非饱和花岗岩残积土力学特性与土水特征参数贝叶斯分析

发布时间：2018-06-02 02:49

本文选题：花岗岩残积土 + 非饱和土　；参考：《南昌大学》2017年博士论文

【摘要】：崩岗侵蚀是南方花岗岩地区土壤侵蚀最为严重的类型,而我省的赣南地区是崩岗侵蚀发育重灾区。影响边坡崩岗侵蚀的因素是多方面的,水是最主要的外因,降雨入渗导致非饱和土边坡含水量增加,从而导致土中的基质吸力也相应地发生变化,土体的强度等特性也相应发生改变。赣南崩岗区以花岗岩风化壳残积土为主,该类土的物理力学性质受水的影响显著,同时还经受干湿循环作用,开展该类土的非饱和土特性研究可以为赣南地区崩岗侵蚀机理研究和防治提供理论支持。本文将以花岗岩残积土为研究对象,采用非饱和土三轴仪和压力板仪等实验仪器研究其强度和变形特征以及土水特征曲线的测定,开展基于贝叶斯理论的土水特征曲线预测研究,建立了采用有限数据对土水特征曲线有效地预测的方法,同时进行土水特征曲线模型参数的不确定性分析,获得土水特征曲线的置信区间,并应用到降雨条件下的边坡稳定性分析中。(1)为探讨崩岗区非饱和花岗岩残积土强度和变形特性,利用非饱和土三轴仪对重塑土样进行控制吸力固结排水剪切试验,研究干密度、净围压和基质吸力等因素对应力-应变关系、强度特性和体变特性的影响,分析非饱和花岗岩残积土在控制吸力条件下的剪切性状,研究表明其应力-应变关系取决于净围压和基质吸力两者组合。通过分析基质吸力对非饱和土强度的贡献可知,基质吸力对黏聚力有较大的影响,且呈线性增长趋势,而对有效内摩擦角影响则较小。同样对石灰改良花岗岩残积土进行控制吸力固结排水剪切试验,以研究石灰含量对非饱和改良土的强度和变形特性的影响,试验分别采用掺灰比0%、4%和8%的改良花岗岩残积土,同时还选用了两种颗粒尺寸(1mm,2mm)花岗岩残积土进行了相关试验,石灰改良土的黏聚力和摩擦角随着石灰含量的增加而增大,其强度也相应提高,且黏聚力提高更加明显,而细颗粒土其颗粒尺寸对强度的影响较小。(2) 土水特征曲线用于描述非饱和土中含水量或饱和度与基质吸力之间的关系,在非饱和土力学中非常重要。本文采用GEO-Experts 土水特征曲线压力板仪对崩岗区花岗岩残积土进行大量的土水特征曲线实验,测定了不同初始干密度、不同颗粒级配、不同石灰掺加量以及干湿循环作用下的重塑非饱土崩岗区花岗岩残积土试样的土水特征曲线。研究表明:初始干密度对土水特征曲线有影响明显,土水特征曲线随着干密度的增大明显下移,当基质吸力较大时,有着相互聚拢的趋势;当干密度相同时,颗粒级配小的土样的吸力更大;土水特征曲线也有随着石灰含量的增加而下移的趋势,表明其水稳定性变好;在干湿循环作用下,土水特征曲线的脱湿曲线和吸湿曲线有显著的滞回现象,滞回圈的面积随着干密度的增大和干湿循环次数的增加而减小,掺入石灰后滞回圈的面积也变小。同时还进行了颗粒级配和竖向应力对土水特征曲线的影响研究,通过施加竖向应力以模拟不同颗粒级配的土壤在受力状态,并获得其相应的土水特征曲线。在无竖向应力时,不同颗粒级配土样的土水特征曲线差别不大,但有竖向应力时,不同颗粒级配土样的土水特征曲线差别明显,且有随着竖向应力增加而增大的趋势。在低基质吸力区竖向应力对土水特征曲线的影响更为明显,而在高基质吸力区的影响则相对更小。根据试验获得的数据利用Fredlund Xing模型、Van Genuchten模型和Gardner模型三种土水特征曲线拟合模型对其进行拟合,通过三种模型的拟合结果对比可知,VanGenuchten模型可以很好地描述和预测崩岗侵蚀区非饱和花岗岩残积土的土水特征曲线。(3)考虑到岩土工程参数和计算模型存在不确定性,根据土水特征曲线预测的自身特点,建立基于贝叶斯更新的土水特征曲线预测框架,将利用现有两种土水特征曲线预测模型Fredlund Xing模型和Van Genuchten模型与试验实测信息相融合,采用贝叶期更新理论框架和前期观测数据对理论计算模型中的计算参数更新,再利用更新后的参数对土水特征曲线进行更为精确的预测,通过模型参数的不断更新实现土水特征曲线动态预测。当采用合理的测量误差的标准差时,采用有限的实验数据更新后预测的土水特征曲线精确度较高,通过贝叶斯更新和实验方法结合方法为快速有效地预测土水特征曲线提供了一种析途径。(4) 土水特征曲线拟合模型中存在多个似合参数,各参数的值通过实验方法获得且具有不确定性。本文探索采用贝叶斯理论分析拟合参数的不确定性,将土水特征曲线拟合参数作为随机变量,采用Fredlund Xing模型、Van Genuchten模型和Gardner模型三种模型以及花岗岩残积土的土水特征曲线试验数据,以马尔可夫链蒙特卡罗方法(MCMC方法)的延迟拒绝适应性算法(DRAM算法)获得模型参数后验分布抽样,获得了不同置信区间的土水特征曲线及其对应参数值,该方法可以有效地估计土水特征曲线模型的参数,并获得土水特征曲线的置信区间。(5) 土水特征曲线在边坡的渗流和稳定性分析中具有非常重要的作用,结合土水特征曲线的不确定性,研究了土水特征曲线的置信度对降雨入渗边坡的渗流场和边坡安全系数的影响,获得非饱和土边坡渗流场和稳定性随置信度的变化规律。不同置信度的土水特征曲线对边坡渗流场影响较大,且孔隙水压力随降雨历时变化最为明显,孔隙水压力随降雨历时的增长速率随着土水特征曲线的置信度百分点的增加而增大;随着置信度百分点的增加,降雨条件下的边坡稳定性在相应减小。
[Abstract]:The erosion is the most serious type of soil erosion in the granite area of southern China, and the southern Jiangxi area is the heavy disaster area in the south of Jiangxi Province. The factors affecting the slope erosion are many aspects. Water is the most important external cause. The water content of the unsaturated soil slope is increased by the rainfall infiltration, thus the matrix suction in the soil is also related. The physical and mechanical properties of this kind of soil are greatly influenced by the water, and the characteristics of the unsaturated soil properties of this kind of soil can be provided for the study and prevention of the mechanism of the erosion of the landslides in the south of Jiangxi. In this paper, we take the granite residual soil as the research object, use the experimental instruments such as unsaturated soil three axis instrument and pressure plate instrument to study the characteristics of its strength and deformation and the measurement of soil water characteristic curve, carry out the study of soil water characteristic curve prediction based on Bayesian theory, and establish the effective preview of the soil water characteristic curve by using the finite data. At the same time, the uncertainty analysis of the parameters of the soil water characteristic curve model is carried out, the confidence interval of the soil water characteristic curve is obtained