Prediction modelling on dynamic modulus of recycled asphalt mixtures based on meso-mechanical analysis

Due to the addition of reclaimed asphalt pavement (RAP), the early proposed two-layered and three-layered models for the prediction of dynamic modulus (E*) are not applicable to the recycled asphalt mixture (RAM), and its properties cannot be accurately characterized due to the different diffusion states between virgin and aged asphalt binders. Therefore, a n-layer inclusion model for the prediction of E* was developed in the study with the attribute of the effect of RAP in the mixture. The effects of constituent material parameters, the diffusion state between virgin and aged asphalt, and the performance transition mode in the virgin-aged asphalt binder on the prediction were investigated. The results showed that the difference between predicted and experimental values could be corrected by a frequency-dependent correction factor. The E* of RAM was mostly sensitive to the variations in the modulus of virgin asphalt, this was because the virgin asphalt provided a better softening effect on the aged asphalt. Moreover, the predicted value of E* generally increased along with the number of layers N of asphalt division (the transition zone between the virgin and aged asphalt binders). Based on the agreement between predicted and experimental values, it was recommended to divide the asphalt layers into 50 layers. Under different degrees of diffusion and different RAP contents, power functions could be used to characterize the performance transition mode between virgin and aged asphalt layers. The E* of RAM was significantly affected by the performance transition mode. The higher the degree of diffusion or the higher the RAP content, the higher the dynamic modulus prediction.