论文部分内容阅读
针对现代民机在排故过程中面对不确定性、多源异类信息时难以进行快速诊断的问题,提出了一种基于贝叶斯Leaky Noisy Or网络的诊断决策模型。采用故障假设-观测-维修操作节点的网络结构,结合专家经验建立贝叶斯网络拓扑结构,将Leaky Noisy Or节点引入网络模型中,同时结合向前多步决策算法,构建多步决策模型。仿真实验表明,与传统方法相比,该模型能够有效解决排故中的不确定性问题,并可最大限度地融合多源异类信息,提高了诊断排故速度。
Aiming at the problem that modern civil aircraft is difficult to diagnose rapidly in face of uncertainty and multi-source heterogeneous information during the troubleshooting, a diagnostic decision model based on Bayesian Leaky Noisy Or network is proposed. Adopting the network structure of the node of fault assumption-observation-maintenance operation, Bayesian network topology is constructed based on expert experience. The Leaky Noisy Or node is introduced into the network model and a multi-step decision model is constructed based on forward multi-step decision algorithm. Simulation results show that compared with the traditional method, this model can effectively solve the uncertainty in troubleshooting and maximize the fusion of multi-source and heterogeneous information, which improves the speed of diagnosis and troubleshooting.