Determination of Complex Modulus Values of Low-Density Polyethylene Modified Bitumen Obtained by Using Two Different Waste Types with Artificial Neural Networks

Abstract

The present study aims to solve an important environmental problem and improve the performance properties of bitumen by using two types of waste low density polyethylene (LDPE). For this purpose, two types of additives, LDPE (A) and LDPE (B), were added to the pure binder at the rates of 1%, 2%, 3% and 4% to obtain modified binders. Then, Dynamic Shear Rheometer experiments were applied on the binders under different temperatures and frequencies, and their behavior under these conditions was investigated. The complex shear modulus values obtained as a result of the experiment were estimated with Artificial Neural Network models created by training with different training algorithms. Experimental results showed that both additives increased the complex modulus values of the binder, with the LDPE (A) additive having higher complex modulus values compared to the LDPE (B) additive. In addition, it was determined that the model obtained with the Levenberg-Marquardt training algorithm gave the best results and it was concluded that the complex module values of asphalt binders can be successfully estimated using Artificial Neural Networks.