飞机发动机故障诊断专家系统研究

发布时间：2018-03-10 21:12

本文选题：故障树分析法　切入点：专家系统　出处：《青岛科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：飞机涡轮风扇式发动机无论是从机械结构还是从系统角度来分析都是特别复杂的,发动机故障问题直接威胁到飞机的飞行安全,还可能会引发其它故障的产生,这决定了涡轮发动机的故障诊断与维修工作是一项艰巨而繁重的任务。建立一个基于故障树分析法,模拟人类专家诊断的涡轮发动机故障诊断专家系统,对发动机故障进行专家级别的快速准确故障诊断,并提出合理化的改进意见,保障发动机和飞机的安全来说意义重大。本论文飞机涡轮发动机故障诊断专家系统,以海军某飞行部队涡轮风扇式发动机异常启动故障作为研究对象。对该型号发动机的故障维修做了长期的跟踪,收集了大量的故障数据和一手的专家诊断资料。利用故障树分析法作为涡轮发动机异常启动故障的故障分析方法,以建立发动机异常启动故障树的形式对系统进行了故障机理和故障原因的分析。同时结合产生式知识表示法,将所收集到的故障数据和专家资料进行知识表示,转换为方便计算机存储的数据库关系结构表格信息。由故障树自动生成基于知识规则的专家系统诊断规则库,同时又对知识库中的诊断知识进行了一致性的校验和完备性的校验,采用正向与逆向混合推理的方法构建系统推理机,实现了对专家知识库中知识的合理利用。该故障诊断专家系统使用Windows操作系统作为系统开发平台,Myeclipse作为系统的编程工具。后台推理机以Java语言作为编程语言开发实现,后台的数据库管理系统使用的是MySQL Workbench 6.3CE,应用人工智能的故障诊断方法和面向对象编程的方法进行设计开发,并开发设计了友好的可视化人机交互界面。最后以海军某部涡轮风扇式发动机启动异常这一典型故障诊断为例,验证了该专家系统的性能和优势,值得在全军开展广泛的应用推广。
[Abstract]:It is very complicated to analyze the turbofan engine from the angle of mechanical structure or system. The problem of engine failure directly threatens the flight safety of aircraft, and may cause other faults. This determines that the fault diagnosis and maintenance of turbine engine is an arduous and arduous task. A fault diagnosis expert system for turbine engine based on fault tree analysis is established to simulate human expert diagnosis. It is of great significance to make expert level rapid and accurate fault diagnosis for engine faults, and to put forward reasonable suggestions to ensure the safety of engine and aircraft. Taking the abnormal start fault of turbofan engine of a naval flying force as the research object, the fault maintenance of this type of engine has been tracked for a long time. A large number of fault data and one-hand expert diagnosis data are collected. The fault tree analysis method is used as the fault analysis method for turbine engine abnormal start fault. The fault mechanism and fault cause of the system are analyzed in the form of establishing engine abnormal start fault tree. At the same time, combined with the production knowledge representation, the collected fault data and expert data are expressed. The fault tree is used to automatically generate the diagnostic rule base of expert system based on knowledge rules. At the same time, the diagnosis knowledge in the knowledge base is checked by consistency check and completeness check, and the system inference engine is constructed by the method of forward and reverse hybrid reasoning. The fault diagnosis expert system uses Windows operating system as the system development platform and Myeclipse as the programming tool of the system. The background inference machine is developed with Java as the programming language. The background database management system uses MySQL Workbench 6.3CE. the fault diagnosis method of artificial intelligence and the method of object-oriented programming are used to design and develop. A friendly visual man-machine interface is developed and designed. Finally, the performance and advantages of the expert system are verified by taking the typical fault diagnosis of a turbofan engine in the Navy as an example. It is worthy of extensive application and popularization in the whole army.
【学位授予单位】：青岛科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：V263.6;TP182

【参考文献】