Scholarlink

<p>This study investigates the mechanical performance of high-strength concrete (HSC) incorporating silica fume (SF) as a supplementary cementitious material and waste glass aggregate (WGA) as a natural coarse aggregate replacement, with potential application in hybrid FRP-strengthened systems. Nine concrete mixtures were prepared with SF (0–20%) and WGA (0–100%) at a constant water-to-binder ratio of 0.38. Compressive strength (CS), tensile strength (TS), and flexural strength (FS) were evaluated at 7 and 28 days following IS standards. An optimal combination of 10–15% SF and 20–40% WGA achieved peak 28-day values of 72.8&#xa0;MPa (CS), 6.02&#xa0;MPa (TS), and 9.35&#xa0;MPa (FS), exceeding M60-grade benchmarks by 10–15%. To predict mechanical properties, ensemble machine learning models including Random Forest (RF) and Gradient Boosting (GB) were developed. Among these, the GB model demonstrated superior predictive performance across all output variables, achieving higher coefficients of determination (R² = 0.9813, 0.9803, and 0.9791 for CS, TS, and FS, respectively) along with lower prediction errors (MAE and RMSE) compared to RF. Feature importance analysis revealed that cement content and SF were the most influential parameters governing strength development. The results indicate that optimized SF–WGA concrete can provide a sustainable alternative to conventional HSC, while GB-based models offer a reliable approach for strength prediction.</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

Experimental and machine learning investigation of hybrid FRP strengthened high strength concrete

  • 期刊论文

  • 发表日期：

    2026年5月13日

查看全文
  • P. Sowmiyadevi,
  • R. Arvind Saravan,
  • C Vinothini,
  • G. Nakkeeran,
  • T. Subbulakshmi

摘要

This study investigates the mechanical performance of high-strength concrete (HSC) incorporating silica fume (SF) as a supplementary cementitious material and waste glass aggregate (WGA) as a natural coarse aggregate replacement, with potential application in hybrid FRP-strengthened systems. Nine concrete mixtures were prepared with SF (0–20%) and WGA (0–100%) at a constant water-to-binder ratio of 0.38. Compressive strength (CS), tensile strength (TS), and flexural strength (FS) were evaluated at 7 and 28 days following IS standards. An optimal combination of 10–15% SF and 20–40% WGA achieved peak 28-day values of 72.8 MPa (CS), 6.02 MPa (TS), and 9.35 MPa (FS), exceeding M60-grade benchmarks by 10–15%. To predict mechanical properties, ensemble machine learning models including Random Forest (RF) and Gradient Boosting (GB) were developed. Among these, the GB model demonstrated superior predictive performance across all output variables, achieving higher coefficients of determination (R² = 0.9813, 0.9803, and 0.9791 for CS, TS, and FS, respectively) along with lower prediction errors (MAE and RMSE) compared to RF. Feature importance analysis revealed that cement content and SF were the most influential parameters governing strength development. The results indicate that optimized SF–WGA concrete can provide a sustainable alternative to conventional HSC, while GB-based models offer a reliable approach for strength prediction.