论文部分内容阅读
针对多部件系统状态维修决策建模存在的不足以及现有模型难以推广到实际应用的问题,提出考虑经济、结构和随机3种依赖性的多部件系统仿真建模和优化方法。首先,利用Gamma退化过程对部件退化进行描述,并给出模型参数估计方法。然后,结合多部件系统的维修作业流程、组成关系和故障历史等信息,分别提出经济依赖性强度矩阵、结构依赖性可达矩阵和随机依赖性概率矩阵对3种依赖性进行建模。最后,同时考虑部件级和系统级决策,构建多部件系统期望周期费用仿真模型,并针对该仿真模型特点,提出了用单纯形算法(NMA)改进的遗传算法(GA)优化求解的过程。某电传系统实例仿真结果表明:多部件系统的3类依赖关系对维修决策的影响不可忽略,考虑依赖性和成组维修的存在能够节省平均维修费用,且使得维修决策优化更加符合实际,验证了所建模型和优化方法的有效性。
In order to solve the existing problems of multi-component system state-based maintenance decision-making and the problems that existing models can not be generalized to practical application, a multi-component system simulation modeling and optimization method considering economy, structure and stochastic is proposed. Firstly, the component degeneration is described by using Gamma degeneration process, and the method of model parameter estimation is given. Then, combined with the information of maintenance work flow, compositional relationship and fault history of multi-component system, three kinds of dependency are modeled as economic dependence intensity matrix, structural dependence reachability matrix and random reliance probability matrix respectively. Finally, considering both component-level and system-level decision-making, a multi-component system expected periodic cost simulation model is constructed. According to the characteristics of the simulation model, a genetic algorithm (GA) improved by simplex algorithm (NMA) is proposed. The simulation results of a telex system show that the influence of three types of dependencies on the maintenance decision can not be neglected. Considering the dependency and the existence of group maintenance, it can save the average maintenance cost and make the maintenance decision optimization more realistic. The validity of the proposed model and the optimization method.