【作者】 曲建军；

【导师】 高亮；

【作者基本信息】 北京交通大学 ， 道路与铁道工程， 2011， 博士

【摘要】 当前,我国既有线提速事业取得了很大的发展,但形成的速密重并举的运输组织方式却使运输与线路养护维修的矛盾日益突出,工务管理部门在建立了科学、先进的轨道检测体系的情况下,主要围绕着提高轨道平顺性来保证运行品质和解决运营与维修间的矛盾。目前随着轨检车和综合检测车的推广应用,利用轨道不平顺动态检测数据对线路养护维修进行科学管理,已成为保障行车安全和取得经济维修效益的重要手段。本论文在应用大量轨道不平顺动态检测数据的基础上,结合工务段维修资料,对轨道不平顺区段评价指标—TQI (Track Quality Index)的时间序列进行了规律分析。应用系统工程的观点,将灰色系统不确定性理论引入轨道不平顺的预测领域,建立了轨道不平顺TITCGM(1,1)-PC灰色非线性预测模型。通过深入挖掘与提取轨道不平顺系统的辨识参数来表征系统发展过程的动态变化,结合大型养路机械的作业效果,分别对TQI时间序列的短期和中、长期预测问题进行了分析与研究。在此基础上,将TITCGM(1,1)-PC灰色预测模型应用在年度综合养护维修计划的编排中,建立基于轨道最优均衡质量状态的综合维修计划模型,为预防性维修体制下的轨道养护维修提供辅助决策指导。论文的主要研究内容和创新成果如下：1.轨道几何不平顺检测数据预处理技术的研究。分析了轨道不平顺动态数据在检测过程中的误差产生原因和特征,针对车载检测数据中存在的粗大误差(异常值)、趋势项和检测里程不对应问题,建立起一套合理的数据预处理方法,保证了TQI时间序列数据源的准确性。2.基于TQI时间序列的新型灰色预测模型的建立。利用灰色系统理论弱化轨道不平顺系统内各影响因素的不确定性,将TQI时间序列作为反映轨道不平顺系统发展的“白色信息”进行充分挖掘。结合TQI时间序列的发展特征,本文将其分离为趋势性成分和随机性成分。通过对灰色核心理论GM(1,1)模型进行系列的优化与改进,建立了TITCGM(1,1)模型模拟趋势性成分,又基于残差序列建立了周期性组合修正模型来模拟随机性成分,最终将两部分模型组合形成TITCGM(1,1)-PC灰色非线性预测模型,并基于MATLAB实现其算法程序。新建模型不仅能够挖掘出TQI随时间稳定发展的趋势特征,还可表现出其随机变化的波动特征。3.基于TQI时间序列的短期预测分析。应用TITCGM(1,1)-PC模型对我国提速线路沪昆线200-250km/h速度等级不同线路类形的实测TQI进行了短期预测分析。在预测分析过程中,利用TITCGM(1,1)-PC模型的后验差和系统发展系数对TQI时间序列的趋势拟合效果、修正拟合效果进行了比较和外推可靠性检验,继而比较验证了模型的预测值与实测值。结合整公里区段的T值管理,将TITCGM(1,1)-PC预测模型应用于整公里区段TQI实测数据的短期预测中,可以合理设置维修时间,科学指导线路的养护维修。4.基于TQI时间序列的中、长期预测模型建立与分析。结合TQI时间序列在大机作业前后的发展模式,利用TITCGM(1,1)-PC模型的时间函数特性对大机作业前维修周期内的TQI时间序列进行系统模拟分析,深入挖掘与提取轨道不平顺系统辨识参数的意义。结合大型养路机械的作业效率,将从已知稳定维修周期内TQI时间序列挖掘出的系统辨识参数作为未知各维修周期内TQI发展的特征参数,从而建立起大机作业影响下的轨道质量的中、长期预测模型,并以提速线路的实测TQI数据进行了预测分析。5.基于轨道最优均衡质量状态的综合维修辅助决策模型的研究。将TITCGM(1,1)-PC灰色预测模型应用在年度综合维修计划的编排中,更为合理地设置了综合维修作业前后轨道不平顺系统的非线性动态发展过程,并确立单元区段轨道质量状态年度综合维修发展模式,取代了养护维修决策计划模型中的轨道不平顺线性恶化模式。结合养护维修的实际情况进行条件约束,建立起基于轨道最优均衡质量状态的综合维修辅助决策模型,使线路不仅具有较好的轨道质量状态,还能保持良好的均衡性发展,科学、经济地安排了大型养路机械的作业资源。更多还原

【Abstract】 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.更多还原