A predictive mimicker for mechanical properties of eco-efficient and sustainable bricks incorporating waste glass using machine learning

Urbanization, industrialization, and economic growth all contribute to the rising demand for bricks. The world is heading towards sustainable, eco-friendly, recyclable materials to enhance the circular economy and mitigate the issues of carbon footprint, overburdened landfills, and waste of natural resources. Therefore, the mix design of bricks has evolved with time by incorporating recycled waste materials, i.e., glass, etc. This paper presents a unique approach to developing machine learning models to predict the mechanical properties of eco-efficient bricks incorporating waste glass, as it requires extensive experimentation to comprehend the properties. This research assesses four essential input parameters affecting mechanical properties. To predict the outcomes, four machine learning models were generated, and their results were compared. These four models include artificial neural networks (ANN), the Gaussian process of regression (GPR), the classification and regression tree (CART), and the support vector machine (SVM). A unique and advanced approach known as the generative adversarial network (GAN) has been employed for augmenting data and enhancing accuracy as the data available in published literature were limited. As a result, artificial neural network (ANN) has the highest accuracy among all models. Therefore, it is the most efficient model with RMSE and R2 of 3.86 MPa and 0.81 for predicting the compressive strength, and RMSE and R2 of 0.82% and 0.995 for predicting the Shrinkage of bricks incorporating glass. This study proposes a distinctive tool using machine learning for the sustainable brick production sector via predicting the mechanical properties of brick incorporating waste glass.