Comparative assessment of pavement fatigue life prediction approaches for use in performance based specification

Abstract

There is a need to incorporate long term pavement performance evaluation into the current specification through development of performance-related specification (PRS). PRS calls for the prediction of pavement life and development of pay adjustment factors based on expected life differences between as-design and as-built conditions. Different distress models have been developed and are in use by different researchers. Testing and analysis features, simplicity, test equipment and knowledge availability play a significant role in selection of the model for use. Nevertheless, to date, a comparative assessment of how the different analysis approaches affect pavement performance evaluation has not been undertaken. In this study fatigue performance evaluation is done on pavements with different surface mixture condition, varying asphalt and air void content to replicate a range of conditions in the field. Material characterization is done by performing complex modulus and fatigue testing. Pavement response is computed using three layer elastic analysis (LEA) programs. Fatigue life evaluation is done using three different fatigue life prediction models. The results from dynamic modulus testing show that there is an increase in stiffness with a decrease in asphalt and air void content which infers a reduction in tensile strain at the bottom of the asphalt layer. The tensile strain outputs from the different LEA programs were comparable and there is about a 20 % increase in strain for a change in bond condition from full bond to full slip. There is a distinction between predicted fatigue life due to the use different models. The highest variation is observed at low air void content level whereas the impact was lower at optimum and high air void content levels.