Predicting compressive and splitting tensile strength of high volume fly ash roller compacted concrete using ann and ann-biogeography based optimization models

Abstract

Roller compacted concrete (RCC) has gained prominence in the construction industry due to its durability, cost-effectiveness, and environmental benefits, particularly with the incorporation of high-volume fly ash (HVFA). However, traditional experimental approaches to evaluating RCC's mechanical properties, such as compressive strength (CS) and splitting tensile strength (STS), are resource-intensive and time-consuming. To address these challenges, this study explores the application of artificial intelligence (AI), specifically artificial neural networks (ANN) and a hybrid ANN-Biogeography-Based Optimization (ANN-BBO) model, to predict the CS and STS of RCC. A dataset comprising 168 RCC mixtures, incorporating various material and process parameters, was analyzed. The ANN-BBO model demonstrated superior predictive accuracy compared to a standalone ANN, with R2 values exceeding 0.98 for both CS and STS, significantly reducing error margins. The findings highlight the effectiveness of AI-driven modeling in optimizing RCC mix designs, minimizing experimental costs, and enhancing the sustainability of concrete production. This research underscores the potential of integrating AI with optimization techniques to refine RCC performance assessment, which enables and facilitates more efficient and sustainable infrastructure development.