Optimization of a ternary geopolymer paste through orthogonal experimental design, ANOVA and entropy-based multi-objective decision analysis

The optimization of geopolymer binders containing multiple aluminosilicate precursors is challenging because the governing mix parameters simultaneously influence mechanical performance, workability, and material cost. Existing studies primarily focus on individual performance indicators, resulting in suboptimal mix designs for practical engineering applications. Therefore, this study proposes an integrated multi-objective optimization framework for ternary Fly Ash, ground granulated blast furnace slag, metakaolin geopolymer paste by combining single-factor experiments, L16 orthogonal experimental design, analysis of variance, and a synthetic weighted decision model incorporating subjective and entropy-based objective weighting. Initial experiments established that curing at 60 °C for 24 h provided the most favourable geo-polymerisation conditions. Subsequently, the effects of GGBS content, MK content, NaOH molarity, alkaline activator ratio, and alkali-to-binder ratio on compressive strength, flowability, and material cost were systematically investigated. Effect analysis indicated that the alkaline activator ratio was associated with the greatest change in compressive strength, consistent with its role in enhanced dissolution and polycondensation of aluminosilicate precursors, and the alkali-to-binder ratio with the greatest change in flowability, although these effects could not be fully separated from their associated two-factor interactions within the design employed; MK dosage was identified as the dominant contributor to material cost. Although the maximum compressive strength of 115.87 MPa was achieved under one experimental condition, the proposed multi-objective optimization identified a balanced mixture comprising 50% GGBS, 10% MK, 6 M NaOH, an alkaline activator ratio of 1:1.5, and an alkali-to-binder ratio of 0.35 as the optimum engineering composition. Repeat testing indicated that the relative ranking of the selected mixtures was broadly preserved, although the three compressive strength specimens per mixture were subsamples from a single batch rather than independent replicates. Furthermore, XRD and SEM analyses indicated microstructural features consistent with enhanced C-(A)-S-H and N-A-S-H gel formation, increased matrix densification, and reduced pore distribution in the optimized mixtures, providing microstructural support for their superior mechanical performance.