In order to improve the lifetime and reliability of equipment operation stage, it is necessary to carry out maintenance for components. The typical selective maintenance model does not consider the maintenance quality uncertainty and the failure effects, which may make the decision-maker overestimates the reliability of completing the next task after system maintenance, resulting in incorrect maintenance decision and prone to serious consequences of failure. In this paper, the potential discrepancy between the target maintenance quality and the actual maintenance quality of the decision maker is considered. At the same time, a multi-state FMECA is proposed to measure the failure effects under different states. Finally, the selective maintenance model of the multi-state series systems is established and solved by the genetic algorithm. The results show that considering the effect of component failure and the uncertainty of maintenance quality has an influence on maintenance decision. The maintenance decision made in this way is more consistent with the engineering practice.

为了提高设备操作台的使用寿命和可靠性，有必要对部件进行维护。典型的选择性维护模型没有考虑维护质量的不确定性和故障影响，这可能会使决策者高估系统维护后完成下一个任务的可靠性，从而导致错误的维护决策并容易导致严重的故障后果。本文考虑了目标维护质量与决策者实际维护质量之间的潜在差异。同时，提出了一种多状态FMECA来测量不同状态下的失效影响。最后，通过遗传算法建立并求解了多状态串联系统的选择性维护模型。结果表明，考虑零件故障的影响和维修质量的不确定性对维修决策有影响。通过这种方式做出的维护决策与工程实践更加一致。