and applied to the slope stability analysis under the rainfall condition. (1) the strength and deformation characteristics of the unsaturated granite residual soil in the landslide area are discussed, and the remolded soil samples are controlled by the unsaturated soil three axis instrument. In the force consolidation drainage shear test, the effects of dry density, net confining pressure and matrix suction on the relationship of stress strain, strength and body variation are studied. The shear properties of unsaturated granite residual soil under the controlled suction condition are analyzed. The study shows that the stress-strain relationship depends on the combination of net confining pressure and matrix suction. The contribution of matrix suction to the strength of unsaturated soil shows that the matrix suction has a great influence on the cohesive force, and it has a linear growth trend, while the effect on the effective internal friction angle is smaller. The same is the control suction drainage shear test on the lime modified granite residual soil to study the strength and change of the lime content to the unsaturated soil. The influence of the shape characteristics is that the improved granite residual soil with ash ratio 0%, 4% and 8% is used respectively. At the same time, two kinds of particle size (1mm, 2mm) granite residual soil have been selected. The cohesive force and friction angle of lime improved soil increased with the increase of lime content, and the strength of the lime soil increased correspondingly, and the cohesiveness increased more clearly. The particle size of fine granular soil has little effect on the strength. (2) the soil water characteristic curve is used to describe the relationship between water content or saturation of unsaturated soil and matrix suction, which is very important in the mechanics of unsaturated soil. This paper uses the GEO-Experts soil water characteristic curve pressure plate instrument to carry out a large amount of soil in the granite residual soil in the landslide granite area. The water characteristic curves of different initial dry density, different grain gradation, different lime content and dry and wet cycle were used to remould the soil water characteristic curve of the granite residual soil specimen under the action of dry and wet cycle. The study shows that the initial dry density has obvious influence on the soil water characteristic curve, and the soil water characteristic curve is increasing with the dry density. When Daming is moved down, there is a tendency to gather together when the suction of the matrix is large; when the dry density is the same, the soil sample with small particle size is more suction; the soil water characteristic curve also moves down with the increase of lime content, indicating that the water stability becomes better, and the dehumidification curve and hygroscopic curve of the soil water characteristic curve under the action of dry and wet circulation. With the increase of dry density and the increase of dry and wet cycle times, the area of hysteresis loop decreases with the increase of dry density and the number of dry and wet cycles, and the effect of grain gradation and vertical stress on the soil water characteristic curve is also studied. The soil of different grain gradation is simulated by applying vertical stress. When there is no vertical stress, the soil water characteristic curves of different particle graded soil samples have little difference, but when there is vertical stress, the soil water characteristic curves of different grain grading soil samples are distinct, and the vertical stress increases with the vertical stress. The