Evaluation of the evolution of the structure of cold recycled mixture subjected to wheel tracking using digital image processing

Abstract

The meso-structural characteristics of cold recycled mixes using asphalt emulsion (CRME) were investigated by wheel tracking tests and two-dimensional digital image processing technology in this study. The trends of contact points, long-axis states, fractional dimensionality, and the rule of aggregate distribution were analysed. Fitted equations of the CRME rutting depth and mesoscopic parameters were established in different loading periods. The number of aggregate contact points appears to increase as the loading time increasing. The long axis of the aggregate tends to be consistent. The long axis inclination and fractal dimension of the aggregate profile gradually decrease over time. The aggregate area ratio increased gradually and then the mixture became dense. The evolution of CRME could be divided into three stages under wheel tracking test: the stage of initial compaction, relative compaction, and destruction. The support vector machine prediction model of rutting depth and mesoscopic parameters was established by K-fold Cross Validation algorithm (K-CV) optimization. The proposed model showed high accuracy for the movement state of the asphalt mixture, which could provide scientific guide for pavement condition evaluation.