Scholarlink

<p>This study applies supervised learning to predict the mechanical properties of biodegradable PLA/wood composites produced by Fused Filament Fabrication (FFF). PLA reinforced with 20&#xa0;wt% wood fibers was printed under varying process conditions layer height, infill density, infill pattern, and raster orientation following a full factorial design of 81 experiments. Compressive strength, hardness, and tensile strength were experimentally evaluated to assess the influence of these parameters. Statistical analysis using ANOVA and the Taguchi method identified the most significant factors and their optimal combinations for improved performance. A Levenberg–Marquardt (LM)-based supervised learning model was developed to capture the nonlinear relationships between printing parameters and mechanical responses. The LM model showed strong predictive capability and consistent convergence, demonstrating its reliability for modelling the complex behavior of FFF-processed composites. Integrating statistical methods with data-driven modelling enables efficient optimization of process parameters and reduces experimental effort. The proposed approach provides a robust framework for enhancing the mechanical performance of sustainable PLA/wood composites and can be extended to other additive manufacturing systems for the design of high-performance, biodegradable materials.</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 biodegradable PLA/Wood composites fabricated by 3D printing: a supervised learning approach

  • 期刊论文

  • 发表日期：

    2025年12月22日

查看全文
  • Nikhil Bharat,
  • D. Veeman,
  • Vijay Kumar

摘要

This study applies supervised learning to predict the mechanical properties of biodegradable PLA/wood composites produced by Fused Filament Fabrication (FFF). PLA reinforced with 20 wt% wood fibers was printed under varying process conditions layer height, infill density, infill pattern, and raster orientation following a full factorial design of 81 experiments. Compressive strength, hardness, and tensile strength were experimentally evaluated to assess the influence of these parameters. Statistical analysis using ANOVA and the Taguchi method identified the most significant factors and their optimal combinations for improved performance. A Levenberg–Marquardt (LM)-based supervised learning model was developed to capture the nonlinear relationships between printing parameters and mechanical responses. The LM model showed strong predictive capability and consistent convergence, demonstrating its reliability for modelling the complex behavior of FFF-processed composites. Integrating statistical methods with data-driven modelling enables efficient optimization of process parameters and reduces experimental effort. The proposed approach provides a robust framework for enhancing the mechanical performance of sustainable PLA/wood composites and can be extended to other additive manufacturing systems for the design of high-performance, biodegradable materials.