不完整滑油数据下发动机磨损的预防性维修决策框架

Preventive Maintenance Decision Framework for Engines Under Wear with Incomplete Oil Monitoring Data

针对不完整滑油监控数据下发动机磨损的预防性维修决策问题,提出了一个完整的建模和分析框架。该框架包括不完整磨粒数据的建模与辨识、基于磨粒数据模型仿真抽样的发动机磨损寿命分布逼近和发动机磨损预防性维修决策的矩阵解析分析三方面内容。通过集成最大熵概率密度估计和迭代递归预测误差算法、动态聚类和有限位相型分布逼近的期望最大化算法、约束优化和粒子群算法,成功解决了上述三个方面的问题;其分析过程提供了视情维修的典型决策流程,可用于基于状态监控数据的装备维修决策。

Aiming at the preventive maintenance decision of engines under wear from incomplete oil monitoring data,this paper puts forward a modeling and analyzing framework.The framework consists of three components,i.e.,the model of incomplete wear metal concentration data and its identification,the life distribution approximation of engines under wear based on the simulation samples of the proposed model,and the matrix analysing method for the preventive maintenance decision of engines under wear.In virtue of integerating the maximum entropy probability density function estimation and the iterative recursive prediction error algorithms,the dynamic clustering and the finite phase type distribution approximation with expectation maximum algorithms,constraint optimization and particle swarm optimization algorithm,problems related to the above components are solved successfully.The analysis process in this paper provides a typical course of condition based maintenance decision-making,which can be applied to equipment maintenance decision with its condition monitoring data.