Evaluating and correlating asphalt binder and mixture fatigue properties considering aging conditions

Asphalt fatigue cracking is a critical distress type for flexible pavement. This study utilized the state of practice binder and mixture cracking tests and parameters to investigate how asphalt binders' and mixtures' fatigue performance vary between the different mix designs and evaluate their changes with aging. The potential relationships between the measured binder properties and mixture performance were also explored. Five asphalt mixtures with disparate mix designs were evaluated and three different aging conditioning levels were conducted. The binder tests include the traditional time and frequency sweep and linear amplitude sweep test, while the mixture evaluations include the semi-circular bending and cyclic fatigue test. The results indicate that the different fatigue parameters from the tests can evaluate and rank the mixes differently. The correlation analysis using the Spearman correlation coefficient shows that the mixture cracking parameters that are more predominant by asphalt mixture's flexibility/brittleness show strong correlations (Spearman coefficient larger than 0.8) with binder cracking indices. The LAS fatigue life-related parameters generally do not correlate well with mixture fatigue indices (Spearman coefficient smaller than 0.6). Furthermore, a prediction model to estimate asphalt mixture's fatigue performance was developed based on linear regression and Artificial Neural Network (ANN) approach, which only requires the mix design info. and binder fatigue property as the inputs, significantly reducing the time and eliminating the effort in preparing and testing the asphalt mixture specimen for mixture fatigue performance measurement.