瞬频特征的谐波小波域提取与应用

Instantaneous frequency extraction in harmonic wavelet domain

为了能精确提取出复杂信号中特征分量的瞬频特征,提出了谐波小波域内的分量瞬频估计方法。该方法首先在分析谐波小波变换的基本原理及其优良特性的基础上,对谐波小波进行加窗光滑处理,改善了谐波小波的时域衰减特性。然后,针对在传统瞬频估计方法的不足之处,将线性回归、自适应滤波、曲线平滑、重采样等方法应用到瞬时相位函数的处理中,提高了瞬时频率特征的提取精度。最后,通过仿真实验与工程实践验证了该方法在信号分量瞬频特征提取中的可行性与有效性。

In order to extract the (IF) of feature component from complex multi-components signals,the component IF extracting method in harmonic wavelet domain was proposed,and its realizing process was given in detail.Based on the analysis of harmonic wavelet transform principle and in light of its excellent properties,the frequency spectrum discontinuity point of harmonic wavelet was smoothed by adding certain windows and the attenuation rate of wavelet in time domain was improved.Aiming at the disadvantages of universal IF estimation,the techniques of linear regression,adaptive filtering,curve smoothing and resampling were applied in the process of deriving instantaneous phase.The simulation results of synthetic and gearbox fault signals were presented to verify the feasibility and validity of the method proposed.