Genetic programming-based multi-objective optimization for enhancing fracture performance in nanolayer-reinforced asphalt mixtures to estimate the initial quality and maintenance life

Abstract

In this study, the fracture behavior of hot mix asphalt (HMA) and warm mix asphalt (WMA) reinforced with two nanolayer additives called molybdenum disulfide nanoparticles (MDSN) and reduced graphene oxide nanoparticles (RGON) was investigated in two time periods, the beginning of exploitation and four years after exploitation. In this regard, the effective fracture toughness (K_eff-I/II), fracture energy (FE), fracture flexibility, including flexibility index (FI), toughness index (TI), and cracking resistance index (CRI), fracture stiffness (FS), including tensile stiffness index (TSI) and tensile strength (TS), and damage factor (DF) were examined. Finally, to ensure the accuracy of the results, the Pearson correlation coefficient (PCC) statistical method was employed to analyze the correlation between the fracture indices. Moreover, the genetic programming (GP) model was used to provide a prediction model for obtaining optimum MDSN and RGON contents in CRI and TSI models using multi-objective optimization (MO). The results showed that RGON and MDSN reinforced mixtures showed acceptable performance against tensile-shear deformations under 0 and 1 freeze–thaw cycles (FTC) cycles. Also, 0.6 and then 0.3 % MDSN and RGON had the best performance regarding fracture toughness, energy, flexibility and stiffness, but, MDSN mixtures were better than RGON mixtures. PCC analysis showed a strong correlation between fracture resistance and stiffness, highlighting the interdependence of these properties. GP model results indicated that the relationship between predicted and actual values of CRI and TSI was appropriately described by R values of 99.23 and 98.42 % for CRI model and 98.96 and 98.57 % for TSI model of HMA, and 99.19 and 98.52 % for CRI and 98.95 and 98.96 % for TSI model of WMA mixtures modified with MDSN and RGON, respectively. MO results showed that 0.539 and 0.514 % for HMA and 0.562 and 0.427 % for WMA were the best optimal MDSN and RGON contents to simultaneously maximize CRI and TSI values, respectively. These findings provide valuable insights for policymakers and project managers, highlighting the potential of MDSN and RGON nanomaterials to significantly enhance the durability and cost-effectiveness of both HMA and WMA pavements.