Investigation of layered elastic theory prediction accuracy for asphalt concrete pavement design using micromechanical viscoelastic finite element modeling

Abstract

In this study, predictions of full-scale micromechanical (MM) finite element (FE) models, developed from X-ray computed tomography images of asphalt concrete samples that were sawn from the accelerated pavement test sections, were used to evaluate the accuracy of layered elastic theory (LET) models that are used in pavement design today. First, MM FE and LET models were both calibrated using the measured strain gauge responses. Predictions of calibrated models were compared to evaluate the reasonableness of LET model outputs at high temperatures. Second, asphalt concrete stiffnesses measured in the laboratory were directly used for LET model development without performing any strain gauge calibration to evaluate the actual predictive capability of LET models in pavement design by using the calibrated MM FE model outputs as the ground truth. Recommendations were also made for future use of the MM FE models to improve the predictive capability of LET models.