Research On Compound Fault Diagnosis Method Of Rolling Bearing Under Varying Speed Conditions

Rolling bearing plays a decisive role in rotating machinery.However,it often causes faults in harsh environments,so it is of great significance to the health monitoring and fault diagnosis research of rolling bearing.This thesis focuses on the common varying speed conditions in engineering applications,and analyzes the characteristics of the compound bearing fault vibration signal and the diagnosis method.Experiment results have proved the effectiveness of the method studied in this thesis for the diagnosis of compound bearing fault under varying speed.The main research contents of the thesis are as follows:Aiming at the problem that the traditional time domain and frequency domain analysis method is difficult to apply to varying speed conditions,the characteristics of fault vibration signals under varying speed conditions are studied.Starting from the mechanism of bearing vibration,the fault vibration model is established in the order of constant speed to varying speed,single fault to compound fault,and the simulation signal is constructed to analyze and compare the characteristics.The characteristics of fault impact components in various domains under variable speed conditions are analyzed,and the effective analysis domain under this condition is pointed out.The characteristics and problems of compound fault in the order domain and time-frequency domain are analyzed in detail,which lays the foundation for the subsequent method research.Aiming at the problem that the compound fault signal components are complex and the mutual interference between the faults makes it difficult for the traditional computation order tracking method to be directly used for the compound fault feature extraction under varying speed,combined with the order domain characteristics of the compound fault signal under varying speed,a compound fault diagnosis method based on the virtual multi-channel signal in angle domain is studied.Firstly,the envelope demodulation of the original signal is carried out,and the equal-angle resampling of the envelope signal is performed by using the computed order tracking to convert the time-domain signal into the angle domain.Then,the variational mode decomposition is introduced to decompose the angle domain signal into multiple modes,and modes with richer fault information are selected through the correlation coefficient to construct the virtual multi-channel signal.Finally,the classical blind source separation algorithm-independent component analysis is used to separate and detect the compound fault.By analyzing the experimental data,this method can effectively detect multiple faults of bearings simultaneously.Aiming at the problem that existing compound fault diagnosis methods based on order tracking and non-stationary fault characteristic demodulation generally need to install an additional tachometer and difficult to apply to complex speed change conditions,combined with the time-frequency domain characteristics of the compound fault signal under varying speed,a novel compound fault diagnosis method based on multiple time-frequency curve classification.Firstly,multiple time-frequency curves with higher amplitude are extracted from the time-frequency distribution of the vibration signal by the local peak search algorithm.Then classify these time-frequency curves according to the proposed classification algorithm to exclude irrelevant interference curves.Finally,the rotational frequency is estimated and the fault diagnosis is performed according to the classification result.Experimental analysis and comparison show that this method has higher speed estimation and fault diagnosis accuracy,and has better practicability than traditional methods in the working condition where the rotational speed is difficult to measure or the speed change is complicated.