【作者】 谷天峰；

【导师】 王家鼎；

【作者基本信息】 西北大学 ， 地质工程， 2007， 博士

【摘要】 郑西客运专线(高速铁路)是我国在黄土地区修建的第一条高速铁路，最高时速达350km，全线铺设无碴轨道。它与普通线路的有碴轨道相比，对路基沉降的变形要求更高、更严。由于客运专线部分经过高烈度的巨厚黄土区，因此黄土震陷是路基抗震设计中必须考虑的重要问题之一。另一方面，列车长期振动引起的黄土地基附加沉降(振陷)是路基抗震设计中必须考虑的另一个重要问题。为此，本文基于国内外有关震(振)陷的研究现状，在综合分析的基础上，结合铁道部重大科研课题开展了高速铁路黄土地基震(振)陷的研究和试验工作。本文在对线路区域的环境地质条件、工程地质特征分析的基础上，开展了多种黄土动、静力学试验，系统的探索了黄土路基在地震和列车振动作用下的震(振)陷变形及其相关问题。研究中采用现场勘察、现场测试与室内试验相结合；定性综合分析与定量、半定量分析相结合；线性与非线性理论相结合；模拟与计算相结合，取得了一些较有价值的研究成果，主要包括：(1)收集整理了郑-西客运专线沿线的环境地质和工程地质资料，并通过野外勘察，全面了解客运专线的工程地质情况，通过土的常规物理、力学试验，获取沿线黄土的一系列物理和静力学参数。(2)在郑-西客运专线沿线进行了一定数量的现场剪切波速测试，测试结果表明剪切波速随深度呈指数关系增大。根据波速测试和震陷试验结果，建立了利用剪切波速预测黄土震陷的回归模型。(3)将长持时的正弦荷载通过动三轴施加在原状黄土上，研究了黄土在这种特殊荷载下的残余应变(振陷)特征。根据试验结果，对不同高度路堤在机车长持时振动荷载下的沉降量进行了简化计算。结果表明，振陷量随路堤增高而减小，随振次的增加而增大。同时，还建立了黄土振陷的预测模型。(4)根据该沿线地震危险性分析结果，对典型黄土场地(华阴、潼关、灵宝、偃师)进行了人工地震波合成。将合成的地震波通过动三轴直接施加到客运专线沿线原状黄土样上来研究黄土的震陷性。根据试验的结果，对四个研究区的黄土震陷量进行了估算，评价了场地震陷性。结果表明：在地震超越概率2％作用下华阴、潼关场地可产生严重震陷，灵宝场地可产生轻微震陷；在地震超越概率10％作用下华阴、潼关、灵宝场地可产生轻微震陷；在以上情况下，偃师场地无震陷性。(5)将模糊信息优化处理技术与神经网络相结合提出了一种模糊神经网络模型，并将其应用于震陷的预测中。这种方法，弥补了神经网络在数据处理方面的不足以及信息扩散在计算方面的缺陷。通过实例验证，这种方法在预测震陷系数上，优于多元回归方法。(6)利用扫描电镜、GIS、分形理论等对黄土震陷前后黄土微结构变化进行分析发现黄土的震陷性主要是由弱胶结粒间孔隙、架空孔隙微结构的变化所造成的，分形理论结果能够较好的反映震陷前后黄土孔隙的变化规律。(7)分析了黄土震陷的影响因素，将黄土结构简化为微观的弹塑性模型，进而对黄土震陷的机理进行了初步探讨。(8)利用FLAC软件建立了潼关、华阴等四个区域的黄土路基弹塑性动力分析模型，输入人工地震波，对黄土地基沉降进行了动力分析，得到了不同地震荷载下的路基黄土震陷量。同时还建立了不同高度路堤的路基分析模型，计算了静沉降以及机车振动荷载下动力沉降。(9)提出了一种有限元与神经网络联合的计算方法，用于计算长持时机车振动荷载下路基的动力沉降。(10)针对郑-西客运专线路基黄土产生的震陷特点，提出强夯、化学灌浆、挤密桩等消除其影响的地基处理措施。更多还原

