Comparison of Hydraulic Pump Faults Diagnosis Methods: Wavelet vs. Spectral Analyses
This paper introduces two faults diagnosis methods, a conventional spectral analysis method and a wavelet transform method, for hydraulic pump applications. The fundamental technologies of both methods, as well as their performance in detecting a few common hydraulic pump defects, are described in this paper. The performance of both diagnoses methods were evaluated based on experimental results. In order to eliminate the effects of border distortion arising from applying wavelet transform to finite-length signals, the pump outlet pressure in this case, a preprocess on the obtained signals is carried to clean up the errors prior to faults diagnosis analysis. Validation results obtained from both methods in analyzing the same data sets indicated that the wavelet transform based method showed a more sensitive and robust detecting capability than that obtained from a spectrum analyses approach.