vertical stress in the low matrix suction area is the same. The influence of the soil water characteristic curve is more obvious, but the influence in the high matrix suction area is relatively smaller. According to the data obtained by the experiment, the Fredlund Xing model, the Van Genuchten model and the Gardner model are fitted to the fitting model of three soil water characteristic curves. The VanGenuchten model can be compared through the comparison of the fitting results of the model type. In order to describe and predict the soil water characteristic curve of the unsaturated granite residual soil in the avalanche erosion area. (3) considering the uncertainty of the geotechnical engineering parameters and the calculation model, according to the characteristics of the soil water characteristic curve predicted, the soil water characteristic curve prediction frame based on the Bayesian updating will be established, and the existing two kinds of soil water characteristics will be used. The line prediction model Fredlund Xing model and the Van Genuchten model are fused with the experimental data. The theoretical framework and the early observation data are used to update the calculation parameters in the theoretical calculation model, and the soil water characteristic curves are more accurately predicted by the updated parameters, and the model parameters are constantly updated. The dynamic prediction of soil water characteristic curve is realized. When the standard deviation of reasonable measurement error is adopted, the accuracy of the soil water characteristic curve predicted by the finite experimental data is higher, and a way of fast and effective prediction of soil water characteristic curve is provided by the combination of Bayesian updating and experimental method. (4) soil water characteristic curve There are multiple coincidence parameters in the fitting model, and the values of each parameter are obtained by the experimental method and have uncertainty. In this paper, the uncertainty of the fitting parameters is analyzed by the Bayesian theory, and the parameters of the soil water characteristic curve are taken as random variables, and the three models of Fredlund Xing model, Van Genuchten model and Gardner model are used. And the soil water characteristic curve test data of the granite residual soil, using the delayed rejection adaptive algorithm (DRAM algorithm) of the Markov Monte Carlo method (MCMC method) to obtain the model parameter posterior distribution sampling, obtained the soil water characteristic curve and its corresponding parameter values of different confidence intervals. This method can effectively estimate the soil water characteristic curve. The parameters of the line model and the confidence interval of the soil water characteristic curve are obtained. (5) the soil water characteristic curve plays a very important role in the seepage and stability analysis of the slope. Combining the uncertainty of the soil water characteristic curve, the influence of the confidence degree of the soil water characteristic curve on the seepage field and the slope safety factor of the rainfall infiltration slope is studied. The soil water characteristic curve of different reliability has great influence on the seepage field of the slope, and the pore water pressure is most obvious with the rainfall duration, and the growth rate of pore water pressure increases with the increase of the confidence degree of the soil water characteristic curve. With the increase of confidence level, the slope stability under rainfall condition decreases correspondingly.
【学位授予单位】：南昌大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TU43

【参考文献】