旋耕后农田模型离散元参数标定

Calibration of discrete element parameters of farmland model after rotary tillage

土壤因其含水率、成分、粒径等因素不同导致接触参数存在差异,为提高新型高密度灌木幼苗移栽机开沟覆土作业过程的准确性,对北方春季耕后农田土壤进行离散元参数标定试验。通过物理试验与仿真试验相结合的方法,采用Hertz-Mindlin(no slip)接触模型以土壤—土壤堆积角和土壤—触土钢滑动摩擦角为响应值,恢复系数、静摩擦系数和滚动摩擦系数为目标值进行参数标定。首先将目标值应用Box-Behenken响应面分析法进行三因素三水平的正交旋转试验得到二阶回归模型;其次以实际堆积角和滑动摩擦角为响应值,通过Design-Expert寻优模块得到最优参数组合:土壤—土壤恢复系数0.4,土壤—土壤静摩擦系数0.8,土壤—土壤滚动摩擦系数0.39;土壤—钢恢复系数0.27,土壤—钢静摩擦系数0.53,土壤—钢滚动摩擦系数0.16;并进行物理试验与仿真试验的对比,结果表明:两者误差分别为0.68%和1.34%,在可接受范围之内,可信度良好,具有一定工程指导意义。

Soil contact parameters are different due to different factors such as soil moisture content,composition,and particle size.In order to improve the accuracy of the trenching and soil covering operation of a new high⁃density shrub seedling transplanter,this study conducted a discrete element parameter calibration experiment on farmland soil after spring cultivation in northern China.Through a combination of physical and simulation experiments,the Hertz-Mindlin(no slip)contact model was used to calibrate the parameters with the soil⁃soil response angle and the soil⁃steel sliding friction angle as the response values,and the recovery coefficient,static friction coefficient,and rolling friction coefficient as the target values.Firstly,the Box-Behenken response surface analysis method was applied to the target value to conduct a three factor and three level orthogonal rotation test,so as to obtain a second⁃order regression model.Secondly,taking the actual stacking angle and sliding friction angle as the response values,the optimal parameter combinations were obtained through the Design-Expert optimization module as follows:soil⁃soil recovery coefficient was 0.4,soil⁃soil static friction coefficient was 0.8,and soil⁃soil rolling friction coefficient was 0.39.Soil⁃steel recovery coefficient was 0.27,soil steel static friction coefficient was 0.53,and soil steel rolling friction coefficient was 0.16.A comparison between physical experiments and simulation experiments was also conducted,and the results showed that the errors of the two methods were 0.68%and 1.34%,respectively,which were within an acceptable range,and had a good reliability and certain engineering guiding significance.