Issue |
MATEC Web Conf.
Volume 396, 2024
8th World Multidisciplinary Civil Engineering - Architecture - Urban Planning Symposium (WMCAUS 2023)
|
|
---|---|---|
Article Number | 02010 | |
Number of page(s) | 9 | |
Section | Construction Materials | |
DOI | https://doi.org/10.1051/matecconf/202439602010 | |
Published online | 24 May 2024 |
Artificial neural network approach to predict asphalt mixtures’ stiffness modulus based on testing frequency and temperature
1 Polytechnic Department of Engineering and Architecture, University of Udine, Via delle Scienze 208, 33100 Udine, Italy
2 Faculty of Civil Engineering, Czech Technical University in Prague, Thákurova 7/2077, 166 29 Prague, Czech Republic
3 Faculty of Civil Engineering, Warsaw University of Technology, Al. Armii Ludowej 16, 00-637 Warsaw, Poland
* Corresponding author: fabio.rondinella@uniud.it
To successfully guarantee the proper durability and serviceability of asphalt pavements, it is crucial to investigate asphalt mixtures’ performance and to design an accurate model to predict their mechanical behaviour. In Machine Learning, Artificial Neural Networks (ANNs) consist of a set of layered and interconnected artificial neurons capable of learning a complex function that maps the input to the target output. This study is specifically aimed at their implementation within the field of pavement engineering. The paper thoroughly discusses the development of an ANN-based methodology capable of predicting the stiffness modulus of an asphalt mixture (AM). The AM under investigation was prepared with spilite aggregate, a 50/70 penetration grade bitumen, and limestone filler. The volumetric properties of each specimen were first determined, and then the sequence of a 4-Point Bending Test was carried out under different conditions. Four testing temperatures (0, 10, 20, 30 °C) and eleven loading frequencies (0.1, 1, 2, 3, 5, 10, 15, 20, 30 and 50 Hz) were selected to investigate the asphalt mixture's mechanical behaviour. The resulting stiffness moduli represented the output of the designed neural model. Prediction accuracy was evaluated utilizing several goodness-of-fit metrics, and the results of this feasibility study proved to be very encouraging. They are certainly limited to the asphalt mixture under investigation. However, the high level of accuracy suggests that trained on a larger dataset, the developed methodology could allow the AMs’ mechanical behaviour to be predicted without the need to carry out the conventional tests that are usually expensive and time-consuming.
© The Authors, published by EDP Sciences, 2024
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.