基于马尔科夫决策的冷贮备串联系统状态维修与备件联合优化

Joint Optimization of Condition-based Maintenance and Spare Parts for Cold Standby Series System Considering Markov Decision

以冷贮备串联系统为研究对象,采用离散过程描述设备的有限退化状态,考虑维修策略与备件库存的耦合效应,以系统长期折扣成本最低为目标,基于马尔科夫决策过程,建立状态维修和备件库存联合优化模型。采用值迭代算法进行求解,得到最优状态维修策略与备件库存策略。研究结果表明,在备件充足时,考虑系统运维状态最佳,对退化部件进行预防维修可降低系统长期折扣成本。简单地选择频繁订货或持有大量库存相较于最优备件策略会产生高额的成本损失,损失最高可达65%,且缺货成本对长期折扣成本影响较不显著。

Taking the cold standby series system as the research object,the discrete process was used to describe the finite degradation state of equipment. Considering the coupling effect of maintenance strategy and spare parts inventory,the joint optimization model of condition-based maintenance(CBM)and spare parts inventory was established based on Markov decision process. The objective was to minimize the long run discount cost of the system. The optimal CBM strategy and spare parts inventory strategy were obtained by using the value iteration algorithm. The results show that when the spare parts are sufficient,the long run discount cost can be reduced by preventive maintenance of degraded components considering the optimal operation and maintenance status of the system. Compared with the optimal spare parts strategy,65% cost increase can be generated by simply choosing frequent orders or holding large quantities of inventory. The impact of shortage cost on long run discount cost is not significant.