Scholarlink

<p>This study optimized the properties of a novel hybrid composite made from plantain and coconut fibers within a polyester resin matrix using artificial intelligence techniques. Three key variables—hybridization ratio (plantain to coconut fibers), total fiber mass fraction, and fiber length—were used as design inputs. The outputs were composite density (g/cm<sup>3</sup>) and three mechanical properties: tensile strength (MPa), impact strength (J), and flexural strength (MPa). To handle this multi-objective problem, Grey Relational Analysis (GRA) was first used to combine the mechanical properties into a single Grey Relational Grade (GRG), representing overall performance. An Artificial Neural Network (ANN) was then trained to predict GRG and density based on the input variables, showing high predictive accuracy. These ANN models were embedded within the Non-Dominated Sorting Genetic Algorithm II (NSGA II) framework to simultaneously maximize GRG and minimize density. From the resulting Pareto front, the optimal solution was selected using the minimum-distance method, yielding a composite with a density of 0.7841&#xa0;g/cm<sup>3</sup>, tensile strength of 48.93&#xa0;MPa, impact strength of 60.09&#xa0;J, and flexural strength of 26.69&#xa0;MPa at a GRG value of 0.6494. These properties corresponded to a hybridization ratio of 1.67, a fiber mass fraction of 4.25&#xa0;wt%, and a fiber length of 40.7&#xa0;mm. The validation experiment yielded the following values: 0.8142&#xa0;g/cm<sup>3</sup>, 47.1000&#xa0;MPa, 59.5000&#xa0;J, and 25.6000&#xa0;MPa for density, tensile strength, impact strength, and flexural strength respectively, with relative errors of 3.8388%, 3.7471%, 0.9794 and 4.0992% respectively. The integrated GRA–ANN–NSGA II approach effectively tailored lightweight, high-performance biocomposites and offers a generalizable tool for sustainable material design. This result represents an important advancement of the role of artificial intelligence in advanced materials engineering and sustainable manufacturing.</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

Multi-objective optimization of plantain/coconut fibres hybrid reinforced polymer composite using ANN, GRA and genetic algorithm

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月21日

查看全文
  • Emmanuel G. Ikenga,
  • Chidozie C. Nwobi-Okoye,
  • Remy Uche

摘要

This study optimized the properties of a novel hybrid composite made from plantain and coconut fibers within a polyester resin matrix using artificial intelligence techniques. Three key variables—hybridization ratio (plantain to coconut fibers), total fiber mass fraction, and fiber length—were used as design inputs. The outputs were composite density (g/cm3) and three mechanical properties: tensile strength (MPa), impact strength (J), and flexural strength (MPa). To handle this multi-objective problem, Grey Relational Analysis (GRA) was first used to combine the mechanical properties into a single Grey Relational Grade (GRG), representing overall performance. An Artificial Neural Network (ANN) was then trained to predict GRG and density based on the input variables, showing high predictive accuracy. These ANN models were embedded within the Non-Dominated Sorting Genetic Algorithm II (NSGA II) framework to simultaneously maximize GRG and minimize density. From the resulting Pareto front, the optimal solution was selected using the minimum-distance method, yielding a composite with a density of 0.7841 g/cm3, tensile strength of 48.93 MPa, impact strength of 60.09 J, and flexural strength of 26.69 MPa at a GRG value of 0.6494. These properties corresponded to a hybridization ratio of 1.67, a fiber mass fraction of 4.25 wt%, and a fiber length of 40.7 mm. The validation experiment yielded the following values: 0.8142 g/cm3, 47.1000 MPa, 59.5000 J, and 25.6000 MPa for density, tensile strength, impact strength, and flexural strength respectively, with relative errors of 3.8388%, 3.7471%, 0.9794 and 4.0992% respectively. The integrated GRA–ANN–NSGA II approach effectively tailored lightweight, high-performance biocomposites and offers a generalizable tool for sustainable material design. This result represents an important advancement of the role of artificial intelligence in advanced materials engineering and sustainable manufacturing.