【作者】 李万清；

【导师】 周又和；

【作者基本信息】 兰州大学 ， 固体力学， 2007， 博士

【摘要】 本论文针对风沙流中混合沙粒-床碰撞开展直接数值模拟研究。首先利用随机方法生成接近自然分布规律的沙床面，进而结合离散单元法(DEM)建立了粒-床碰撞数值模拟系统。在此基础上研究了混合沙粒-床碰撞的起跳特征，并且提出一套统计方法对碰撞结果进行统计处理。从而进一步分析了平床面上粒径尺度对粒-床碰撞起跳特征的影响，给出各粒径尺寸下的入射与起跳物理量之间的关系。最后将平床面碰撞模拟系统推广到倾斜床面，讨论了床面倾斜对粒-床碰撞特征的影响，给出了同时包含粒径尺度和床面倾角参数的粒-床碰撞规律。主要工作如下：1、考虑到天然沙漠沙粒径分布具有一定规律这一特点，本文提出了近似自然床面的生成方法，建立了基于混合粒径平床面的粒-床碰撞二维离散动力学(DEM)模拟系统。2、由于粒-床碰撞具有高度的随机性，我们提出了一套统计方法对结果进行处理。探讨了床面颗粒数、样本空间大小和碰撞位置等情况下统计结果的稳定性，找到了使其稳定的参数区间。3、在上述建立的模拟程序和统计方法基础上，我们针对多粒径、多速度、多角度入射，采用大样本进行碰撞模拟。研究表明，模拟结果与已有单一粒径结果差别明显，混合粒径模拟结果与实验数据更为吻合，得出应以随机混合粒径床面的碰撞规律为基础进行风沙流和沙波纹的模拟。有力的证明了进行多粒径粒-床碰撞模拟实验研究的必要性。4、在混合粒床相互作用模拟基础上，对粒-床碰撞过程中不同粒径粒子的运动行为进行了深入细致的研究。首次发现并提出粒子的起跳特征与粒径的依赖关系，给出了包含粒径参数的入射与起跳关系，模拟结果与实验结果吻合良好。5、由于沙漠表面是起伏不平且呈近似周期性的波纹分布，我们沿一个波纹不同点处取其切平面，生成具有不同坡度的倾斜床面实施粒-床碰撞模拟实验，从而研究波纹表面不同位置处的起跳特征的异同。研究发现，波纹不同位置处的起跳规律具有较大的差异，随着坡度的变化，某些重要的起跳物理量具有明显的变化规律。本文首次系统的给出了各起跳物理量与入射速度、入射粒径和波纹倾角的变化关系。更多还原

【Abstract】 This thesis focuses on the key problem in the wind blown-sand transport, the sand particle-bed collision. Firstly, a 2D mixed sand bed is prepared by stochastic method and the direct numerical simulation program is created based on Discrete Element Method (DEM). And then we insight the grain-bed collision on flat mixed size bed using this program and deal with the results by a statistic approach. Secondly, the effect of particle sizes on grain-bed impact is discussed, meanwhile the relationship between incident and lift-off quantities for different particle size are presented. Finally, we extend our program to different slope to estimate the effect of slope angle on grain-bed impact. The relationships of incident and lift-off quantities proposed by this thesis, including impact speed, impact grain size and slope angle, will be significant for windblown sand transport and sand ripple simulation. The main works are concluded as follows:1. Considering the Probability Density Function (PDF) of natural sand particle size has unimodal distribution, this thesis presents one approach to propagation an approximate natural sand bed used to target bed. Moreover, a program code is created based on Discrete Element Method (DEM) or Particle Dynamic Method (PDM) in two dimensional to realize the mixed size grain-bed impact process.2. Due to grain-bed interactions are stochastic processes, we propose a statistical method to deal with results, and also find the appropriate coefficients of steady state program and statistical method when the number consisted of sand bed, number of sample and impact points vary.3. Using the grain-bed impact program and statistical method, to each different incident particle size, different incident speed and angle a lot of samples will be used to simulate impact. We find that there are distinct deviations between mixed grain size and uniform grain size simulation results, and the results of this thesis well agree with the experimental data of the multiple size sand grains, which suggests us that the simulations of windblown saltation and sand ripple should be based on our results. It is also justified the necessary to study mixed grain-bed impact.4. Based on the simulation on flat bed, we insight the effect of particle size on the impact processes. For the first time, we argue the relationship between particle size and random physical quantities of lift-off particles. These simulation results well consist with measurements. 5. The efforts have been done on the flat bed are prepared for us to discuss the grain-bed impact on more natural sand beds which always are undulate in field desert. In order to find the effect of slope on different location of one sand ripple, we simulate the impact processes on a series of slopes. The results demonstrate that the random physical quantities of lift-off particles on different slop have markedly deviations. Some random physical quantities change regularly varying with the angle of slope. For the first time this thesis presents the relationship between incident and lift-off physical quantities of particles effected by incident speed, incident particle size and slope angle.更多还原