Scholarlink

<p>The rising environmental impact from natural aggregate extraction necessitates the adoption of sustainable alternatives like Recycled Coarse Aggregate (RCA). However, challenges linked to RCA’s variable quality and porosity limit its mainstream use in structural concrete. This study presents a dual-framework evaluation, experimental characterization and machine learning (ML) modeling of concrete incorporating RCA from 25 to 100% under two water–cement ratios (0.6 and 0.65). Mechanical characteristics were assessed, with the 100% RCA mix at 0.6 w/c ratio demonstrating peak compressive strength (CS) of 30.33&#xa0;MPa. Concurrently, seven ML algorithms, including XGBoost and Random Forest, were trained on an enriched dataset (experimental and literature) to forecast CS based on mix design parameters. XGBoost achieved superior performance (R<sup>2</sup> = 0.929, RMSE = 2.71&#xa0;MPa), identifying curing days, cement content, and w/c ratio as key features. A web-based Streamlit application was developed to enable real-time compressive strength forecasting from user inputs. The integration of empirical testing and advanced ML modeling demonstrates a robust and scalable approach to optimize RCA concrete for sustainable structural applications.</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

Hybrid assessment of concrete with recycled aggregate using experiment and machine learning techniques for structural performance optimization

  • 期刊论文

  • 发表日期：

    2025年9月22日

查看全文
  • Sagar Paruthi,
  • Rupesh Kumar Tipu,
  • Nadeem Nisar,
  • Nakeem Nisar,
  • Mohammad Waris,
  • Afzal Husain Khan

摘要

The rising environmental impact from natural aggregate extraction necessitates the adoption of sustainable alternatives like Recycled Coarse Aggregate (RCA). However, challenges linked to RCA’s variable quality and porosity limit its mainstream use in structural concrete. This study presents a dual-framework evaluation, experimental characterization and machine learning (ML) modeling of concrete incorporating RCA from 25 to 100% under two water–cement ratios (0.6 and 0.65). Mechanical characteristics were assessed, with the 100% RCA mix at 0.6 w/c ratio demonstrating peak compressive strength (CS) of 30.33 MPa. Concurrently, seven ML algorithms, including XGBoost and Random Forest, were trained on an enriched dataset (experimental and literature) to forecast CS based on mix design parameters. XGBoost achieved superior performance (R2 = 0.929, RMSE = 2.71 MPa), identifying curing days, cement content, and w/c ratio as key features. A web-based Streamlit application was developed to enable real-time compressive strength forecasting from user inputs. The integration of empirical testing and advanced ML modeling demonstrates a robust and scalable approach to optimize RCA concrete for sustainable structural applications.