Influence of the MEPDG Binder Characterization Input Level on the Predicted Dynamic Modulus for Idaho Asphalt Concrete Mixtures

摘要

The current version of the Mechanistic-Empirical Pavement Design Guide (MEPDG)includes two different models for the hot-mix asphalt (HMA) dynamic modulus (E*) prediction.In MEPDG levels 2 and 3 HMA stiffness characterization; users can choose either one of thesetwo models. The first model is the NCHRP1-37A viscosity-based model while the second one isthe NCHRP1-40D binder shear modulus-based model. The main difference between the twomodels is the binder stiffness characterization parameter. Moreover, MEPDG allows users tochoose between three hierarchical input levels for binder stiffness characterization.This paper focuses on the evaluation of the influence of the binder characterization inputlevel on the prediction accuracy of the MEPDG E* predictive models. In this study, a total of 22HMA mixtures commonly used in Idaho with six binder grades were investigated. Bindercharacterization data were obtained from Brookfield viscosity testing for MEPDG level 1analysis for conventional binders, Dynamic Shear Rheometer (DSR) testing for MEPDG level 1analysis for Superpave binders, and typical MEPDG values (MEPDG level 3 analysis).Comparison between laboratory measured and predicted E* values along with the threeinvestigated binder characterization cases, revealed that both MEPDG E* predictive modelsresulted in reasonable E* predictions. The 1-37A E* model along with MEPDG level 3 binderinputs yielded the most accurate and least biased E* estimates for Idaho mixtures. Nevertheless,the lowest prediction accuracy and highest bias in the E* estimates, especially at the highertemperatures, resulted from the 1-40D model with MEPDG level 3 binder characterization.