Recognition Of Railway Wheel Polygon Based On EEMD-WVD Time Frequency Method

Wheel polygon is a common form of wheel damage on high-speed railways.It can aggravate wheel-rail shock vibration and damage,seriously affect passenger comfort,and threaten train safety.The monitoring and identification of wheel damage status is of great significance for maintaining the safe and efficient operation of trains.Traditional detection method can achieve wheel damage detection under stationary or low speed operation,but it is greatly affected by human factors,the detection efficiency is low,the cost is high,and the train wheels under high speed operation cannot be detected in time.It is an effective method to detect the wheel polygon damage based on the vibration of the axle box of the vehicle.Among them,the signal analysis and processing is the key technology to ensure the accurate identification of the wheel state.Therefore,this paper proposes an improved ensemble empirical mode decomposition and Wigner-Ville distribution time-frequency analysis method to analyze the vibration acceleration signals of the axle box caused by the wheel polygon obtained by the simulation model to identify the wheel polygon damage.The main contents include:(1)EEMD-WVD time-frequency analysis method.According to the algorithm principles and characteristics of EEMD and WVD,EEMD is used to suppress the cross term of WVD without reducing the time-frequency resolution of WVD,which provides a theory for the identification of acceleration signals of axle boxes.(2)Dynamic model of rigid-flex coupling system for vehicle track.In order to study the effect of wheel polygonization on the vibration characteristics of the axle box and its state identification,a dynamic simulation model of the vehicle-track rigid-flexible coupling system considering wheelset elasticity and track elasticity was established by using ANSYS and Universal Mechanism software.(3)Response characteristics of axlebox vibration caused by wheel polygon.Analyze the response characteristics of the axle box vibration caused by the wheel polygon during high-speed running of the train,and establish the relationship between the train speed,wheel polygon harmonic order and wave depth and the time-frequency domain characteristics of the axle box acceleration by using the simulation model.(4)Wheel polygon state recognition based on EEMD-WVD time-frequency analysis.The proposed EEMD-WVD combined time-frequency analysis method is used to analyze and process the axle box acceleration simulation signals under the effect of wheel polygons to realize the wheel polygon state recognition and damage assessment.The reliability and advantages of the proposed EEMD-WVD time-frequency analysis method are verified by analyzing field measured data and comparing it with other time-frequency analysis methods.