Mileage Errors Estimation And Correction Model For Track Geometry Data From Track Inspection Cars

Track irregularity is the excitation source of wheel-rail system,which has adverse effect on the vibration of vehicle and passenger comfort.Besides,it aggra vates the dynamic impact between wheel and rail,causes great interaction force and leads to the damage of railway and vehicle components.Therefore,high-speed railway track should have high smoothness for guaranteeing the safe,stable and continuous operation of high-speed vehicle.Nowadays,track inspection car is the most common tool for measuring the track dynamic irregularity.However,the system of mileage identification in track inspection car is negatively affected by wheel size error,relative sliding between wheel and rail,wheel-axle grating encoder faults,GPS limitation etc..The detection data inevitably has mileage errors,which has significantly adverse impacts on the evaluation accuracy of track quality,labor intensity and maintenance cost.Therefore,studying on the mileage error of track dynamic irregularity is vital for the accurate evaluation of track geometry,the improvement of maintenance time utilization and the in-depth study of the evolution law of track geometry.The main research work of this paper is as follows:(1)The phenomena and effect of mileage errors of track geometry data from track inspection car were analyzed.According to practical engineering,the feasibility of Pearson correlation coefficient in processing waveform matching algorithm for track geometry data was analyzed.Based on the analysis of curve information an absolute mileage error evaluation model of track geometry was established,which can increase the number of mileage correction equipment,shorten the distance between each absolute mileage error evaluation equipment and improve the accuracy of the model.(2)The mileage errors between each channel of time history track geometry detection data was analyzed.Considering the data of multiple channels and poor repeatability of track geometry data,a relative mileage error evaluation model for track geometry data was established based on the multiple waveform matching.The model was solved by Lagrange multiplier method to obtain more reliable relative mileage error values.Based on linear interpolation algorithm and cubic polynomial interpolation,a mileage error correction model was established.According to the data of track geometry data from one railway line,the model can effectively avoid the waveform distortion caused by the processing of mileage errors based on single data.Proposed model has good reliability and stability in dealing with poor waveform repeatability.In addition,the value of window length was suggested by analyzing the influence of various window lengths on the accuracy of correction.(3)Based on the standard deviation of time history data,an index to measure the characteristics of track geometry fluctuation was proposed,which can be used to locate the regions with large track irregularity fluctuation quickly.According to engineering examples,the fluctuation characteristics of track irregularity in subgrade,bridge and tunnel were analyzed.It was found that the fluctuation characteristics were easily affected by outliers.(4)Based on the time history detection data and statistical methods,the characteristics of change rate of track geometry data were analyzed.Besides,the threshold range of track irregularity feature was proposed.An outlier recognition model for track geometry data was established based on the time history data,and an outlier correction model was established based on linear prediction algorithm.According to engineering example and dynamics,it was verified that the recognition model and correction model in this paper had higher calculation accuracy,processing models can effectively avoid errors in identifying outliers and make the detection data closer to the actual geometrical state of the track.