Abstract: In order to solve the problem of lack of accurate simulation parameters in the discrete element simulation of stem mustard seed, the discrete element parameters of “Fuza 2” stem mustard seed were determined by direct measurement and virtual calibration method. First of all, the triaxial size, equivalent radius, thousandkernel weight and density of the stem mustard seeds were measured by using digital vernier caliper, electronic scale and other physical and mechanical property measuring tools. Crash ramp test, threepoint sliding test, and ramp test were conducted, and coefficient of restitution, coefficient of static friction and coefficient of rolling friction between seed and the polymethyl methacrylate (PMMA) were attained to be 0.561, 0.420, and 0.155, respectively. The dynamic repose angle and static repose angle of stem seeds were 35.235 2° and 14.483 0°, respectively through rotating drum test and the cylinder lifting test combining with MATLAB image processing techniques. Then, seedseed rolling friction coefficient and seedseed static friction coefficient were proved to be significant factors for both kinds of repose angle by a discrete element simulation test based on PlackettBurman design method. Therefore, taking seedseed rolling friction coefficient and seedseed static friction coefficient as control factors, and dynamic repose angle and static repose angle as evaluation index, a simulation test based on Central Composite Design (CCD) test method was carried out. Quadratic regression equations of dynamic repose angle and static repose angle with control factors were established respectively. The regression model was used to optimize the control factor combination, and two sets of optimal factor combinations were obtained as 0.666 1, 0.023 7 and 0.341 6, 0.036 0. In the end, the simulation test parameters were respectively set as two sets of optimal factor combinations to compare with actual values measured in physical test. The results indicate that the relative errors between the simulated values and measured values of dynamic repose angle and static repose angle are 0.477%, 3.590% and 1.820%、4.950%. The results show that when the seedseed static friction coefficient and the seedseed rolling friction coefficient are 0.666 1 and 0.023 7 in the simulation, the performance of the stem mustard seeds are more in line with the actual physical characteristics. This conclusion can provide a reference for the research of sowing process of stem mustard seed and the design of related devices.

Key words: stem mustard seed, discrete element method, repose angle, parameter calibration

摘要： 为解决茎瘤芥种子离散元模拟中缺乏准确仿真参数的问题，通过直接测量和虚拟标定方法确定“涪杂2号”茎瘤芥种子离散元参数。利用数显游标卡尺、电子秤等物理机械特性测定工具，得到茎瘤芥种子的三轴尺寸、等效半径、千粒重和密度；通过斜面碰撞法、三点滑动法和斜面法分别得到茎瘤芥种子与有机玻璃之间的恢复系数、静摩擦系数和滚动摩擦系数分别为0.561、0.420、0.155；采用旋转鼓试验和圆筒提升试验结合MATLAB图像处理技术获得茎瘤芥种子的动态堆积角和静态堆积角分别为35.235 2°和14.483 0°。基于PlackettBurman试验筛选出对动态堆积角和静态堆积角影响显著的参数因子均为种子与种子间的滚动摩擦系数、种子与种子间的静摩擦系数；通过中心复合设计分别获得显著性因子与动态堆积角、静态堆积角的二次回归方程，并以动态堆积角和静态堆积角的实测值为目标，基于回归模型对因子参数组合进行优化，得到两组种子与种子间的静摩擦系数、种子与种子间的滚动摩擦系数最优参数组合为0.666 1、0.023 7以及0.341 6、0.036 0。对两组最优参数组合进行仿真验证，结果显示两组参数组合下动态堆积角、静态堆积角的仿真值与实测值相对误差分别为0.477%、3.590%和1.820%、4.950%。结果表明，仿真中种子与种子间的静摩擦系数、种子与种子间的滚动摩擦系数分别为0.666 1、0.023 7时，更符合茎瘤芥种子实际的物理特性。本研究能为茎瘤芥播种过程及相关装置设计优化提供参考。

关键词: 茎瘤芥种子, 离散元, 堆积角, 参数标定