Scholarlink

<p>The need for sustainable construction practices has intensified interest in integrating industrial byproducts into concrete mixtures to reduce environmental impact without compromising structural performance. This study proposes a hybrid artificial intelligence framework for predicting compressive strength in concrete incorporating fly ash and ground granulated blast furnace slag. A dataset of 1030 mix designs was analyzed using artificial neural networks (ANN) optimized with three metaheuristic algorithms: Chicken Swarm Optimization (ANN-CSO), Moth Flame Optimization (ANN-MFO), and Whale Optimization Algorithm (ANN-WOA). These hybrid models were evaluated based on their predictive capabilities. Performance metrics included the coefficient of determination (R<sup>2</sup>), root mean square error (RMSE), and mean absolute error (MAE). A 12-metric score analysis and Anderson–Darling residual tests were conducted to assess predictive quality and statistical robustness. Taylor plots provided visual comparisons in terms of correlation, standard deviation, and centered RMSE. Specifically, ANN-WOA achieved the highest accuracy (testing R<sup>2</sup> = 0.855, RMSE = 0.076) and ranked highest overall (score = 63). Its residuals also most closely approximated a normal distribution (<i>p</i> = 0.127). While ANN-WOA was most accurate, ANN-CSO offered a resource-efficient balance suitable for limited environments. Beyond technical performance, this study contributes to Sustainable Development Goals (SDGs), particularly SDG 9 (Industry, Innovation, and Infrastructure), SDG 11 (Sustainable Cities and Communities), and SDG 13 (Climate Action), by promoting low-carbon materials and intelligent modeling. The results affirm that hybrid data-driven modeling enhances predictive efficiency and supports sustainable decision-making in modern concrete technology.</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

Metaheuristic-enhanced ANN framework for compressive strength prediction in concrete with supplementary cementitious materials

  • 期刊论文

  • 发表日期：

    2025年8月26日

查看全文
  • Abba Bashir,
  • Sadi I. Haruna,
  • Yasser E. Ibrhim,
  • Sani I. Abba

摘要

The need for sustainable construction practices has intensified interest in integrating industrial byproducts into concrete mixtures to reduce environmental impact without compromising structural performance. This study proposes a hybrid artificial intelligence framework for predicting compressive strength in concrete incorporating fly ash and ground granulated blast furnace slag. A dataset of 1030 mix designs was analyzed using artificial neural networks (ANN) optimized with three metaheuristic algorithms: Chicken Swarm Optimization (ANN-CSO), Moth Flame Optimization (ANN-MFO), and Whale Optimization Algorithm (ANN-WOA). These hybrid models were evaluated based on their predictive capabilities. Performance metrics included the coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE). A 12-metric score analysis and Anderson–Darling residual tests were conducted to assess predictive quality and statistical robustness. Taylor plots provided visual comparisons in terms of correlation, standard deviation, and centered RMSE. Specifically, ANN-WOA achieved the highest accuracy (testing R2 = 0.855, RMSE = 0.076) and ranked highest overall (score = 63). Its residuals also most closely approximated a normal distribution (p = 0.127). While ANN-WOA was most accurate, ANN-CSO offered a resource-efficient balance suitable for limited environments. Beyond technical performance, this study contributes to Sustainable Development Goals (SDGs), particularly SDG 9 (Industry, Innovation, and Infrastructure), SDG 11 (Sustainable Cities and Communities), and SDG 13 (Climate Action), by promoting low-carbon materials and intelligent modeling. The results affirm that hybrid data-driven modeling enhances predictive efficiency and supports sustainable decision-making in modern concrete technology.