Scholarlink

<p>When disposed of in landfills, printed circuit boards (PCBs) release hazardous substances. Furthermore, the global sand crisis has gained considerable attention among environmentalists over the last few years, and the United Nations has proposed some initiatives to reduce the use of river sand. Despite the existence of several promising sustainable alternatives to alluvial sand, there has been little effort to implement those initiatives in the construction industry. This study explores the use of recycled non-metallic PCB (NM-PCB) as a partial replacement for fine aggregate in mortar, combined with silica fume and marble powder. No data on their combined effects on mortar properties has been discovered yet which restricted large-scale application of NM-PCB in the construction industry. Therefore, this study investigates the potential of machine learning (ML) to predict the compressive strength (CS) of mortar containing NM-PCB, silica fume, and marble powder, as CS is the most critical property of concrete. A comprehensive experimental dataset of 270 samples was developed with eight (8) input variables, with compressive strength as the output. Due to the significant pozzolanic activity of silica fume and the micro filler effect of marble powder, their optimal dosage in mortar was determined using machine learning. The highest Compressive Strength (CS) achieved was 17.6&#xa0;MPa in a mix containing 5% SF, 5% MP and 3% NM-PCB. Linear Regression, Random Forest, and Extreme Gradient Boosting models were applied to predict compressive strength, with RF and XGB optimized via grid search and validated using k-fold cross-validation. Model performance was evaluated using R², RMSE, MAE, and MAPE. The Random Forest model was the most accurate, achieving an R² of 0.96, while XGBoost also performed well with R² = 0.90. SHAP analysis showed that silica fume (7–12&#xa0;kg/m³) and NM-PCB (13–22&#xa0;kg/m³) enhance compressive strength when combined with over 360&#xa0;kg/m³ of cement. ICE and PDP analyses highlighted curing age as the most influential factor, with silica fume, water–cement ratio, and superplasticizer dosage also significantly affecting strength. A graphical user interface was developed as a decision-support tool for researchers and preliminary mix design optimization for practitioners and externally validated with experimental results demonstrating that recycling NM-PCB in concrete promotes sustainable construction.</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

Compressive strength of mortar incorporating recycled nonmetallic printed circuit board waste: experimental and interpretable machine learning analysis

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年5月9日

查看全文
  • Asif Shahzad,
  • Sardar Kashif-ur-rehman,
  • Muhammad Faisal Javed,
  • Kashif Mehmood,
  • Hisham Alabduljabbar,
  • Furqan Ahmad

摘要

When disposed of in landfills, printed circuit boards (PCBs) release hazardous substances. Furthermore, the global sand crisis has gained considerable attention among environmentalists over the last few years, and the United Nations has proposed some initiatives to reduce the use of river sand. Despite the existence of several promising sustainable alternatives to alluvial sand, there has been little effort to implement those initiatives in the construction industry. This study explores the use of recycled non-metallic PCB (NM-PCB) as a partial replacement for fine aggregate in mortar, combined with silica fume and marble powder. No data on their combined effects on mortar properties has been discovered yet which restricted large-scale application of NM-PCB in the construction industry. Therefore, this study investigates the potential of machine learning (ML) to predict the compressive strength (CS) of mortar containing NM-PCB, silica fume, and marble powder, as CS is the most critical property of concrete. A comprehensive experimental dataset of 270 samples was developed with eight (8) input variables, with compressive strength as the output. Due to the significant pozzolanic activity of silica fume and the micro filler effect of marble powder, their optimal dosage in mortar was determined using machine learning. The highest Compressive Strength (CS) achieved was 17.6 MPa in a mix containing 5% SF, 5% MP and 3% NM-PCB. Linear Regression, Random Forest, and Extreme Gradient Boosting models were applied to predict compressive strength, with RF and XGB optimized via grid search and validated using k-fold cross-validation. Model performance was evaluated using R², RMSE, MAE, and MAPE. The Random Forest model was the most accurate, achieving an R² of 0.96, while XGBoost also performed well with R² = 0.90. SHAP analysis showed that silica fume (7–12 kg/m³) and NM-PCB (13–22 kg/m³) enhance compressive strength when combined with over 360 kg/m³ of cement. ICE and PDP analyses highlighted curing age as the most influential factor, with silica fume, water–cement ratio, and superplasticizer dosage also significantly affecting strength. A graphical user interface was developed as a decision-support tool for researchers and preliminary mix design optimization for practitioners and externally validated with experimental results demonstrating that recycling NM-PCB in concrete promotes sustainable construction.