AI-BASED PREDICTIVE MODELING OF SUSTAINABLE GEOPOLYMER CONCRETE USING AGRICULTURAL WASTE MATERIALS

The construction industry is one of the largest contributors to global carbon emissions due to the extensive use of cement in conventional concrete production. Sustainable alternatives such as geopolymer concrete have gained significant attention because they incorporate agricultural and industrial waste materials while reducing environmental impact. This research presents an Artificial Intelligence (AI)-based predictive modeling framework for sustainable geopolymer concrete utilizing agricultural waste materials including Sugarcane Bagasse Ash (SBA), Banana Peel Ash (BPA), and Fly Ash Type C polymer. The proposed framework integrates machine learning algorithms with a lightweight web-based application to predict key concrete performance metrics including compressive strength, flexural strength, and initial and final setting times. Four regression-based machine learning modelsRidge Regression, Elastic Net Regression, Partial Least Squares Regression (PLS), and Support Vector Regression (SVR)were trained and evaluated using experimental geopolymer concrete datasets. Results demonstrated that SVR significantly outperformed the other models, achieving high predictive accuracy with R2 values reaching 0.979 for certain output variables.