Compressive Strength Estimation of Fly Ash/Slag Based Green Concrete by Deploying Artificial Intelligence Models

Area of Science:

• Materials Science
• Civil Engineering
• Artificial Intelligence

Background:

• Cement production releases significant carbon dioxide, necessitating sustainable alternatives like green concrete.
• Incorporating industrial wastes such as ground-granulated blast furnace slag (GGBFS) and fly ash (FA) is crucial for reducing cement usage and environmental impact.
• Accurate prediction of concrete compressive strength is vital for structural integrity and performance.

Purpose of the Study:

• To investigate the optimal ratios of GGBFS and FA in binder content for green concrete production.
• To evaluate and compare the performance of artificial intelligence models, including adaptive neurofuzzy inference system (ANFIS), gene expression programming (GEP), and gradient boosting tree (GBT), in predicting concrete compressive strength.
• To enhance the accuracy of compressive strength prediction models by detailing parameter effects and optimizing model performance.

Main Methods:

• Employed ANFIS, GEP, and GBT models to analyze the influence of GGBFS and FA on concrete compressive strength.
• Optimized hyperparameters for GEP (200 chromosomes, 5 genes, 12 head sizes) and ANFIS (aspect ratios 0.5, 0.1, 7, 150).
• Validated the trained models using 40% of experimental data, performing parametric and sensitivity analyses.

Main Results:

• The Gradient Boosting Tree (GBT) model demonstrated superior accuracy in predicting concrete compressive strength compared to ANFIS, GEP, and existing literature models.
• GBT model achieved a correlation coefficient (R) of 0.95 for both training and validation, with low mean absolute error (MAE) and root mean square error (RMSE).
• Sensitivity analysis identified concrete aging as the most influential parameter, followed by GGBFS content, aligning with established literature findings.

Conclusions:

• The GBT model offers a highly accurate and reliable approach for optimizing green concrete mix designs.
• Artificial intelligence, particularly GBT, provides a powerful tool for predicting concrete compressive strength and guiding the use of industrial wastes in sustainable construction.
• The study highlights the significant impact of material composition and aging on concrete performance, offering valuable insights for developing eco-friendly construction materials.