Acoustic Emission Signal Monitoring And Characteristics Analysis Of Rolling Bearings For Large Amusement Facilities

In recent years,while the amusement facility industry has developed rapidly,its safety issues have been receiving much attention.As a key component of largescale amusement facilities,rolling bearing failure may directly cause accidents and casualties.There is no effective detection method due to its difficult disassembly characteristics.Therefore,it is an urgent need for a non-destructive testing method that is in-service and non-disassembly for rolling bearings of large amusement facilities.Acoustic emission technology is a dynamic and non-destructive monitoring technology,and it also has the ability of early forecasting to meet this demand.Research on rolling bearings with different operating conditions based on acoustic emission technology and the fault diagnosis method is proposed,which lays a foundation for on-line acoustic emission monitoring and fault diagnosis of rolling bearings in large amusement facilities.The main work is as follows:(1)Building a ferris wheel simulation device with a height of 2.9 m and a diameter of 2 m.The bearing is a 23138 CA rolling bearing.Acoustic emission(AE)tests were carried out to study the characteristics of the AE source of the bearing with rolling element and inner ring faults.(2)Acoustic emission signals of rolling bearings without defects,rolling element failures and inner ring failures are collected.Firstly,the signal is denoised by singular value decomposition technique,then the signal is decomposed by empirical mode decomposition method.The effective IMF component is extracted by correlation coefficient method,and the IMF component with the largest energy is selected for approximate entropy calculation.The experimental results show that the approximate entropy of the fault bearing is obviously larger than that of the normal rolling bearing.(3)Combining the energy of the IMF component with the approximate entropy of the largest energy IMF component as the feature vector is input to the support vector machine for classification and recognition,which is higher than the recognition accuracy of the single feature.Compared with the classification and recognition effect of support vector machine and BP neural network,support vector machine is more suitable for classification and recognition of rolling bearing faults with small sample data.(4)The acoustic emission monitoring tests was carried out on the ferris wheel and wave rolling of an amusement park in Zhengzhou and the wave rolling of an amusement park in Liuzhou.The parameters and waveform spectrum of the acoustic emission signals were analyzed,and the approximate entropy method was used to judge the bearing without defects.