Influence of coarse aggregate meso-fabric on dynamic modulus of porous asphalt mixtures: An X-ray CT voxel reconstruction approach

Abstract

Coarse aggregate contacts and orientations govern load-transfer paths in porous asphalt (PA) mixtures, yet their respective relationships with dynamic modulus (| E *|) remain unclear. Industrial X-ray computed tomography (X-CT) and digital image processing were used to reconstruct air void and coarse aggregate skeleton voxel models for six PA-13 gradations. Fabric tensors were employed to quantify anisotropy of contact normals and major axis orientations, and the resulting parameters were correlated with | E *| of PA mixtures. The results show that reconstructed air void contents agreed with measurements, yielding a mean relative error (MRE) of 5.15%. Spatially, the air voids followed a U-shaped distribution along the specimen height. Coordination number increased with particle size, while major axis inclinations were concentrated between 40° and 140° and approximated a normal distribution. Additionally, contact normal and major axis orientation fabrics showed distinct anisotropy, with greater horizontal dispersion for the latter. Across most temperature-frequency combinations, contact normal fabric parameters showed stronger associations with | E *| than major axis orientation fabrics, particularly under low-temperature and intermediate-temperature low-frequency conditions. Within this six gradation dataset, contact networks better explained differences in | E *| among gradations than particle orientation. Consequently, the findings offer valuable insights for assessing skeleton load-bearing capacity and optimizing gradation design in PA mixtures.