Scholarlink

<p>This study proposes a machine learning (ML)-based approach to predict the flexural strength of short carbon fiber-reinforced polylactic acid (SCFR-PLA) composites. The relevance of this research lies in the increasing demand for advanced materials in construction, where the accurate prediction of mechanical properties is crucial for performance optimization. Using ML algorithms, the model incorporated printing parameters such as nozzle temperature, layer thickness, bed temperature, and strain measurements as input features. The flexural strength values from mechanical testing served as the target variables for model training. A comprehensive dataset, consisting of 2625 samples, was split into 70% for training, 15% for validation, and 15% for testing. The models were evaluated using four regression techniques: Artificial Neural Networks (ANN), Support Vector Machines (SVM), Regression Trees (RT), and Linear Regression (LR). The results indicate that ANN and SVM outperformed the other models, achieving a validation R-squared value of 0.98 and a Root Mean Square Error (RMSE) of 1.68. These findings demonstrate the potential of ML-based models for predicting the flexural strength of SCFR-PLA composites under various processing conditions. The findings of this study are significant for the construction industry, where material performance prediction can enhance the design and manufacturing processes of composite materials.</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

A machine learning approach for predicting flexural strength in short carbon fiber reinforced PLA composites

  • 期刊论文

  • 发表日期：

    2025年5月14日

查看全文
  • Yunus Emre Karabacak,
  • Bahri Barış Vatandaş,
  • Recep Gümrük

摘要

This study proposes a machine learning (ML)-based approach to predict the flexural strength of short carbon fiber-reinforced polylactic acid (SCFR-PLA) composites. The relevance of this research lies in the increasing demand for advanced materials in construction, where the accurate prediction of mechanical properties is crucial for performance optimization. Using ML algorithms, the model incorporated printing parameters such as nozzle temperature, layer thickness, bed temperature, and strain measurements as input features. The flexural strength values from mechanical testing served as the target variables for model training. A comprehensive dataset, consisting of 2625 samples, was split into 70% for training, 15% for validation, and 15% for testing. The models were evaluated using four regression techniques: Artificial Neural Networks (ANN), Support Vector Machines (SVM), Regression Trees (RT), and Linear Regression (LR). The results indicate that ANN and SVM outperformed the other models, achieving a validation R-squared value of 0.98 and a Root Mean Square Error (RMSE) of 1.68. These findings demonstrate the potential of ML-based models for predicting the flexural strength of SCFR-PLA composites under various processing conditions. The findings of this study are significant for the construction industry, where material performance prediction can enhance the design and manufacturing processes of composite materials.

Graphical Abstract