Reciprocating Compressor Fault Diagnosis Based on Airborne Acoustic Signal: Comparison Study Between Modulation Signal Bi-spectrum and Teager Energy Spectrum Method

Reciprocating air compressors are the one of the most complex machines in industrial application that involves a number of rotating and reciprocating parts. Maintenance and proper fault diagnosis of these machines are quite important. Airborne acoustic analysis can be a very good choice due to the less number of sensors associated with it, making it a cost effective method and also having a wide dynamic range, the microphones can record more detailed information compared to the other traditional methods. The acoustic signals of the compressor are nonstationary. The non-linear properties of the valve fluttering, fluid-mechanical interaction, resonance due to the turbulence flow through pipe and a huge background noise make it difficult to analyse the acoustic signals for detection of compressor faults. Depending on the nature of the acoustic signals obtained from a compressor two methods have been found to be very affective from the previous study. This paper reviews the application of the modulation signal bi-spectrum (MSB) method and Teager Kaiser Energy (TKE) method in order to compare the performances of these two methods for diagnosis of faults in a two stage reciprocating single acting air compressor.