SudoDEM: Unleashing the predictive power of the discrete element method on simulation for non-spherical granular particles

This paper presents a novel open-source discrete element code, SudoDEM, for efficient modeling of both 2D and 3D non-spherical particles under a GPL v3 or later license. Built upon a popular open-source code YADE, our code inherits the core of a classic DEM framework empowered by OpenMP acceleration, and further offers unique features of a rich library of prime particle shapes, including poly-superellipsoids, superellipsoids, cylinders, cones, polyhedrons for 3D and disks and superellipses for 2D. Unlimited choices of more complex particle shapes can be readily generated by clumping these prime shapes. Efficient modeling of complex shaped particles hinges on contact detection. In SudoDEM, we have developed three generic and efficient contact detection algorithms, the parametric common normal (PCN) algorithm, the Gilbert–Johnson–Keerthi (GJK) algorithm, and the hybrid PCN–GJK algorithm, to handle contacts among complex-shaped particles during a typical DEM simulation. The new DEM code is validated and further showcased by multiple examples, including granular packing, triaxial compression, and landslide, its robustness, efficiency and versatility in providing realistic solutions to granular mechanics problems. The project is hosted at an open-source page at https://sudodem.github.io, while the source codes are freely available at a GitHub repository (https://github.com/SudoDEM). We foresee a great capability and potential for SudoDEM in advancing future progress in granular physics and granular mechanics and in fostering advanced simulations of critical engineering and industrial processes pertaining to granular media.

本论文提出一款全新的开源离散元代码SudoDEM，可在GPL v3及后续版本许可协议下，高效实现二维与三维非球形粒子的建模工作。本代码基于主流开源代码YADE开发，继承了经典离散元框架的核心功能，并通过OpenMP实现并行加速；同时新增了丰富的基础粒子形状库，涵盖三维场景下的多超椭球、超椭球、圆柱、圆锥、多面体，以及二维场景下的圆盘与超椭圆。通过组合这些基础粒子形状，可灵活生成任意复杂的粒子形态。复杂形状粒子的高效建模核心在于接触检测。在SudoDEM中，我们开发了三种通用且高效的接触检测算法：参数公共法线（Parametric Common Normal, PCN）算法、吉尔伯特-约翰逊-基尔蒂（Gilbert–Johnson–Keerthi, GJK）算法，以及混合PCN-GJK算法，用于在典型离散元仿真过程中处理复杂形状粒子间的接触问题。本新型离散元代码通过多个示例完成了验证与展示，涵盖颗粒堆积、三轴压缩及滑坡场景，其在颗粒力学问题的求解中展现出优异的鲁棒性、高效性与通用性，可生成贴合实际的求解结果。该项目的开源主页为https://sudodem.github.io，源代码可于GitHub仓库（https://github.com/SudoDEM）免费获取。我们认为，SudoDEM具备推动颗粒物理学与颗粒力学领域未来发展的强大能力与潜力，同时可助力针对颗粒介质相关的关键工程与工业过程开展先进仿真研究。

搜集汇总

背景与挑战

背景概述

SudoDEM是一个开源离散元方法代码数据集，专注于非球形颗粒的模拟，提供丰富的2D和3D基本粒子形状库和高效的接触检测算法。该数据集支持颗粒材料力学问题的仿真，适用于计算物理和工程领域，基于GPLv3许可证发布，包含源代码和相关示例。