Scholarlink

<p>In this study, the hardness and tensile strength of polypropylene/carbon nanotubes (PP/CNT) and low-density polyethylene/carbon nanotubes (LDPE/CNT) composite materials were predicted using machine learning (ML). The composites were fabricated through a microwave-assisted manufacturing process, and their mechanical properties were evaluated at an elevated temperature of 100&#xa0;°C, which remains below the melting points of both PP and LDPE. Hardness and tensile tests were conducted using a Rockwell hardness tester and a universal testing machine, respectively. To predict these properties, various ML algorithms, i.e., extreme gradient boosting (XGBoost), K-nearest neighbors (KNN), and random forest (RF), were employed. Among these, the RF model demonstrated the highest accuracy and consistency, particularly for PP/CNT composites. Performance metrics, including root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (<i>R</i><sup>2</sup>), indicated that the RF model outperformed other algorithms. For PP/CNT composites, it exhibited minimal variations of ± 0.5 for hardness and ± 2 for tensile strength, highlighting strong predictive accuracy. In contrast, LDPE/CNT composites showed greater variations of ± 1 and ± 3, respectively, due to a weaker correlation between predicted and experimental values. Overall, this study underscores the potential of ML, particularly the RF algorithm, in providing reliable predictions of composite material properties.</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

Data-driven insights into the high-temperature behavior of polymer/carbon nanotubes nanocomposites

  • 期刊论文

  • 发表日期：

    2025年9月16日

查看全文
  • Harshit Sharma,
  • Gaurav Arora,
  • Raj Kumar,
  • Suman Debnath,
  • Suchart Siengchin

摘要

In this study, the hardness and tensile strength of polypropylene/carbon nanotubes (PP/CNT) and low-density polyethylene/carbon nanotubes (LDPE/CNT) composite materials were predicted using machine learning (ML). The composites were fabricated through a microwave-assisted manufacturing process, and their mechanical properties were evaluated at an elevated temperature of 100 °C, which remains below the melting points of both PP and LDPE. Hardness and tensile tests were conducted using a Rockwell hardness tester and a universal testing machine, respectively. To predict these properties, various ML algorithms, i.e., extreme gradient boosting (XGBoost), K-nearest neighbors (KNN), and random forest (RF), were employed. Among these, the RF model demonstrated the highest accuracy and consistency, particularly for PP/CNT composites. Performance metrics, including root mean square error (RMSE), mean absolute error (MAE), and coefficient of determination (R2), indicated that the RF model outperformed other algorithms. For PP/CNT composites, it exhibited minimal variations of ± 0.5 for hardness and ± 2 for tensile strength, highlighting strong predictive accuracy. In contrast, LDPE/CNT composites showed greater variations of ± 1 and ± 3, respectively, due to a weaker correlation between predicted and experimental values. Overall, this study underscores the potential of ML, particularly the RF algorithm, in providing reliable predictions of composite material properties.

Graphical abstract