Mechanistic-Empirical Rut Prediction Model for In-Service Pavements

Abstract

Rutting is a major mode of failure in flexible pavements. Development of accurate predictive rut performance models is an ongoing pursuit of the pavement engineering community. This has resulted in a plethora of rut prediction models ranging from purely mechanistic to empirical. Presented is the development of a mechanistic-empirical rut prediction model that uses data from 39 in-service flexible pavements from Michigan. The proposed model accounts for the rut contribution of the subgrade, subbase, base, and asphalt concrete layers. The model addresses inventory-type variables like pavement cross section, ambient temperature, and asphalt consistency properties. The applicability of the model was validated by using data from 24 Long-Term Pavement Performance–Global Positioning System (GPS) sites. For 19 of the 24 GPS sites, the predicted rut depth was within 5 mm of the measured rut depth.