Author(s): Chinedu A Okafor, Fatima L Abubakar and Tunde O Balogun

Abstract: The study presents an advanced artificial intelligence (Artificial Intelligence (AI))-based framework for optimizing high-performance concrete (High-Performance Concrete (HPC)) mix design to achieve superior mechanical strength, durability, and cost efficiency. Traditional empirical mix design methods often fail to capture the nonlinear interactions among materials and performance parameters, leading to suboptimal mixtures and increased resource consumption. To address this limitation, the research integrates predictive machine learning models—including Artificial Neural Network (ANN), Support Vector Machine (SVM), and Gradient Boosting Regression (GBR)—with a multi-objective optimization algorithm, the Non-Dominated Sorting Genetic Algorithm-II (NSGA-II). Experimental and literature-based datasets comprising key mix parameters such as cementitious content, water-to-binder ratio, supplementary cementitious materials (silica fume, fly ash, and GGBS), and superplasticizer dosage were analyzed to develop robust predictive models. Among all tested algorithms, GBR exhibited the highest accuracy, achieving an R² of 0.95 and the lowest RMSE for compressive strength and durability indices. Feature importance analysis identified the water-to-binder ratio, silica fume percentage, and superplasticizer dosage as dominant predictors of HPC performance. The integrated AI-optimization framework generated Pareto-optimal designs that achieved up to 94 MPa compressive strength and a rapid chloride permeability below 1, 100 C at a 10% lower cost than conventional designs. Validation experiments confirmed the close agreement between predicted and actual results, emphasizing the model’s reliability and potential for real-world implementation. The study concludes that AI-driven predictive-optimization methodologies provide a powerful alternative to empirical design approaches, enabling the development of sustainable, cost-effective, and high-performance concrete mixtures suitable for modern construction.

Pages: 43-48 | Views: 4 | Downloads: 2

Download Full Article: Click Here