Study On Analysis Method Of Track Geometry Data Based On The Theory Of Control Chart

The quality of the rail track geometry directly affects the safety,stability and comfort of the traffic.In recent years,with the increase of railway lines and the speed of trains,higher requirements have been put forward for the maintenance and management of track geometry.The track inspection car is one of the most commonly used tools for the inspection of the geometric state of the track,which plays an important role in the maintenance and management of the track.In-depth analysis of the inspection data of the track inspection car is of great significance for the realization of the fine management of the track geometry.In this paper,the control chart technology is introduced into the field of railway track state analysis.Based on the data of rail track inspection car,individual control chart and Z-MR control chart model are established to deeply analyze the data of the track geometry.The main research work is as follows:First of all,the autocorrelation of the track geometry data is analyzed.The independence of data statistics is the premise of the statistical process control theory.The autocorrelation of data will invalidate the control chart theory.Using the autocorrelation function(ACF)and partial autocorrelation function(PACF)to test and identify the autocorrelation of track geometry inspection data in spatial and time domain respectively.The reason for the autocorrelation of track geometry inspection data is analyzed,and several measures to resolve the autocorrelation of state data are given.Secondly,the statistical(probability distribution)characteristics of track geometry data are analyzed.In view of the randomness of track irregularity,applying mathematical statistics principles and methods,statistical analysis is performed on track geometry inspection data,frequency distribution histogram is plotted,and the main probability distribution function is selected for goodness-of-fit test.The characteristics of probability distribution of track irregularity are studied.Thirdly,the individual control chart model for spatial data analysis of track geometry state is established.The track geometry data acquired by the track inspection car has strong autocorrelation.The control limits based on the empirical percentile are used to widen the control limits,the ratio of receiving false alarms is obtained,and an individual control chart model for analyzing spatial data is established.Taking the twist irregularity as an example,the actual data of Shenmu-Shuozhou Railway was used to verify the effectiveness of the individual control chart model.Finally,the Z-MR control chart model for analyzing the spatiotemporal data of the track geometry is established.Since the individual control chart method based on the static warning limit only uses spatial data to diagnose geometry irregularity of the track,in order to achieve more accurate monitoring purposes,a Z-MR control chart model for simultaneously analyzing spatial-temporal data is established.Data Binning is first used to overcome the impact of track inspection car positioning errors on data analysis,and Box-Cox model is used to take "treatment" measures for data that does not meet the normal distribution to improve the normality of the data.Finally,the effectiveness of the Z-MR control chart model was verified using the actual data of the Shenmu-Shuozhou Railway.The study indicates that that the approach could be used for condition assessment of tracks.The two control charts used give earlier fault warnings compared to the traditional approach,providing decision support for planned condition-based maintenance actions.This gives the possibility to increase the operational availability of track and ensures that the rails are always in a controlled and reliable state.