Study On The Track Irregularity Prediction And Decision-aided Technology Based On TQI Of Raising Speed Lines

ABSTRACT:At present, great development has been achieved for the speed raising of the existing lines in our country, but meanwhile, the speed-density-weight developing transportation organization form makes the transportation and track maintenance become more and more contradictory. With the establishment of the scientific and advanced track inspection system, the maintenance manage department has made great efforts in the improvement of the track irregularity managements to ensure the running quality and to solve the contradiction between transportation and track maintenance. As the track inspection vehicles and comprehensive monitoring trains are putting into popularization and application, using the track dynamic inspection data to carry out scientific track maintenance management has become a significant way to ensure the safety of the running vehicles and to achieve economic benefits.Based on the massive track dynamic inspection data and the track maintenance data, the rules of the time sequence of the track irregularity section evaluation index-TQI (Track Quality Index) have been analyzed. From the view of system engineering, the thesis introduces the grey system theory into the track irregularity prediction field, and the grey nonlinear prediction model TITCGM(1,1)-PC for track irregularity has been established. Then the dynamic variation of the development process is represented by deeply digging and extracting the significance of the identification parameters of the track irregularities development system. With the operation work of the large track maintenance machine, the short-term, mid-term and long-term TQI prediction are analyzed and studied. On this basis, the TITCGM(1,1)-PC grey prediction model are applied into the annual rehearsals of the comprehensive maintenance plans, and the predictive comprehensive maintenance model basing on the optimized balancing track quality state is set up, thus providing the decision-aided guidance for track maintenance of the predictive maintenance system. The main contents and innovations of this thesis are outlined as follows:(1) Research on the track geometric irregularity inspection data preprocessing technology. The causes and characteristics of the errors of the track irregularities dynamic data in the inspection process are studied. According to the gross error (outliers), trend item and mileage errors within different track inspection procedures, a set of reasonable data preprocessing method is put forward, which ensures the accuracy of the TQI time sequence data source based on the inspection data.(2) Establishment of the track irregularities prediction model based on the TQI time sequence data. The uncertainty of each influence factor within the track irregularity system itself is weakened with the grey system theory, and the TQI time sequence data are adopted as the whiten information which reflect the track irregularity system. The TQI time sequence data are divided into the trend composition and random composition according to the developing characteristic of TQI time sequence data itself. Through a series of improvements and optimizations to the core prediction theory GM(1,1) model, the trend composition is simulated by the TITCGM(1,1) model, and then the periodic combination correction model is established basing on the residual series for the simulation of the random composition. With the above two models, the TITCGM(1,1)-PC grey nonlinear prediction model is finally established. Its algorithm program is realized via MATLAB. The new model can not only dig the stable developing characteristics of the TQI, but also can represent the fluctuation characteristics of the random trend.(3) The analysis for short term prediction based on TQI time sequence data. Using the TITCGM(1,1)-PC model, the short term prediction analysis to the different line types of the Shanghai-Kunming raising speed railway line with the speed level of 200-250km/h is made. In the prediction process, using the posterior error and the system developing coefficient of the TITCGM(1,1)-PC model, comparisons and extrapolation reliability tests are made for the TQI time sequence trend fitting effect and correction fitting effect. Then the prediction data and its true values are compared. With the T value management of the whole kilometer section and through the application of the TITCGM(1,1)-PC prediction model for short-term prediction with the whole kilometer section TQI inspection data, the scientific maintenance period and maintenance work plan can be set accordingly.(4) The analysis model for mid-term and long-term prediction was built ba sed on TQI time sequence data. With the developing mode of TQI time sequenc e data before and after operation by the large track maintenance machine, TQI ti me sequence during maintenance cycle before large track maintenance machine o peration is simulated and analyzed, using the time function character of the TIT CGM(1,1)-PC prediction model. Then the significance of the system identification parameters of the track irregularity developing system is deeply dug and extract ed. Considering the operation efficiency of the large track maintenance machine, identification parameters of the grey theory excavated from TQI time sequence d uring known and stable maintenance cycle are used for the characteristic paramet ers development of TQI in each unknown maintenance period. The track irregula rity mid-term and long-term prediction model is established considering the influe nce of the large track maintenance machine. In this way, the prediction is made using the TQI inspection data obtained from the raising speed line.(5) Research on the comprehensive maintenance decision-aided model based on the optimized balancing quality state. The grey prediction model TITCGM(1,1)-PC is applied in the annual rehearsals of the comprehensive maintenance plan, which sets the nonlinear dynamic developing process of the track irregularity system before and after the comprehensive maintenance work more reasonably. Then the unit section track quality state annual comprehensive developing mode is determined, which replaced the linear deterioration developing mode in the decision-aided maintenance plan model. With the actual maintenance situation as the condition restriction, the preventive comprehensive maintenance plan model of the optimized balancing track quality is established. In this way, the line can not only serve at a relatively high track quality state, but also maintain a balancing development, which makes the arrangements for the working resources of the large machine both scientific and economical.