Scholarlink

<p>With the escalating threat of global warming, the construction sector is shifting towards sustainable practices to mitigate its significant carbon footprint. A key challenge, however, is minimizing dependence on resource-intensive laboratory testing while advancing the use of eco-friendly materials. To address this, the study explores the application of machine learning to accurately predict the compressive strength of eco-efficient bricks made by partially replacing sand with stone dust. Bricks were manufactured with varying combinations of fly ash and stone dust, followed by compressive strength testing. Experimental data were utilized to train four machine learning models: artificial neural networks (ANN), support vector machines (SVM), classification and regression trees (CART), and Gaussian process regression (GPR), using five input parameters—cement, sand, fly ash, stone dust, and curing time. The maximum compressive strength of the bricks was achieved with 40% fly ash and 80% sand replaced by stone dust after 28&#xa0;days of curing. The results of machine learning models show that ANN outperformed and ranked as the most reliable model across the training and testing phases with Root Mean Square Error (RMSE), Pearson’s correlation coefficient (R), and Objective Function (OBJ) values of 1.3769, 0.9564, and 1.2227, respectively. Integrating various evaluation indices, the model performance can be ranked as ANN &gt; SVM &gt; CART &gt; GPR. Sensitivity analysis identified fly ash content and curing time as the most influential variables. This research presents the machine learning tool to advance sustainable construction efforts by enhancing the precision and efficiency of eco-brick production, paving the way for innovative and environmentally responsible building solutions.</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

Prediction of compressive strength of stone dust-based fly ash bricks: experimental study and advanced machine learning algorithm

  • 期刊论文

  • 发表日期：

    2025年8月13日

查看全文
  • Mehtab Alam,
  • Adil Khan,
  • Muhammad Umar,
  • Tauqir Ahmed,
  • Ahmad Yar Mazari,
  • Muhammad Usman Arshad

摘要

With the escalating threat of global warming, the construction sector is shifting towards sustainable practices to mitigate its significant carbon footprint. A key challenge, however, is minimizing dependence on resource-intensive laboratory testing while advancing the use of eco-friendly materials. To address this, the study explores the application of machine learning to accurately predict the compressive strength of eco-efficient bricks made by partially replacing sand with stone dust. Bricks were manufactured with varying combinations of fly ash and stone dust, followed by compressive strength testing. Experimental data were utilized to train four machine learning models: artificial neural networks (ANN), support vector machines (SVM), classification and regression trees (CART), and Gaussian process regression (GPR), using five input parameters—cement, sand, fly ash, stone dust, and curing time. The maximum compressive strength of the bricks was achieved with 40% fly ash and 80% sand replaced by stone dust after 28 days of curing. The results of machine learning models show that ANN outperformed and ranked as the most reliable model across the training and testing phases with Root Mean Square Error (RMSE), Pearson’s correlation coefficient (R), and Objective Function (OBJ) values of 1.3769, 0.9564, and 1.2227, respectively. Integrating various evaluation indices, the model performance can be ranked as ANN > SVM > CART > GPR. Sensitivity analysis identified fly ash content and curing time as the most influential variables. This research presents the machine learning tool to advance sustainable construction efforts by enhancing the precision and efficiency of eco-brick production, paving the way for innovative and environmentally responsible building solutions.