Singular Signal Feature Extraction and Identification Based on Empirical Orthogonal Function for Module Maxim and Phase Matrix of Wavelet

The singular signal feature extraction and identification based on empirical orthogonal function for module maximum and phase matrix of wavelet is proposed.By using the method,the singular characteristic value vector and the plot of maximum and phase are obtained as the signal feature.The experiments on the application of fault diagnosis for rolling bearings shows that singular characteristic value vector has the merit of time translation invariability,the plot of denoised module maximum and phase matrix of wavelet clearly reflects the singular feature of vibration.