Researchers Leverage AI to Enhance Asphalt Durability with Graphene Oxide

The University of Transport Technology in Vietnam has developed an innovative approach to enhance the durability of asphalt pavements by utilizing Graphene Oxide (GO) as a modifier. In their study, titled “Application of Extreme Gradient Boosting in Predicting the Viscoelastic Characteristics of Graphene Oxide Modified Asphalt at Medium and High Temperatures,” researchers employed a machine learning technique to accurately predict key viscoelastic properties of GO-modified asphalt.

As the demand for robust infrastructure increases due to rising traffic loads and the adverse effects of climate change, traditional asphalt pavements struggle to meet performance standards. The incorporation of Graphene Oxide has shown significant improvements in the mechanical properties of asphalt. Notably, it enhances anti-aging performance and high-temperature stability, which are crucial for maintaining road quality.

Understanding the viscoelastic characteristics of GO-modified asphalt, specifically the complex modulus (G*) and phase angle (φ), is vital for engineers. However, conventional experimental methods to measure these properties are often complex, costly, and time-consuming. To address this challenge, the researchers constructed an extreme gradient boosting (XGB) model that predicts G* and φ based on data collected from previous experiments.

Promising Results from Machine Learning Application

The study analyzed two datasets comprising 357 samples for G* and 339 samples for φ. By integrating nine input parameters representing various asphalt binder components, the XGB model demonstrated exceptional predictive capabilities. The coefficient of determination (R2) values for G* and φ were recorded at 0.990 and 0.9903, respectively, while the root mean square error (RMSE) values stood at 31.499 and 1.08.

The predictive accuracy of the XGB model was further validated through comparisons with experimental results and five alternative machine learning models, including artificial neural networks and random forests. The findings confirmed that the XGB model outperformed its counterparts in terms of accuracy, establishing it as a valuable tool for predicting the properties of GO-modified asphalt.

Additionally, the researchers conducted a Shapley additive explanations (SHAP) value analysis to evaluate the influence of each input parameter on the viscoelastic properties. This analysis revealed that the initial properties of asphalt, the content of Graphene Oxide, and the mixing temperature are the most significant factors affecting G* and φ.

Significance of the Research

The implications of this research extend beyond academic interest, as improved asphalt properties can lead to enhanced road durability and reduced maintenance costs. The findings provide a pathway for the construction industry to adopt more sustainable practices by utilizing advanced materials like Graphene Oxide, which can withstand increasing loads and climate challenges.

The full text of the study, authored by Huong-Giang Thi Hoang, Hai-Van Thi Mai, Hoang Long Nguyen, and Hai-Bang Ly, is available for those interested in the detailed methodologies and findings. For further reading, the paper can be accessed at https://doi.org/10.1007/s11709-024-1025-y. As the world seeks innovative solutions to infrastructure challenges, studies like this highlight the potential for technology to enhance construction materials and methods effectively.