Akademik Veri Yönetim Sistemi
Canadian Metallurgical Quarterly, 2025 (SCI-Expanded, Scopus)
Reprocessing iron ore tailings (IOTs) provides a sustainable route to minimise environmental impact while extracting residual value from mining waste. This study evaluates the technical and data-driven feasibility of recovering iron from fine-grained IOTs sourced from the Otlukilise beneficiation plant in Türkiye. High-intensity magnetic separation was employed to upgrade Fe content, and a Random Forest regression model was developed to predict the Fe grade of magnetic concentrates using inputs, such as feed Fe grade, magnetic field strength, product weight, and separation stage. The model achieved a coefficient of determination (R²) of 0.735, with a mean absolute error (MAE) of 1.176 and a root mean square error (RMSE) of 1.559, confirming reliable predictive capability. Feature importance analysis identified feed Fe grade and magnetic field intensity as the most influential parameters. Although both gravity separation and magnetic concentration produced limited recoveries due to the ultra-fine size of the tailings, an integrated flowsheet generated a concentrate with 59.2% Fe. Despite the low overall recovery, the study offers valuable insight into the operational boundaries of tailings reprocessing and highlights the role of machine learning in supporting process prediction and decision-making. These findings advance data-driven approaches to sustainable and context-specific mining waste valorisation.