Fault Feature Extraction of Reciprocating Compressor Based on Adaptive Waveform Decomposition and Lempel-Ziv Complexity

Abstract

The multi-source impact signal of reciprocating compressor often represents nonlinear and non-stationary. For this reason, the fault features of the signal are difficult to quantitatively describe using conventional signal processing methods. In this paper, a novel adaptive waveform decomposition method was proposed to convert the strong non-stationary multi-component signal into stationary single-component signal. Subsequently, the signal was denoised with threshold-based mutual information to protect from being interfered by the noise. Finally, to measure the nonlinearity of reciprocating compressor signals in four states (normal valve sheets, gap valve sheets, fractured valve sheets, and bad spring), the normalized Lempel-Ziv complexity indexes were employed. The results reveal that the proposed method is capable of extracting the different faults states of reciprocating compressor accurately, which supplies a measure of fault diagnosis and maintenance strategy for the reciprocating compressor.