砾石土特征分析及数字岩心重构

Analysis of Gravelly Soil Characteristics and Digital Core Reconstruction

摘要: 砾石土中的孔隙结构对土体的渗透性和稳定性有很大影响，砾石土微观结构特征直接影响其宏观性能，为了更深刻认识砾石土内部细观结构，为砾石土中的管涌、流土等渗透侵蚀破坏防治提供参考。对于砾石土制取样品后进行CT扫描获取图像，对图像进行预处理，基于图像重建出三维数字岩心。通过其孔隙度、孔隙连通性、孔径和粒径等对砾石土的结构特征进行分析，从而得到砾石土的细观结构特征。选取砾石土体孔隙充分发育的一张二维图像，采用模拟退火法进行重构，以两点相关函数、线性路径函数作为约束函数。成功重构出与真实土体孔隙结构特征相似的数字岩心。模拟退火法建模所需建模资料要求较低，是一种经济、高效的建模方法。采用两点相关函数和线性路径函数对于重构图像进行评价，可以得到模拟退火算法重构的砾石土孔隙结构与真实土体细观结构较为相似，重构出的砾石土数字岩心效果较好。

Abstract: The pore structure of gravelly soil has a great influence on the permeability and stability of soil mass, and the microstructural characteristics of gravelly soil directly affect its macroscopic properties, which provides a reference for preventing and controlling the seepage erosion damage of piping and flowing soil in gravelly soil. For the gravelly soil samples after the acquisition of a CT scan image, image pre-processing is based on the image reconstruction of a three-dimensional digital core. The microstructure characteristics of gravelly soil are analyzed by porosity, pore connectivity, pore size and particle size. A two-dimensional image of gravelly soil with well-developed pores was reconstructed by simulated annealing, and the two-point correlation function and linear path function were taken as constraint functions. The digital core is reconstructed successfully, which is similar to the pore structure of real soil. Simulated annealing (SA) is an economical and efficient modeling method, which requires less modeling data. By using the two-point correlation function and linear path function to evaluate the reconstructed image, the pore structure of gravel soil reconstructed by simulated annealing algorithm is similar to that of real soil, and the reconstructed gravel soil digital core has a better effect.