Kinetics of healing of asphalt mixtures

Promotion of healing of asphalt mixtures is significant to enhance the resiliency and extend the service life of asphalt pavements. Extensive laboratory measurements of effects of healing and field observations have proven its impact. Many approaches have been developed to explain and model healing of asphalt mixtures based on its mechanisms and/or mechanical features. However, due to the complexity caused by the co-existence of several coordinated rate processes during healing, there are still unclear aspects in terms of mechanisms and appropriate representation of each rate process. Target this issue, this study proposes to investigate healing of asphalt mixtures using kinetic and thermodynamic quantities and formulations, aiming at clarifying the relationships among the coordinated processes involved in healing and developing an approach to predict the rate of each process. Healing is found to contain two coordinated rate processes: free energy change and bond restoration. Healing rate is proportional to the rates of these two processes, and the healing activation energy is the product of the activation energy of the free energy change and that of the bond restoration.In order to calculate each activation energy and rate constant, the recoverable strain energy (RSE) and recoverable pseudo strain energy (RPSE) measured in the rest time of a creep and step-loading recovery (CSR) test are used. The RPSE in the intact material and the surface energy constitute the total free energy change for healing. A three-parameter model is proposed to simulate a rate process like the change of the RSE or free energy change with time, and a dimensionless logarithmic rate is defined to obtain a representative change rate. This representative rate is ideal to be used as the rate constant in the Arrhenius equation to compute the activation energy. The activation energies for viscoelastic recovery, healing, free energy change, and bond restoration are determined for two types of asphalt mixtures at unaged and aged states. The relations between different activation energies are clarified so as the relationships among different rate processes. Furthermore, the healing activation energy obtained from this kinetic approach is proven to be the same as that calculated using the damage density from a mechanistic approach. A method to generate a damage density-based healing curve at any rest time at any temperature using just energy measurements is proposed in this study.