Granular materials are widely used for bases or subbases in pavement structures. They typically exhibit strong anisotropic properties which relate to stress states and particle characteristics. The conventional design procedure for flexible pavements underestimates the anisotropy of resilient moduli. This study established an anisotropic resilient modulus model for granular materials that considered gradation and particle shape characteristics. Vertical and horizontal resilient moduli of certain granular materials were measured in self-developed triaxial tests to obtain corresponding model parameters and anisotropic coefficients. Gradation and particle shape models were established to quantify the granular material characteristics, and the parameters were regressed. Particle shapes were obtained via image processing, and the ratio (η) of particle sphericity to roundness was chosen as a shape parameter. Results show that η increases with the decrease in particle size, and the average values of η for graded gravel and natural laterite are 0.54 and 0.63, respectively. The η distribution curves indicate that the proportion of relatively anisotropic particles, rather than extremely anisotropic particles, results in the differences in particle shape characteristics. The regression relationship between the anisotropic calculation parameters and the model parameters of vertical resilient modulus, gradation, and particle shape was established. Thus, the horizontal resilient modulus and the anisotropic coefficient can be predicted via conventional resilient modulus tests and gradation, and particle shape analysis. This study shows that the anisotropy of granular materials decreases with the increase in coarse particles and the uniformization of the particle size distribution, and it increases with the increase in anisotropic particles and the polarization of the η distribution.

粒状材料广泛用于路面结构中的基层或底基层。它们通常表现出与应力状态和颗粒特性相关的强各向异性特性。柔性路面的传统设计程序低估了弹性模量的各向异性。本研究建立了考虑级配和颗粒形状特征的颗粒材料各向异性弹性模量模型。自行研制的三轴试验测量了某些颗粒材料的纵向和横向弹性模量，得到相应的模型参数和各向异性系数。建立级配和颗粒形状模型来量化颗粒材料的特性，并对参数进行回归。颗粒形状是通过图像处理获得的，比例（η ) 的颗粒球形度到圆度被选为形状参数。结果表明，η随着粒径的减小而增大，级配砾石和天然红土的η平均值分别为0.54和0.63。该η分布曲线表明，相对各向异性粒子的比例，而不是极端各向异性粒子，导致了粒子形状特征的差异。建立了各向异性计算参数与竖向回弹模量、级配、颗粒形状等模型参数的回归关系。因此，水平弹性模量和各向异性系数可以通过传统的弹性模量测试和分级以及颗粒形状分析来预测。本研究表明，颗粒材料的各向异性随着粗颗粒的增加和粒度分布的均匀化而降低，随着各向异性颗粒的增加和η分布的极化而增加。