基于逻辑图的国产民机液压系统故障诊断与定位

Fault diagnosis and location of hydraulic system of domestic civil aircraft based on logic data

为研究国产民机液压系统典型故障诊断与故障部件定位技术,依据故障形成条件构建典型故障逻辑图,收集典型故障运行数据,采用贝叶斯网络实现故障诊断与故障部件定位。针对某型国产民机机组操作手册中的故障形成条件,参考逻辑图构建方法,以液压系统典型故障为例,建立国产民机故障逻辑图,直观反映故障形成逻辑关系;依据逻辑图,考虑故障形成条件,建立与逻辑图对应的贝叶斯网络,并将故障形成的逻辑关系以条件概率分布值表示;获取国产民机快速存取记录器(quick access recorder,QAR)数据,依据逻辑图输入信息获取QAR参数信息;依据建立的贝叶斯网络与获取QAR数据,应用正向推理实现液压系统典型故障诊断,在完整信息与部分信息条件下,应用反向推理实现液压系统故障部件定位。研究表明所提方法可准确实现故障诊断,并可在完整信息下精准进行故障部件定位,在部分信息下对潜在可能故障部件给出发生的概率,有效辅助故障部件定位工作。研究工作对完善国产民机的机载健康管理系统与地面健康管理系统的故障诊断功能具有一定的参考意义。

To study the typical fault diagnosis and fault location technology of the hydraulic system of the domestic civil aircraft,the logic data of the typical fault is constructed according to the formation conditions of the fault.The operation data of the typical fault is collected,and the Bayesian network is used to realize the fault diagnosis and fault components position.First,according to the fault formation conditions in the unit operation manual of a certain type of domestic civil aircraft,referring to the construction method of the logic data,taking the typical fault of the hydraulic system as an example,the fault logic data of the domestic civil aircraft is established to intuitively reflect the logical relationship of the fault formation;secondly,based on the constructed logic data,considering the forma⁃tion conditions of the fault,a Bayesian network corresponding to the logic data is established,and the logical rela⁃tionship formed by the fault is represented by the value of the conditional probability distribution;obtain quick access recorder(QAR)data and its parameter information according to the input information of the logic data;finally,according to the established Bayesian network and the obtained QAR data,apply forward reasoning to real⁃ize the diagnosis of typical faults of the hydraulic system.Under the condition of partial information,reverse reason⁃ing is applied to locate the faulty components of hydraulic system.The research shows that the proposed method can accurately diagnose faults,and can accurately locate faulty components in complete information,and give the prob⁃ability of occurrence of potentially faulty components under partial information,which can effectively assist in the location of faulty components.The research work has certain reference significance for improving the fault diagnosis function of the airborne health management system and the ground health management system of domestic civil air⁃craft.