Performance evaluation of fly-ash and silica-fume based geopolymer concrete at different alkaline molarities with machine learning-based strength prediction.

The high environmental impact of conventional Portland cement has driven the need for sustainable alternative binders with improved performance characteristics. Geopolymer concrete (GPC), particularly using industrial by-products such as fly ash and silica fume, has emerged as a promising solution; however, the influence of alkaline activator molarity on its combined mechanical and durability performance remains insufficiently understood. This study investigates the effect of sodium hydroxide molarity (8 M-14 M) on fly ash-silica fume based geopolymer concrete under ambient curing conditions, aiming to establish relationships between activator concentration, microstructure, and performance. Experimental results show that increasing molarity significantly enhances performance up to an optimum level. The 12 M mix exhibited approximately 30-38% higher compressive strength, 23-28% higher split tensile strength, and 24-30% higher flexural strength compared to the 8 M mix. Durability performance also improved substantially, with sorptivity reduced by about 26-29% and rapid chloride permeability (RCP) reduced by 44-52%. Microstructural analysis confirmed the formation of a dense and homogeneous geopolymer gel matrix at optimum molarity, while excessive alkalinity led to localized microcracking. Statistical analysis validated the significance of molarity effects, and machine learning models achieved high predictive accuracy (R² > 0.95). Overall, this study provides an integrated experimental-statistical-data-driven framework for optimizing geopolymer concrete, contributing to the development of sustainable and high-performance cementitious materials.