Evaluation of Dynamic Modulus Predictive Equations for Southeastern United States Asphalt Mixtures

Mechanistic–empirical pavement design has recently come to the forefront of design, with many states looking to the Mechanistic–Empirical Pavement Design Guide (MEPDG) as their future primary design method. Central to any flexible pavement mechanistic–empirical design framework is the characterization of hot-mix asphalt (HMA) through the use of dynamic modulus E^*. Because expensive and specialized equipment is required to measure E^* in the laboratory, predictive E^* equations that use binder and mixture properties to estimate E^* at various frequencies and temperatures must be evaluated. There are currently three global E^* predictive equations, two of which are used at Levels 2 and 3 in the MEPDG. These three models (Witczak 1–37A, Witczak 1–40D, and Hirsch) were evaluated with the use of 18 HMA plant-produced, lab-compacted mixtures (representative of general-use mixtures used in the southeastern United States) that were placed at the 2006 National Center for Asphalt Testing test track. E^* predictions were made at three temperatures and three frequencies for direct comparison with measured values. The Witczak models had the greatest deviation from measured values, and the Witczak 1–40D model overestimated E^* by approximately 61%. The Hirsch model most accurately predicted the moduli for the 2006 test track mixtures. Calibration of the Hirsch model for these mixtures indicated that the Poisson ratio selected for the asphalt binder had little effect on its prediction capabilities. The little improvement resulting from calibration proves that this step is unnecessary.