Scholarlink

<p>Roller Compacted Concrete (RCC) has garnered substantial attention in the field of pavement engineering as a result of its affordability and expeditious construction process in comparison to conventional concrete. Although the challenge of accurately predicting the compressive strength (CS) and refining the mix design remains challenging. This study presents a machine learning (ML) based method for very accurate CS of RCC prediction with the utilization of models such as Linear Regression (LR), Random Forests (RF), Decision Trees (DT), and Artificial Neural Networks (ANN). The finding shows that the DT model exhibited the highest performance, achieving an R<sup>2</sup> value of 0.93 and RMSE of 2.167&#xa0;MPa on the testing set. A Partial Dependence Plot (PDP) analysis was conducted to evaluate the influence of various input parameters on CS. The results revealed that cement and age strength content had the greatest beneficial influence on CS, whereas fine aggregate had the least influence. This research demonstrates the endless possibilities of machine learning models, particularly the DT model, in forecasting the CS of RCC and provides valuable insights into how various input datasets affect RCC strength. The study suggests future research avenues, including multi-objective optimization and expanding the model’s application to a greater variety of RCC mixtures for improved applicability.</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

Machine learning based prediction of compressive strength in roller compacted concrete: a comparative study with PDP analysis

  • 期刊论文

  • 发表日期：

    2025年3月18日

查看全文
  • Charuta Waghmare,
  • Mohammad Gulfam Pathan,
  • Syed Aamir Hussain,
  • Tripti Gupta,
  • Anshul Nikhade,
  • Monali Wagh,
  • Khalid Ansari

摘要

Roller Compacted Concrete (RCC) has garnered substantial attention in the field of pavement engineering as a result of its affordability and expeditious construction process in comparison to conventional concrete. Although the challenge of accurately predicting the compressive strength (CS) and refining the mix design remains challenging. This study presents a machine learning (ML) based method for very accurate CS of RCC prediction with the utilization of models such as Linear Regression (LR), Random Forests (RF), Decision Trees (DT), and Artificial Neural Networks (ANN). The finding shows that the DT model exhibited the highest performance, achieving an R2 value of 0.93 and RMSE of 2.167 MPa on the testing set. A Partial Dependence Plot (PDP) analysis was conducted to evaluate the influence of various input parameters on CS. The results revealed that cement and age strength content had the greatest beneficial influence on CS, whereas fine aggregate had the least influence. This research demonstrates the endless possibilities of machine learning models, particularly the DT model, in forecasting the CS of RCC and provides valuable insights into how various input datasets affect RCC strength. The study suggests future research avenues, including multi-objective optimization and expanding the model’s application to a greater variety of RCC mixtures for improved applicability.