Abstract
Two parameters are currently used to characterize the crack propagation resistance of asphalt mixtures at intermediate temperatures using the semi-circular bending (SCB) test: the critical strain energy release rate (Jc) and the flexibility index (FI). Both parameters are calculated from the total energy to failure defined as the area under the load, load-line displacement curve. However, this area includes an unknown amount of dissipated energy which is not useful to resist cracking. Also, the amount of dissipated energy may vary between specimens. Hence, either Jc or FI cannot be used as single parameters to characterize crack propagation resistance of asphalt mixtures. In this study, a new method was introduced to characterize the crack propagation resistance of asphalt mixtures using two parameters that are not dependent on the total energy to failure. In this method, mixtures are tested at multiple intermediate temperatures. A mathematical model is used to fit the relationship between the load and the load-line displacement for each SCB sample from the beginning to the end of loading. Two parameters are derived from the mathematical model: initial bending stiffness, defined as the load divided by the load-line displacement just before loading, and normalized Shannon entropy, which is a statistical parameter that is used to indicate the mechanical response of the specimen. Mixtures that have higher initial bending stiffness require higher stresses to initiate cracking, while higher Shannon entropy indicates higher resistance to crack propagation. At intermediate temperatures, initial bending stiffness decreases while the Shannon entropy increases by increasing testing temperature. So, based on initial bending stiffness and Shannon entropy definitions, as temperature increases, it is easier to initiate but harder to propagate a crack. Minimum acceptable initial bending stiffness and minimum acceptable Shannon entropy at a temperature of interest can be selected and used as pass/fail criteria. Finally, the relationship between initial bending stiffness and normalized Shannon entropy can be used to distinguish between mixtures. The sensitivity of the new method to different parameters that affect crack propagation resistance of asphalt mixtures was studied and compared to the FI measured at 25°C according to AASHTOTP124-16. The selected parameters were: binder source, binder PG, binder content, reclaimed asphalt pavement content, recycled asphalt shingles content, aging, warm mix asphalt, and aggregate type. The new method was also applied to plant-produced mixtures with known field cracking performance. Results showed that the new method correlated better than the FI with the field data.