Mechanistic Design Guide Calibration for Pavement Rehabilitation

Project Details









Federal Highway Administration State Planning and Research Funding
Oregon Department of Transportation

Principal Investigator
Chris Williams

Director, AMPP

About the research

The Oregon Department of Transportation (ODOT) is in the process of implementing the recently introduced AASHTO Mechanistic-Empirical Pavement Design Guide (MEPDG) for new pavement sections. The majority of pavement work conducted by ODOT involves rehabilitation of existing pavements. Hot mix asphalt (HMA) overlays are preferred for both flexible and rigid pavements. However, HMA overlays are susceptible to fatigue cracking (alligator and longitudinal cracking), rutting, and thermal cracking. This study conducted work to calibrate the design process for rehabilitation of existing pavement structures. Forty-four pavement sections throughout Oregon were included. A detailed comparison of predictive and measured distresses was made using MEPDG software Darwin M-E (Version 1.1). It was found that Darwin M-E predictive distresses did not accurately reflect measured distresses, calling for a local calibration of performance prediction models. Darwin M-E over predicted total rutting compared to the measured total rutting and most of the rutting predicted by Darwin M-E occurs in the subgrade. For alligator (bottom-up) and thermal cracking, Darwin M-E underestimated the amount of cracking considerably as compared to in-field measurements. A high amount of variability between predicted and measured values was observed for longitudinal (top-down) cracking. The performance (punch-out) model was also assessed for continuously reinforced concrete pavement (CRCP) using Darwin M-E’s default (nationally calibrated) coefficients.

Four distress prediction models (rutting, alligator, longitudinal, and thermal cracking) of the HMA overlays were calibrated for Oregon conditions. It was found that the locally calibrated models for rutting, alligator, and longitudinal cracking provided better predictions with lower bias and standard error than the nationally (default) calibrated models. However, there was a high degree of variability between the predicted and measured distresses, especially for longitudinal and transverse cracking, even after the calibration. It is believed that there is a significant lack-of-fit modeling error for the occurrence of longitudinal cracks. The Darwin M-E calibrated models of rutting and alligator cracking can be implemented, however, it is recommended that additional sites be established and included in the future calibration efforts to improve the accuracy of the prediction models.