摘要
在发动机故障诊断、特征提取中时频分析是一种重要的方法。利用包含调频波和谐波的仿真信号对Gabor变换、小波变换、Wigner分布、平滑Wigner分布、Hilbert-Huang变换以及压缩小波等时频分析方法,在时间分辨率、频率分辨率以及多分量信号识别能力上进行了比较,并用发动机实测信号进行了分析验证。结果表明,压缩小波能提供较高的时频分辨率,而且对特征频段能够实现精细重构。
Time-frequency analysis is widely used in engine fault diagnosis and feature extraction. In this paper,the simulated signals with FM and harmonic wave are used to compare the time-frequency analyzing methods in their capability of identifying time resolution,frequency resolution and the multi-component signals. The methods compared are： short time fourier transform（ STFT）,gabor transform,continuous wavelet transform（ CWT）,Wigner-Ville distribution（ WVD）,smoothing pseudo-Wigner-Ville distribution（ SPWVD）,Hilbert-Huang transform and synchrosqueezed wavelet transform（ SWT）. The the result of the comparison is verified with the real signals of the engine,which shows that SWT is capable not only of providing high time-frequency respresentation but also of refining the reconstruction of frequency band.
出处
《军事交通学院学报》
2016年第4期35-40,共6页
Journal of Military Transportation University
基金
总后勤部重点项目(BS311C011)
关键词
时频分析
特征提取
压缩小波
time-frequency analysis
feature extraction
synchrosqueezed wavelet
