Application of Artificial Neural Networks to Predict the Workability of Self-compacting concrete

Abstract

In a context of constant innovation in construction techniques, self-compacting concrete (SCC) has emerged as an innovative material, requiring an optimized formulation to ensure its fresh-state properties. This study proposes the use of artificial neural networks (ANN) to predict SCC formulations containing limestone filler. Based on a large database drawn from more than 60 studies, ANN models connect formulation parameters (water/cement ratio, aggregates, superplasticizer, limestone filler) to three key performance indicators (slump-flow diameter, V-funnel time, and L-box ratio). The developed ANN models have demonstrated their accuracy (via R2 and MSE) in predicting experimental results with good agreement. This approach significantly reduces the number of laboratory tests required, thereby saving time and money while reducing CO2 emissions. Thus, the proposed method combines artificial intelligence and data analysis to offer an innovative and sustainable solution for SCC formulation.