Predicting the Impact of Crumb Rubber Size on the Rutting Resistance of Crumb Rubber Modified Asphalt Mixtures Using Gene Expression Programming

Crumb rubber modified asphalt (CRMA) offers several advantages such as increased rutting resistance and enhanced stability. However, few studies have explored the relation between CR size and rutting resistance of CRMA mixtures at different temperatures as well as the development of prediction models to assess the performance of CRMA mixtures. Therefore, this paper presents a novel prediction model that can be used to estimate the impact of crumb rubber size on the rutting resistance of CRMA mixtures at different temperatures. Two different base binders (60/70 and 80/100) and three different crumb rubber particle sizes (mesh # 20, mesh # 40, and mesh # 80) were utilized in this study to ascertain and predict the impact of crumb rubber size on the rutting resistance of asphalt mixtures at three different test temperatures (25°C, 40°C, and 60°C). Marshall stability test and Hamburg Wheel Tracking Test (HWTT) were conducted to ascertain modified asphalt mixtures’ overall stability and rutting resistance. Evolutionary machine learning technique, gene expression programming (GEP) was used to build the predictive model. The GEP model was developed based on experimental results of HWTT for CRMA incorporating different crumb rubber sizes. The data was divided into two validation and testing sets and different input variables such as percent of binder replaced (PBR), voids filled with asphalt (VFA), density of CRMA mixtures, percentage of virgin binder (PBR) used, mesh size, binder type, and temperature at which the rutting test was conducted were used to develop the GEP model. The developed prediction model was found to be highly precise as depicted by the R, root mean squared error (RMSE), and mean absolute error (MAE) values. The sensitivity analysis confirmed that the CR particle size along with the testing temperature had the most significant impact on the rutting resistance of CRMA mixtures. Finally, parametric analysis was conducted which further justified the potential use of the developed GEP model to predict the rutting resistance CRMA mixtures.