二代小波降噪与盲分离结合应用于航空发动机振动信号分析  

Signal Processing of Aeroengine Vibration with Second Generation Wavelet Denoising and Blind Source Separation

作　　者：马建仓[1] 刘小龙[1] 陈静[1] 

机构地区：[1]西北工业大学电子信息学院,西安710129

出　　处：《机械科学与技术》2010年第1期7-11,共5页Mechanical Science and Technology for Aerospace Engineering

基　　金：国家自然科学基金项目(60672184)资助

摘　　要：传感器采集的飞机发动机振动信号必然是各个部件振源信号以及周围环境强烈干扰的混合信号,传统的振动信号处理方法抗扰去噪效果并不理想,很难得到振源信号。笔者介绍了利用第二代小波阈值降噪、并结合盲源分离得到振源信号的方法。在对某型航空发动机空中停车振动信号进行二代小波分解、利用阈值处理并重构取得了理想的降噪效果的基础上,进一步应用盲源分离技术(FastICA)分离得到转子振动源信号,这样可以有效的提取故障特征,提高故障诊断的准确性。The aeroengine rotor vibration signals inevitably contain interference and other components. With the traditional vibration signal processing methods, the effect of removing other components and interference is not satis- factory. The combination of second generation wavelet de-noising with blind source separation is applied to obtai- ning the rotor vibration signals. For vibration signals of shutdown failure of an in-flight aeroengine, the wavelet de- composition and the threshold dealing are adopted, the ideal denoising effects are achieved after processing of wave- let reconstruction and blind source separation. Finally, to obtain pure rotor vibration signal, the effective fault fea- ture can be extracted, and fault diagnosis is more accurate.

关 键 词：航空发动机 第二代小波 盲源分离 故障诊断 阈值函数 

分 类 号：TN911.7[电子电信—通信与信息系统]