Hot Mix Asphalt Dynamic Modulus Prediction Models Using Neural Networks Approach

The primary objective of this study is to develop a simplified Hot Mix Asphalt (HMA) dynamic modulus (|E*|) prediction model with fewer input variables compared to the existing regression based models without compromising prediction accuracy. ANN-based prediction models were developed using the latest comprehensive |E*| database that is available to the researchers containing 7,400 data points from 346 HMA mixtures. The ANN model predictions were compared with the existing regression-based prediction models which are included in the latest Mechanistic-Empirical Pavement Design Guide (MEPDG). The ANN based |E*| models show significantly higher prediction accuracy compared to the existing regression models although they require relatively fewer inputs. The findings of this study present a “paradigm shift” in the way the hot-mix asphalt material characterization has been handled by pavement materials engineers.