Multi-Scale Performance Evaluation of M30 Concrete Incorporating Nano-TiO₂, Nano-SiO₂, and Glass Fibers: A 90-Day Experimental and Machine Learning Study

This study explores the enhancement of concrete performance through the integration of nano-titanium dioxide (TiO₂), nano-silicon dioxide (SiO₂), and alkali-resistant glass fibers. These materials were introduced to simultaneously improve the concrete matrix at nano- and macro-scales by partially replacing cement and incorporating discrete reinforcement. The experimental program evaluated mechanical and durability properties at 7, 14, 28, and 90 days of curing. Compressive strength was tested using cube specimens, while tensile and flexural strengths were measured using standard cylindrical and beam specimens, respectively, in accordance with IS codes. Durability assessments included water absorption, sorptivity, RCPT, and UPV, conducted following relevant ASTM standards. SEM and XRD analyses were carried out at 90 days to examine microstructural developments. The results indicated significant improvement in all evaluated parameters. The ternary-modified concrete exhibited notable gains in compressive, tensile, and flexural strengths by day 90, while durability indicators showed reduced water absorption, chloride permeability, and increased ultrasonic pulse velocity. Microstructural analysis confirmed a denser matrix and enhanced C–S–H formation. To support the experimental findings, machine learning models using Random Forest and XGBoost algorithms were developed to predict key performance indicators based on mix composition and curing age. These models achieved R² values above 0.90, demonstrating high predictive reliability. Overall, this research confirms that the combined application of TiO₂, SiO₂, and glass fibers offers a comprehensive, scalable solution for developing high-performance, durable concrete. The integration of experimental and machine learning approaches offers a forward-looking strategy for sustainable infrastructure development.