中文摘要：

为真实反映沥青混合料内部非均质（粗集料、砂浆和空隙）多层次（矿料级配）的结构特性,通过X射线断层扫描技术获取试件的多层图像,运用MATLAB对沥青混合料CT图片进行了集料、胶浆和空隙的识别与分离;借助离散元方法,实现了由二维可视化图形向二维数值化模型的转换,并提出了空隙特征的还原方法;确定合适间距挑选代表性CT图片生成二维模型,将每个二维模型复制一定倍数联结形成整体,建立了包含不规则形状集料、级配特征和空隙的沥青混合料三维离散元模型。研究结果表明,该算法及程序操作简便,切实可行,弥补了传统的数值方法采用统计学原理或采用对试件损伤研磨建立三维细观模型的不足,为开展基于不连续介质力学的虚拟试验奠定了重要基础。

英文摘要：

In order to truly reflect the heterogeneous（coarse aggregates,asphalt mastic and air voids）and multilayer（aggregate gradation）morphological features in the asphalt mixture,multilayer images are obtained using X-ray tomography to scan the real specimen,and the aggregates,the mastics and the air voids in the Computerized Tomography（CT）images are recognized and separated using MATLAB.Each slice of the two dimensional（2D）image is then converted to 2Dnumerical model and the presentation of the air voids is proposed.Several representative CT images are chosen by determining aproper interval to generate the 2Dmodel.Each 2Dplane is duplicated several times to form a three dimensional（3D）discrete element model of asphalt mixture considering irregular shape coarse aggregates,gradation,and air voids.The results indicate that the algorithm is convenient and practical,and could make up the shortage of traditional digital method based on statistics theory or 3Dmicromechanical damage model.The results also provide the significant foundation for virtual test based on the discontinuous medium mechanics.