Abstract: In order to characterize the degradation state of hydraulic pump, the degradation feature extraction method and degradation state identification method are studied. The degradation feature extraction method for pump which is named S transform relative spectrum entropy (STRSE) is proposed based on S transform (ST) for analyzing non-stationary vibration signal and relative entropy (RE)for characterizing the probability distribution difference among different signals. The analysis results of simulation signal demonstrate the availability and rationality of the proposed ST relative energy spectrum entropy (STRESE) and ST relative singular spectrum entropy (STRSSE) used as degradation features. The degradation feature vector is composed of the two features. The practical vibration signals of pump with piston shoe wear-out and loose faults are analyzed, and the results demonstrate the effectiveness of the proposed two features. The weighted grey relation method is used in the pump degradation state identification. A standard degradation mode matrix is built, and the degradation feature vectors of the samples to be identified are extracted. The grey relational analysis of degradation feature vectors and standard degradation mode matrix are performed. The grey correlation degrees are used to judge the degradation state of pump, and the analysis results demonstrate the favorable performance of the proposed method.

Key words: ordnance science and technology, hydraulic pump, degradation state identification, S transform, relative spectrum entropy, weighted grey relational analysis