Prediction and modeling of mechanical properties of concrete modified with ceramic waste using artificial neural network and regression model

Over two centuries, concrete has been crucial to building. Thus, eco-friendly concrete is being developed. Emulating these tangible traits has recently gained popularity. Ceramic waste concrete's mechanical properties were modeled in this study. Ceramic waste percentages ranged from 5 to 20%. Compressive and tensile concrete strengths were modeled. To predict concrete hardness, regression modeling and artificial neural network (ANN) were used. Model performance was evaluated using prediction coefficients and root-mean-square error (RMSE). ANN models outperformed linear prediction with a coefficient for determination (R2) of 0.97. ANN models achieved root-mean-square errors (RMSEs) of 1.22 MPa, 1.21 MPa, and 1.022 MPa after 7, 14, and 28 days of retraining, respectively. Linear regression model showed RMSE values of 1.21, 1.32, and 1.27 MPa at 7, 14, and 28 days, respectively. In determining the compressive and tensile strength, the R2 was 0.70, meanwhile the ANN model achieved 0.87. Given its accuracy in predicting the strength qualities of ceramics cement and structural stiffness, the ANN model presents a promising tool for representing various types of concrete.