In this paper, the shear resistance of asphalt mixtures, which accounts for the permanent deformation characteristics of flexible pavements to a large extent, is analyzed based on the discrete element (DE) method from a microscopic perspective. This study first considered the processes used to obtain the microscopic parameters for the DE model, which typically simulated an asphalt mixture based on its three components. Then the study employed Burger's model to simulate the rheological behavior of asphalt sand mastics (fine aggregates, fines, and asphalt binder). A random generation algorithm was also developed to generate coarse aggregate elements in the DE model complying with the realistic gradations of asphalt mixtures. So as to more precisely model the rheological characteristics of asphalt sand mastics, the microscopic parameters of Burger's model were calibrated via simulations of uniaxial tests in the DE model. Finally, meaningful conclusions were achieved by analyzing the simulation result and the laboratory result. The simulation result was consistent with the laboratory test result, so the use of the established DE model to evaluate the shear resistance property of asphalt mixtures is feasible.

Granular materials, such as ballast and aggregate, are widely used nowadays in civil and transportation engineering. Discrete element method has been extensively used to simulate the behavior of granular materials. In this study, properties of granular materials used in pavement and railway were investigated by laboratory tests and discrete element modeling. Firstly, the bonding performance between pavement layers was evaluated. Open-graded friction course pavement and traditional dense asphalt mixture pavement were both explored. Two dense asphalt mixtures (D, BM) and one open-graded friction course (OGFC) mixture were selected for the comparison. The laboratory test results show that, for traditional dense samples, the interlock effect between layers played an important role in pavement layer bonding. For specimens composed of OGFC and a dense mixture (D or BM), OGFCBM showed a better shear performance than OGFC-D, due to the double effects of a larger interface contact area and a larger interface roughness than OGFC-D. The DEM modeling was focused on the interlock effect between pavement layers by conducting the direct shear box test. Results from DEM modeling show that D-BM gave a higher shear strength, which agreed with the laboratory test. Secondly, laboratory tests were conducted to investigate the shear fatigue performance between OGFC and underlying layer. Results indicate that contact area between OGFC and underlying layer play the critically important role. The larger the contact area, the better the shear fatigue performance. Thirdly, a full scale laboratory test was conducted to investigate the pressure distributions under a single steel or timber crosstie. It is found that pressure distribution was different for steel and timber crossties. Cyclic loading could change the pressure distribution under both steel and timber crossties, but the effect of cyclic loading was more obvious on steel crosstie than on timber crosstie. Last, one coupled framework between discrete element method (DEM) and finite element method (FEM) was developed to investigate the ballast-tie interaction. The normal contact condition and the stress distribution beneath the steel crosstie and timber crosstie were obtained from the simulation. Stress distribution obtained by DEM-FEM simulation was consistent with the findings from the laboratory test.

Proceedings of the Third International Conference on Discrete Element Methods, held in Santa Fe, New Mexico on September 23-25, 2002. This Geotechnical Special Publication contains 72 technical papers on discrete element methods (DEM), a suite of numerical techniques developed to model granular materials, rock, and other discontinua at the grain scale. Topics include: DEM formulation and implementation approaches, coupled methods, experimental validation, and techniques, including three-dimensional particle representations, efficient contact detection algorithms, particle packing schemes, and code design. Coupled methods include approaches to linking solid continuum and fluid models with DEM to simulate multiscale and multiphase phenomena. Applications include fundamental investigations of granular mechanics; micromechanical studies of powder, soil, and rock behavior; and large-scale modeling of geotechnical, material processing, mining, and petroleum engineering problems.