PURPOSES : The feasibilities of cohesive elastoplastic contact model and discrete element method (DEM) for asphalt concrete mixture compaction process were evaluated. METHODS : The contact models that were used to simulate the dynamic behavior of construction materials were reviewed. The characteristics of cohesive elastoplastic models were discussed from the perspective of integration with existing contact models. Two asphalt mixtures that were fabricated with specific aggregate gradations and binder contents were compacted according to the Superpave gyratory compaction specification. The parameters for the model were determined via trial-and-error method. The heights of the specimens were plotted with respect to number of gyrations. The results of the laboratory tests were compared to those of numerical simulations. The displacement of particles in asphalt mixture specimen was also visualized to understand the effect of gyratory compaction on asphalt mixture specimen. RESULTS : The DEM model exhibited a significant friction coefficient dependency on compaction degree and slop. The DEM model with parameters determined through trial and error demonstrated reasonable simulation results in terms of specimen height at a gyration number. CONCLUSIONS: Even though a little discrepancy was observed between the results of the experimental test and numerical simulation, a combination of DEM with cohesive elastoplastic contact model seems to be applicable for the simulation of asphalt mixture compaction in laboratory. However, the model needs to be enhanced to be used for more realistic compaction processes, including heat transfer, phase change, and vibratory loading.

PURPOSES: This study proposes a cohesive shrinkage particle model that can be used to simulate a variety of dynamic behaviors and phase changes of construction materials, including road subsidence and debris flow, and phase change curing, via discrete element method (DEM). METHODS : From the perspective of DEM modeling, the water-content-dependent characteristics of soil particles and related modeling techniques are reviewed from literature. The static friction, cohesion, and particle size change are considered as the major parameters that should be reflected in DEM modeling for a more realistic simulation. The relationships of water content with cohesive force and particle radius, as determined from experimental test results in the relevant study, are utilized to develop the cohesive shrinkage model. For each water content value, the snapshot in simulation is compared to that in the experimental study. RESULTS: The numerical simulation shows very good agreement with the experimental test in terms of overall sample radius and thickness change due to drying. However, the local curling of soil sample in the DEM simulation does not perfectly match that in the experimental test. CONCLUSIONS : The cohesive shrinking particle model seems to be good enough for simulating the volumetric and phase changes of soil samples due to drying. However, it seems necessary to consider both bonding and cohesive contact models in DEM modeling because the only cohesive contact model exhibited limitations in the simulation of curling and crack development.