Research On Fault Dianosis Of Automotive Transmission Bearings

The rolling bearing is the key to the automotive transmission.The rolling bearing fault signal of transmission is nonstationary,nonlinear,non-gaussian,and the signal-to-noise ratio is low.So it is difficult to extract the fault features,thus increasing the difficulty of the rolling bearing fault diagnosis.Therefore the fault diagnosis and condition monitoring to the automotive transmission bearing is important to ensure the safe and improve the safety performance of transmission and the car.For fault diagnosis and trend prediction of the automotive transmission bearings,this paper carried out a research into the following several aspects:(1)Build the 3d modeling of bearing model,and then through the Ansys Workbench software to simulate the working condition of rolling bearing in the transmission.And introduce the simulation method in detail.It is concluded that when the rolling bearing is working,the signal is nonstationary,nonlinear,non-gaussian.(2)Chose the CEEMD for signal processing according to the characteristic of the signal.And expound the advantages of CEEMD compared with EMD and EEMD,then use the signal to verify it.(3)By comparing the advantages and disadvantages of the time-domain indexes,determine the kurtosis index as characteristics for bearing fault.Then solve the kurtosis index of IMF component after CEEMD decomposition,constitute the bearing fault feature vector and higher dimensional array.Then use the method of Laplace score for the selection of the IMF component,namely dimension reduction,and then form a characteristic matrix,to prepare for the later data classification.(4)According to the characteristics of the support vector machine(SVM),and use the particle swarm optimization algorithm to optimize the parameters of support vector machine.And finally test samples,the support vector machine model has good classification performance.Not only can recognize the fault of the inner ring,outer ring and rolling body,but also can classify fault degree,and the accuracy is acceptable.