【作者】 赵刚；

【导师】 陶夏新；

【作者基本信息】 哈尔滨工业大学 ， 岩土工程， 2010， 博士

【摘要】 我国绝大多数寒冷地区属季节性冻土区,春季路堑边坡浅层滑塌失稳是一种多见的工程病害,一直受到国内外寒冷地区公路工程、岩土工程界的关注。入冬,大地封冻由表及里,土中水分向地表迁移;春季融化亦由表及里,表层融水不能渗入下部未融化的土层中,使上部土体含水量增加、强度减小,使原来的均质边坡成为非均质的边坡;有时冻融就导致浅层滑塌,显然地下水的活动起了关键作用。本文紧紧抓住冻融过程中地下水分的迁移这个核心问题,以工程应用为目的,土样试验和大型模型试验相结合,深入研究,建立冻融过程土中水分迁移模型,与数值分析结合,探讨路堑边坡冻融失稳的机理。研究先从各方面条件均能够很好控制的土样试验入手,针对路堑边坡含水量低、地下水埋藏深的特点,采取封闭系统,底部设隔水界面,单向冻结。在土样侧面布置温度传感器,通过顶板的温度控制冻结的速率,实时观测温度沿深度的变化,控制冻结、融化的进程,测试不同深度土中的含水量。原状土试验模拟天然状态,重塑土试验便于区分冻结、融化两个过程的作用,研究土中水分在冻融过程迁移的规律。试验中同步记录的各深度处的温度,测定冻、融过程中含水量随深度的变化的数据,与初始含水量比较,为进一步建立冻融过程中水分迁移模型奠定了基础。试验中,选取相同含水量的土样,设定了5种温度模式,测定不同温度梯度下土中水分的迁移量;按五种设定的含水量调制土样,在相同温度梯度条件下,对不同初始含水量的土样,分别测定土中水分的迁移量。对含水量均匀的土样,分别测试完全冻结和冻结后从顶面融化到设定位置的含水量分布,研究融化到设定位置后,冻结面上水分聚集层的位置及其含水量的变化情况。试验结果表明,冻结并融化到设定位置后,土样未融化部分出现横向和竖向的裂缝、明显的冻融分界线;控制不同顶板温度,不同含水量的试样中各点处温度的变化趋势基本相同;不同初始含水量,冻融后含水量的分布曲线的趋势是相同的,上部土层的含水量增加、下部土层的含水量减少;顶板控温相同,初始含水量大的土样,冻结要消耗的潜热量大,冻结速度慢,迁移的水分多,水分集聚亦多;不同冻结温度和融化温度对试样中水分的迁移影响非常明显,顶板冻结温度越高,水分积聚越明显。重塑的粉质粘土含水量均匀的土样,自上而下冻结后出现含冰量不均匀现象,有的形成冰夹层;冻结后自上而下融化到设定位置后,在冻融交界面之上水分聚集,出现一个含水量最大的薄层。针对现有的水分迁移机制仍然存在的疑问,冻融过程中水分迁移模型偏于复杂,有的参数值不能直接通过试验获得,对土体工程特性考虑不足,一些建模尚缺乏实验论证,应用的结果往往不够理想的局面,建模不追求对多孔多相介质带相变的固、液、汽、热耦合的统一研究,暂不考虑水分迁移对温度场的反馈影响,强调水气扩散在冻结过程中迁移的重要性,从工程角度出发,宏观上把握水分迁移量与温度梯度的关系,分别考虑冻结、融化过程中水分迁移的机理,借助Fick和Darcy定律进一步改进建立模型。(1)假定非饱和土中水分在冻结过程中主要以水汽扩散形式的迁移,在此过程中液态水通过转化为水汽参加迁移,未具体考虑转化过程。推导、计算中,设定了冻结锋面以下土层中水汽的浓度方程,建立了浓度-深度随冻结锋面发展的移动坐标系。为考虑温度梯度的影响,在Fick定律公式的右边乘上温度梯度,得到了修正的非饱和土的Fick定律,作为冻结过程的水分迁移模型,发展了计算方法和程序。本文采用一个综合的扩散项,而不是另增加一个由温度梯度引起的扩散项。从物理上看,表达了温度梯度直接影响扩散率,而不是在原来的水气扩散量的基础上再增加一部分;从数学上看,少了一个参数,便于参数的估计。(2)假定融化过程中,融土中水分迁移主要是在重力作用下的竖向渗流。用增加温度梯度引起的水分迁移项来修正非饱和土达西定律,作为融化过程中的水分迁移模型。接着,以试验得到的冻结、融化后含水量沿试样深度的分布曲线为目标函数,以本文所建立的冻、融中水分迁移模型为正演模块,借助遗传算法反演了上述两个模型的参数。分别把反演结果及其平均值代入模型中计算含水量沿土样深度的分布曲线与试验得到的各个曲线对比,论证了模型可靠性及本文思路的可行性。本文将较大比例尺的室内边坡冻融失稳模型试验设计为土样试验和现场原位观测之间的一个既有一定体积、又可以很好控制各种边界条件的中间环节,弥补土样尺寸小受边界影响大的缺陷。在归纳国内外类似试验的方案、确定模型相似律的基础上,设计、制作了60厘米高的边坡模型以及温度控制条件,在表层密集布设温度、含水量观测传感器,每隔5分钟采集一组数据。试验采用了根据电磁脉冲反射方法研制的时域反射仪TDR,对量值随温度的变化做了标定。经过35天,历时840小时,5个冻融循环,试验显示了水分迁移、冻裂、分凝冰层生成、坡面局部滑塌等现象。通过记录的数据分析,说明温度场控制得比较好,含水量变化比较复杂。对第一个冻融循环的仔细分析,说明了冻结中水分扩散向上迁移,融化过程中向下渗流的规律。依据试验结果阐述了坡面滑塌的主要原因是土层含水量提高引起抗剪强度的降低。为了研究冻融对土的抗剪强度的影响,定量地揭示含水量提高所引起土层抗剪强度降低的程度,通过制备不同含水量的土样,经过冻融过程,采用三轴不排水剪切试验,测得土的抗剪强度参数。结果表明抗剪强度指标确实下降很大,验证了本文试验中边坡发生滑塌主要是由于浅层土体含水量增加很多,导致土的抗剪强度大幅下降是重要因素的结论。最后,本文对试验的边坡模型,进行了冻融稳定性数值分析。在有限元分析软件ABAQUS的基础上二次开发,建立了冻融边坡模型的有限元模型,设置了边界条件和单元参数。在垂直坡面3.0米的深度内,按0.3米划分网格为冻融网格,以下为非冻融网格,借助ABAQUS/CAE生成inp形式输入文件。根据模型试验中观测到的温度场,冻结和融化过程中水分的迁移,含水量和三轴试验结果统计关系修正冻融单元的抗剪强度参数,计算了冻结后融化到5个不同深度时模型的变形和内力,结果显现了浅层滑动的可能性。更多还原

【Abstract】 Most of the cold regions in China are seasonal frozen regions where shallow slide of highway cutting slope is a popular phenomenon in spring. Attentions have been paid to the issue in the areas of road engineering and geotechnical engineering in cold region for many years. In winter, the soil moisture migrates upwards as the ground frozen from the surface to the deep. In spring, the ground thawing starts from the surface to the deep, the water at the shallow layer permeates downwards by gravity and cannot infiltrate into the frozen soil beneath, and then it accumulates in a thin soil layer just above the interface. This results in the increase of the water content and the decrease of the shear strength of the soil layer, and making the original homogeneous slope become inhomogeneous, sometimes which makes the shallow slide. Obviously, the water migration in soil during freeze and thaw process plays a key role.The soil moisture migration in the freeze-thaw process is taken as the core point in this dissertation; Combining with soil sample and large scale model experiments, two mathematical models are built up separately for the migrations in freeze and thaw processes for engineering application purpose. The instability mechanism of highway cutting slope during freeze-thaw process is studied with also numerical analysis.The study begins with soil sample experiment, which conditions are well controlled. A close system with watertight boundary at the bottom of the soil sample is adopted, the samples are frozen and thawed in one-way to the practical situation of low water content and deep ground water table in all cutting slope. The freezing rate is handled through the temperature of the top plate, temperature variation with depth is observed regularly time to time from the thermal sensors set at a side of the soil sample, and then the frozen or thaw depth is controlled. Undisturbed soil test