Scholarlink

<p>Bricks are fundamental materials extensively utilized in masonry construction. However, the conventional production of clay bricks necessitates substantial extraction of natural clay, leading to environmental degradation and unsustainable land use. To mitigate these issues, the incorporation of alternative materials as partial substitutes for clay is imperative. In this regard, Pond Ash (PA), a by-product of thermal power plants, emerges as a viable replacement when blended with Cement and Stone Dust (SD), offering a sustainable solution for brick manufacturing without compromising essential mechanical and physical properties. This study investigates the application of Machine Learning (ML) algorithms for predicting the 28-day Compressive Strength (CS) of PA-based bricks reinforced with Cement and SD. The experimental mix design maintained a constant PA content of 70%, while the remaining 30% consisted of varying proportions of Cement and SD. The predictive modeling was based on an experimental dataset comprising 100 samples, with mixture proportions as input features and CS as the target variable. To capture the nonlinear and complex relationships inherent in the dataset, four supervised regression models were employed: Random Forest (RF), Gradient Boosting Regressor (GBR), AdaBoost Regressor (ADA), and a Stacked Ensemble (SE) model. Model performance was rigorously evaluated using the coefficient of determination (R²) and validated against a separate dataset of 8 unseen experimental values. Additionally, a 10-fold cross-validation strategy was implemented to ensure model generalizability and minimize overfitting. The R² values obtained for RF, GBR, ADA, and SE were 0.989, 0.989, 0.982, and 0.989, respectively, indicating a high degree of accuracy and consistency in strength prediction. These results underscore the effectiveness of ML-based approaches in modeling the compressive behavior of PA-based bricks. The integration of ML techniques into the analysis and design process demonstrates significant potential in optimizing sustainable brick formulations, thereby contributing to the advancement of eco-efficient construction practices. Microstructural studies also have been carried out.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predictive modeling of compressive strength in pond ash bricks reinforced with cement and stone dust using machine learning

  • 期刊论文

  • 发表日期：

    2025年7月16日

查看全文
  • Mohammad Sufian Abbasi,
  • Vikash Singh,
  • Zishan Raza Khan,
  • Syed Aqeel Ahmad

摘要

Bricks are fundamental materials extensively utilized in masonry construction. However, the conventional production of clay bricks necessitates substantial extraction of natural clay, leading to environmental degradation and unsustainable land use. To mitigate these issues, the incorporation of alternative materials as partial substitutes for clay is imperative. In this regard, Pond Ash (PA), a by-product of thermal power plants, emerges as a viable replacement when blended with Cement and Stone Dust (SD), offering a sustainable solution for brick manufacturing without compromising essential mechanical and physical properties. This study investigates the application of Machine Learning (ML) algorithms for predicting the 28-day Compressive Strength (CS) of PA-based bricks reinforced with Cement and SD. The experimental mix design maintained a constant PA content of 70%, while the remaining 30% consisted of varying proportions of Cement and SD. The predictive modeling was based on an experimental dataset comprising 100 samples, with mixture proportions as input features and CS as the target variable. To capture the nonlinear and complex relationships inherent in the dataset, four supervised regression models were employed: Random Forest (RF), Gradient Boosting Regressor (GBR), AdaBoost Regressor (ADA), and a Stacked Ensemble (SE) model. Model performance was rigorously evaluated using the coefficient of determination (R²) and validated against a separate dataset of 8 unseen experimental values. Additionally, a 10-fold cross-validation strategy was implemented to ensure model generalizability and minimize overfitting. The R² values obtained for RF, GBR, ADA, and SE were 0.989, 0.989, 0.982, and 0.989, respectively, indicating a high degree of accuracy and consistency in strength prediction. These results underscore the effectiveness of ML-based approaches in modeling the compressive behavior of PA-based bricks. The integration of ML techniques into the analysis and design process demonstrates significant potential in optimizing sustainable brick formulations, thereby contributing to the advancement of eco-efficient construction practices. Microstructural studies also have been carried out.