【作者】 王祥；

【导师】 徐鹏；

【作者基本信息】 北京交通大学 ， 交通运输规划与管理， 2021， 硕士

【摘要】 季节性冻土地区的路基冻胀导致轨道几何尺寸发生变化(称为轨道变形),降低了线路平顺性,甚至构成列车运行安全隐患。我国东北地区位于季节性冻土地区,每年存在明显的冻胀和融沉现象,而高速铁路对轨道不平顺要求极高,因此,针对运营高速铁路,掌握路基冻胀过程中轨道变形随温度变化的过程,不仅能够为铁路工务部门管控季冻区域内轨道几何尺寸变化提供依据,也能够大大降低行车安全隐患。本文针对无砟高速铁路,利用综合检测列车数据、气温数据和13处路基冻胀监测数据,以每处路基冻胀为对象,研究了路基冻胀量和轨道变形的关系,确定了轨道变形的梯形变化过程,并对梯形变化过程的特征进行了分析。根据得到的轨道变形的梯形过程,量化了路基冻结深度和冻胀量的关系,建立了路基冻胀处轨道变形随温度变化的梯形函数关系,基于贝叶斯变点识别方法和马尔可夫链蒙特卡罗方法提出了估计梯形函数中参数的方法。将建立的梯形函数关系和参数估计方法应用到HSR-A高速铁路上,利用3年的数据分析了估计出的轨道变形随温度的梯形关系和轨道变形预测值的误差。结果显示:轨道变形与温度间的梯形函数关系与实际变化过程相符,在没有维修作业的情况下,用y=x拟合轨道变形预测值和实际值的p值远小于0.05。这表明:建立的梯形函数关系是有效的,估计梯形函数中参数的方法是可靠的。对轨道变形随温度变化的梯形函数关系中的系数进行了聚类分析,由于K均值聚类算法受初始解影响较大,本文利用基于粒子群的K均值聚类算法对其进行聚类分析,该算法通过加入粒子群算法,使得聚类算法具有较好的全局搜索能力,根据得到的聚类结果,进一步表明了本文所建立的轨道变形随温度变化的函数关系必须是以每处路基冻胀为对象。图38幅,表3张,参考文献66篇。更多还原

【Abstract】 Frost heaves of the subgrade in regions with seasonal freezing can cause changes in the geometric dimensions of a railway track(known as track deformation),which reduces the smoothness of the track and even poses a hidden danger to train operation.Northeast China is located in a seasonally frozen area,and there are obvious frost heaves and thaw settlements every year.High-speed railways have extremely high requirements for track irregularities.Therefore,for the operation of high-speed railways,grasping the change of track deformation with temperature in the process of frost heave can not only provide a basis for the railway public works department to control the changes in the geometric dimensions of the track in the seasonal freezing area,but also greatly reduce the hidden dangers of driving safety.Aiming at the ballast-less high-speed railway,this paper uses Track Geometry Measurements,atmospheric temperature data and 13 subgrade frost heave monitoring data,with each subgrade frost heave as the object,studies the relationship between subgrade frost heave and track deformation,and determines the track deformation as Trapezoidal change process,and analyzes the characteristics of the trapezoidal change process.According to the trapezoidal process of track deformation,the relationship between the frozen depth of subgrade and the amount of frost heave was quantified,and establishes the trapezoidal function relationship of track deformation depended temperature at the position of frost heave.Based on the Bayesian change point identification method and Markov Chain Monte Carlo method proposes a method to estimate the parameters in the trapezoidal function.The established functional relationship and parameter estimation method are applied to a high-speed railway,and the errors in forecasting track deformation are analyzed using the trapezoidal relationship estimated from three years of data.The results show that the trapezoidal functional relationship between track deformation and temperature is consistent with the actual changes in the process.In the absence of maintenance operations,the p-value in fitting the predicted track deformation value and the actual value to y = x was less than0.05.This shows that the established relationship is valid and the method of estimating the functional parameters is reliable.Cluster analysis is used on the coefficients in the trapezoidal function relationship of track deformation depended temperature.Since the K-means clustering algorithm is greatly affected by the initial solution,this paper uses the K-means clustering algorithm based on Particle Swarm to implement clustering analysis.By adding the Particle Swarm algorithm,this algorithm makes the clustering algorithm have better global search ability.According to the obtained clustering results,it further shows that the functional relationship of track deformation depended temperature established in this paper must be based on the frost heave of each subgrade as the object.38 pictures,3 tables,66 references.更多还原