An investigation of different fitting functions to accurately model the compressive relaxation modulus master curve of asphalt mixes

Abstract

This research was carried out to rank the different fitting functions utilised to construct the compressive relaxation modulus master curve of asphalt mixes. To this end, 72 cylindrical dense-graded asphalt mixture specimens containing crushed stone aggregates with 60/70 penetration asphalt binder were fabricated using two aggregate gradations, two binder contents, two air void levels, and three ageing conditions with three replicates for each experimental combination. Uniaxial compressive relaxation modulus tests were conducted on the specimens at four temperatures using the trapezoidal loading pattern at a low level of input strain. Different functions, that is, Pure Power Law (PPL), Generalised Power Law (GPL), Modified Power Law (MPL), Modified Power Law Series (MPLS), Standard Sigmoid (SS), Generalised Logistic Sigmoid (GLS), and Prony Series (PS), were utilised as the fitting models, and the compressive relaxation modulus master curves of all the experimental combinations were constructed using the numerical shifting technique. The compressive relaxation modulus values resulted by the mentioned fitting functions were graphically and statistically compared to each other, and PS was found to be the most accurate function, followed by MPLS, GLS, MPL, SS, GPL, and PPL, respectively. Finally, the rankings determined in compression were compared to those previously obtained in tension for the same experimental combinations.