基于Plackett-Burman试验设计与响应面法优化玉米秸秆离散元模型

Optimization of discrete element model of corn stalk based on Plackett-Burman design and response surface methodology

关键词:  玉米秸秆  离散元  临界载荷  Plackett-Burman试验设计  响应面

Key words:  corn stalk  discrete element  critical load  Plackett-Burman design  response surface methodology

摘要:

为了减小免耕播种机防堵装置数值模拟的误差，结合物理试验与仿真试验进行玉米秸秆离散元参数的标定。基于Hertz-Mindlin with bonding 接触模型建立了玉米秸秆离散元模型，以物理试验与仿真试验中临界载荷的相对误差为试验指标进行仿真单轴压缩试验；采用Plackett-Burman试验筛选出具有显著性影响的试验因素；采用Box-Behnken响应面法对玉米秸秆离散元模型进行优化。试验结果表明:接触半径和单位面积切向刚度对试验指标影响极显著(P<0.01)，单位面积法向刚度对试验指标影响显著(0.01<P<0.05)；玉米秸秆的最优离散元参数为，接触半径1.2 mm、单位面积法向刚度9.361×10⁷ N/m³、单位面积切向刚度9.845×10⁷ N/m³，在此条件下，仿真压缩试验的临界载荷为950.2 N，与物理试验值935.4 N的相对误差为1.58%，验证了参数的可靠性。标定的玉米秸秆离散元模型可用于免耕播种机防堵装置的数值模拟，为防堵装置结构优化改进提供了依据。

Abstract:

In order to improve the accuracy of the numerical simulation of the anti-blocking device, the discrete element parameters of corn stalk were calibrated by combining with physical and simulation experiments. The model of the simulated compression test was established based on Hertz-Mindlin with bonding contact model. Simulation experiments were designed by using the relative error of critical load between physical and simulation experiments as the experimental index. Factors that have positive effects to the experimental index were screened by Plackett-Burman design. The experiments were designed by Box-Behnken experimental design to optimize the discrete element model of corn stalk. The experiment result shows that contact radius and shear stiffness per unit area had an extremely significant influence on the index(P<0. 01), and the normal stiffness per unit area had a significant influence on the index(0. 01<P<0. 05). Statistical analysis was carried out to obtain optimum parameters of discrete element model of corn stalk, which were contact radius of 1. 2 mm, normal stiffness per unit area of 9. 361×10⁷ N/m³, and shear stiffness per unit area of 9. 845×10⁷ N/m³. Under this condition, the critical load of corn stalk was 950. 2 N in the simulated experiment. The critical load of corn stalk was 935. 4 N in physical uniaxial compression experiments. The relative error between simulated and physical experiments was 1. 58%. In coclusion, this study showes that the discrete element model of corn stalk is reliable and can be used to simulate the operation of the anti-blocking device of the no-tillage seeder.