Application of Sparse Component Analysis-Empirical Mode Decomposition in Compressor Fault Diagnosis

Abstract

The key of mechanical fault vibration diagnosis is to obtain vibration state information in an all-round way, and the key point is fault signal separation and feature extraction. Early fault signals and weak signals are submerged in strong background noise, which directly affects the effect of signal extraction and brings difficulties to mechanical fault diagnosis. In this paper, a method of mechanical fault vibration diagnosis based on Sparse Component Analysis (SCA) and Empirical Mode Decomposition (EMD) is proposed. Combining the advantages of the two methods, SCA is used to remove interference signals quickly, extract useful signals, and EMD is used to decompose fault feature information efficiently. Compared with other methods, this method can extract complex fault signals and weak symptoms early, comprehensively and accurately, and extract signals with higher similarity and more accurate separation accuracy. It has not only theoretical research value, but also practical engineering significance.