基于BN的某型航空发动机燃油系统故障诊断

发布时间：2018-04-03 11:55

本文选题：航空发动机　切入点：燃油系统　出处：《中国民航大学》2015年硕士论文

【摘要】：随着现代民航运输业的迅速发展,民航飞机的故障诊断技术的准确性和全面性也越来越受到人们的关注,航空发动机是飞机动力的直接提供单元,对飞机的安全运行起到决定性的作用。同时,燃油系统则是航空发动机能源的供给单元,是飞机安全运行的保障,所以航空发动机燃油系统的诊断与维修就显得至关重要。其中对于航空发动机燃油系统故障诊断的关键问题在于如何准确、快速、全面地判断故障发生的部件。本文利用贝叶斯网络(BN,Bayesian Network)诊断方法对CFM56-5B燃油系统进行深入分析,并引入诊断代价函数来分析诊断成本,对该型发动机进行诊断决策。文章首先对CFM56-5B发动机燃油系统进行具体的结构介绍及原理分析,并通过历史故障的信息和实践经验给出燃油系统的部分常见故障;其次介绍了概率论知识和贝叶斯理论,通过建立BN网络拓扑结构的形式来构建BN诊断模型,并通过发动机燃油系统故障产生机理和故障征兆,来建立发动机燃油系统故障的BN网络拓扑结构,分析其定性信息和定量信息;然后将诊断代价函数引入诊断决策,在经计算获取贝叶斯故障概率数据的前提下考虑诊断的安全性指标和经济性指标,对系统诊断步骤进行合理排序,使概率相对高、诊断代价相对低的诊断步骤得以优先进行。本文最后通过对CFM56-5B型发动机燃油系统的特定故障进行分析和诊断决策应用;在进行具体诊断代价函数的建立时利用修正熵权法对各项指标权重进行赋值,并利用实际的时间、费用值来确定最终的诊断代价,最终确定了基于贝叶斯网络诊断的诊断决策,并与TSM(故障隔离手册)诊断程序进行比较,应用TOPSIS综合评价方法,以诊断代价为标准,验证了BN的优越性。
[Abstract]:With the rapid development of modern civil aviation transportation industry, people pay more and more attention to the accuracy and comprehensiveness of the fault diagnosis technology of civil aviation aircraft.It plays a decisive role in the safe operation of the aircraft.At the same time, fuel system is the supply unit of aero-engine energy and the guarantee of aircraft safe operation, so the diagnosis and maintenance of aero-engine fuel system is very important.The key problem in the fault diagnosis of aero-engine fuel system is how to judge the fault components accurately, quickly and comprehensively.In this paper, the Bayesian network diagnosis method is used to analyze the CFM56-5B fuel system in depth, and the diagnostic cost function is introduced to analyze the diagnosis cost and make the diagnosis decision for the engine.In this paper, the structure and principle of CFM56-5B engine fuel system are introduced, and some common faults of fuel system are given through historical fault information and practical experience, and then the knowledge of probability theory and Bayesian theory are introduced.The BN network topology structure of engine fuel system fault is established by establishing BN network topology structure and the mechanism and symptom of engine fuel system fault.The qualitative and quantitative information is analyzed, and then the diagnostic cost function is introduced into the diagnosis decision, and the safety index and economic index of diagnosis are considered on the premise of obtaining Bayesian fault probability data by calculation.The reasonable ranking of the system diagnostic steps makes the diagnostic steps with relatively high probability and relatively low diagnostic cost take precedence.In the end, through the analysis of the specific faults of the fuel system of CFM56-5B engine and the application of the diagnosis decision, the paper uses the modified entropy weight method to assign the weight of each index, and makes use of the actual time when establishing the specific diagnostic cost function.The cost value is used to determine the final diagnosis cost, and finally the diagnosis decision based on Bayesian network diagnosis is determined, and compared with the diagnosis program of TSM (Fault isolation Manual), the comprehensive evaluation method of TOPSIS is applied, and the diagnostic cost is taken as the standard.The superiority of BN is verified.
【学位授予单位】：中国民航大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：V263.6

【相似文献】