EVALUATION AND CALIBRATION OF DYNAMIC MODULUS PREDICTION MODELS OF ASPHALT MIXTURES FOR HOT CLIMATES

Abstract

The dynamic modulus (E*) of asphalt mixtures is considered a primary entry in Mechanistic-Empirical (ME) pavement design and analysis. Various models have been published and aimed to estimate the modulus on the basis of the mixture volumetrics and material features. This study aims to review four commonly incorporated dynamic modulus prediction models of Hirsch, Alkhateeb, Witzack 1-37A, Witzack 1-40D and validate and calibrate Hirsch and Alkhateeb models for use in Qatar. Based on the study outcomes, the Hirsch model showed a high prediction accuracy of asphalt mixture moduli before calibration with a coefficient of determination (R2) of 87.2% between predicted and measured values. This R2 value is improved after calibration to 89.2%. Alkhateeb model, on the other hand, had a R2 of 70.8% before calibration, which also improved to 89.2% after calibration. Based on the study results, it is recommended to use the calibrated Hirsch or Alkhateeb model in Qatar instead of the uncalibrated version of the models. The moduli predicted by the Hirsch model before and after calibration were employed in this study to perform a mechanistic-empirical analysis of typical pavement structures in Qatar. According to the findings, the percent change in the predicted fatigue due to the use of the calibrated Hirsch model reached more than 50% with an average value of 17.33%, while the percent change in rutting reached 14% with an average value of 3.65%. These results highlight the importance of using locally calibrated models to improve dynamic modulus predictions performance.