A RECIPROCATING COMPRESSOR FAULT FEATURE EXTRACTION METHOD BASED ON LMD AND MPE

This paper presents a feature extraction method based on local mean decomposition (LMD) and multiscale permutation entropy (MPE) for the fault diagnosis of reciprocating compressor at different bearing clearance states. The vibration signal was decomposed into a set of PFs by LMD, and the highlighted PFs which contain the main information of fault state were chosen according to the correlation coefficient. MPE of the selected PF components were calculated, and the optimized scale factor was selected based on the average Euclidean distances. The feature vectors of different bearing clearance fault states were extracted by using this method, and the comparison with results of LMD and sample entropy verifies the superiority of this method.