Keywords

Abstract

Self-compacting geopolymer concrete; 

 Scanning electron microscope;

 Manufactured sand; 

Metakaolin; 

Ground-granulated blast furnace slag; 

Molarity

This work conducted experimental research into the flow and mechanical characteristics of self-compacting geopolymer concrete (SCGC) made from ecologically beneficial byproducts of industry such as ground granulated blast furnace slag (GGBFS) and metakaolin (MK). Through trial and error, the mix's proportion of self-compacting geopolymer concrete (SCGC) was determined. The mass fraction of GGBFS with metakaolin was varied by 0%, 10%, 20%, and 30% by mass for all molarities, including 8M, 10M, and 12M. The superplasticizer (S.P.) dose of 1.5% and the fluid-to-binder (F/B) ratio of 0.37 by mass were held constant for each mix percentage, with the extra water content being changed correspondingly. Workability properties were assessed in addition to mechanical properties, which comprised compressive strength, split tensile strength, flexural strength, shear strength, and impact strength. The application of machine learning algorithms to forecast the compressive strength of SCGC is the focus of this study. Specifically, Random Forest (RF), Gradient Boost, and Extreme Gradient Boost (XGB) are utilized. Different success measures, like Root mean squared error (RMSE), mean squared error (MSE), mean absolute error (MAE), and R-squared (R2), are used to judge these methods. The Gradient Boost model outperforms the others, achieving an R2 score of 0.934 on training data and 0.929 on test data, showcasing its precision and accuracy. The success of the Gradient Boost model can be attributed to its incorporation of randomness and ensemble diversity, making it a powerful tool for predicting compressive strength in various scenarios.

© 2024 MIM Research Group. All rights reserved.