Exploring the influence of waste glass granular replacement on compressive strength in concrete mixtures: a normalization and modeling study

In many third-world countries, effective waste material management has become a crucial concern due to the escalating quantity of waste materials. Among these materials, waste glass holds significance due to its widespread usage in various daily human functions. Numerous researchers have explored the feasibility of using waste glass as a substitute for cement or sand in concrete. This paper focuses on employing waste glass granules as a replacement for coarse aggregate in concrete, as coarse aggregate constitutes a significant portion of the concrete mix. A comprehensive review of prior studies is conducted to elucidate the impact of utilizing waste glass granules as a substitute for coarse aggregate on both the fresh and mechanical properties of concrete. To achieve this objective, experimental data from previous research are gathered, and statistical models are developed using Gaussian progress regression (GPR), Support vector machine (SVM), Ensemble boosting tree (EBT), artificial neural network (ANN), along with multi-linear regression (MLR). These models are employed to predict the compressive strength of concrete. Various statistical parameters are utilized to evaluate and compare the efficiency of these models, with the aim of identifying the most effective one. Among the models considered, the artificial neural network is deemed the most efficient, as it demonstrates a higher R 2 and lower values of RMSE, MAE, and SI. Specifically, the SI value of the ANN model is higher by 265%, 118%, 333%, and 113% compared to MLR, GPR, SVM, and EBT models, respectively.