航空学报 > 2007, Vol. 28 Issue (1): 68-71

基于小波过程神经网络的飞机发动机状态监视

钟诗胜,李洋   

  1. 哈尔滨工业大学 机电工程学院
  • 收稿日期:2005-08-10 修回日期:2006-03-11 出版日期:2007-01-10 发布日期:2007-01-10
  • 通讯作者: 钟诗胜

Condition Monitoring of Aeroengine Based on Wavelet Process Neural Networks

ZHONG Shi-sheng,LI Yang   

  1. Department of Mechanical and Electrical Engineering, Harbin Institute of Technology
  • Received:2005-08-10 Revised:2006-03-11 Online:2007-01-10 Published:2007-01-10
  • Contact: ZHONG Shi-sheng

摘要:

针对飞机发动机状态监视问题,提出了小波过程神经网络模型。其隐层和输出层为过程神经元,隐层激活函采用小波函数。该模型结合了过程神经网络可以处理连续输入信号的特点及小波变换良好的时频局域化性质,有更强的学习能力和更高的预测精度。文中给出了相应的学习算法,并以飞机发动机状态监视中排气温度裕度的预测为例,分别利用3层前向过程神经网络和小波过程神经网络进行预测。结果表明,小波过程神经网络结构更简单,收敛速度更快,优于过程神经网络,因而为飞机发动机状态监视提供了一种有效的方法。

关键词: 过程神经元, 小波过程神经网络, 学习算法, 飞机发动机状态监视

Abstract:

Aimed at the problem of aeroengine condition monitoring, wavelet process neural networks (WPNN) model is proposed. Its hidden layer and output layer are composed of process neuron and hidden layer function consists of wavelet function. The network not only has the capability to deal with continuous input signals, but also has the time-frequency local property of wavelet analysis. The learning ability of WPNN is better and the predictive precision is high.

Key words: process , neuron,  , wavelet , process , neural , networks,  , learning , algorithm,  , condition , monitoring , of , aeroengine

中图分类号: