Asphalt materials exhibit intrinsic healing capabilities. Researchers have carried out numerous investigations to characterize various aspects related to healing in asphalt materials since 1960s. Nevertheless, due to the complexity associated with the healing mechanism, a diverse and sometimes contradicting understanding in terms of experimental methods, models and numerical analysis exists. This paper attempts to give an overview of various asphalt healing studies with an emphasis on mechanical testing and analysis, and compares theories on the healing mechanism and their reported dependence on wide variety of factors including rest periods, temperature, aging, and moisture on the healing process. Bitumen, mastic, and asphaltic mixture levels are hereby included and the differences in the defined healing indices are discussed. The review shows that many of the commonly used analysis approaches can be challenged due to the co-existence of different phenomena during fatigue and healing. Possible post-processing methods that consider the non-linearity of the material response and decomposition of various modes of energy dissipation at multiple scales hold promise for unbiased quantification of healing.

In the present study, three-point bending tests are carried out using single-edge notched beam specimens of bitumen and mastic to quantify healing. Experiments are conducted at a controlled displacement rate of 1 mm per minute at -15 ℃. After the crack propagation, samples are given a rest period of 2 hours at 10 ℃ to promote healing before re-testing them. Two different analysis approaches appealing to linear elastic fracture mechanics and viscoelastic fracture mechanics are compared. In order to perform analysis based on viscoelastic fracture mechanics, the elastic-viscoelastic correspondence principle is used. The amount of healing after the rest period is quantified using various healing indices based on the recovery of stiffness, peak load, fracture toughness, fracture energy, and J-integral. From the analysis performed on bitumen and mastic samples, the study illustrates that the quantum of healing is different when comparing different healing indices. While the stiffness-based healing index demonstrated the healing ability of bitumen, other healing indices used in the study confirmed the higher healing potential of mastic. The healing based on critical value of J-integral shows a distinct difference in the healing of bitumen and mastic. The study emphasizes that the quantification of healing capacity when using different healing indices should be closely linked to its measured conditions.

The definition and quantification of healing of bituminous materials is a complex process that has given rise to a wide range of publications on healing test methods and post-processing procedures. The main issue in such investigations is that most of them still include the recovery due to the viscoelasticity of material into its healing potential. In this study, a new approach is used to quantify the healing devoid of viscoelastic recovery. The method is based on the assumption that the total strain in the material additively decomposes into viscoelastic and viscoplastic strains. The present study defines healing as the recovery of the quantum of viscoplastic strain. Creep and recovery experiments are carried out on bitumen and mastic at a wide range of stress levels at 5 and 15 ◦C. A 70/100 penetration grade bitumen and silica-based filler passing 75 µm sieve are used.The proposed method is applied to these experimental data sets to separate the viscoelastic and viscoplastic strains from the total strain. The healing measure used in this study considers the change in viscoplastic strain during the recovery period to the maximum value of viscoplastic strain generated during the creep. At any stress levels, the extent of viscoplastic strain and recovery is found to be considerably different between binder and mastic. The maximum values of viscoplastic deformation observed for bitumen and mastic are 31.7 and 6.2 %, respectively. The corresponding healing measures at the end of recovery period are 15.8 and 21.7 % for bitumen and mastic, respectively. At both test temperatures, the amount of healing is found to increase with the decrease in stress levels for both bitumen and mastic. Future work will focus on the separation of damage from the viscoplastic strain to validate a constitutive model that is able to simulate the damage and healing process.