【作者】 张静；

【导师】 王清；

【作者基本信息】 吉林大学 ， 地质工程， 2010， 博士

【摘要】 本文以吉林省西部地区分散性冻土为研究对象,结合国家自然科学基金项目“东北季冻区路基土水分迁移的微观机理研究”,通过对不同季节、不同深度的土样的室内常规物理试验、化学试验、微观结构测试及分散性鉴定试验研究,分析了土的物化性质和分散性在土体剖面上的分布特征及其随季节的变化。另外对土样进行了室内水分迁移模拟试验,通过对冻融前后土的含水率、含盐量、微观结构、分散性的对比分析,研究了多个冻融循环过程中的水分迁移特征及迁移的机理,盐分随水分的运移特征及微观结构、分散性的变化,总结出了冻融循环过程中的盐分迁移和微结构变化对分散性的影响机理。在以上试验数据的基础上,利用灰色关联分析法,对土的物理化学参数与分散性参数—分散度之间的关联度进行了计算,探讨了物理化学性质对分散性的影响大小,确定了影响分散性的主要参数,研究了吉林省西部地区分散性冻土的分散机理,并结合当地的地形地貌、气候、水文特征,分析其成因。最后,采用分散度和已确定的对分散性最具影响意义的5个参数:交换性钠百分比、阳离子交换容量、pH值、有机质含量、粘粒含量,建立了BP神经网络分散性模型,并通过4组土样的试验值和模型计算值的对比对模型进行了验证,对吉林省西部地区水利工程建设和土质改良具有一定的理论指导意义。更多还原

