Predicting the mechanical properties of sustainable concrete containing waste foundry sand using multi-objective ANN approach

Abstract

The amount of waste materials obtained from industries is increasing every day, which has been identified as one of the crucial issues in many countries. Waste foundry sand (WFS) is a by-product of the foundry industry, which can be used as a partial replacement for fine aggregate in concrete. The aim of this study is to predict the mechanical properties of concrete containing WFS using an artificial neural network (ANN) assisted by multi-objective multi-verse optimizer (MOMVO) algorithm. In the proposed model, both network error and complexity were considered as multi-objective optimization problems which were solved using MOMVO. To develop the proposed model, a comprehensive database including effective parameters on the mechanical properties of concretes were gathered and modeled in MATLAB environment. For compressive strength, splitting tensile strength, modulus of elasticity and flexural strength of concrete containing WFS, several optimal ANN models were achieved and the performances of the two selected models for each mechanical property were compared. The results showed the potential of acceptable accuracy of the developed ANN model assisted by MOMVO algorithm in estimation of the studied mechanical properties. Finally, a parametric study was carried out to investigate the contribution of each input variable on the mechanical properties of concrete containing WFS. The results inferred that the ratios of water to cement, fine aggregate to total aggregate, and coarse aggregate to cement had the most effect on the mechanical properties.