基于离散元的土壤模型参数标定方法
发布时间：2019-01-01 16:19
【摘要】：离散元法(EDEM)建立土壤模型过程中部分土壤颗粒参数直接测量难度较大,若基于间接测量的土壤参数值建立离散元土壤模型进行仿真,导致仿真结果误差较大。本文结合代理模型基本理论,提出一种离散元土壤模型的参数标定及优化方法,步骤如下:根据基本试验测定的参数建立离散元土壤模型;结合堆积角及剪切试验,利用模型仿真进行模型参数敏感性分析;以敏感性参数为变量,以真实试验测量值为目标值构造代理模型;通过高斯-牛顿迭代法进行参数优化。由敏感性分析结果知,代理模型自变量为土壤颗粒半径、颗粒间静摩擦因数及滚动摩擦因数,目标量为土壤堆积角、黏聚力、内摩擦角。以涿州保护性耕作试验站土壤(砂壤土)为原型,经优化建立的土壤模型变量参数值分别为:颗粒半径5.7 mm,颗粒间静摩擦因数0.45,滚动摩擦因数0.21。将建立的离散元土壤模型进行轮胎-土壤相互作用仿真模拟,分析轮胎-土壤接触面最大应力、平均应力,并通过田间试验进行验证,将接触面最大应力值、平均应力的仿真值与实际测量值进行比较,结果表明:虚拟仿真与实测值之间数值差异在5.1%以内,标定优化后的土壤模型能够近似代替真实土壤进行仿真。
[Abstract]:In the process of establishing soil model by discrete element method (EDEM), it is difficult to measure some soil particle parameters directly. If the discrete element soil model is established based on indirectly measured soil parameters, the error of simulation results will be large. In this paper, based on the basic theory of agent model, a method of parameter calibration and optimization of discrete element soil model is proposed. The steps are as follows: the discrete element soil model is established according to the parameters measured by basic experiments; Combined with the stacking angle and shear test, the sensitivity analysis of model parameters is carried out by model simulation; the agent model is constructed by taking sensitive parameters as variables and real test measurements as target values; and the parameters are optimized by Gao Si Newton iteration method. According to the results of sensitivity analysis, the independent variables of the agent model are the radius of soil particles, the static friction coefficient and the rolling friction coefficient of the particles, and the target amount is the soil accumulation angle, the cohesion force and the internal friction angle. Using the soil (sandy loam soil) of Zhuozhou Conservation Tillage Station as the prototype, the parameters of the optimized soil model are as follows: static friction coefficient of particle radius 5.7 mm, rolling friction coefficient 0.21. The discrete element soil model is used to simulate the tire soil interaction, and the maximum stress and average stress of the tire soil interface are analyzed, and the maximum stress values of the contact surface are verified by field experiments. The results show that the numerical difference between the simulated and measured values is less than 5.1%, and the calibrated and optimized soil model can approximate replace the real soil for simulation.
【作者单位】： 中国农业大学工学院;
【基金】：国家重点研发计划项目(2016YFD0700301)
【分类号】：S151.9
，
本文编号：2397805
[Abstract]:In the process of establishing soil model by discrete element method (EDEM), it is difficult to measure some soil particle parameters directly. If the discrete element soil model is established based on indirectly measured soil parameters, the error of simulation results will be large. In this paper, based on the basic theory of agent model, a method of parameter calibration and optimization of discrete element soil model is proposed. The steps are as follows: the discrete element soil model is established according to the parameters measured by basic experiments; Combined with the stacking angle and shear test, the sensitivity analysis of model parameters is carried out by model simulation; the agent model is constructed by taking sensitive parameters as variables and real test measurements as target values; and the parameters are optimized by Gao Si Newton iteration method. According to the results of sensitivity analysis, the independent variables of the agent model are the radius of soil particles, the static friction coefficient and the rolling friction coefficient of the particles, and the target amount is the soil accumulation angle, the cohesion force and the internal friction angle. Using the soil (sandy loam soil) of Zhuozhou Conservation Tillage Station as the prototype, the parameters of the optimized soil model are as follows: static friction coefficient of particle radius 5.7 mm, rolling friction coefficient 0.21. The discrete element soil model is used to simulate the tire soil interaction, and the maximum stress and average stress of the tire soil interface are analyzed, and the maximum stress values of the contact surface are verified by field experiments. The results show that the numerical difference between the simulated and measured values is less than 5.1%, and the calibrated and optimized soil model can approximate replace the real soil for simulation.
【作者单位】： 中国农业大学工学院;
【基金】：国家重点研发计划项目(2016YFD0700301)
【分类号】：S151.9
，
本文编号：2397805
本文链接：https://www.wllwen.com/kejilunwen/nykj/2397805.html
