Innovative soft computing techniques including artificial neural network and nonlinear regression models to predict the compressive strength of environmentally friendly concrete incorporating waste glass powder

Since concrete and mortar productions (industry) are the biggest users of natural resources, its sustainability is under threat. The environmental and economic concern is the most important challenge that the construction industry is facing. The usage of waste materials in recent years has become an interesting subject for researchers. One of these waste materials is waste glass which is used as cement replacement in concrete. This article deals with proposing models to predict the compressive strength of concrete modified with waste glass powder as partial replacement of cement. In this paper, 153 data were collected from previous studies, and they were statically analyzed and represented in three models (linear regression model, nonlinear regression model, and ANN model) for predicting the compressive strength. Variables with significant effect on the compressive strength, such as curing time, w/b, cement content, sand content, waste glass content, and coarse aggregate content were considered. Various statistical assessments such as root mean squared error, mean absolute error, scatter index, and the coefficient of determination (R²) were used to evaluate the efficiency and performance of the proposed models. The obtained results showed that the ANN model showed better efficiency for predicting the compressive strength of normal concrete mixtures containing waste glass powder compared to other models. Finally, the applications of these developed models were limited to forecasting the compressive strength of normal concretes containing various dosages of waste glass powders.