A Fundamental Study on Suction and Hysteresis of Soil-Water Characteristic Relation of Cohesive Soil

这笔赠款为部分饱和粘性土壤中毛细管效应的基础研究提供资金。具体来说，土壤吸力的大小和分子来源，以及粘土颗粒之间的空气-水毛细管弯月面的曲率和接触角是令人感兴趣的。这些课题的研究将为粘性土中极高吸力(高达100倍大气压)的来源和土-水特性关系滞后的来源等悬而未决的基本问题提供答案。这项研究将借助数值和实验方法进行。在数值方法方面，将使用被称为离散单元法的基于粒子的建模技术和被称为分子动力学的基于原子的建模技术。在实验方法上，将采用压板吸力测试法、傅里叶变换红外光谱分析方法和环境扫描电子显微镜方法。数值研究是计算密集型的。在过去，获得无限和容易获得的计算能力的困难阻碍了这种计算密集型研究。将图形处理单元(GPU)用于计算目的的最新发展使得以负担得起的价格在办公室工作站上拥有万亿次浮点计算能力(基本上是一台并行超级计算机)成为可能。上述最新发展(离散元模拟、分子动力学和基于GPU的计算的进展)为解决这项研究中具有挑战性的基础研究提供了机会。如果研究成功，这项研究将带来设计涉及非饱和细粒土的岩土结构所需的基础科学知识和方法。例如，这项研究将导致控制由暴雨和/或地震引发的部分饱和山坡上发生的山体滑坡的合理方法。

This grant provides funding for fundamental study on the capillary effects in a volume of partially saturated clayey soils. Specifically, the magnitude and molecular sources of soil suction, and curvature and contact angles of air-water capillary meniscus between clay particles are of interest. The studies on these topics will provide answers to outstanding fundamental questions concerning the sources of very high suction in clayey soils (as high as 100 times the atmospheric pressure) and the sources of hysteresis of soil-water characteristic relation. The study will be conducted with the aid of numerical and experimental methods. In regard to the numerical methods, a particle-based modeling technique known as the discrete element method and an atom-based modeling technique known as the molecular dynamics will be used. Concerning the experimental methods, the pressure-plate suction testing method, Fourier Transform Infrared spectroscopic method and Environmental Scanning Electron Microscopic method will be employed. The numerical studies are computationally intensive. The difficulty of obtaining unlimited and readily-available computational power has hindered such computationally intensive studies in the past. The recent developments in the use of Graphic Processing Units (GPUs) for computational purposes make possible having teraflop-computational power (basically a parallel supercomputer) right on an office workstation at an affordable price. The confluence of the above recent developments (the advancements in discrete element modeling, molecular dynamics and GPU-based computing) provides an opportunity for tackling the challenging fundamental research of interest in this study.If successful, the study will lead to fundamental scientific knowledge and methodologies needed for designing geotechnical structures involving unsaturated fine-grained soils. For example, the study will lead to rational methods for controlling landslides occurring on partially saturated mountainous slopes triggered by torrential rain and/or earthquakes.