摘要
针对传统的航空发动机故障诊断方法正确率较低,并且对异常数据不敏感的问题,将智能诊断算法引入航空发动机气路故障诊断领域。以涡轴发动机为例,分析了常见气路部件故障类型的成因和表现,并在Tensorflow上建立基于深度信念网络的故障诊断模型。与传统的故障诊断方法相比,具有更高的故障诊断正确率。
Because low accuracy and insensitivity to abnormal data exist in the traditional method of fault diagnosis, the intelligent diagnosis algorithm is introduced into the fault diagnosis field of aero-engine gas circuit. Taking the turboshaft for example, the causes and manifestations of common types of gas circuit component faults are analyzed, and a fault diagnosis model based on deep belief network is established on Tensorflow. Compared with the traditional fault diagnosis method, it has a higher fault diagnosis accuracy rate.
作者
林嘉琦
徐建国
刘星怡
LIN Jiaqi;XU Jianguo;LIU Xingyi(College of Energy and Power, Nanjing University of Aeronautics and Astronautics, Nanjing 210001, China)
出处
《机械制造与自动化》
2019年第5期179-182,共4页
Machine Building & Automation
关键词
航空发动机
气路部件
深度信念网络
故障诊断
aero-engine
gas circuit component
deep belief network
fault diagnosis
