针对导弹退化监测信息不充分条件下的维修决策问题, 结合导弹实际, 对以维纳模型描述退化过程的导弹部件进行研究。利用贝叶斯方法, 综合历史寿命数据和在线退化监测信息, 建立可更新的剩余寿命预测模型。采用粒子群算法求解极大似然函数, 得到寿命分布参数的先验估计, 利用递推法得到参数的后验估计, 从而更新了剩余寿命分布。在此基础上分别以可用性目标和经济性目标建立维修决策模型;再综合考虑两个目标, 采用极值交互式方法建立多目标规划模型;最后将模型应用于导弹惯导系统的核心部件——激光陀螺仪, 实现了陀螺仪的多目标状态维修优化决策。

Aiming at the maintenance decision based on insufficient monitored degeneration informationand considering the actual situation of missileswe made study on missile components by using Wiener model to describe their degeneration process.Bayes method was used to establish a renewable remaining life prediction model on the basis of historical life data and online monitored degeneration information.Particle swarm arithmetic was used to solve the maximum likelihood functionand a priori estimate of lifetime distribution parameter was obtained.The posteriori estimate was obtained by using recursive method.Thereforethe remaining life distribution was updated.Thenthe maintenance decision models were built up based on usability and economic objectives respectively.And a multiobjective model was established by using interaction method with integration of the two objectives.The model was used in a laser gyroscopethe critical component of the missile inertial navigation systemand realized the optimized multiobjective maintenance decision.