Scholarlink

<p>The motivation for advanced multiscale hybrid bio-polymer nanocomposites stems from the growing and current need to develop superior thermal, mechanical, dielectric, and viscoelastic materials across industries such as automotive, electronics, and biomedicine. Research on bio-polymer nanocomposites, such as those integrating polylactic acid (PLA), graphene nanoplatelets (GNPs), carbon nanotubes (CNTs), and cellulose nanofibers, has so far demonstrated encouraging improvements in material properties. However, current models often struggle to accurately predict and optimize properties, mainly because of the complex interplay among various nanoscale constituents and processing conditions. A multicriteria methodology proposed here will be able to address this limitation using contemporary machine learning techniques in conjunction with a wide variety of optimization methodologies. The use of the proposed multicriteria methodology is dependent upon its ability to model complex behaviors of materials as well as estimate uncertainties associated with these behaviors. The model is useful for predicting thermal, mechanical, dielectric, and viscoelastic properties. Based on data captured through experiments and simulations, Random Forest Regression (RFR) shall analyze the data to identify the parameters relevant to determining the values of these properties. To find multi-objective trade-offs, paper adopt the particle swarm optimization (PSO) framework for search and also perform efficient fine-tuning of processing parameters via Bayesian optimization. The respect of physical laws is ensured through physics informed neural networks (PINNs), thereby improving accuracy in multi-scale modeling. Research carried out an analysis at three distinct levels: single effects of each nanocomponent, binary combinations of any of the four nanocomponents, and simultaneous effects of all four nanocomponents. The results reveal a 10–15% improvement in mechanical and thermal properties, along with a 30–40% reduction in simulation time. Hence, these results are beneficial towards further optimization of high durability and enhanced functional properties for bio-polymer nanocomposites.</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

Mechanical and thermal properties prediction and optimization of multiscale high performance hybrid biopolymer nanocomposites

  • 期刊论文

  • 发表日期：

    2026年7月6日

查看全文
  • Vaishali Mendhe,
  • Chetan D. Karekar,
  • Shradhesh Marve,
  • Lowlesh N. Yadav,
  • Hirkani Padwad,
  • Nischal Puri,
  • Latika Pinjarkar,
  • Rakesh Patel,
  • Aseel Smerat,
  • P. Jagadesh,
  • Karan Kumar Chaudhari,
  • Vikrant S. Vairagade

摘要

The motivation for advanced multiscale hybrid bio-polymer nanocomposites stems from the growing and current need to develop superior thermal, mechanical, dielectric, and viscoelastic materials across industries such as automotive, electronics, and biomedicine. Research on bio-polymer nanocomposites, such as those integrating polylactic acid (PLA), graphene nanoplatelets (GNPs), carbon nanotubes (CNTs), and cellulose nanofibers, has so far demonstrated encouraging improvements in material properties. However, current models often struggle to accurately predict and optimize properties, mainly because of the complex interplay among various nanoscale constituents and processing conditions. A multicriteria methodology proposed here will be able to address this limitation using contemporary machine learning techniques in conjunction with a wide variety of optimization methodologies. The use of the proposed multicriteria methodology is dependent upon its ability to model complex behaviors of materials as well as estimate uncertainties associated with these behaviors. The model is useful for predicting thermal, mechanical, dielectric, and viscoelastic properties. Based on data captured through experiments and simulations, Random Forest Regression (RFR) shall analyze the data to identify the parameters relevant to determining the values of these properties. To find multi-objective trade-offs, paper adopt the particle swarm optimization (PSO) framework for search and also perform efficient fine-tuning of processing parameters via Bayesian optimization. The respect of physical laws is ensured through physics informed neural networks (PINNs), thereby improving accuracy in multi-scale modeling. Research carried out an analysis at three distinct levels: single effects of each nanocomponent, binary combinations of any of the four nanocomponents, and simultaneous effects of all four nanocomponents. The results reveal a 10–15% improvement in mechanical and thermal properties, along with a 30–40% reduction in simulation time. Hence, these results are beneficial towards further optimization of high durability and enhanced functional properties for bio-polymer nanocomposites.