Scholarlink

<p>Material extrusion additive manufacturing presents a promising method for fabricating polymer components, with the optimization of tensile strength and elongation as key objectives. This study applies machine learning techniques to predict these properties using a dataset comprising 171 samples and six input variables: layer thickness, infill pattern, infill density, print speed, and nozzle and bed temperatures. Three models are compared: Linear Regressor, Decision Tree Regressor, and Support Vector Regressor. Among these, the Decision Tree Regressor demonstrates superior performance, achieving R2 values of 0.9955 (training), 0.8866 (testing), and 0.9751 (validation) for tensile strength, with corresponding MAE values of 0.4747, 2.9259, and 1.2769. For elongation, the R2 values reach 0.9627 (training), 0.8534 (testing), and 0.9990 (validation), alongside MAE values of 0.2273, 0.6170, and 0.0741. Feature importance analysis identifies bed temperature and layer thickness as the primary influencers. These findings underscore the effectiveness of decision tree regression for property prediction, thereby supporting data-driven optimization in additive manufacturing.</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

Predicting Mechanical Properties of Polymers in Additive Manufacturing: A Machine Learning Multioutput Regression Approach

  • 期刊论文

  • 发表日期：

    2025年12月17日

查看全文
  • Kunal Dewangan,
  • T. Ram Prabhu,
  • Srinivasu Gangi Setti

摘要

Material extrusion additive manufacturing presents a promising method for fabricating polymer components, with the optimization of tensile strength and elongation as key objectives. This study applies machine learning techniques to predict these properties using a dataset comprising 171 samples and six input variables: layer thickness, infill pattern, infill density, print speed, and nozzle and bed temperatures. Three models are compared: Linear Regressor, Decision Tree Regressor, and Support Vector Regressor. Among these, the Decision Tree Regressor demonstrates superior performance, achieving R2 values of 0.9955 (training), 0.8866 (testing), and 0.9751 (validation) for tensile strength, with corresponding MAE values of 0.4747, 2.9259, and 1.2769. For elongation, the R2 values reach 0.9627 (training), 0.8534 (testing), and 0.9990 (validation), alongside MAE values of 0.2273, 0.6170, and 0.0741. Feature importance analysis identifies bed temperature and layer thickness as the primary influencers. These findings underscore the effectiveness of decision tree regression for property prediction, thereby supporting data-driven optimization in additive manufacturing.

Graphical Abstract