航空发动机转子系统早期裂纹碰摩故障特征提取方法

Comparative study on feature extraction method of the early stage crack and rub-impact fault in aero engine rotor system

碰摩故障是航空发动机的典型故障。早期碰摩的故障特征混在噪声中,通常不易被发现。分别应用小波变换、小波包分解和EMD方法对早期裂纹碰摩故障的特征提取方法进行了对比研究。其中,小波变换在时域和频域均具有很好的局部化特征,能够提供目标信号各个频率子段的频率信息;小波包分解提高了时频分辨率,它可以保存信号的能量并根据特征进行精确的重构;EMD方法能够自适应地把非平稳信号分解成一系列零均值的调频调幅的总和。详细论述了三种方法对早期裂纹碰摩故障特征提取的过程和实验结果。实验结果表明:对于早期裂纹碰摩故障,EMD法能够准确地提取出倍频处的故障特征,该方法更适用于早期裂纹碰摩故障的特征提取。

Rub-impact fault is a typical fault of aero engines. It is difficult to detect the features of early rubimpact fault since they are mingled in the noise. In this paper,the comparative study is conducted on the wavelet transform,wavelet packet decomposition and EMD method applied to the feature extraction of the early stage crack rub-impact fault. Among them,wavelet transform has good localization features in both time domain and frequency domain,which provides the frequency information of the target signal at different frequencies. Wavelet packet decomposition improves the time-frequency resolution. It saves the energy of the signal and reconstructs accurately by its characteristics. The EMD method adaptively decomposes the nonstationary signal into the sum of a series of zero mean frequency modulated amplitude modulation. The process and the experimental results of the feature extraction about early rubbing fault based on the three methods are discussed in detail. Experimental results show： For the early crack failure,the EMD method can accurately extract the fault features of the frequency doubling,and the method is more suitable for the feature extraction of the early crack rubbing.