Development of a Novel Prediction Model for Interface Shear Strength in Asphalt Pavement Using the CART Model

Interface bonding between asphalt layers plays a vital role in ensuring the proper functionality of pavement structures. Interlayer Shear Strength (ISS) is recognized as an indicator quantifying the interface bonding quality. Consequently, accurate evaluation and prediction of the ISS is imperative in determining the performance of asphalt pavement structures. By conducting laboratory experiments and employing machine learning (ML) techniques, this research aims to predict and assess the ISS in asphalt pavement. In this regard, the classification and regression trees (CART) model was proposed based on measured data collected from laboratory experiments. Three experimental factors of curing temperature, normal stress, and tack coat application rate were selected as variables. The findings showed that the developed CART model explained over 98% of the experimental data in a relatively short period. The curing temperature was found to have the most significant influence on the ISS, followed by normal stress and tack coat dosage. Moreover, a parametric analysis of the interaction effects of input parameters on the ISS revealed that higher curing temperature and lower normal stress reduced the ISS. In contrast, a high tack coat application rate and low normal stress corresponded to a lower ISS of the asphalt pavement. The outcomes of this study could pave the way for the realization of a reliable and efficient design of interlayer bonding between asphalt pavement layers.