Development and validation of particle models for DEM simulations of the dynamical behavior of saturated soils

为了更好地理解土壤液化等自然现象，土壤模型起到了至关重要的作用。土壤液化是指饱和水的沉积物的结构在地震等动力荷载作用下失去强度的过程，可造成灾难性后果：建筑物、围护结构或堤坝倒塌、快速粘土区的滑坡(Saint-Jean-Vianney，魁北克，1971)。因此，由于魁北克省有大量的水电大坝，因此它是魁北克省主要关注的问题。然而，目前最先进的技术还不能正确地预测土壤液化。该研究项目的总体目标是开发和验证用于描述沙粒之间相互作用的颗粒尺度的改进模型，将有来自蒙特利尔理工学院和魁北克水电研究所(IREQ)的合作者参与。更具体地说，它的目标是改进和验证颗粒的椭球模型，提供颗粒的高分辨率自适应表示，并开发和验证描述任意形状颗粒的随机接触模型。新模型将在高性能计算机平台(SiGran)上实现和测试，用于在颗粒和孔隙尺度上模拟颗粒介质与流体在静态和动态加载下的相互作用。拟议的研究计划将涉及三名博士生，两名理科硕士。学生和几名本科生，他们将接触到当代计算工程和科学的挑战，并接受数值分析、科学计算、数学建模、概率、贝叶斯推理和模型验证方面的培训。为了更好地了解液化过程，开发先进的建模和模拟工具，将为加拿大和世界各地在土壤沉积物附近建造的抗震结构的设计和维护提供宝贵的技术。

Soil modeling has played an essential role in order to better understand natural phenomena such as soil liquefaction. Soil liquefaction, which refers to the process by which the structure of water-saturated sediments loses its strength due to dynamical loadings such as during earthquakes, can have catastrophic consequences: collapse of buildings, retaining structures, or embankment dams, landslide in quick clay areas (Saint-Jean-Vianney, Quebec, 1971). It is therefore a source of major concerns in the province of Quebec due to its large number of hydroelectric dams. Yet, the current state-of-the-art does not allow one to correctly predict soil liquefaction. This research project, whose general objective is the development and validation of improved models at the grain scale for describing the interactions between sand particles, will involve collaborators from Ecole Polytechnique Montreal and Hydro-Quebec's Research Institute (IREQ) in Varennes, Quebec. More specifically, it will aim at improving and validating ellipsoidal models of grains, at providing high-resolution adaptive representations of grains, and at developing and validating a stochastic contact model to describe particles of arbitrary shape. The new models will be implemented and tested in the high-performance computer platform (SiGran) for the simulation of interaction between granular media with fluid flow under static and dynamic loading at the particle and pore scale. The proposed research program will involve three Ph.D. students, two M.Sc. students, and several undergraduate students, who will be exposed to contemporary challenges in computational engineering and sciences and trained in numerical analysis, scientific computing, mathematical modeling, probabilities, Bayesian inference, and model validation. Development of advanced modeling and simulation tools, for a better insight of liquefaction processes, would provide an invaluable technology for the design and maintenance of earthquake-proof structures constructed near soil deposits in Canada and around the world.********