基于离线状态监测的复杂装备预知维修决策及优化  

Predictive Maintenance Decision and Optimization of Complicate Equipment Based on Off-line State Monitoring

作　　者：邢厚刚[1] 邓力[2] 肖楚琬[3] 

机构地区：[1]91404部队 [2]海军航空工程学院兵器科学与技术系 [3]海军航空工程学院接改装训练大队

出　　处：《海军航空工程学院学报》2015年第3期269-276,共8页Journal of Naval Aeronautical and Astronautical University

基　　金：海军航空工程学院青年基金资助项目(HYQN201417)

摘　　要：基于免疫粒子群优化算法,对可进行离线状态监测故障呈现渐发特性的复杂装备预知维修决策问题进行了研究。分析了维修决策的必备条件,包括研究对象的前提假设、决策内容、剩余维修时间预测以及维修相关费用;将复杂装备维修决策问题转化为带有约束条件的目标函数求最小解的问题,并构建了数学模型;运用结合粒子群优化算法明确方向性搜索和免疫系统多样性保持能力优势的免疫粒子群算法,对模型进行求解;通过某航空机电设备的实例应用验证,该方法可以有效地解决复杂装备预知维修决策的问题,结果与实际情况基本一致。Aiming to the predictive maintenance decision of complicated equipment, of which state could be monitored at discrete time and the fault showed gradual failure, a new method based on artificial immune particle swarm optimization al- gorithm was presented. First, the basic conditions of maintenance decision was analyzed, which included presupposition of subject investigated, decision content, remaining maintenance time, and maintenance cost. Second, the maintenance deci- sion problem was transformed to an object function with the characteristics constraints which sought a minimum, and then its mathematical model was built. Third, the mode was solved by artificial immune particle swarm optimization algorithm, which had the advantages of ability to maintain diversity of the immune system and specific directional search of particle swarm algorithm. Finally, the application in a certain aerial electromechanical device validated the method was available to solve the predictive maintenance decision of complicated equipment, and the results were consistent with practical situa- tion.

关 键 词：状态监测 健康指数 剩余维修时间 人工免疫系统 粒子群算法 

分 类 号：TJ07[兵器科学与技术—兵器发射理论与技术] TP18[自动化与计算机技术—控制理论与控制工程]