小波分析在发动机早期故障识别中的应用研究  

Exploring Effective Early Identification of Aero-Engine Rotor Faults

作　　者：王仲生[1] 何红[1] 陈钱[1] 

机构地区：[1]西北工业大学航空学院,陕西西安710072

出　　处：《西北工业大学学报》2006年第1期68-71,共4页Journal of Northwestern Polytechnical University

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

摘　　要：在对飞机发动机早期故障进行分析的基础上,提出了利用虚拟仪器和小波分析相结合对发动机早期故障进行识别的原理与方法。文中对如何从检测信号中提取早期故障特征信号和对早期故障特征信号进行分离、放大、识别等进行了分析和研究,并通过实验证明了所提方法的有效性。结果表明,虚拟仪器强大的图形化功能与小波分析良好的多分辨率时频局部化特性,能够从复杂的微弱信号中提取出早期故障特征信号,并能有效地消除噪声,对早期故障进行快速识别。Existing methods, in our opinion, are not effective in early identification of aero-engine rotor faults, owing to the difficulty encountered in suppressing the dominant noise signal. We aim to present the promising results of our exploration of an effective method of such early identification. Our method for identifying early faults of aero-engine is based on analysis of aero-engine faults and makes use of Wavelet Analysis and Virtual Instrument. In the full paper, we explain our method in much detail; here, we give only a briefing. We extract early fault characteristic signal from detected signal and analyze the separation, enlargement and identification of early fault characteristic signal. We demonstrate the effectiveness of our method in the following three respects： （1） we show that not only early fault can be quickly identified but also we can determine the tnoment it happens; （2） noise can be effectively suppressed and singular signal of early fault can be detected by third-order Daubechies wavelets; （3） early fault characteristic signal can be extracted from complex weak signal by exploiting the strong graphical visualization power of Virtual Instrument and the high local magnification power of time-frequency wavelet analysis.

关 键 词：小波分析 飞机发动机 早期故障 特征提取与识别 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]