Scholarlink

<p>Structural optimization may imply in many instances the presence of categorical, discrete, and continuous variables. This is particularly the case for composite structures due to the stacking sequence definition which involves not only structural considerations, but also a set of complex manufacturing- and cost-related constraints. While approaches ranging from relaxation of all variables to evolutionary methods have been proposed in the literature, it is still challenging to deal consistently with all manufacturing and structural constraints, especially in high-dimensional problems (hundred or more design variables). This work proposes a mixed-variables optimization formulation that can integrate a large variety of manufacturing and structural constraints in a monolithic approach based on a modified Bi-level Outer Approximation algorithm. The proposed framework allowed the treatment of composite optimization problems accounting for catalog selection of stacking sequences, manufacturing compatibility of adjacent elements, as well as variability control schemes. Various optimization formulations have been proposed tailored both for conventional laminates and for double–double laminates. The proposed approaches were tested on both an academic use-case, a multi-cell panel and, on an industrial use-case, a vertical tail plane front spar.</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

A unified mixed-variables framework for manufacturing-constrained optimization of composite laminates applied to conventional and double–double layups

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年9月10日

查看全文
  • Alejandro Garcia-Pis,
  • Christian Gogu,
  • Laurent Michel,
  • Stéphane Grihon,
  • Simone Coniglio

摘要

Structural optimization may imply in many instances the presence of categorical, discrete, and continuous variables. This is particularly the case for composite structures due to the stacking sequence definition which involves not only structural considerations, but also a set of complex manufacturing- and cost-related constraints. While approaches ranging from relaxation of all variables to evolutionary methods have been proposed in the literature, it is still challenging to deal consistently with all manufacturing and structural constraints, especially in high-dimensional problems (hundred or more design variables). This work proposes a mixed-variables optimization formulation that can integrate a large variety of manufacturing and structural constraints in a monolithic approach based on a modified Bi-level Outer Approximation algorithm. The proposed framework allowed the treatment of composite optimization problems accounting for catalog selection of stacking sequences, manufacturing compatibility of adjacent elements, as well as variability control schemes. Various optimization formulations have been proposed tailored both for conventional laminates and for double–double laminates. The proposed approaches were tested on both an academic use-case, a multi-cell panel and, on an industrial use-case, a vertical tail plane front spar.