Quantitative characterization of aggregates contact networks in asphalt mixtures through topological methods

The macroscopic properties of asphalt mixtures are predominantly governed by their meso-structure, particularly the particle contact networks among aggregates. For particulate systems composed of multi-size aggregates, the precise and efficient quantitative characterization of aggregates contact structures has become a critical challenge in the design of high-performance asphalt mixtures, owing to the randomness of the mixture formation process and the diversity in aggregates. To address this challenge, this study combined industrial computed tomography scanning, digital image processing, and deep learning techniques to extract meso-structural features, while the contact structures between aggregates were simplified using a point-line model and a quantitative characterization method was established based on topological theory. Aggregates with particle sizes exceeding 4.75 mm were classified into central coarse particles (CCPs) and skeleton particles (SCPs) according to their particle size and functional characteristics. Quantitative characterization of the meso-structures of three asphalt mixtures, namely AC-16, SMA-16, and OGFC-16, was achieved using degree distribution characteristic (DDC), assortativity coefficients (AsC), and average clustering coefficient (ACC). Statistical analysis yielded the following parameter ranges: the proportion of invalid aggregates within the mesoscopic contact networks of CCPs and SCPs; the average number of contacts and the concentration range of the AsC and the ACC among valid aggregates.