Neural network-based model to predict compressive strength of concrete incorporating supplementary cementitious materials and recycled aggregates

ABSTRACT

The behavior of concrete incorporating recycled aggregates in total or partial replacement of natural aggregates is different from the behavior of conventional concrete. Likewise, the partial substitution of cement by supplementary cementitious materials also affects the behavior of concrete. Despite the fact that both partial substitutions have undoubted benefits from the point of view of sustainability, their effect on the compressive strength of concrete is difficult to model. This document aims to predict the compressive strength of these special concretes through the use of artificial neural networks (ANN). Training and testing data for the development of the ANN model were prepared using 309 dosages published in 22 different literature sources. The results indicated that the ANN model is an efficient approach for predicting the compressive strength of concretes that incorporate recycled aggregates and supplementary cementitious materials.

Keywords
K-fold validation; ANN; UHPC; supplementary cementitious materials; 7-day compressive strength