Artificial Neural Network Approach to Estimating Stiffness Behavior of Rubberized Asphalt Concrete Containing Reclaimed Asphalt Pavement

Feipeng Xiao, Serji N. Amirkhanian

Journal of Transportation Engineering · 2009 · 60 citations · 22 references

Concepts

TL;DR

Accurately predicting the stiffness of asphalt pavements is difficult due to the complex behavior of materials under various loading, pavement structure, and environmental conditions. This study explores the utilization of the artificial neural network (ANN) in predicting the stiffness behavior of rubberized asphalt concrete mixtures with reclaimed asphalt pavement (RAP). The authors constructed 296 asphalt mixture beams varying rubber type, RAP source, and rubber content, tested at 5 °C and 20 °C, and used a regression approach with seven engineering‑property inputs to build ANN models predicting stiffness. Results showed the ANN techniques to be more effective in predicting the fatigue life of the modified mixture than traditional regression‑based prediction models.

Abstract

Accurately predicting the stiffness of asphalt pavements is difficult due to the complex behavior of materials under various loading, pavement structure, and environmental conditions. This study explores the utilization of the artificial neural network (ANN) in predicting the stiffness behavior of rubberized asphalt concrete mixtures with reclaimed asphalt pavement (RAP). A total of 296 asphalt mixture beams were constructed from two different rubber types (ambient and cryogenic), two different RAP sources, and four rubber contents (0, 5, 10, and 15%). All samples were tested at two different testing temperatures of 5 and 20°C. The regression statistical method was used to predict the stiffness behavior of these mixtures via the 7 input variables covering the material engineering properties of the asphalt beams. In addition, the data were organized into 5 independent variables and one dependent variable (the stiffness values of the modified mixture beams) in ANN models. Results showed the ANN techniques to be more effective in predicting the fatigue life of the modified mixture than traditional regression-based prediction models.

References

22