Abstract: A jointly policy of condition-based maintenance (CBM) and spare ordering is proposed for a single-unit system/critical component subject to a two-stage delay-time process. A non-periodic inspection policy with various intervals is presented, in which the system is detected with the relatively longer interval T₁ during the initial deterioration process, but with a shorted interval T₂ after the defective state is firstly found by an inspection. While, the spare is ordered and imperfect maintenance (IPM) is carried out. In the subsequent deterioration process, IPM is required and performed once the defective state is revealed until the number of IPM exceeds the threshold N_max; then, a preventive replacement is needed. If a failure occurs within the inspection interval, a corrective replacement arises. According to the system's state and the spare's availability, all possible renewal events and the corresponding joint decisions are discussed and then the objective function to minimize the expected cost per unit time is established using the renewal award theorem for optimizing T₁,T₂, N_max. The results fromthe numerical example indicate that the proposed policy decreases the expected cost per unit time compared to the comparative models.

Key words: condition-based maintenance, non-periodic inspection, imperfect maintenance, spare ordering, delay-time

摘要： 针对单部件系统/关键部件提出视情维修与备件订购联合策略,其中系统退化服从两阶段延迟时间过程且采用非周期检测策略,退化初期以检测间隔T₁检查系统状态,而在第一次识别缺陷状态时,缩短检测周期为T₂、订购备件且进行不完美维修;若系统在随后的退化中被识别处于缺陷状态,执行不完美维修直至超过阈值次数N_max并采取预防性更换,但若在检测周期内发生故障则进行更换。根据系统状态和备件状态分析各种可能更新事件及相应的联合决策,利用更新报酬理论构建最小化单位时间内期望成本的目标函数,优化T₁,T₂, N_max。与对比模型策略相比,算例结果表明所提出的联合策略能有效降低单位时间内的期望成本。

关键词: 状态维修, 非周期检测, 不完美维修, 备件订购, 延迟时间