Local calibration and Implementation of AASHTOWARE Pavement ME performance models for Oklahoma pavement systems

Abstract

ABSTRACT AASHTOWare Pavement ME Design software was developed as an advanced pavement design tool based on the mechanistic-empirical pavement design guide. Existing performance prediction models utilised in this design guide are nationally calibrated, and it is essential to calibrate these models based on the specific local materials and environmental conditions, and pavement performance information. In this study, Pavement ME Design prediction models, including IRI, rutting, top-down and bottom-up fatigue cracking, and transverse cracking models were calibrated for the state of Oklahoma. The predictions from distress and international roughness index models were evaluated and compared with the measured distress values, and the accuracy and bias terms of each model were determined. Nationally calibrated models showed large errors and significant bias values, which asserts the need for local calibration. Locally calibrated coefficients of distresses and international roughness index models for the Oklahoma pavement system were determined. Predictions from calibrated models show that the use of calibrated coefficients improves Pavement ME predictions and the design of flexible pavements in the state. The impact of the local calibration of Pavement ME on flexible pavement designs was analysed. Locally calibrated models lead to lower IRI prediction and higher fatigue cracking and rutting in the designs.