机械故障诊断中微弱信号处理特征的提取

Feature Extraction of Weak Signal in Machine Fault Diagnosis

论述了门限小波变换的四阶累积量在微弱信号特征提取中的功能、实现和应用条件。比较了连续小波变换(CWT)、短时傅里叶变换(STFT)、维格纳-威利(WVD)利用高阶累积量和没有利用高阶累积量情况下,提取微弱信号特征的特性。仿真表明。

This paper describes the feature extraction of weak transient signal corrupted by heavy additive noise in machine fault diagnosing. A new method is proposed to solve the problem and how it is realized as well as its application condition. A comparison is conducted between the forth cumulant of the threshold of CWT(Continuous Wavelet Transform)and other methods based on CWT, WVD(Wigner-Ville Distribution) and STFT(Short Time Fourier Transform)with or without cumulant analysis. The result shows that the combination of Higher Order Cumulant and CWT offers a significantly better detection probability in poor signal-to-noise ratio scenarios.