A New Fault Diagnosis Method of Adaptive Demodulated Resonance Technique Based on Wavelet Packet in Multi-Information Domains

Abstract

Fault information is incomplete while using a single information domain fault feature parameters to construct fault feature vector, and demodulated resonance technique have to predetermine resonant frequency and fixed center frequency also has its shortcomings , in order to solve these problems, a new fault diagnosis method is proposed of adaptive demodulated resonance technique based on wavelet packet in multi-information domains. The fault feature vector extracted from multi- information domains is described, signal processing flow of envelope demodulation based on denoising and filtering of wavelet packet is analyzed, the fault diagnosis method of adaptive demodulated resonance technique based on wavelet packet is given, and the method is applied to fault diagnosis of axial piston hydraulic pump. Experiment results show that multi-domain feature vector increases the completeness of the fault information, it is able to obtain good diagnosis effect, and the new fault diagnosis method is able to identify known and unknown faults resonance frequency automatically, the frequency range is narrow, the rate of diagnosis is high.