<![CDATA[Concrete contributes approximately 8% of global CO₂ emissions. The incorporation of Supplementary Cementitious Materials (SCMs) as partial cement replacements is widely recognized as an effective strategy to reduce the carbon footprint of concrete. However, accurately quantifying the relationship between mix composition and carbon emissions remains challenging. This study develops a machine learning model to predict the carbon footprint of SCM-based concrete using material composition data. A global dataset comprising 1,456 mix designs collected from 136 publications across 27 countries was compiled, resulting in 1,294 valid samples after preprocessing. Four regression algorithms were evaluated: Support Vector Regression (SVR), Random Forest Regression (RFR), Decision Tree Regression (DTR), and Gradient Boosting Regression (GBR), with hyperparameter tuning using 5-fold cross-validation. All models achieved high predictive accuracy (R² > 0.998), with GBR demonstrating the best performance (R² = 0.9996; RMSE = 1.7452 kg CO₂/m³; MAE = 1.2779 kg CO₂/m³). Feature importance analysis identified cement as the dominant contributor (>99.8%) to emissions. Sensitivity analysis confirmed a strong linear relationship between cement content and CO₂ emissions (~0.82 kg CO₂ per kg cement). These findings support emission-reduction strategies in sustainable concrete design.]]>
Copyrights © 2026
