Optimising Hybrid Asphalt Mixtures and Enhancing Prediction Models using Response Surface Methodology

Abstract Although waste materials in asphalt mixtures are promising in improving pavement performance, they have attracted considerable attention owing to ecological issues and the imperative for sustainable infrastructure. Thus, this article assessed the Marshall stability efficiency of hybrid asphalt mixtures containing palm oil fuel ash (POFA), garnet waste, and sawdust. Approximately 0%, 3%, 6%, and 9% of hybrid materials by asphalt weight were employed to assess the conventional features of hybrid asphalt mixtures. A central composite design (CCD) within the response surface methodology (RSM) framework was also utilised to examine the impact of several hybrid materials on the mechanical features of hybrid asphalt mixtures. This process involved considering three input response variables (POFA, garnet waste, and sawdust contents), along with four Marshall stability parameters stability, flow, bulk density, and void in total mix (VTM). Consequently, a substantial correlation was yielded between the hybrid asphalt mixture and hot mix asphalt features. The hybrid contents presented a notable effect on the performance of the hybrid asphalt mixture. Among the samples, the 6% hybrid asphalt mixture demonstrated superior stability, with a Marshall stability value of 22.33 kN. Substantial correlations in the quadratic model were also observed between hybrid asphalt content and the mechanical performance of the hybrid asphalt mixture. The optimal percentages for hybrid materials were then determined using numerical optimisation and the resulting quadratic model. These values were 6.5%, 7.3%, and 6.5% for POFA, garnet waste, and sawdust. Given that all responses yielded a percentage error below 5%, this model revealed strong agreement and effectiveness. Compared to the commercial asphalt modification in terms of Marshall stability, this study confirmed the excellent precision of the prediction model while demonstrating that the optimised hybrid content formulations were better at similar amounts.