The viscoelastic properties of epoxy asphalt and straight asphalt mixtures were investigated through uniaxial compression, dynamic loading, and creep tests. Epoxy asphalt exhibited higher stiffness and compressive strength than straight asphalt, showing predominantly elastic behavior, while straight asphalt showed barrel-shaped deformation, indicating a viscous response. Based on these results, a new viscoelastic model was proposed by connecting the Kasai model, which captures the peak phase angle response, and the fractional Maxwell model, which reproduces residual deformation, in series. The proposed model reproduced static and dynamic behaviors of epoxy and straight asphalt more accurately than the Burgers and FDM4 models, and with comparable accuracy to the GFDZ model. Moreover, it achieved this with fewer parameters, providing a more efficient framework for engineering applications. Temperature dependence was evaluated using the time–temperature conversion ( TTC ). For epoxy asphalt, a master curve was constructed at a reference temperature of 40 °C, and temperature-dependent coefficients were introduced to unify static and dynamic properties. While TTC is typically applied to dynamic properties, this study proposed a framework to extend its applicability to static properties. This enabled description of both static and dynamic temperature dependence using material properties at the reference temperature. Although the unified coefficients showed slightly larger discrepancies than fitting at each temperature, they captured trends within the tested range. These findings provide a laboratory-scale constitutive basis for subsequent structural analyses, limited to the tested 0–500 s linear viscoelastic conditions and temperature range, and requiring validation under longer-duration loading and different confinement states. • A new viscoelastic model for epoxy asphalt is proposed using fractional calculus. • The model combines the Kasai and fractional Maxwell models in series. • Captures temperature-dependent static and dynamic behaviors with fewer parameters. • Temperature dependence is incorporated via a master curve and shift factors.
Nakamura et al. (Wed,) studied this question.