基于离散单元法和人工神经网络的近壁颗粒动力学特征研究

颗粒与壁面的相互作用往往对颗粒流动具有显著影响. 为研究颗粒与壁面作用机理, 对滚筒内颗粒流动过程进行离散单元法(DEM)数值模拟. 基于模拟结果统计分析靠近壁面处颗粒的运动特征, 结果表明, 小摩擦系数时颗粒平动和旋转速度均近似满足正态分布, 但由于壁面影响, 摩擦系数增大时颗粒沿滚筒轴向的旋转速度偏离正态分布, 颗粒动力学理论推导壁面边界条件时应考虑速度正态分布的修正及速度脉动的各向异性. 采用人工神经网络(ANN)构建了颗粒无因次旋转温度、滑移速度和平动温度之间的函数模型, 进而可以在常规双流模型壁面边界条件中考虑颗粒旋转的影响. 基于DEM模拟及结果分析可以为壁面边界条件的理论构造和半经验修正提供基础数据和封闭模型. 

CHARACTERIZATION OF NEAR-WALL PARTICLE DYNAMICS BASED ON DISCRETE ELEMENT METHOD ANDARTIFICIAL NEURAL NETWORK

The interactions between the particles and the walls often have significant effects on the particle flows. In order to study the mechanism of the interactions between the particles and the walls, the discrete element method (DEM) simulations of the particle flow in the rotating drum are carried out. Based on the statistical analysis of the simulation results, the characteristics of the near-wall particle motion are shown. The results indicate that the particle translational and rotational velocity approximately satisfy the normal distribution when the friction coefficient is small. However, due to the wall effects, the axial rotational velocity deviates from the normal distribution when the friction coefficient increasing. The kinetic theory of granular flow should consider the correction of the velocity normal distribution and also the anisotropic of the velocity fluctuation when deriving the wall boundary conditions. An artificial neural network (ANN) is used to construct a function model between dimensionless particle rotational temperature and particle slip velocity and particle translational temperature, and then the influence of particle rotation can be incorporated in the conventional boundary conditions within the two-flow model. Through comprehensive DEM simulation and result analysis could provide basic data and closure models for the theoretical construction and semi-empirical correction of the wall boundary conditions.