ABSTRACT

Artificial neural networks (ANNs) and G_b*-based regression models were used for the prediction of the dynamic modulus (|E*|) of South Carolina’s hot mix asphalt mixtures (HMAs) the majority of which contained recycled asphalt pavement (RAP). Models’ training and testing were done using a database that contained 1656 |E*| values from 93 HMA mixtures. G_b*-based models included the Hirsch, revised Hirsch, Bari-Witczak, revised Bari-Witczak, Al-Khateeb 1, Al-Khateeb 2, NCHRP 1-40D, and the simplified global models. The results showed that G_b*-based regression models had a significant bias in prediction; Coupling VMA and G_b* had the most influence on |E*|; four-layer ANNs generally had a better performance than three-layer ANNs on using Hirsch model’s related inputs; ANN 3-15-15-1 and ANN 8-15-15-1 (developed with similar input variables as the Hirsch and Witczak regression models respectively) showed very high performance of R² > 0.994 on testing. Therefore, ANNs could be considered to capture the influence of the binders’ rheological properties, mixture’s volumetric properties, and RAP on |E*| of HMA mixtures far better than regression-based models.

摘要

人工神经网络 (ANN) 和基于 G _b * 的回归模型用于预测南卡罗来纳州热拌沥青混合料 (HMA) 的动态模量 (|E*|)，其中大部分包含再生沥青路面 (RAP) . 模型的训练和测试是使用包含 1656 |E*| 的数据库完成的。来自 93 种 HMA 混合物的值。基于G _b * 的模型包括 Hirsch、修订的 Hirsch、Bari-Witczak、修订的 Bari-Witczak、Al-Khateeb 1、Al-Khateeb 2、NCHRP 1-40D 和简化的全局模型。结果表明，基于_Gb *的回归模型在预测上存在显着偏差；耦合 VMA 和 G _b* 对 |E*| 的影响最大；在使用 Hirsch 模型的相关输入时，四层人工神经网络通常比三层人工神经网络具有更好的性能；ANN 3-15-15-1 和 ANN 8-15-15-1（分别使用与 Hirsch 和 Witczak 回归模型相似的输入变量开发）在测试中显示R ²  > 0.994 的非常高的性能。因此，可以考虑使用人工神经网络来捕捉粘合剂的流变特性、混合物的体积特性和 RAP 对 |E*| 的影响。HMA 混合物的性能远优于基于回归的模型。