Enhancing geopolymer concrete strength with sodium hydroxide using Gegenbauer graph neural networks approach

Abstract

An eco-friendly and inventive alternative to cement-based concrete is geopolymer concrete (GPC) due to its reduced carbon footprint, as it completely replaces cement. Despite their environmental benefits, the mechanical performance of GPC is highly sensitive to the mix of proportions and curing conditions, presenting significant challenges in achieving consistent strength. To enhance the strength and mechanical properties of GPC with sodium hydroxide (NaOH), a unique approach using Gegenbauer graph neural networks (GGNN) is presented in this work. The main objectives of this study include reducing CO₂ emissions. The strength of GPC with NaOH is predicted using GGNN. The GGNN method is also used to analyze the mechanical properties of GPC under different NaOH molarities and different ratios of sodium silicate to NaOH. The proposed method is simulated in MATLAB and is compared with existing methods like long short-term memory (LSTM), artificial neural network (ANN), and back propagation neural network (BPNN). It is found that the oven-cured GPC achieved better mechanical strength compared to the ambient-cured GPC. The proposed model attained the highest compressive strength (CS) of 75.42 MPa along with a high correlation coefficient of 0.9871 compared to the previous studies. In contrast to the existing methods, the proposed model achieved a high prediction accuracy of 98.5% along with a low CO₂ emission of 7% demonstrating its superior performance in accurately predicting the mechanical strength and reducing carbon footprints. This indicates the robustness and reliability of the proposed model for optimizing material properties and advancing the field of sustainable construction materials.