On-line Condition Monitoring And Fault Diagnosis For Rolling Bearing Of Urban Rail Vehicle Gearbox

Urban rail vehicles,as a tool to support the normal travel of urban residents,are required to reflect the advantages of fast,efficient and convenient.The gearbox is an indispensable part in the running process of city planning vehicle,and it needs to ensure its normal operation at all times.In order to achieve this target,it is necessary to carry out real-time condition monitoring of rolling bearing.In this dissertation,the research object is the rolling bearing of the gearbox of the city planning vehicle,and the effective information collection,feature extraction,fault diagnosis model establishment and sensor information fusion are studied.The fault signals are pre-processed by noise reduction,feature extraction by ensemble empirical mode decomposition(EEMD),and fault recognition by Support Vector Machine.The main contents are as follows:(1)Firstly,the vibration mechanism of urban rail vehicle gearbox and the causes of rolling bearing failure are analyzed.On this basis,an intelligent online monitoring platform of rolling bearing is built to collect the vibration and temperature signals of rolling bearing during the operation of urban rail vehicle.The requirements of the monitoring system are analyzed.The hardware circuit design and chip selection are completed.Each module is tested,which can be used in the monitoring system In the running process of wheel box rolling bearing,accurate and real-time state information is collected.(2)In view of various conditions of urban rail vehicles during operation,there are various interferences in the process of fault signal extraction,in which noise is the main factor.Before fault diagnosis,the operation fault signal is de-noised by wavelet packet analysis method,and the horizontal noise reduction comparison effect is made,which verifies the advantages of wavelet packet de-noising.(3)The method of ensemble empirical mode decomposition(EEMD)is used to decompose the vibration signal.EEMD can also effectively eliminate the noise in the decomposition process.EEMD can decompose the signal into different IMF components,introduce the signal energy,and build a comprehensive evaluation index to screen IMF.(4)On the basis of signal analysis,the parameters of support vector machine(SVM)were optimized by genetic algorithm(GA),and a GA optimized SVM model was established to diagnose the gearbox fault signals with the accuracy of 98.44%.And the fault analysis when the shaft temperature rises is added in the rolling bearing fault diagnosis,so as to comprehensively identify the bearing fault.This fault identification method is higher than the traditional support vector machine fault identification.