Machine learning for efficient CO2 sequestration in cementitious materials: a data-driven method

Abstract

Extensive experimental work has proved that CO₂ sequestration by cementitious materials offers a promising venue for addressing the rising carbon emissions problem. However, relying merely on experiments on specific materials or some simple empirical methods makes it difficult to provide a comprehensive understanding. To address these challenges, this paper applies three advanced machine-learning techniques (Decision Tree, Random Forest, and eXtreme Gradient Boosting (XGBoost)), with existing datasets coupling with data collected from the literature. The results show that the XGBoost model significantly outperforms traditional linear regression approaches. In addition, aiding in the SHapley Additive exPlanations(SHAP), apart from the widely recognized factors, cement type was also investigated and shown its crucial role in affecting carbonation depth. CEM II/B-LL and CEM II/B-M are two types having high carbonation potential. The results enable the identification of key factors influencing CO₂ sequestration through cement and provide insights into optimizing experimental design.