Abstract

Thirteen hot asphalt mixtures were tested for fatigue in four-point beam configuration. Theresults were analysed using the MEPDG fatigue model and a modified version of it, which included binders' thermal susceptibility as an additional forecasting variable. Seven mixtures were dense-graded, three were stone mastic asphalts and three were semi-open. Three mixtures used unmodified asphalt binders UB, while the rest had SBS polymer modified binders PMB. The fatigue lives of PMB mixtures were one order of magnitude higher than UB ones. The results indicated that the higher the degree of modification in the binder, the higher the fatigue life. No significant differences in fatigue life were observed between the three aggregate gradations. The MEPDG fatigue model parameters were calibrated using the test results. The predictions showed very low accuracy. When including the binders' thermal susceptibility as an additional forecasting variable, significant improvements were achieved in the accuracy of the predictions.