Research On Analysis Method Of Rolling Bearing Compound Fault Based On Resonance Demodulation

As the core component of rotating equipment,rolling bearings have complex and harsh working environment,frequent failures,and different faults are coupled with each other.Once faults occur,it often brings great economic losses and even threatens people’s life safety.In order to solve the above problems,based on the damage theory and feature extraction methods of rolling bearings inland and overseas,this paper studies the analysis method of rolling bearing compound fault based on resonance demodulation.Firstly,the basic principle of the resonance demodulation is introduced,and the effectiveness of the resonance demodulation is verified by a simulation experiment.Then the influence of noise,low frequency harmonics and the difference of resonance frequency band characteristics on the detection result of resonance frequency band is analyzed.Finally,two difficulties,determination of the optimal resonance frequency band and the fault impact quantification index,in the existing resonance demodulation are analyzed,which lays a foundation for the following chapters.Secondly,to solve the problem of resonance frequency band detection of rolling bearing compound fault,this paper proposed an adaptive cycle shock factor Infogram(CSFI)based on the idea of fast kurtogram(FK),and the CSF index that can reflect both the transient energy and periodicity of the fault impact is proposed,so that the fault impact component can be quantified more accurately.In addition,the Welch spectrum estimation is utilized to determine the boundary between the modes in the EWT,which can improve the accuracy of the detected modes and the robustness of CSFI.Finally,the effectiveness of CSFI is verified by a simulation experiment,and the results show that the CSF can well quantify the fault impact,and the combination of Welch spectrum estimation and EWT can more accurately detect the optimal resonance band,so as to highlight the fault features.Thirdly,the PVMD-DS is proposed to solve the problem that the resonance band with relatively insignificant characteristic is covered up due to the obvious difference in resonance frequency band of rolling bearing compound faults.By constructing the incomplete decomposition condition and the iteration termination condition,a progressive decomposition VMD algorithm is proposed,and the problem of determining the number of modes K in VMD was successfully solved by extracting the modes one by one.In addition,the penalty factor α of VMD is optimized by the proposed energy fluctuation factor(EFF),and combining the indexes of kurtosis and EFF,a double screening criterion for fault mode is proposed to achieve accurate screening of fault mode.Finally,the effectiveness of PVMD-DS is verified by a simulation experiment,and the results show that PVMD-DS can solve the problem of determining the key parameters of K and α in traditional VMD algorithm,and the double screening criterion can select the fault mode accurately.At the same time,the PVMD-DS is more excellent when considering the performance of extracting fault characteristics and algorithm running efficiency.Finnally,the two datasets,HRB6205 bearing inner and outer race compound fault and outer race and rolling element compound fault,from ABLT-1A testbench are used to verify the effectiveness of the CSFI and PVMD-DS.By comparing with existing algorithms such as FK and VMD-CSA,the superiority of CSFI and PVMD-DS in practical application is further verified.