Scholarlink

<p>Additive manufacturing enables the fabrication of lightweight polymer-composite components with tailored architectures, but their temperature-dependent thermo-viscoelastic performance remains challenging to predict. Short carbon fiber-reinforced nylon composites fabricated by fused filament fabrication (FFF) are particularly relevant for semi-structural applications where stiffness retention and damping must be maintained under thermal exposure. In this study, 27 FFF-printed specimens were fabricated by varying layer height (0.15–0.25&#xa0;mm), printing temperature (275–285&#xa0;°C), and infill density (90–100%), and their thermo-viscoelastic response was evaluated by dynamic mechanical analysis from 30 to 180&#xa0;°C at 1&#xa0;Hz. Storage modulus (M′) decreased from ~ 650–820&#xa0;MPa at low temperatures to ~ 100–180&#xa0;MPa at 180&#xa0;°C, corresponding to ~ 70–85% stiffness loss due to thermal softening. Lower layer height and higher infill density improved stiffness retention by ~ 15–25%, while damping was reduced by ~ 10–18%. Loss modulus (M″) decreased from ~ 70–100&#xa0;MPa to ~ 15–25&#xa0;MPa, whereas Tan δ showed a broad U-shaped response. Machine learning models were developed to predict M′, M″, and Tan δ from process parameters and temperature. Under conventional data splitting, MLP achieved R<sup>2</sup> ≈ 0.997 and 0.986 for M′ and M″, respectively, while ensemble models achieved R<sup>2</sup> ≈ 0.88–0.90 for Tan δ. Strict leave-one-process-combination-out (LOPCO) validation gave R<sup>2</sup> = 0.898 for M′, but lower values of R<sup>2</sup> = 0.456 for M″ and R<sup>2</sup> = 0.542 for Tan δ, indicating that prediction of damping-related responses is more challenging for fully withheld process combinations. This study establishes a process–property prediction framework for FFF-printed composites, with practical relevance to lightweight housings, brackets, fixtures, and semi-structural components where thermal stiffness retention and damping are critical.</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-assisted dynamic mechanical analysis of additively manufactured short carbon fiber-reinforced nylon composites

  • 期刊论文

  • 发表日期：

    2026年8月4日

查看全文
  • Dushyant Dubey,
  • Satinder Paul Singh,
  • Bijoya Kumar Behera

摘要

Additive manufacturing enables the fabrication of lightweight polymer-composite components with tailored architectures, but their temperature-dependent thermo-viscoelastic performance remains challenging to predict. Short carbon fiber-reinforced nylon composites fabricated by fused filament fabrication (FFF) are particularly relevant for semi-structural applications where stiffness retention and damping must be maintained under thermal exposure. In this study, 27 FFF-printed specimens were fabricated by varying layer height (0.15–0.25 mm), printing temperature (275–285 °C), and infill density (90–100%), and their thermo-viscoelastic response was evaluated by dynamic mechanical analysis from 30 to 180 °C at 1 Hz. Storage modulus (M′) decreased from ~ 650–820 MPa at low temperatures to ~ 100–180 MPa at 180 °C, corresponding to ~ 70–85% stiffness loss due to thermal softening. Lower layer height and higher infill density improved stiffness retention by ~ 15–25%, while damping was reduced by ~ 10–18%. Loss modulus (M″) decreased from ~ 70–100 MPa to ~ 15–25 MPa, whereas Tan δ showed a broad U-shaped response. Machine learning models were developed to predict M′, M″, and Tan δ from process parameters and temperature. Under conventional data splitting, MLP achieved R2 ≈ 0.997 and 0.986 for M′ and M″, respectively, while ensemble models achieved R2 ≈ 0.88–0.90 for Tan δ. Strict leave-one-process-combination-out (LOPCO) validation gave R2 = 0.898 for M′, but lower values of R2 = 0.456 for M″ and R2 = 0.542 for Tan δ, indicating that prediction of damping-related responses is more challenging for fully withheld process combinations. This study establishes a process–property prediction framework for FFF-printed composites, with practical relevance to lightweight housings, brackets, fixtures, and semi-structural components where thermal stiffness retention and damping are critical.