Scholarlink

<p>This study investigates the low-velocity impact performance of epoxy-based composites reinforced with short carbon fibers (SCF), cenospheres (CN), and their hybrid combinations, through a comprehensive approach involving experimental testing, finite element (FE) simulations, and machine learning (ML) predictions. Energy absorption (EA), specific energy absorption (SEA), and peak impact force were evaluated to understand the influence of reinforcement type and composition on impact resistance and weight efficiency. Results revealed that neat epoxy exhibited the lowest impact tolerance due to its brittle nature, while SCF-reinforced composites (EP-SCF20) achieved high energy absorption (35.52&#xa0;J) and peak force (692&#xa0;N) but moderate SEA due to increased density. Conversely, CN-filled composites (EP-CN20) showed similar energy absorption (30.38&#xa0;J) but superior SEA (28.66&#xa0;J&#xa0;cm<sup>3</sup>/g) owing to reduced density and energy dissipation via microcracking and particle crushing. Hybrid composites outperformed individual reinforcements, with EP-HYB2 (10% SCF + 10% CN) exhibiting the highest SEA (33.78&#xa0;J&#xa0;cm<sup>3</sup>/g) and significant energy absorption (38.18&#xa0;J), attributed to synergistic mechanisms including fiber bridging, pull-out, and cenosphere-induced crack deflection. The novelty of this work lies in the integrative exploration of hybrid reinforcement effects on both impact resistance and mass-specific performance, combining experimental mechanics with validated finite element simulations and data-driven machine learning models. The study establishes a clear phenomenological understanding of how hybrid filler ratios affect failure modes, peak force response, and energy dissipation under dynamic loading. FE simulations closely matched experimental trends, validating the numerical model’s reliability. ML-based regression models further demonstrated the potential for predictive material design, with the linear model showing reasonable accuracy for energy absorption (<i>R</i><sup>2</sup> = 0.512), while the decision tree model underperformed for peak force prediction due to data limitations. This multi-scale, hybridized methodology offers a novel framework for developing lightweight, impact-resistant polymer composites tailored for aerospace and automotive structural applications.</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

Integrated experimental, numerical, and machine learning approach to impact resistance of SCF-cenosphere epoxy composites

  • 期刊论文

  • 发表日期：

    2026年1月6日

查看全文
  • Vishwas Mahesh,
  • Dineshkumar Harursampath,
  • S. Sreelakshmi

摘要

This study investigates the low-velocity impact performance of epoxy-based composites reinforced with short carbon fibers (SCF), cenospheres (CN), and their hybrid combinations, through a comprehensive approach involving experimental testing, finite element (FE) simulations, and machine learning (ML) predictions. Energy absorption (EA), specific energy absorption (SEA), and peak impact force were evaluated to understand the influence of reinforcement type and composition on impact resistance and weight efficiency. Results revealed that neat epoxy exhibited the lowest impact tolerance due to its brittle nature, while SCF-reinforced composites (EP-SCF20) achieved high energy absorption (35.52 J) and peak force (692 N) but moderate SEA due to increased density. Conversely, CN-filled composites (EP-CN20) showed similar energy absorption (30.38 J) but superior SEA (28.66 J cm3/g) owing to reduced density and energy dissipation via microcracking and particle crushing. Hybrid composites outperformed individual reinforcements, with EP-HYB2 (10% SCF + 10% CN) exhibiting the highest SEA (33.78 J cm3/g) and significant energy absorption (38.18 J), attributed to synergistic mechanisms including fiber bridging, pull-out, and cenosphere-induced crack deflection. The novelty of this work lies in the integrative exploration of hybrid reinforcement effects on both impact resistance and mass-specific performance, combining experimental mechanics with validated finite element simulations and data-driven machine learning models. The study establishes a clear phenomenological understanding of how hybrid filler ratios affect failure modes, peak force response, and energy dissipation under dynamic loading. FE simulations closely matched experimental trends, validating the numerical model’s reliability. ML-based regression models further demonstrated the potential for predictive material design, with the linear model showing reasonable accuracy for energy absorption (R2 = 0.512), while the decision tree model underperformed for peak force prediction due to data limitations. This multi-scale, hybridized methodology offers a novel framework for developing lightweight, impact-resistant polymer composites tailored for aerospace and automotive structural applications.