Evaluation of dynamic modulus predictive equations for asphalt mixtures with recycled asphalt pavement material

Abstract

The NCHRP 1-37A and NCHRP 1-40D models have been incorporated into the PMED program to predict dynamic modulus (|E*|) of asphalt mixes for Level 2 and Level 3 inputs to design and analyze flexible pavement structures. These empirical predictive equations were developed using databases that did not include asphalt mixes with reclaimed asphalt pavement (RAP) material. The objective was to evaluate the predictive capability of the existing |E*| predictive equations for Indian asphalt mixes with RAP. Fourteen mixes were used in the study; two dense-graded mixtures containing 0%, 15%, 25%, and 35% RAP and two gap-graded mixes containing 25%, 35%, and 45% RAP. The concordance correlation coefficient was used as an index to evaluate the predictive capability of existing |E*| equations compared to measured |E*|. The NCHRP 1-37A and Hirsch models showed very good precision in predicting |E*|. However, the prediction bias was very high, resulting in a poor overall agreement of predicted |E*| with measured values. The NCHRP 1-40D model showed very high precision and low bias, which resulted in reasonably good overall |E*| predictions compared to measured values. A new |E*| prediction equation was developed; the validation data showed a good overall agreement at all test temperatures and frequencies. The newly developed |E*| model was used to predict the phase angle of asphalt mixtures using the first-order derivative of the |E*| master curve, this method showed a good overall agreement with measured values at all test temperatures and frequencies, except for high temperatures and low frequencies.