燃气涡轮发动机故障诊断的人工神经网络法  

AN ARTIFICIAL NEURAL NETWORK APPLIED TO AIRCRAFT GAS TURBINE DIAGNOSIS

作　　者：陈大光[1] 刘福生[1] 

机构地区：[1]北京航空航天大学

出　　处：《航空动力学报》1994年第4期344-348,共5页Journal of Aerospace Power

基　　金：航空科学基金

摘　　要：介绍了人工神经网络专家系统在航空涡轮风扇发动机故障诊断上的应用。通过一个高涵道比涡扇发动机故障诊断的实例分析，验证了三层逆传播网络的可行性。非常令人鼓舞的诊断结果证明：神经网络的模糊联想特点、分布存储和并行处理能力以及对随机误差的抑制作用，使其成为一个有效的和快速的方法。它有很好的应用前景，并可广泛用于航空发动机的故障诊断。An artificial neural network based expert system has been developed and is applied to aircraft turbofan engine fault diagnosis.The case study of a high bypass ratio turbofan engine has verified the feasibility of a three-layer back propagation network.The very encouraging diagnostic results prove that the fuzzy association feature, the distributed memorizing and parallel processing ability,and the random error depressing capability make it an effective and quick method which has a very bright perspective and can be used extensively for aircraft engine diagnosis.

关 键 词：燃气轮机 航空发动机 故障诊断 神经网络 

分 类 号：V235.13[航空宇航科学与技术—航空宇航推进理论与工程]