Scholarlink

<p>This study presents a comprehensive experimental and computational investigation into the mechanical performance optimization of epoxy-based composites reinforced with alkali-treated mulberry fibers and alumina micro-fillers. The primary objective was to enhance tensile strength, flexural strength, and energy absorption while minimizing tensile failure strain, through a hybrid methodology combining experimental design, machine learning, and multi-objective optimization. Composite laminates were fabricated using vacuum-assisted resin transfer molding (VARTM) based on a Taguchi L16 orthogonal design, varying fiber length (0.5-150&#xa0;mm), fiber content (30-60 wt.%), and filler content (3-12 wt.%). Mechanical characterization was performed per ASTM standards, and predictive modeling was implemented using a stacked ensemble framework integrating Random Forest, Gradient Boosting, and a polynomial meta-learner. The model achieved high predictive accuracy with R<sup>2</sup> values of 0.91-0.94 across all responses. Multi-objective Moth Swarm Algorithm (MOMSA) was employed to explore the Pareto front, identifying optimal configurations with superior performance. The knee-point solution yielded a tensile strength of 171.15&#xa0;MPa, flexural strength of 131.71&#xa0;MPa, energy absorption of 10.45&#xa0;J, and tensile failure strain of 1.02%, closely matching model predictions. Fractographic SEM analysis revealed hybrid failure modes including fiber pull-out, filler debonding, and matrix shear yielding, supporting the observed enhancements. The developed framework demonstrates the effectiveness of machine learning-guided multi-objective optimization for tailoring natural fiber-reinforced composites, offering a scalable approach for sustainable material design.</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

Modeling and Optimization of Alumina Filler-Enhanced Epoxy/Mulberry Fiber Composites Using Stacked Ensemble Modeling and Multi-objective Moth Swarm Algorithm

  • 期刊论文

  • 发表日期：

    2026年4月27日

查看全文
  • N. Shanmugasundaram,
  • P. Hariharasakthisudhan,
  • M. Jayaraj,
  • K. Logesh

摘要

This study presents a comprehensive experimental and computational investigation into the mechanical performance optimization of epoxy-based composites reinforced with alkali-treated mulberry fibers and alumina micro-fillers. The primary objective was to enhance tensile strength, flexural strength, and energy absorption while minimizing tensile failure strain, through a hybrid methodology combining experimental design, machine learning, and multi-objective optimization. Composite laminates were fabricated using vacuum-assisted resin transfer molding (VARTM) based on a Taguchi L16 orthogonal design, varying fiber length (0.5-150 mm), fiber content (30-60 wt.%), and filler content (3-12 wt.%). Mechanical characterization was performed per ASTM standards, and predictive modeling was implemented using a stacked ensemble framework integrating Random Forest, Gradient Boosting, and a polynomial meta-learner. The model achieved high predictive accuracy with R2 values of 0.91-0.94 across all responses. Multi-objective Moth Swarm Algorithm (MOMSA) was employed to explore the Pareto front, identifying optimal configurations with superior performance. The knee-point solution yielded a tensile strength of 171.15 MPa, flexural strength of 131.71 MPa, energy absorption of 10.45 J, and tensile failure strain of 1.02%, closely matching model predictions. Fractographic SEM analysis revealed hybrid failure modes including fiber pull-out, filler debonding, and matrix shear yielding, supporting the observed enhancements. The developed framework demonstrates the effectiveness of machine learning-guided multi-objective optimization for tailoring natural fiber-reinforced composites, offering a scalable approach for sustainable material design.