【Abstract】 Zhengzhou-Xi’an passenger express railway is the first high-speed (350 km/h at best ) railway construction in loess area in China. Compared with common railway of ballast track,it is required to paved by ballastless track on the whole line and has stricter control of roadbed deformation. As the railway pass through high earthquake intensity region with huge thick loess, loess seismic subsidence should be considered in roadbed seismic design, in addition dynamic loess additional settlement caused by the long term vibration of the train should also be considered. Based on the present research situation about seismic subsidence and dynamic additional settlement of soil at home and abroad, in addition to comprehensive analysis of characteristics of environmental geology and loess property along the line, the research project on loess seismic subsidence and dynamic loess additional settlement caused by the long term vibration of the train are carried out, which is supported by China-MOR(Ministry Of Railway).Based on the analysis of environmental geology and geological condition along the railway line, loess dynamic and static experiments are carried out. Through field investigation as well as field and lab test, semiquantitative and quantitative method, linear and nonlinear analysis, simulation and computation loess seismic subsidence, dynamic loess additional settlement caused by the long term vibration of the train as well as their correlations are studied. The research work has accomplished with the following achievements: 1.By collecting environmental and engineering geological materials along the passenger express railway line and field investigation; engineering geological characteristics of the passenger express railway have been comprehensively understood. And by means of routine physics and mechanics tests, a series of loess physical and static parameters have been obtained.2.Field shear wave velocity test of loess along the passenger express railway line have been carried out. The test results shows that as depth rises, shear wave velocity increases in exponential trend. According to result of shear wave velocity and lab seismic subsidence experiment of loess, the regression model has been established through which seismic subsidence of loess can be predicted by shear wave velocity.3.By imposing long time-sustained sinusoid loading on natural loess specimens in dynamic triaxial test, characteristic of loess residual strain(dynamic subsidence) has been studied. Based on the results, the settlement of the embankment in different height under long time-sustained vibration of train has been computed. The result shows that dynamic loess additional settlement has negative relationship with the height of embankment and positive relationship with frequency of vibration. At the same time, predicting model of dynamic additional settlement of loess has been proposed.4.Based on seismic hazard analysis results, seismic wave has been artificially synthesized in typical loess field (Huayin Tongguan Lingbao Yanshi). By imposing synthesized seismic wave on natural loess specimens along the passenger express railway line in dynamic triaxial test, seismic subsidence is studied. According to test results, seismic settlement quantity of loess in four research area has been evaluated. Which indicates that severe seismic subsidence may appear in Huayin as well as Tongguan and slim seismic subsidence takes place in Lingbao under seism exceedance probability of 2%; slim seismic subsidence may appear in Huayin Tongguan and Lingbao under seism exceedance probability of 10%; under both probability above, there is no seismic subsidence in Yanshi. 5. A fuzzy neural network model is proposed through the combination of fuzzy information optimization technology and neural network, with which loess seismic subsidence can be predicted. The method overcomes the shortage in data processing of neural network and disadvantage of computation of information diffusion. By examples, this method is superiorer to multiple regression analysis in prediction of loess seismic subsidence.6.By SEM GIS and Fractal Theory, Loess micro-structure of pre-and-post seismic subsidence have been studied, the study shows that subsidence is mainly induced by the changes of weakly cemented void and space void. Therefore, change principle of the loess porosity of pre-and-post seismic can be described better by Fractal Theory.7.Factors influencing loess seismic subsidence have been analyzed; through simplifying loess structure to micro-elastic model, loess dynamic subsidence mechanism has been deeply discussed.8.Loess dynamic elastic-plastic models of Tongguan and Huayin area etc are established by FLAC software, and through inputting artificial seismic wave data, loess roadbed settlement have been dynamically analyzed. Through numerical simulation loess seismic subsidence quantity under different seismic loading are obtained. At the same time, the model for roadbed in different height is established through which static subsidence settlement as well as dynamic subsidence settlement under the train vibration can be calculated.9. A new method combining finite element analysis with artificial neural networks is proposed, which is used for calculating dynamic subsidence settlement induced by long term vibration of train.10. According to characteristics of loess seismic subsidence along the railway, measures of dynamic tamping, chemical grouting, compaction pile etc are proposed to improve the roadbed properties.更多还原