Application of AI on moisture damage of modified asphalt binders

Damage of asphalt pavements relating to moisture is being researched with many decades. But the exact reason for the moisture damage and the mathematical expression is unknown. To ascertain such effects concerning adhesive forces, a nanoscale experiment with an atomic force microscope (AFM) is going to be performed in this study. For making samples, Styrene-butadiene-styrene (SBS) polymer were mixed with base asphalt binder, which is tested under AFM on the glass substrate. In both dry and wet conditions, asphalt samples are processed. The correlation of moisture content damage in SBS-modified asphalts and lime can be predicted in this study through an artificial intelligence rule. In dry condition, asphalt base binders have shown greater adhesion/cohesion binding values than polymer-modified asphalt sample. In wet conditions, it shows an adverse effect. Asphalt base binders are more responsive to damp than polymer-modified asphalt binders. Based on these points, several artificial intelligences (AI) were applied and the ANFIS model (in contrast to MLP and SVM) showed great assurance. Relative error average was to be very low: 0.02 and 0.03, correspondingly, for the observed and projected data, also showing the stable presentation of the model. The dry sample was run for all three neural network models for making a statistical study, and it was detected that MLP is better than the other two models.