Journal
JOURNAL OF MATERIALS IN CIVIL ENGINEERING
Volume 30, Issue 4, Pages -Publisher
ASCE-AMER SOC CIVIL ENGINEERS
DOI: 10.1061/(ASCE)MT.1943-5533.0002242
Keywords
Recycled asphalt shingles; Artificial neural networks; Dynamic modulus
Categories
Funding
- National Natural Science Foundation of China [51278188]
- Fundamental Research Funds for the Central Universities
- Young Core Instructor Foundation from the Education Commission of Hunan Province
Ask authors/readers for more resources
The difference between hot-mix asphalt (HMA) containing recycled asphalt shingles (RAS) and virgin HMA in terms of composition and properties causes difficulties when engineers try to predict the performance of HMA containing RAS. This study develops a prediction model based on artificial neural networks (ANN) to predict the dynamic modulus (E*) of HMA containing RAS. The E* database used in this study to develop the ANN model contains 1,701 sets of experimental data, which were obtained from four different demonstration projects. In order to train and test the model, the data were randomly divided into two different subsets: one is for training, containing 1,361 data points, and the other is for testing, containing 340 data points. The input parameters of the proposed model included percent passing a #200 sieve (200), cumulative percent retained on a #4 sieve (4), cumulative percent retained on a 9.5-mm sieve (38), cumulative percent retained on a 19-mm sieve (34), air voids (Va), effective binder content (Vbeff), viscosity of the asphalt binder (), loading frequency (f), and RAS contents (pa, by weight of total aggregate). The sensitivity analysis of these parameters was performed by correlating each model parameter with E*. The proposed ANN model was compared with the Iowa model and showed significantly higher prediction accuracy than the Iowa model. It can be concluded that the proposed ANN model has great potential to be used as a tool to predict the E* of asphalt mixtures containing RAS.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available