Research on void transformation behavior of asphalt mixture based on pore network model and skeletonization

Voids are closely related to the compaction state and macroscopic properties of asphalt mixtures. This research employed progressive gyratory compaction and X-ray computed tomography (CT) scanning to obtain meso-structural images of asphalt mixtures under different compaction states. By utilizing digital image processing and Three-Dimensional (3D) reconstruction, Pore Network Model (PNM) and topological skeleton of voids were constructed. Quantitative indicators for the size, morphology, and distribution characteristics of various void types were proposed. A thorough analysis was conducted on the void transformation pathways and spatial distribution behaviors under different compaction states. The results revealed that during compaction, both AC-13 and SMA-13 exhibit a transformation sequence from connected voids to semi-connected voids and subsequently to isolated voids, while OGFC-13 maintains a void connectivity rate consistently above 99%. Compaction simplifies and linearizes the connected void network in AC-13 and SMA-13, whereas OGFC-13 retains relatively high network complexity. When the compaction degree reached 96%, the meso-structure of AC-13 and SMA-13 achieved reliable stability. Furthermore, compaction facilitates a more uniform distribution of isolated voids, particularly in SMA-13. These findings offered theoretical references for the design and compaction control of asphalt mixtures.