The Flat Scar Fault Detection Of Urban Rail Train Based On Trackside Vibration Signal

With the national railway speeding up,The failure rate of vehicles shows an upward trend.To ensure the operation safety of train,Thus,detection of wheel flat timely and effective is of great significance.In this paper,taking the wheel flat as the research object and the multi-body dynamics software SIMPACK as the simulation foundation,by building Wheel flat model,The wheel-rail impact caused by Wheel flat is discussed.In order to detect the wheel flat,a method combine Empirical mode decomposition and Multi-scale entropy is proposed.The experimental results indicate that this method is effective and feasible.First,the formation mechanism and prophylactic measure of wheel flat scars are summarized.A detailed classifcation of the wheel flat is carried out.The formation mechanism and external influence factor of every kind of flat are discussed.Then the prophylactic measure is giving.next the vehicle dynamic indicators related to Wheel flat is discussed and the damage caused by flat is pointed out.Then,wheel-Rail impact analysis based on Multibody dynamics software is discussed.Firstly,a complete and meticulous introduction of wheel-rail model is introduced.The mechanism of two-dimensional and three-dimensional flat-impact mode is discussed from the perspective of wheel-rail kinematics.In order to analysis the flat impact from the wheel-rail dynamics,the vehicle-rail coupling model is built using the SIMPACK.Then the influence of length and new or old of flat were investigated using wheel-rail impact force as the evaluation standard.The results shows there are not correlation between the wheel-rail impact force and vehicle velocity,the length and new or old of wheel flat which proves the vibration acceleration of rail is a better parameter to detect wheel flat.Next,a method combined Empirical mode decomposition and Multi-scale Entropy is proposed to detect the wheel flat.First,the common methods to process vibration signal were introduced.Then EMD and MSE were emphatically introduced.Considering the advantages of multiscale entropy in the process of dealing biological signals,an innovative method combined MSE and EMD is applied to the fault detection of wheel flat.Besides,the partial mean value of MSE is used to represent the multi-scale entropy.By contrast,the method proposed is feasible and effective.Finally,the online monitoring system for vibration is designed.The overall design and the hardware design of the system were introduced in detail.Through the analysis of the field measured data,the method proposed in this paper can identify the wheel flat effective which provides a new solution for detection of wheel flat.