In-silico manufacturing of asphalt concrete

Abstract

In this paper we present an algorithm to produce realistic virtual asphalt concretes. The inputs are the geometry of aggregates, and the composition of the asphalt concrete. The outputs are the air voids content and, a 3-dimensional representation of the aggregate skeleton. The air voids that are predicted by the algorithm have been compared to those measured from asphalt made in the laboratory. The key parameters that influence the compaction of an asphalt concrete are the shape of the aggregates, number of aggregates and, volume of mastic, i.e. bitumen, filler, and aggregates smaller than 2 mm. An equation based on a Weibull function, with independent variables the number of aggregates and force required to compact the virtual aggregates, has been developed. This equation can predict the force that the algorithm needs to compact the virtual aggregates to a level equivalent to that reached by a laboratory compactor.