Akademik Veri Yönetim Sistemi
Materials, cilt.18, sa.17, 2025 (SCI-Expanded, Scopus)
This study explores the integration of copper oxide (Cu2O) into bitumen and leverages Artificial Intelligence (AI) to evaluate and optimize the binder’s performance across multiple scales. Comprehensive laboratory tests, including conventional binder properties, rheological analysis, aging simulations, low-temperature cracking, and moisture susceptibility, were conducted on base and Cu2O modified asphalt binders. The results were used to train predictive models using gradient boosting regressors for each performance category. Optimization identified ideal Cu2O ratios for different engineering goals, offering practical recommendations. Based on this integrated cost-performance analysis, a Cu2O concentration of 2.3% was recommended as the most efficient trade-off point. AI modeling using Gradient Boosting Regressor (GBR) achieved high predictive performance, with R2 values reaching 0.98 for BBR prediction and 0.78 for rheology, and mean absolute error (MAE) values as low as 4.21. This demonstrates the model’s robustness in capturing complex nonlinear binder behaviors.