is to simulate the natural state, while reshaped soil test is to distinguish the effects of freezing and thawing, the both are exploring the soil moisture migration rules in the two processes.Temperatures at different depths were measured and recorded simultaneously during the test, and the distribution of water content with depth were measured after the sample frozen or thawed at a given depth, and compared with the original water content. All the data acquired in the tests are building a firm base for the mathematic models.In the experiments, 5 temperature conditions for samples with the same initial water content were adopted to measure the moisture migration amount under different temperature gradients; while 5 reshaped samples with different given initial water contents were adopted under the same temperature condition also to measure the water migration amounts. The water content distributions were measured when the samples completely frozen or thawed from the top to a given depth afterwards from the initial uniform distribution. The position of the soil layer with the richest water content above the freeze-thaw interface and the variation of the water content at that depth were studied for the thawed cases.The test results show that some transverse and longitudinal cracks appear in unthawed portion after the sample completely frozen and then thawed to the given depth, a distinguish freeze-thaw interface is also observed; the temperatures at various depths show the similar trend with different top temperatures even with different water contents; the water content curves after freeze-thaw are the similar on the whole, i.e. the water content increases in the upper soil layer and decreases in the lower layer; the more initial water content, the more potential heat consumption in freezing, the slower freezing rate, the more moisture migrated and the more water aggregated above the interface; the effect of temperature condition on moisture migration is quite obvious, the higher freezing temperature at top plate, the more water aggregates above the interface; the inhomogeneous ice content, even segregated ice, appear in reshaped silt-clay samples with initial homogeneous water content after frozen from top to bottom; a thin soil layer with the richest water content appears just above the interface after the sample frozen completely from top and then thawed to the given depth.There are still queries now to the existing water migration mechanism theories: the moisture migration models are too complicated for engineering practical purpose, values of some parameters in the models could not be obtained from experiment directly, engineering characteristics of soil mass is not taken into account enough, almost of the models are lack of validation, application results of these models are not satisfactory in some cases. From this state of the art, the modeling is not in pursuit of the comprehensive description of real solid, liquid and gas phase coupled with heat in porous media, and the feedback effect of moisture migration on temperature field is not taken into account temporarily in this paper. The importance of moisture diffusion in water migration during freezing is emphasized, and the relationship between water migration amount and temperature gradient is handled in addition, from the engineering point of view. The water migration mechanisms for two processes of freezing and thawing are considered separately, and two models were worked out by modification of the Fick law and the Darcy law.(1) Assuming the moisture in unsaturated soil migrates in the way of diffusion during freeze process. In this process, liquid water takes part in the migration by transformed into moisture, and the detailed transformation is not considered. The concentration equation of the moisture in the soil beneath the frozen front is presented, and a moving coordinate system is established for the concentration-depth relationship with the freezing front development. In order to take the effect of temperature gradient into account, temperature gradient is multiplied to the right side of the Fick law, thus an improved Fick law for unsaturated soil is obtained as the water migration model of freeze process. A set of calculation procedure and program are developed at the same time.A comprehensive diffusion term is adopted in this paper, instead of adding a diffusion term caused by temperature gradient. From physics point of view, it describes the fact that temperature gradient influent the diffusion rate directly, not adding an additional amount on the original moisture diffusion. From mathematics point of view, eliminating one parameter is convenient to parameters estimation.