The Applicationof Of Waveletridge Method In Fault Diagnosis Based On EMD

With the development of social production, continuously improve the efficiency ofproduction equipment. It puts forward higher requirements at the hydraulic system and thebearing element performance. The hydraulic pump and the bearing element of real-timefault diagnosis is more important. Axial plunger pump with rolling bearing fault signal is akind of typical non stationary, nonlinear signal, so it requires a suitable for processingnon-stationary signals, nonlinear methods to accurately extract the fault characteristics.EMD decomposition and wavelet ridge in the processing of such signals has its uniqueadvantages.Empirical mode decomposition (EMD) can decomposes the original signal into aseries of intrinsic mode function (IMF). Each IMF component amplificate thecharacteristics of data effect,and the research of each IMF component can be clearlyfound fault feature. Wavelet ridge line based on the wavelet transform theory, it is madein time---frequency meeting each time point wavelet coefficients modulus maximumpoints formed by collection. These tend can characteriza fault feature information to bemore clear. The empirical mode decomposition and wavelet ridge combined for faultsignal analysis can be more clear analysis of fault information.In order to verify the validity of the method and superiority, this paper analyzes theswash plate axial piston pump sampled vibration signals and American Case WesternReserve University bearing fault simulation test bench for acquisition of bearing faultvibration signals. Through the contrast of marginal spectrum which is extracted by faultsignal EMD decomposition, sensitive IMF components of the wavelet ridge envelopedemodulation analysis. Accurate extraction of the hydraulic pump and the rolling bearingvarious state when the sensitive frequency. Through the sensitive component IMFwavelet ridge demodulated spectrum and Hilbert transform demodulation spectrumcomparation, by contrast of Hilbert transform demodulation it proves that the method hashigher time-frequency localization accuracy.This paper presents a EMD decomposition based on wavelet ridge demodulation signal energy feature extraction method. Through the EMD decomposition sensitive IMFcomponents of the wavelet ridge demodulated signal envelope,then use EMD methodagain after the envelope of the signal be reduced the sampling frequency. Use thesecond decomposition of the IMF component signal energy efficient extraction of faultfeature vector. Using K means clustering method for the hydraulic pump with bearingvarious status of fault pattern recognition, with Hilbert transform demodulation extractedfeature vector comparison shows that the method has some advantages.