when to order spare parts and how long the repair is postponed after triggering the maintenance decision. A nonlinear Wiener process with randomness is established to characterize the degradation trend, along with updating online parameters under Bayesian framework at each inspection point. Unlike traditional discrete models, this strategy optimizes both ordering and maintenance, relying on a comprehensive cost rate indicator. Furthermore, based on residual useful life (RUL), this model adopts predictive maintenance to avoid resource waste of scheduled maintenance. Additionally, due to the timely maintenance accompanying monitoring reduces the availability of logistics resources, this model adopts a delay interval determined by RUL's expectation and a delay coefficient, and then which is optimized through order time and delay coefficient. Ultimately, the applicability of the proposed policy is verified by the actual case study of high-speed train bearings.