ABSTRACT
This paper summarizes a study dealing with the calibration/validation of the Texas mechanistic-empirical (TxME) flexible pavement design approach. In doing so, Texas-specific data were used, sourced from DSS, a Texas DOT maintained pavement performance database. Calibration coefficients were established for the layer rutting, bottom-up fatigue cracking and transverse (thermal) cracking models. For each distress, analysis of variance was conducted to establish the factors that affect prediction errors. Calibration factors were estimated through different optimization techniques such as the generalized reduced gradient (GRG) and curve fitting algorithms. The calibration coefficients resulted in significant quality of fit improvements compared to those obtained using the default calibration factors.
