Probabilistic modelling and experimental investigation of variability in compaction of asphalt mixtures

Asphalt mixture compaction is a critical phase of pavement construction and is essential for pavement quality and longevity. Owing to material heterogeneity—the random distribution of aggregates, binders and air voids—compaction of asphalt mixtures exhibits substantial variability, which in turn influences pavement performance. This paper investigates compaction variability through a combined theoretical and experimental approach. The experimental work involves a set of laboratory gyratory compaction tests on specimens with different sizes and aggregate gradations. Multiple replicates were used to quantify the variability in compaction behaviour. A recently developed mechanistic model for gyratory compaction is extended into a probabilistic framework using Monte Carlo simulation, with a Gaussian random field representing the initial density variability induced by material heterogeneity. The experimental results show that the variability in the gyratory compaction curves increases with decreasing specimen size and increasing aggregate size. The probabilistic model simulation closely matches the experimental trends, and provides a theoretical explanation for the observed behaviour. A parametric study based on the numerical model further reveals the influence of aggregate size and specimen size on compaction variability. This study offers valuable insights for controlling variability in asphalt pavement construction.