Verification of Rutting Predictions from Mechanistic–Empirical Pavement Design Guide by Use of Accelerated Loading Facility Data

As a part of FHWA's Accelerated Loading Facility (ALF) experiment, asphalt pavement lanes were built with six different asphalt binder types and two different thicknesses and were tested for rutting and fatigue cracking. Loose asphalt mixtures were taken from each lane at the time of construction and compacted in the laboratory. These plant-produced specimens compacted in the laboratory and other laboratory-produced specimens were subjected to dynamic modulus and flow number tests with a simple performance tester (SPT). Field rutting measurements of the ALF lanes were conducted during and after the loading. A good agreement between the permanent strains measured by the flow number test and the ALF rutting measurements was observed for the initial rutting. The rutting of asphalt mixture layers of the ALF lanes was also predicted by using Level 1 and Level 3 analyses with Mechanistic–Empirical Pavement Design Guide (MEPDG) software by using the detailed data that were available for the ALF lanes. Comparison of the predicted rutting with the rutting measured on the ALF indicated that the rutting predicted by Level 3 analysis with MEPDG software, based on the asphalt stiffness equation, agreed somewhat with the actual rutting on the ALF lanes. However, the rutting predicted by Level 1 analysis, based on the actual SPT-measured dynamic modulus, significantly over-predicted the amount of rutting. The NCHRP Project 1-37A permanent deformation prediction models used in MEPDG software were also used separately to predict the rutting of the ALF lanes by using the predicted and the measured stiffness values. Results similar to those obtained by Level 1 and Level 3 analyses with MEPDG software were obtained. The overprediction of rutting by Level 1 analysis with MEPDG software may be related to the calibration of the NCHRP Project 1-37A permanent deformation equation. The model seems to have been calibrated by using the stiffness predicted from the NCHRP Project 1-37A equation, which is proven to overpredict stiffness at high temperatures, rather than the tested SPT stiffness and, thus, can provide reasonable rutting estimates only for Level 3 analyses (by use of the predicted stiffness).