Investigation of Mechanical Behavior, Plastic Shrinkage, and Microstructure of Self-Compacting Geopolymer Concrete and The Effects of Single-Component and Two-Component Fibers

Abstract

Substituting aluminosilicate pozzolans for cement offers a stable and environmentally sustainable mixture. The primary innovation lies in incorporating polypropylene and steel fibers into self-compacting mixtures to enhance mechanical properties and durability, coupled with artificial intelligence modeling (artificial neural networks and decision trees) for compressive strength prediction. This integrated experimental and computational methodology enables more precise optimization of mix designs.The methodology involved producing 16 mix designs with alkaline concentrations of 8 and 12 M, variations in pozzolan quantities, and fiber volumes (0.5 to 1% by volume). Samples underwent ambient and heat curing, with evaluations of fresh properties (slump flow, L-box, V-funnel), mechanical properties (compressive and tensile strength), water absorption, shrinkage, and microstructure (SEM and XRD). Modeling was performed using experimental data to predict compressive strength. Results indicated that increasing molarity from 8 to 12 densifies the concrete microstructure, boosting compressive strength by up to 25% (from 40 to 96.83 MPa at 28 days), although alkaline solution costs may rise. Steel fibers, by forming resilient internal networks, improved tensile and flexural strength by up to 30% and reduced drying shrinkage by 11% . Hybrid fiber combinations enhanced high-temperature durability, retaining 80% strength at 600°C. The AI models achieved accuracies exceeding 88% , positioning this concrete as a sustainable and efficient alternative to Portland cement concrete.