Electrical conductivity, microstructures, chemical compositions, and systematic multivariable models to evaluate the effect of waste slag smelting (pyrometallurgical) on the compressive strength of concrete

Piro, Nzar Shakr; Mohammed, Ahmed Salih; Hamad, Samir M.; Kurda, Rawaz

Electrical conductivity, microstructures, chemical compositions, and systematic multivariable models to evaluate the effect of waste slag smelting (pyrometallurgical) on the compressive strength of concrete

2022

Piro, Nzar Shakr | Mohammed, Ahmed Salih | Hamad, Samir M. | Kurda, Rawaz

Concrete is a composite material widely used in construction. Waste slag smelting (pyrometallurgical) (steel slag (SS)) is a molten liquid melt of silicates and oxides created as a by-product of steel production. It is a complex solution of silicates and oxides. Steel slag recovery conserves natural resources and frees up landfill space. Steel slag has been used in concrete to replace fine and coarse particles (gravel). Three hundred thirty-eight data points were collected, analyzed, and modeled. It was determined which factors influenced the compressive strength of concrete with steel slag replacement in the modeling phase. Water/cement ratio was 0.3–0.872, steel slag content 0–1196 kg/m³, fine aggregate content 175.5–1285 kg/m³, and coarse aggregate content (natural aggregate) 0–1253.75 kg/m³. In addition, 134 data were collected regarding the electrical conductivity of concrete to analyze and model the effect of SS on electrical conductivity. The correlation between compressive strength and electrical conductivity was also observed. This research used a linear regression (LR) model, a nonlinear regression (NLR) model, an artificial neural network (ANN), a full quadratic model (FQ), and an M5P tree model to anticipate the compressive strength of normal strength concrete with steel slag aggregate substitution. For predicting the electrical conductivity, the ANN model was performed. The compressive strength of the steel slag was raised based on data from the literature. Statistical techniques like the dispersion index and Taylor diagram showed that the ANN model with the lowest RMSE predicted compressive strength better than the other models.

اظهر المزيد [+]