Volume 24, Issue 1 (2024)
MCEJ 2024, 24(1): 7-23 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
UMAR MOHAMAD SALEH I, Effati M, Jalali M, Mirabdolazimi S M. Evaluation of the self-healing behavior of warm mix asphalt with the help of failure mechanism and in terms of induction characteristics using Artificial Neural Networks. MCEJ 2024; 24 (1) :7-23
URL: http://mcej.modares.ac.ir/article-16-67195-en.html
URL: http://mcej.modares.ac.ir/article-16-67195-en.html
1- Master's student in civil engineering, road and transportation, GUilan University-Rasht
2- Assistant Professor, Department of Civil Engineering, Faculty of Engineering, University of Guilan ,Meysameffati@guilan.ac.ir
3- University of Guilan
4- Assistant Professor, Department of Civil Engineering, Faculty of Engineering, University of Guilan
2- Assistant Professor, Department of Civil Engineering, Faculty of Engineering, University of Guilan ,
3- University of Guilan
4- Assistant Professor, Department of Civil Engineering, Faculty of Engineering, University of Guilan
Abstract: (1333 Views)
Improving the self-healing performance of warm mix asphalt depends on several factors and parameters that are highly interdependent and have significant complexity. In this study, the self-healing performance of warm mix asphalt was investigated using artificial intelligence and artificial neural network of multi-layer perceptron and radial base with two hidden layers. To conduct this study, two additives Sasobit and Zycotherm were used. The three-point bending test was performed at two temperatures of 25 and -16 degrees Celsius and with two crack lengths of 10 and 20 mm, and the fracture toughness, fracture energy and critical load indices were determined for each of the states. Asphalt samples were subjected to induction heating at two frequencies of 87 and 88 kHz and three induction times of 60, 90 and 120 seconds. The results of sensitivity analysis in two artificial neural network models showed that in the MLP network, the fracture toughness parameter had the greatest impact on the output. It was also observed that the test temperature parameter had the highest sensitivity coefficient in the RBF network. The results showed that in the perceptron neural network with two layers in the test section, the root mean square error (RMSE) values increased from 10.46 in the first model to 4.27 in the fourth model. The results of the basic radial artificial neural network showed that the addition of input parameters reduced the root mean square error (RMSE) value of the test section from 10.56 to 4.35. The results of MLP and RBF network estimation have shown that the addition of input variables to the model has increased NS in all three parts of test, train and validation. In this way, in the MLP network, the value of NS in the test section has reached from 0.45 to 0.90 and the estimation accuracy has doubled. In the RBF network, similarly to MLP, with the addition of the NS parameter, the NS value has increased from 0.44 to 0.90. Also, the results of this study showed that in both types of MLP and RBF networks, the value of R2 in the second group was higher than the first group in all test, train and validation sections. In general, the results of this study showed that the artificial neural network has appropriate performance and accuracy due to the nature of learning and the ability to train from the previous laboratory results in estimating the self-healing ability and modeling the complex relationship of the influential input variables, and the use of the proposed intelligent model by reducing the of experiments and cost can be effective in evaluating the self-healing behavior of warm mix asphalt.
Keywords: Self-healing, Sasobit, Zycotherm, Multilayer Perceptron Artificial Neural Network, Basis Radial Function
Article Type: Original Research |
Subject:
Roads and Transportation
Received: 2023/01/31 | Accepted: 2023/10/19 | Published: 2024/02/29
Received: 2023/01/31 | Accepted: 2023/10/19 | Published: 2024/02/29
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |