Utilizing machine learning to integrate silica-based production waste material in ceramic tiles manufacturing: Progressing toward sustainable solutions

The production of ceramic tiles is an intricate process of combining modern technology with ancient craftsmanship, which creates versatile and durable finished surfaces for various construction applications. The manufacturing of ceramic tiles causes the production of significant waste during different production stages such as shaping, firing, and finishing leading to environmental degradation and depletion of resources. Therefore, this research work was planned to investigate the possibility of utilizing ceramic waste in the manufacturing of ceramic tiles. Two types of ceramic waste, namely glaze waste and body waste were used with 5–10% replacement of different materials in the production of ceramic tile. The results revealed that up to 10% of the waste produced during the manufacturing of ceramic tiles can successfully be reused in the manufacturing of ceramic tiles. The ceramic tiles produced with 10% replacement of feldspar satisfied the minimum criteria of flexural strength, shrinkage, and water absorption supported by scanning electron microscopy findings. Moreover, a machine learning model was developed using six input variables, i.e., feldspar (F), three different types of clays (C1, C2, and C3), body waste (BW) and glaze waste (GW), while three variables, i.e., firing shrinkage (S_F), flexure strength (FS) and water absorption (WA), as output. Applying a machine learning approach to data revealed that the model, i.e., artificial neural network (ANN), exhibited strong and consistent performance, displaying notable generalization capabilities (with an R² greater than 0.95 on the test set). This study aims to shape sustainable waste management approaches and establishes a blueprint for harnessing ceramic waste with the potential to be a primary resource in the production of ceramic tiles.