Microstructural assessment and supervised machine learning-aided modeling to explore the potential of quartz powder as an alternate binding material in concrete

Concrete composites with quartz powder (QP) are promising for sustainable buildings because of their superior mechanical properties and sustainability. The present study proposes a novel approach to predict the mechanical properties of quartz powder-infused concrete (QPC) using machine learning (ML) approaches. In this paper, prediction models were established, and the optimum model was determined using seven ML algorithms: extreme gradient boosting (XGB), adaptive boosting, bagging regressor, categorial boosting (CatB), gradient boosting, k-nearest neighbor, and random forest. For experimental validation, QPC samples were prepared by replacing cement with 10 % fly ash and 5 %, 10 %, 15 %, and 20 % QP content. Compressive strength (CS) test was conducted to assess the mechanical strength, and microstructural analysis was performed to highlight the interplay between QP and the concrete matrix. Furthermore, parametric modeling was applied for detailed parametric analysis and investigation of feature importance. The results indicate that for the prediction of CS, the CatB model provided the best R 2 values of 0.999 and 0.857, respectively, on both the train and test stages. With corresponding R 2 values of 0.999 and 0.857 on both the training and testing sets, the XGB model also offers reliable performance for predicting CS. The SHAP analysis identified that the superplasticizers, W/C ratio, and fly ash had the most significant impact on the predicted strength of QPC. Furthermore, Microstructural analysis revealed that the incorporation of QP and fly ash showed denser microstructure and greater C-S-H bonds, which enhanced compressive strength. The outcomes of this research can be effective for quick and sustainable manufacturing of concrete and optimizing input factors in the construction industry.