Scholarlink

<p>Concrete remains a cornerstone of the construction industry, however, the depletion of natural aggregates and the associated environmental hazards presents significant environmental challenges, while alternative solutions remain costly. This study evaluates the strength and durability of eco-friendly concrete by incorporating marble dust (MD: 5%, 10%, and 15%) and glass fiber (GF: 0.5%, 0.75%, and 1%). The optimal experimental design mix (MD10% and GF1%) enhanced compressive strength by 3.5%, 4.1%, and 5.3% at 7, 14, and 28&#xa0;days, respectively. Similarly, MD10% + GF0.75% improved flexural strength (4.4%, 6.3%, 8.5%), split tensile strength (7.8%, 5%, 8.8%), rebound hammer results (5.6%, 6.1%, and 7.2%) and Sulfate exposure was reduced by 5%–8%, 9%–12% at 28 and 90&#xa0;days respectively. The Random Forest (RF) Approach and Shapley method identified cement and the number of curing days as the dominant controls of compressive strength. Machine learning models (Extreme Gradient Boost (XGB), Random Forest (RF) and Support Vector Machines (SVMs)) effectively predicted compressive strength. The XGB model exhibited the highest predictive accuracy (R<sup>2</sup> = 0.9691, NSE = 0.994, ubRMSE = 0.047), followed by Random Forest (RF) model (R<sup>2</sup> = 0.986, NSE = 0.9544, ubRMSE = 0.0423) and the Support Vector Machine determine the accuracy of the model's predicted versus observed prediction (R<sup>2</sup> = 0.9972, NSE = 0.9232, ubRMSE = 0.0542). XGB model has the scalability to fit complex relationships within big datasets, thus improving accuracy. The optimum combination, MD10% and GF0.75% is a cost-effective and sustainable concrete option that addresses environmental concerns.</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 and strength optimization of sustainable concrete incorporating glass fiber and marble dust using machine learning techniques

  • 期刊论文

  • 发表日期：

    2025年5月16日

查看全文
  • Anmol Manhas,
  • Shalika Mehta,
  • Sahil Sharma

摘要

Concrete remains a cornerstone of the construction industry, however, the depletion of natural aggregates and the associated environmental hazards presents significant environmental challenges, while alternative solutions remain costly. This study evaluates the strength and durability of eco-friendly concrete by incorporating marble dust (MD: 5%, 10%, and 15%) and glass fiber (GF: 0.5%, 0.75%, and 1%). The optimal experimental design mix (MD10% and GF1%) enhanced compressive strength by 3.5%, 4.1%, and 5.3% at 7, 14, and 28 days, respectively. Similarly, MD10% + GF0.75% improved flexural strength (4.4%, 6.3%, 8.5%), split tensile strength (7.8%, 5%, 8.8%), rebound hammer results (5.6%, 6.1%, and 7.2%) and Sulfate exposure was reduced by 5%–8%, 9%–12% at 28 and 90 days respectively. The Random Forest (RF) Approach and Shapley method identified cement and the number of curing days as the dominant controls of compressive strength. Machine learning models (Extreme Gradient Boost (XGB), Random Forest (RF) and Support Vector Machines (SVMs)) effectively predicted compressive strength. The XGB model exhibited the highest predictive accuracy (R2 = 0.9691, NSE = 0.994, ubRMSE = 0.047), followed by Random Forest (RF) model (R2 = 0.986, NSE = 0.9544, ubRMSE = 0.0423) and the Support Vector Machine determine the accuracy of the model's predicted versus observed prediction (R2 = 0.9972, NSE = 0.9232, ubRMSE = 0.0542). XGB model has the scalability to fit complex relationships within big datasets, thus improving accuracy. The optimum combination, MD10% and GF0.75% is a cost-effective and sustainable concrete option that addresses environmental concerns.