Creep compliance is one of the fundamental tests of Mechanistic- Empirical flexible pavement design procedure in the AASHTO 2002 design guide. In this research, a new Artificial Neural Network model for estimating the HMA creep compliance with the generalization ability of R=0.949 has been developed successfully using feed forward multi layer perceptron Artificial Neural Networks (ANNs) with Levenberg-Marquardt backpropagation training algorithm. The ANN model has 14 inputs including selected percent passings from aggregate gradation curve, asphalt content, asphalt penetration degrees at 77 and 125 Fahrenheit , asphalt penetration index and test temperature. The model has 8 outputs consist of creep compliances of Hot Mix Asphalt at 1, 2, 5, 10, 20, 50, 100 seconds and Poisson coefficient. All needed data has been extracted from LTPP database 2007 by developing many programs by SQL in Access 2000. As the result, 975 creep compliance tests data and 975 associated aggregate tests data series and 975 associated asphalt tests data series has been extracted successfully for ANNs training and evaluation.
Hasani, A., & Heydaripanah, A. (2010). Estimation of Creep Compliance of Hot Mix Asphalt by Artificial Neural Networks. Journal of Civil and Surveying Engineering, 44(1), -.
MLA
Abolfazl Hasani; Ali Heydaripanah. "Estimation of Creep Compliance of Hot Mix Asphalt by Artificial Neural Networks". Journal of Civil and Surveying Engineering, 44, 1, 2010, -.
HARVARD
Hasani, A., Heydaripanah, A. (2010). 'Estimation of Creep Compliance of Hot Mix Asphalt by Artificial Neural Networks', Journal of Civil and Surveying Engineering, 44(1), pp. -.
VANCOUVER
Hasani, A., Heydaripanah, A. Estimation of Creep Compliance of Hot Mix Asphalt by Artificial Neural Networks. Journal of Civil and Surveying Engineering, 2010; 44(1): -.
)