Micromechanics-based dynamic modulus prediction of polymeric asphalt concrete mixtures

Hot-mix asphalt (HMA) mixtures are widely used in the surface layer of flexible pavements. HMA mixtures are a kind of polymer matrix composite composed of polymeric asphalt mastic with the inclusion of particulate-filled media and air voids. Nowadays, more and more petroleum-based polymers (epoxy, resin) have been widely used in the HMA mixtures, in which the polymer matrix accounts for around 10-20% by volume. The properties of polymer matrix play a key role in the performance of flexible pavements. Dynamic modulus (|E*|) of HMA mixtures is one of the fundamental engineering properties measured by the simple performance tester (SPT) and has been incorporated as a basic input in the American Association of State Highway and Transportation Officials (AASHTO) Mechanistic-Empirical Design Guide for flexible pavements. Although direct laboratory testing and empirical equations (such as the Witczak model and the Hirsch model) provide two ways to obtain the values of dynamic modulus of HMA mixtures, a predictive model based on the microstructure of HMA mixtures is more desirable. This paper presents a micromechanical model to predict the dynamic modulus of HMA mixtures. In this model, HMA mixtures are treated as a composite by embedding the mastic-coated aggregate particles into the equivalent medium of HMA mixtures. Based on the proposed model, closed-form equations were derived to predict the dynamic modulus value of HMA mixtures. The equations were able to take into account aggregate gradation and air void size distribution. In addition, laboratory experiments were conducted to verify the developed model. The dynamic modulus values of mastics and HMA mixtures were obtained through direct laboratory testing. The dynamic modulus of mastic was then used to predict the dynamic modulus of laboratory-prepared HMA mixtures with the newly developed model. It was found that the predicted dynamic moduli agreed reasonably well with the measured ones at high frequencies. The reasons for the discrepancy between measured and predicted dynamic moduli and the factors affecting the dynamic modulus were also explored in the paper. (C) 2007 Elsevier Ltd. All rights reserved.