基于半马尔可夫模型的民用航空发动机热端组件系统检查间隔优化

根据采集的民用航空发动机热端组件系统检修信息和专家对系统退化状态的判别，在系统状态退化过程为离散半马尔可夫链过程的假设前提下，分别建立了基于专家估计数据、基于检查数据以及基于融合数据的各宏观退化状态驻留时间估计模型，并应用最大似然函数法和MCMC（Markov chain Monte Carlo）法对模型参数进行估计，得到基于不同数据源的各宏观退化状态下驻留时间估计值和状态转移系数，并以一定使用周期内的检修费用最优为目标建立状态转移概率模型，仿真得到3个典型宏观退化状态下的最优检查间隔分别为1750、350、70循环。该仿真结果与目前的民航运行生产工程实际情况非常接近，可以为民航运输企业的检修决策提供客户化的决策支持并提高经济效益。 

Optimization of inspection interval for civil aero-engine hot section subassembly system based on semi Markov model

According to the acquisition of inspection and maintenance information of civil aero engine hot section subassembly system and typical degradation state judgment by experts, the models of sojourn time estimation of each typical macro degradation state based on expert estimation opinion, inspection information and their fusion data were set up under the assumptions of the system state degradation process obeying a discrete semi Markov chain. The methods of maximum likelihood estimation and MCMC(Monte Carlo Markov Chain) were applied to estimate the model parameters, the sojourn time and the state transition coefficients in each macro degradation state based on these three kinds of data. Meanwhile, the state transition probability models were built in a certain service cycle for optimal inspection and maintenance cost, and then the optimal inspection intervals in three typical macro degradation states, ie 1750, 350, 70cycles were obtained by simulation. The results showed the optimal inspection intervals were relatively close to the reality situations in the civil aviation operation and production, providing the technical supports in the aspect of customer oriented inspection and maintenance decision making and improving economic benefits for civil aviation transport enterprises.