ANN-based prediction of compressive strength in cement and polystyrene-stabilized earth bricks

Abstract

The growing demand for environmentally-friendly building materials has led to a renewed interest in earth bricks. To facilitate their use, it would be ideal to know their physical properties, in particular their compressive strength. This study utilizes ANN and regression techniques to predict the compressive strength of cement-stabilized compressed earth bricks (CEB) with expanded polystyrene (EPS) granules, facilitating their application in the construction industry. The input parameters used for training our models include optimum cement and water content, polystyrene content, hardening age, and amount of clay content. The database was created through literature reviews experiments in order to propose equations for predicting compressive strength. The best model was achieved using ANN, with performance metrics of R² = 0.97, RMSE = 0.82 MPa, and MAE = 6.62 MPa, outperforming multivariate regression. This study provides us with an artificial intelligence model for predicting the compressive strength of CEBs with EPS granules.