(2) Assuming the water migration in thawed soil is mainly the vertical permeation from the gravity. An additional term of water permeation caused by the temperature gradient is added to the Darcy law for unsaturated soil, as a water migration model for the thaw process.Furthermore, values of parameters in the two models are inversed by means of the Generic Algorithm, where the objective function is taken as the curve of water content– depth after the frozen and thawed processes respectively obtained from the experiments, and the forward calculation modules is built from the programs for the two models. The water content - depth curves are then calculated from the inversed parameter values and their mean respectively. The resulted curves are compared with those from experiments, thus the reliability of the model and the feasibility of the approach of this paper are validated.A large scale slope model experiment for instability after frozen-thaw is designed as a link between the soil sample experiment and the in situ experiment in this dissertation, with well-controlled boundary conditions and certain volume to eliminate the effect of boundaries in soil sample test. The designs of similar model tests both at home and abroad are summarized, the similarity law for model is determined. A slope model with 60 centimeters high is designed and made in laboratory, and the temperature condition is preset. Thermal sensors and sensors for water content are buried densely in the upper soil layers of the model. The data are acquired every 5 minutes during the experiment. Time domain reflector (TDR) is adopted in the experiment, which is developed from electromagnetic pulse reflection method, and the measured values are scaled to eliminate the shifting of readings with temperature. During 35 days, 840 hours which consists of 5 freeze-thaw cycles, the experiment shows the phenomenon such as water migration, frozen crack, segregated ice, and local slides on the slope, etc. The analysis on the recorded data shows that the temperature field is well-controlled and the water content variation is complex. The detailed data analysis of the first freeze-thaw cycle indicates rules that the moisture moves upwards by diffusion during the freezing, and permeates downwards by gravity during the thawing. From the results, it can be concluded that the main reason of shallow slide on slope is the significant decrease of shear strength in the thin soil layer just above the freeze-thaw interface resulted from the large increase of water content.In order to deal with the effect of freeze– thaw on shear strength of soil, to reveal the degree of decrease of soil shear strength from the increase of the water content quantitatively, soil samples with different water contents are reshaped, and are further frozen and then thawed, the shear strength parameters of the soil are tested by means of triaxial undrained shearing experiment. The result shows that the values of shear strength parameters decease certainly to a large extent, thus the above conclusion is validated that the main reason of shallow slides on the model slope is the large decrease of the shear strength of the shallow soil resulted from the increase of water content there.Finally, numerical analysis of the slope model for stability in the freeze-thaw experiment is carried out. A second development is made by means of ABAQUS software for finite element analysis, a finite element model is constructed for the model in the experiment, and the boundary conditions and element parameters are assigned. In the depth of 3.0 meters vertical from the surface, grids with 0.3 meters distance are freeze-thaw grids whose parameters vary with water content during the test, and others are not. The input data files with“inp”type are generated by ABAQUS/CAE. The shear strengths of freeze– thaw grids are modified from the statistical relationship with water content from the triaxial test mentioned above. The water content data measured in the experiment, the deformations and stresses of elements after frozen and then thawed to 5 depths are calculated respectively. The results show the potential of shallow slide on the slope.更多还原