https://doi.org/10.3221/IGF-ESIS.79.08
The quasi-brittle nature of Geopolymer Concrete (GPC) severely limits its broader structural viability. This study investigates the optimization of a Ground Granulated Blast-furnace Slag (GGBS) geopolymer concrete through sequential Styrene-Butadiene Rubber (SBR) latex modification and alkali-resistant (AR) glass fiber reinforcement. Polymer dosages from 0% to 25% were evaluated, validating a 15% SBR concentration as the microstructural baseline. Scanning Electron Microscopy and Energy Dispersive X-Ray Analysis (SEM/EDAX) confirmed that this dosage facilitates a monolithic interpenetrating polymer network and enhances Calcium Aluminosilicate Hydrate (C-A-S-H) gel precipitation. Subsequently, AR glass fibers were incorporated (0% to 2.0%) into the optimized SBR-GPC matrix. Empirical testing isolated 1.5% as the absolute optimal fiber dosage, yielding peak 28-day compressive (31.70 MPa), flexural (5.39 MPa), and bearing (62.26 MPa) capacities. Furthermore, single-point loading fracture mechanics testing revealed that the 1.5% integration maximized both fracture energy and the critical stress intensity factor, while also achieving a 50.00% enhancement in ultimate impact energy absorption and reducing chloride permeability to 1350.20 Coulombs (Low). Dosages exceeding 1.5% induced severe agglomeration, systematically degrading structural integrity. Finally, mathematical regression models were compared against advanced Machine Learning architectures like Artificial Neural Networks (ANN) and Gaussian Process Regression (GPR). While simple empirical equations reliably predicted tensile and flexural capacities, the GPR framework demonstrated exceptional superiority in mapping the non-linear compressive strength behavior, achieving a Leave-One-Out Cross-Validation of 0.950.
