Machine learning modelling to design micro-structured concrete absorber of carbon dioxide (CO2)

This work introduces the key aspects involved in optimizing binder-based microstructures using machine-learning techniques, with the primary goal of enhancing their uptake properties—particularly their capacity to absorb carbon dioxide (CO₂). The utilization of CO₂ in concrete and its constituents is currently achieved by converting gaseous CO₂ into solid carbonates through mineral carbonation.

The proposed approach comprises several essential experimental and numerical steps to design a micro-structured concrete absorber guided by machine-learning procedures. These include data collection and model training, preprocessing of microstructural variables to ensure compatibility with machine-learning algorithms, selection of suitable algorithms, model validation, optimization, and prediction.

By enabling the exploration of a wide range of microstructures and mineral carbonation pathways, this methodology provides a route to engineer concrete materials with superior CO₂ absorption capacities, contributing to the development of sustainable and environmentally responsible construction materials. The synergy between machine learning and concrete science holds significant potential to transform how the construction industry enhances material performance while reducing environmental impacts.