飞行器发动机结构系统早期故障分类识别方法研究  

Study of an Early Fault Classification and Recognition Method for Aircrafts′ Engine Structural Systems

作　　者：何红[1] 王仲生[1] 赵佐[1] 

机构地区：[1]西北工业大学,西安710072

出　　处：《机械科学与技术》2006年第7期793-796,共4页Mechanical Science and Technology for Aerospace Engineering

基　　金：航空科学基金项目(03I53068);国家自然科学基金项目(60472116)资助

摘　　要：在对飞行器发动机结构系统早期故障进行分析的基础上,提出应用小波变换、分形与模糊神经网络相结合的早期故障识别分类方法。该方法首先利用小波变换和分形技术对早期故障的微弱信号进行处理并提取早期故障信息,然后利用模糊神经网络(FNN)对早期故障特征信息进行聚类分析识别。实验结果表明,该方法是一种识别早期故障的有效方法。The early fault of aircrafts＇ engine structural systems was analyzed. The early fault classification and recognition method that combines wavelet transformation and fractual theory with fuzzy neural network（FNN） was put forward. Wavelet transformation was applied to the processing. Fractal theory was used to extract early fault information. Then FNN was utilized to recognize and classify the early fault feature information. Experimental results indicate that the classification and recognition method is an effective method for recognizing early faults.

关 键 词：结构系统 小波变换 分形 模糊神经网络 

分 类 号：V23[航空宇航科学与技术—航空宇航推进理论与工程]