Fault Diagnosis of Reciprocating Compressor Using Empirical Mode Decomposition-Based Teager Energy Spectrum of Airborne Acoustic Signal

Due to the presence of several rotating and reciprocating components, the acoustic signal obtained from the compressor normally involves transient impact and noise. The nonlinear and non-stationary response also contribute to the corruption of the useful information and lead to the diagnosing difficulty using the traditional methods. Due to the high nonstationary characteristic, time-frequency domain analysis can be a good choice for processing the airborne signal. In this paper, a new method based on empirical mode decomposition (EMD) and the Teager energy operator (TEO) is proposed to extract the fault features from the airborne acoustic signal for the reciprocating compressor fault diagnosis. The very unique advantage of TEO in detecting the transient response from a signals is exploited, and be applied to track the total mechanical energy that is responsible for generating the acoustic signal. Firstly, the signal is decomposed in the monocomponents by using EMD to satisfy the monocomponent criterion of the Teager energy operator. Next, the appropriate intrinsic mode function (IMF) is selected based on the correlation coefficient for obtaining the instantaneous energy by TEO. Finally, the spectrum analysis is done on the energy series for identifying the repeating frequency of the periodic impulses and thereby to diagnose the induced compressor faults under broad range of discharge pressures. The proposed method was also compared with the existing state of the art techniques like Hilbert energy spectrum and traditional spectral analysis. The comparison study shows the effectiveness of the proposed method over the existing methods in diagnosis of reciprocating compressor faults based on airborne acoustic signal analysis.