Evaluation of the correlations between laboratory measured material properties with field cracking performance for asphalt pavement

Cracking is one of the primary distresses for asphalt pavements. There are many types of laboratory tests developed to evaluate the cracking performance of asphalt material, however, relatively little attention has been dedicated to evaluating and correlating the laboratory measured properties with the actual pavement performance. The main objective of this study is to investigate the relationships between the various laboratory measured binder/mixture properties with the actual pavement cracking (including both fatigue and thermal cracking) performance while also considering the important mix design and pavement structure parameters. Field pavement performance data were collected from 23 project sites with one control HMA section and at least one WMA section at each site. Laboratory testing was performed on the field cores taken from these sections as well as the corresponding extracted and recovered binders. Advanced statistical analysis method including the Pearson correlation and Spearman rank correlation coefficient were then employed to evaluate the correlations between the disparate laboratory measurements and actual pavement performance data. Results show that T*D (thickness of HMA/WMA layer*vertical failure deformation) parameter shows the good correlation with the length of field fatigue cracking, while the Pb%*ε_b-low (mixture binder content*binder failure strain) parameter shows the good correlation with the length of field thermal cracking. These correlations do not change with varying the pavement type (HMA or WMA). Based on the database generated in this study, a preliminary threshold value of 50 for T*D parameter and 10 for Pb%*ε_b-low parameter are proposed to minimize and control the cracking problem of asphalt mixtures in general.