基于小波神经网络的直升机主减速器故障诊断系统  

Fault diagnostics system for helicopter main gearbox using wavelet neural network

作　　者：刘立生[1,2] 杨宇航[2] 

机构地区：[1]南京理工大学自动化学院,南京210094 [2]中国人民解放军总参谋部陆航研究所,101121

出　　处：《航空动力学报》2012年第6期1255-1260,共6页Journal of Aerospace Power

基　　金：重点科研项目

摘　　要：应用离散小波变换(DWT)和神经网络相结合构建直升机主减速器速器故障诊断系统:DWT对振动信号进行特征提取,神经网络对故障进行辨识和分类。阐述了DWT、帕塞瓦尔定理和广义回归神经网络(GRNN)基本理论,提出了直升机主减速器的故障诊断系统流程图,最后用某型直升机飞行时主减速器上的振动数据对该系统进行验证。实验使用了BPNN(back-propagation neural network)和GRNN两种神经网络,结果表明:提出的故障诊断系统能对主减速器故障进行较好的辨识和分类,这将为直升机主减速器故障诊断系统的进一步开发提供新的技术参考。A fault diagnostics system for helicopter main gearbox （MGB） was construc- ted using discrete wavelet transform （DWT） and neural network, and the DWT was used to extract feature vectors and neural network was used for fault distinguish and classification. Firstly, the theory of DWT, Parseval theorem and GRNN were introduced, and then, the flowchart of fault diagnostics system for helicopter MGB was proposed, finally, the system was verified by the vibration data which were collected during helicopter flying, and both generalized regression neural network （GRNN） and back-propagation neural network （BPNN） were used and compared on the experimental platform. Experimental results indicate that the proposed diagnostics system can distinguish the MGB fault better, which would provide a new technical reference for the further development of helicopter MGB fault diagnostics system.

关 键 词：故障诊断 离散小波变换(DWT) 广义回归神经网络(GRNN) 帕塞瓦尔定理 主减速器 

分 类 号：V233[航空宇航科学与技术—航空宇航推进理论与工程]