Research On Improved Blind Source Separation Method Of Gear Box Fault Vibration Signal

As an important mechanical transmission equipment,gearbox faults must be early predicted to avoid the occurrence of major accidents.It is the key step of gear box fault diagnosis how to extract fault features timely and effectively.Among many fault feature extraction methods,modern vibration signal analysis technology has been widely used in gear box fault feature extraction.However,in order to avoid devanning,the mixed vibration signals contain different vibration signals and environmental noise,which results in the fault feature information to be covered by non-fault information.Thus,it is difficult for a single or traditional signal analysis method to extract fault feature.On basis of fully investigating modern vibration signal analysis,gearbox as the main research object,The de-noising wavelet and independent component analysis are deeply studied.Combineing with the advantages of self-adaption wavelet transform(SAWT),improved fixed-point algorithm(LFastICA)and envelope spectrum,SAWT-LFastICA-envelope spectrum is proposed to extract fault features from gearbox.The research contents mainly include the following five aspects:(1)The vibration mechanism and fault signal characteristics of gearbox are studied,which paves the way for subsequent accurate extraction.(2)Investigating the vibration signal analysis methods of gearbox,the advantages of different signal methods are summarized for providing strong basis to select comprehensive method.(3)SAWT is elaborated in detail,and the superiority of SAWT denoising is verified by comparing with the traditional wavelet,which provides the technical guarantee for the pre-processing of the observed signal.(4)It is deduced that LFastICA algorithm with better stability.Then it is proved that LFastICA has better decoupling effect by simulation signals analysis.(5)Based on the above research and the introduction of kurtosis and envelope spectrum,SAWT-LFastICA-envelope spectrum is proposed.Finally,the extracted fault features of gearbox further verify the validity and accuracy of the comprehensive analysis method by experimental data analysis.