Self compacting concrete with recycled aggregate compressive strength prediction based on gradient boosting...

Area of Science:

• Materials Science and Engineering
• Civil Engineering
• Computational Intelligence

Background:

• Self-compacting concrete (SCC) requires high workability, posing challenges in densely reinforced or complex formwork.
• Estimating 28-day compressive strength of SCC, especially with recycled aggregates for sustainability, relies on time-consuming laboratory tests.
• Recycled aggregates in concrete mixtures promote eco-friendly and sustainable construction practices.

Purpose of the Study:

• To develop and evaluate a novel hybrid machine learning model for accurately and efficiently estimating the 28-day compressive strength of SCC containing recycled aggregates.
• To provide a cost-effective and faster alternative to conventional laboratory testing methods for SCC strength prediction.
• To enhance mix design optimization and quality control in real-life construction projects through an accessible decision-support tool.

Main Methods:

• A hybrid Gradient Boosting Regressor Tree (GBRT) model integrated with Bayesian Optimization was developed.
• The model's performance was evaluated using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and R-squared ([Formula: see text]).
• SHAP (Shapley Additive Explanations) was employed to interpret the black-box model and identify predictor importance.

Main Results:

• The hybrid GBRT model achieved an average RMSE of 6.000, MAE of 3.968, and R-squared of 0.806 in five-fold cross-validation.
• The developed model demonstrated superior prediction accuracy and robustness compared to single learner models like Support Vector Regression (SVR) and K-Nearest Neighbors (KNN).
• SHAP analysis revealed key predictors influencing compressive strength and their trends, aiding in model interpretability.

Conclusions:

• The developed hybrid GBRT model is an accurate, fast, and economical substitute for predicting the 28-day compressive strength of self-compacting concrete with recycled aggregates.
• The model's strong predictive capability and robustness highlight its potential to replace conventional laboratory testing.
• An easy-to-use graphical interface was created, serving as a valuable decision-support tool for civil engineers and practitioners in mix design and quality control.