Research On Data-driven Rationality Of Track Dynamic Irregularity Management Value Of High Speed Railway

Track irregularity is the main source of wheel-rail system and the main cause of various vibration of the train.It has a huge impact on the safety of train operations,the comfort of passengers and the economy of maintenance and repair.Under the background of the continuous development of high-speed railways in our country,there are problems not only about high speed brought by the increase of speed but also high density brought by the passenger and freight mixed lines.This puts forward higher requirements for the smoothness of the track.Therefore,it is particularly important to make a scientific and reasonable evaluation of the smoothness of the track.At present,the main evaluation methods for line maintenance in China are still the mean value management and the peak value management.The mean value management is based on five types of data,including Surface(left and right rails),Align(left and right rails),Gauge,Xlevel and Twist.The standard deviations are calculated and summed with 200 meters as the unit section,namely as a track quality index(Track Quality Index,TQI)for evaluating track status.The peak value management includes not only seven irregularity indexes,but also vertical acceleration,horizontal acceleration and gauge change rate.The peak value management takes the measured data half-peak value as the management value,sets four levels of thresholds and deducts points according to different levels of exceeding thresholds,as the standard of maintenance.Therefore,the rationality and relevance of these two evaluation methods are directly related to the work intensity and on-site cost of maintenance.This paper studies the rationality of track dynamic irregularity management value,which is of great significance to find the correlation between these two kinds of managements.It will also rationalize and improve the existing management standard.The main research work of this paper is listed as follows:(1)Based on the anomaly detection and correction model of the time series,we explored the relationship between the half-peak value and the standard deviation of each index through quantile regression and put forward the corresponding mean management recommended values.The rationality of the existing TQI management value was analyzed and evaluated based on the measured data of track inspection vehicles with different track slabs and operating speeds in southwest and east China;(2)In order to study the rationality of the current peak value management,this paper adopted the peak-over-threshold method based on extreme value theory,took the over-limit data in the historical inspection data as the modeling object,and used the generalized Pareto distribution to fit the over-limit data samples.Finally,we proposed a recommended value for peak management by combining the division rule and probability of over-limit peak value.Therefore,the peak management value obtained focuses on reflecting the degraded section of the line,which is more in line with the actual needs of maintenance;(3)In order to study the relative importance of different indexes of track quality index management value in track dynamic irregularity,the classification and linear regression models between horizontal acceleration,vertical acceleration and five kinds of irregularity were established by using the measured data of track inspection vehicles of different track slabs.By using dominance analysis method,the relative importance of dynamic irregularity index under different track slabs and operating speeds were calculated.Combined with the engineering example,the reliability of this method was verified,which can be used to guide the maintenance of the line.