Authors

Abstract

To improve the accuracy and stability of discrete element method (DEM) parameter calibration for soybean seeds under sowing conditions, a soybean seed DEM model was developed in EDEM based on the intrinsic mechanical properties of seeds at sowing-stage moisture content, and contact parameter calibration between seeds and the seed metering device was conducted. Free-fall collision, inclined-plane sliding, and rolling tests were performed to determine the initial ranges of contact parameters between soybean seeds and Somos8000 resin as well as stainless steel. The experimentally measured angle of repose (23.02°) was used as the response variable. Key factors were screened using a steepest ascent test, and parameter optimization was carried out by combining response surface methodology (RSM) with a GA-BP-GA optimization framework. The results showed that the optimal parameter combination consisted of a coefficient of restitution of 0.27, a static friction coefficient of 0.23, and a rolling friction coefficient of 0.056, yielding a simulated angle of repose of 22.35° with a relative error of 2.91%, which was lower than that obtained by RSM (4.39%). Bench tests further confirmed the reliability of the model and calibrated parameters.

Abstract in Chinese

为提高播种工况下大豆离散元参数标定的精度与稳定性，基于播种期含水率条件下的大豆种子本征力学参数，在 EDEM 中构建了大豆种子离散元模型，并开展了种子与排种器接触参数标定研究。通过自由落体碰撞、斜面滑动和斜面滚动试验，确定了大豆种子与 Somos8000 树脂及不锈钢材料之间的初始接触参数范围。以物理试验测得的大豆堆积角（23.02°）为响应指标，采用最陡爬坡试验筛选关键因子，并结合响应面法（RSM）和 GA-BP-GA 优化框架进行参数优化。结果表明，最优参数组合为碰撞恢复系数 0.27、静摩擦系数 0.23 和滚动摩擦系数 0.056，对应仿真堆积角为 22.35°，相对误差为 2.91%，优于 RSM 的 4.39%。台架试验验证了模型与参数的可靠性。