The Study of Local Wave Noise Reduction Based on the Correlation Analysis

Abstract

The signal can be divided into a number of intrinsic mode components (IMF) through local wave decomposition. The decomposition process is equivalent to the adaptive filter for signal. The frequencies of each IMF have reduced with the decomposition order. This paper established judgment criteria of signal, noise and pseudo component based on the local wave decomposition, used the characteristics of cross-correlation coefficient and autocorrelation sequence, combined with the power spectral density. The dominant mode function was extracted effectively and the final effects are used gear reducer in the JZQ250 of fault rolling bearing inner ring. It provides a new method for non-stationary signal adaptive noise reduction and fault diagnosis of rotating machinery.