土体的细观孔隙结构非常复杂,且对其渗流特性有着重要的影响.为了研究土体的细观孔隙特征及流体在土体孔隙内流动的详细信息,基于土体CT扫描的切片图像,采用数字图像处理技术提取出土体的孔隙率及孔隙的两点自相关函数.根据孔隙率随机生成了土体的细观孔隙结构,然后将土体孔隙的两点自相关函数作为目标函数,采用模拟退火算法对重构土体的细观孔隙结构进行优化运算,从而得到了与真实土体孔隙结构特征相似的细观模型.最后,联合用于多孔介质细观渗流数值模拟的格子Boltzmann方法,对重构土体模型进行了渗流场的数值计算,计算结果表明:土体在渗流过程中主要沿着孔隙连通性较好的通道进行,在局部封闭的孔隙内渗流流速较小,且最大渗流流速通常发生在整体贯通通道的较窄部位.
The mesoscopic pore structure of soil is very complex,which has important influence on the seepage characteristics. In order to study the pore characteristics of soil and the details of fluid flow in the pore of soil,based on the slice images of soil obtained by CT scanning,the porosity and two-point autocorrelation function of soil are obtained with digital image processing technology. The pore structure of soil is generated randomly according to the porosity,and the two-point autocorrelation function is used as the objective functions to optimize the reconstructed mesoscopic pore structure with simulated annealing algorithm. The mesoscopic model is thus obtained,which is similar to the real soil for the pore structure characteristics. The lattice Boltzmann method,which is used to simulate mesoscopic flow of porous media,is employed to calculate numerically the seepage field of reconstructed model. The calculation results show that the soil seepage is mainly along the pore passage with good connectivity,the seepage velocity is small in the partially closed pores,and the maximum flow velocity is usually in the narrow part of the whole-through-channel.