Bonded discrete element method (DEM) models utilize networks of bonds between discrete particles to simulate continuum behaviors of real materials, most notably large deformations and failure which are difficult to simulate with mesh-based methods. However, the process of calibrating the bond parameters to produce specific macroscale properties remains an active area of research. Current calibration methods typically demonstrate the applicability of calibrations on the geometry they were created from. Our research utilizes an energy-based method to produce calibrations for monodisperse randomly packed bonded particles without running simulations. These calibrations are then utilized to simulate networks of bonded particles that vary from the calibration set in one of three ways: (1) number of particles; (2) aspect ratio; and (3) cross section shape. This work quantifies the standard deviation in the mean Young’s modulus and Poisson’s ratio produced by a general DEM calibration for randomly packed bonded particle specimens if the coordination number, particle diameter, shape, or aspect ratio is varied.

粘结离散元法 (DEM) 模型利用离散颗粒之间的粘结网络来模拟真实材料的连续行为，尤其是难以使用基于网格的方法进行模拟的大变形和失效。然而，校准键参数以产生特定宏观特性的过程仍然是一个活跃的研究领域。当前的校准方法通常证明校准对其所创建的几何体的适用性。我们的研究利用基于能量的方法来对单分散随机堆积的粘合颗粒进行校准，而无需运行模拟。然后利用这些校准来模拟粘合粒子的网络，这些粒子以以下三种方式之一与校准集不同：(1) 粒子数量；(2)长宽比；(3)截面形状。这项工作量化了在配位数、颗粒直径、形状或纵横比变化的情况下，随机填充的粘合颗粒样本的一般 DEM 校准产生的平均杨氏模量和泊松比的标准偏差。