Collaborative Research: Multi-scale Modeling and Measurement of Clay Aggregate Behavior

本研究的目的是促进对粘土作为工程材料的力学行为的理解。粘土由非常细小（2微米）、带电、具有化学活性的颗粒组成，这些颗粒的行为和组装受到在较小（即分子和界面）尺度上作用于单个颗粒的物理化学力的强烈影响。不同种类的粘土矿物广泛存在于自然地质环境中，通常以聚集体或团簇的形式存在，它们的相互作用控制着基础设施设计和建设中使用的富含粘土的宏观工程性质。虽然基于微观力学的实验和计算研究已经大大提高了对砂（粒径大于75um）行为的理解，但粘土颗粒性质的重要性尚未得到广泛研究。该合作项目旨在开发和验证一个新的多尺度框架，通过研究基本粘土颗粒的微观聚集以及由此产生的粘土集合体之间的相互作用，来理解粘土的宏观（即连续尺度）力学特性。该项目将为开发下一代粘土力学特性本构模型提供基础知识，这些模型可用于地下工程，最终降低与地基设计、地下建筑和地质能源（如粘土页岩等）使用相关的风险和成本。通过创建与粘土行为的小规模测量和多尺度建模相关的在线模块，该研究将直接影响未来岩土工程和地质力学学生的教育。研究工作包括对一种常见粘土矿物伊利石进行紧密集成的多尺度实验、原子和粗粒度多尺度模拟。主要工作包括：1)模拟和实验验证不同孔隙水化学条件下单个粘土集料的行为；2)开发新的建模技术和实验方法，用于研究聚集体-聚集体相互作用和结构（即粒子取向和取向分布函数的定量测量）；3)通过与粘土性质宏观测量结果的比较，验证了多尺度模型的预测结果。实验测量将使用马萨诸塞大学阿默斯特分校（UMass）现有的实验室设备以及阿贡国家实验室的分析设备，而麻省理工学院（MIT）的数值模拟将利用美国国家科学基金会（NSF）极端科学与工程发现环境（XSEDE）国家超级计算资源。通过建立一个理解基于颗粒的粘土微观力学的框架，该研究旨在提供一个创新的多尺度视角来解释基于连续体的粘土特性的潜在基础，如黏聚和蠕变，这些特性迄今为止只能通过宏观方法从现象上观察到。长期目标是推广该方法，使其能够自下而上地预测复杂自然土壤的力学特性，这些土壤包括不同粘土矿物与粉砂大小的颗粒的混合物，以及不同的孔隙水化学。

The goal of this research is to advance the understanding of the mechanical behavior of clays as engineering materials. Clays comprise very fine ( 2 um), electrically charged, chemically active particles whose behavior and assembly are strongly affected by physico-chemical forces operating on individual particles at the smaller (i.e., molecular and interfacial) scales. Clay minerals of different varieties occur widely in natural geological environments and are generally found in the form of aggregates or clusters whose interactions control the macroscopic engineering properties of clay-rich soils used in the design and construction of infrastructure. While micromechanics-based experimental and computational research has advanced significantly the understanding of the behavior of sands (particle sizes greater than 75um), the importance of the particulate nature of clays has not been extensively investigated. This collaborative project aims to develop and validate a new multiscale framework for understanding the macroscopic (i.e., continuum scale) mechanical properties of clays by studying the microscale aggregation of elementary clay particles and the interactions between the resultant clay aggregates. The project will provide the fundamental understanding needed to develop the next generation of constitutive models for mechanical properties of clays that can be used for subsurface engineering, ultimately reducing risks and costs associated with the design of foundations, underground construction and use of geological energy resources (e.g., clay shales, etc.). The research will impact directly upon the education of future geotechnical engineering and geomechanics students through the creation of online modules related to the small-scale measurement and multiscale modeling of clay behavior. The research work comprises a closely-integrated program of multiscale experimentation, atomistic and coarse-grained multiscale simulations for one common clay mineral, illite. The research involves the following main tasks: 1) modeling and experimental validation of individual clay aggregate behavior under a range of porewater chemistry conditions; 2) development of new modeling techniques and experimental methods for investigating aggregate-aggregate interactions and fabric (i.e., quantitative measurement of particle orientation and of orientation distribution function); and 3) validation of multiscale model predictions through comparison with macroscopic measurements of clay properties. Experimental measurements will use existing laboratory facilities at the University of Massachusetts Amherst (UMass) as well as the analytical facilities at the Argonne National Laboratory, while numerical simulations at the Massachusetts Institute of Technology (MIT) will take advantages of NSF eXtreme Science and Engineering Discovery Environment (XSEDE) national supercomputing resources. By developing a framework for understanding particulate-based clay micromechanics, the research aims to provide an innovative multiscale perspective for explaining the underlying basis of continuum-based clay properties such as cohesion and creep, which have to date only been observed phenomenologically via macroscopic approaches. The long-term goal is to generalize the methodology to enable bottom-up prediction of mechanical properties for complex natural soils that comprise mixtures of different clay minerals with silt- and sand-sized particles as well as varying porewater chemistry.