Prediction of compressive strength of bauxite residue-based geopolymer mortar as pavement composite materials: an integrated ANN and RSM approach

Geopolymers are cementitious materials that can be produced using industrial waste materials, such as bauxite residue (BR) and phosphogypsum (PG). In this research, a comprehensive experiment was conducted to collect a dataset comprising various mix proportions of bauxite residue and phosphogypsum geopolymer mortars (GPM), along with their corresponding compressive strength measurements. The dataset was then utilized to train an artificial neural networks (ANN) model and response surface methodology (RSM), which serves as a predictive tool for estimating the compressive strength of geopolymer mortars. The results demonstrate that the ANN model successfully predicts the compressive strength of GPM with high accuracy. The highest achieved compressive strength for GPM was recorded at 31.53 MPa. Moreover, the achieved strength can be an option for creating pavement composite materials. In terms of predictive accuracy, the ANN approach demonstrates an impressive R 2 value of 0.94, whereas RSM shows a slightly lower R 2 value of 0.92. The superior R 2 value obtained through the ANN method signifies its ability to offer more precise estimations of the desired outcomes in comparison to RSM. © 2023, The Author(s), under exclusive licence to Springer Nature Switzerland AG.