Scholarlink

<p>The construction industry’s extensive use of natural resources and significant contribution to global waste production necessitates the investigation of sustainable alternatives such as recycled aggregates. This study explores the reuse of construction and demolition concrete waste by optimizing and predicting the mechanical properties of concrete containing recycled coarse aggregates (RCA). RCA replaced natural coarse aggregates at varying percentages ranging from 25 to 100%. The mechanical properties investigated include compressive strength, flexural strength, split tensile strength, impact strength, bond strength, and overall concrete quality through non-destructive testing on conventional concrete (CC) and recycled aggregate concrete (RAC). The SEM/EDS, XRD and FTIR were employed to explore the hydration phases, bonding between RCA and slurry, pore structure, morphology and mineralogy of the hardened RAC. Response surface methodology and artificial neural network (ANN) is an optimization and prediction technique utilized to RAC. Experimental results showed that up to 25% RCA replacement yields mechanical performance comparable to CC, with RAC25 achieving 7.12&#xa0;MPa in flexural strength versus 7.98&#xa0;MPa for CC, and showing only a slight reduction in impact strength and bond characteristics. Beyond 50% replacement, noticeable reductions were observed, particularly in RAC100, due to increased porosity and micro-cracks. However, RAC mixes maintained acceptable quality in rebound hammer and UPV tests. RSM model optimization and prediction of the data from experimental results significantly show an <i>R</i><sup>2</sup> confirming (<i>R</i><sup>2</sup> &gt; 0.9380), (<i>R</i><sup>2</sup> &gt; 0.9941), and (<i>R</i><sup>2</sup> &gt; 0.9622). Furthermore, the ANN model captured the same data from RSM optimization as indicated by the high R threshold-(<i>R</i> &gt; 0.9997), (<i>R</i> &gt; 0.99121), (<i>R</i> &gt; 0.99963). This study reveals that RAC maintains comparable strengths to CC at 28&#xa0;days. Machine learning permits reliable prediction of concrete's mechanical properties, which helps civil engineers maintain valuable time, labour, and resources.</p> Graphical abstract <p></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

An investigation into the mechanical and microstructural properties of concrete utilizing recycled aggregate, incorporating optimization and prediction

  • 期刊论文

  • 发表日期：

    2025年6月8日

查看全文
  • S. S. Balaji,
  • S. Prakash Chandar,
  • P. T. Ravichandran

摘要

The construction industry’s extensive use of natural resources and significant contribution to global waste production necessitates the investigation of sustainable alternatives such as recycled aggregates. This study explores the reuse of construction and demolition concrete waste by optimizing and predicting the mechanical properties of concrete containing recycled coarse aggregates (RCA). RCA replaced natural coarse aggregates at varying percentages ranging from 25 to 100%. The mechanical properties investigated include compressive strength, flexural strength, split tensile strength, impact strength, bond strength, and overall concrete quality through non-destructive testing on conventional concrete (CC) and recycled aggregate concrete (RAC). The SEM/EDS, XRD and FTIR were employed to explore the hydration phases, bonding between RCA and slurry, pore structure, morphology and mineralogy of the hardened RAC. Response surface methodology and artificial neural network (ANN) is an optimization and prediction technique utilized to RAC. Experimental results showed that up to 25% RCA replacement yields mechanical performance comparable to CC, with RAC25 achieving 7.12 MPa in flexural strength versus 7.98 MPa for CC, and showing only a slight reduction in impact strength and bond characteristics. Beyond 50% replacement, noticeable reductions were observed, particularly in RAC100, due to increased porosity and micro-cracks. However, RAC mixes maintained acceptable quality in rebound hammer and UPV tests. RSM model optimization and prediction of the data from experimental results significantly show an R2 confirming (R2 > 0.9380), (R2 > 0.9941), and (R2 > 0.9622). Furthermore, the ANN model captured the same data from RSM optimization as indicated by the high R threshold-(R > 0.9997), (R > 0.99121), (R > 0.99963). This study reveals that RAC maintains comparable strengths to CC at 28 days. Machine learning permits reliable prediction of concrete's mechanical properties, which helps civil engineers maintain valuable time, labour, and resources.

Graphical abstract