【Abstract】 The dispersive soil refers to a kind of cohesive soil, which has a tendency to disperse in low-salinity water(or pure water)due to its more mutual repulsion of ions than mutual attraction. Dispersive soil will disperse into the original particle in low-salinity water and it is susceptible to be carried away by flowing water. The erosion phenomenon is serious than fine sand or silt, which is the important reason for erosion and piping failure of dams, channel slope, roads and slope. The western region of Jilin Province lies in the south of Songnen Plain, and it is one of the concentrated distribution area of soda saline soil. To improve grain production capacity, three large water conservancies of inducing Nenjiang to Baicheng, Da’an irrigation and Hadashan hydro project have been built to irrigate and treat alkaline wasteland. However, there have emerged serious erosion damage and collapse phenomenon in the project construction process. Through the investigation and experiments found that the soil in this area belong to dispersive soil, and rich in sodium ions which increases the thickness of particle surface diffusion layer, so that soil particles dispersed in water. Therefore, researching on the dispersive soil, mastering its dispersion characteristics have some theoretical significance to the water conservancies construction and soil improvement in the western Jilin Province.Associated with the project of“study on the microscopic mechanism of moisture migration in roadbed of seasonal frozen area in northeast China”sponsored by the national natural scientific fund, the basic physical and chemical properties, microstructure and dispersion of dispersive seasonal frozen soil of different seasons and different depths were studied firstly. Then analyze the distribution characteristics of physicochemical properties and dispersion in the soil profile and their changes with seasons. In addition, used freeze-thaw method to simulate water migration, and through comparing the water content, salt content, microstructure, dispersion of soil before and after freezing, researched on the water and salt migration characteristics and the changes of microstructure and dispersion in the freeze-thaw process, summarized the impact of salt migration and the change of microstructure on the dispersion in process of freezing and thawing. Based on the above test data, adopted gray relational analysis method to calculate the correlation between physicochemical parameters and the dispersion parameter-the dispersive percent, determine the impact degree of physical and chemical parameters on the dispersion and the main parameters which impact dispersion, discussed the dispersion mechanism of dispersive seasonal frozen soil in the western region of Jilin Province, combined with the local topography, climate, hydrological characteristics, analyzed the causes of dispersive soil. Finally, used the dispersion percent and five parameters that impact on the dispersion significantly (exchangeable sodium percentage, cation exchange capacity, pH, organic matter content, clay content) to establish the BP neural network model of dispersion, and validated the model by the comparing the calculated values to experimental value of 4 groups of soil samples. Through the study, the following conclusions were obtained:1. It makes a test on the properties such as granulometric composition, mineral composition, soluble salts, cation exchange capacity and micro-structure of soil samples from the western of Jilin Province, which shows that: it is dominated by silt in western of Jilin Province and clay content is higher. As the dispersion of the soil is strong, silt and clay content has not changed much with and without the dispersant in analytic tests of a particle deposition of still water, so the role of dispersant is not obvious. Mineral composition dominated by primary minerals and clay mineral include Eamon mixed-layer minerals, illite and kaolinite, and none montmorillonite. There is much more capillary pore at the main than coarse pores. The percentage of micro-porosity (<0.001mm) content is more as well. The total amount of soluble salt in the soil is greater than 0.3%, and soluble salt is mainly dominted by Na+and HCO3?. The pH value is greater than 7, and is alkaline. So the soil in this area belongs to weak carbonic acid (soda alkaline saline soil). This area is part of continental monsoon climate, so salty content of soil in the variation with depth is closely related with the seasonal changes. It is always high temperature and rainfall in summer, due to rainfall leaching the salt content in surface soil (generally 30cm or less) increased with depth and then decreases as the depth increases;For dry and windy in autumn, salt influenced by eaporation with will migrate to the surface with water, resulting in the decrease of salt content with the increase of depth; In winter, because of the exist of soil temperature gradient, moisture will migrate from unfrozen soil to the permafrost and salt also will move up; Permafrost gradually melt and moisture evaporate strongly until spring , which make salt remained in the surface soil. So their contents should be larger than any other seasons.2. Through the dispersion identification tests of soil in different parts found that the soils in Xianghai and Xinglin are non-dispersed soil, and the soils in zhenlai demonstration area, Sanzhigou, Da’an, Nilin and Momoge are dispersive soil, while their dispersion distribution in the soil profile are the same as the soluble salts change with depth. In spring and autumn, due to the strongly evaporation effect, salt content of the soil at surface is high, so that dispersion decreases with the depth increases. As the salt content of soil in spring is more than that in autumn, the dispersion of soil in spring is also higher than that in autumn. In the summer, due to leaching effects, the dispersion in the range of 0-30cm depth increase with the depth increase, and decrease with the increase of depth below 30cm. This also explains the soluble sodium ions in the soil of this area are the main reasons for the dispersion.3. The test results of capillary water rise in direct method show that there is almost no capillary water arising of Da’an and Sanzhigou soil samples, that is mainly due to the less coarse pores, high content of micro-void in the soil and the large amounts of exchangeable sodium ions of soil which form thick combination water film around the soil particles, and the water combination film overlap expand to full pore the last, block the channels of capillary water arising, result in very low arising height of capillary water. Furthermore, due to the strong dispersion of two sets of soil samples, the soil particle surface energy is quite high, and most of the water strongly bound by particles, so that capillary resistance is considerably, which making it difficult for water migration.The rule of moisture changing with height is not obvious in the water migration test of freeze-thaw method. There is only small amount of increase with height at the temperature changing boundary, which is the film water migration caused by temperature gradient, not the rise of capillary water. Part of the salt crystallize in freezing process resulting in solute potential gradient between the surface and certain depths below it, so the salt in soil layers below migrate to the frozen layer with apparently. The salt content increases with the soil column height. This is the reason for the significant salinity increase of soda-alkali soil at surface in process of freezing and thawing. The dispersion of soil samples increase significantly after freezing and thawing, partly because the sodium ions migrate with water in the freezing process causing the content increase of exchangeable sodium ion surrounding soil particles, increasing the thickness of diffusion layer, reducing the attractiveness of soil particles; on the other hand, the frost heaving generated in the freezing process increase the distance between the particles and undermine the link force between the particles, thereby increase the dispersion of soil.4. The analysis on the relationship between physicochemical parameters and dispersion by gray relational analysis method indicate that the order of correlation between seven parameters and the dispersion percent is: cation exchangeable capacity> exchange sodium ion percentage> organic matter content > clay content> pH> specific surface area> structural damage rate. The linear fitting of the physicochemical parameters and the dispersion found that exchangeable sodium ion percentage, cation exchange capacity, clay content, organic matter content and pH are significantly linear positive correlation with the dispersion percent, and the correlation between specific surface area, structural damage rate and dispersion percent is not obvious. Therefore, the parameters that impact on the dispersion greatest are the exchangeable sodium ion percentage, cation exchange capacity and organic matter content, followed by are pH and clay content, while the specific surface area and structural damage rate have little effect on the dispersion.5. Based on the above analysis, the dispersion mechanism of dispersive soil in the western region of Jilin Province which obtained the following aspects:①montmorillonite unit cell is composed of two silicon-oxygen tetrahedron chip with aluminum oxygen octahedral. And the two adjacent cell layers of oxygen connected to the connectivity are very weak, there is a good cleavage. The water molecules and exchangeable cations can enter each filmso that crystals have a greater volume change with strong expansion. Montmorillonite which is almost no connectivity after absorbing water can be dispersed into some very small scale-like particles. The area of soil contains a certain amount of illite and mixed layer minerals imon. And the groundwater in the area is rich in sodium. So the illite crystal replaces with the role of adsorption of sodium ion, which has a strong dispersion.②Large amounts of exchangeable sodium ions in the soil which is increasing the surface of soil particles double layer thickness, increasing the distance between particles and weakening the link between the particle force, that is why the soil in this region has dispersion.③The pH of the soil is greater than 7, or even greater than 9 which is alkaline .Because the pore solution pH, and illite, montmorillonite presence other large power difference between the pH value which can form a thick diffusion layer. Thus the soil losses the cohesion which caused by particles scattered and high pH value of soil in the study area provided the dispersion of environmental conditions.④While the region contains a certain amount of soil organic matter and clay with cementation playing a certain role, this effect is not sufficient to inhibit the spread of soil.⑤The area is frost area. Salt (mainly sodium bicarbonate) with the upward migration of water in the process of freezing and thawing, so exchangeable sodium in the soil ions is increasing which reinforcing the dispersion of the soil; In addition, frost effect can increase the distance between the particles, reduce the link between the particle. Then the soil is looser and more susceptible to erosion damaged.6. Choose the dispersion of soil parameters - dispersion and 5 affect the dispersion of the important parameters to quantify the extent, and establish three layer BP neural network model. The parameters contain the percentage of exchangeable sodium, cation exchange capacity, pH, organic matter content and clay content. With 30 groups of soil samples as training samples in the western region of Jilin Province, the BP neural network model is trained, and then use another 4 groups of soil samples to verify the model. The test results from the model, and the error between the actual value and model output is relatively small. The relationship is more accurate by establishing the model which is between the exchangeable sodium percentage, cation exchange capacity, pH, organic matter content, clay content and the dispersion. Also show that five parameters are the root cause of the dispersion, you can use this model predicts the dispersion of soil.更多还原