装备剩余寿命预测与维修保障策略研究

Study on Equipment Remaining Life Prediction and Maintenance Support Strategy

本文针对装备管理中的装备剩余寿命预测和维修保障问题给出了一种普适简洁、可操作性强的模型,可根据所给模型预测装备的剩余寿命及其最优的维修保障策略。首先根据装备的性能指标退化情况,对其性能指标进行分析,得出每次检测的性能指标分布函数。之后根据此分布函数建立了聚类模型。利用聚类模型进行剩余寿命的区间估计,用神经网络模型进行点估计对区间估计结果进行检验。最终根据实际情况中的更换与库存相关数据和可靠度要求,确定约束条件,利用遗传算法求解出一种现实可行的装备寿命预测与维修保障策略。

Concerning the issue of equipment remaining life prediction and maintenance support in equipment management,this paper presents a universal,concise and operable model,based upon which the remaining life of equipment and its optimal maintenance strategy can be predicted.Firstly,the performance indexes of equipment are analyzed as per their degradation,in order to obtain the performance indexes distribution function of each testing.Secondly,a cluster model is established based on this distribution function to estimate the range of the remaining life,and a neural network model is applied to make point estimation which tests the range estimation result.Finally,the constraints are determined by the actual replacement,related inventory data and reliability requirement,and a practical and feasible equipment life prediction and maintenance support strategy is acquired using the genetic algorithm.