Predicting the mechanical properties of spent foundry sand concrete (SFSC) using artificial neural network (ANN)

Abstract

As a by-product of industry, various sorts of waste are produced. The dumping and disposal of this waste is a critical issue that causes environmental difficulties and harms the environment. The escalating paucity of dumping sites and the high cost of disposal have sparked a need for an alternative method of disposing of this enormous industrial waste. In the case of foundry industries, the final waste material is sand, which is widely known as waste foundry sand (WFS) or spent foundry sand (SFS). This spent foundry sand is available in huge amounts and is cheaper than natural sand. As Natural sand is the primary component used in the building industry to make concrete. Also, natural sand is expensive to transport and consume and results in the depletion of natural resources since its formation takes hundreds of years. In this paper, the fine aggregate is partially replaced with the SFS of various zones in different proportions, such as 10%, 20%, and 30%. The uniaxial compressive strength and splitting tensile strength of SFS-based concrete are predicted using the artificial neural network (ANN) technique. Further, these mechanical properties are compared with the experimental results. The obtained results demonstrate the model's high robustness in predicting strength with a lower absolute error and high correlation with the data sets.