In this work, a multi-scale optimization strategy for lightweight structures, based on a global-local modelling approach is presented. The approach is applied to a realistic wing structure of a civil aircraft. The preliminary design of the wing can be formulated as a constrained optimization problem, involving several requirements at the different scales of the structure. The proposed strategy is characterized by two main features. Firstly, the problem is formulated in the most general sense, by including all design variables involved at each problem scale. Secondly, two scales are considered: (i) the structure macroscopic scale, where low-fidelity numerical models are used; (ii) the structure mesoscopic scale (or component-level), where enhanced models are involved. In particular, the structural responses are evaluated at both global and local scales, avoiding the use of approximated analytical methods. To this end, fully parametric global and local finite element models are interfaced with an in-house genetic algorithm. Refined models are created only for the most critical regions of the structure, and linked to the global one by means of a dedicated sub-modelling approach.

中文翻译：

---

通过全局/局部建模方法进行翼盒的多尺度最小重量设计

在这项工作中，提出了一种基于全局-局部建模方法的轻量级结构的多尺度优化策略。该方法应用于民用飞机的现实机翼结构。机翼的初步设计可以表述为一个有约束的优化问题，涉及结构不同尺度的几个要求。所提出的策略具有两个主要特征。首先，问题是从最一般的意义上表述的，包括在每个问题尺度上涉及的所有设计变量。其次，考虑两个尺度：（i）结构宏观尺度，使用低保真数值模型；(ii) 结构细观尺度（或组件级），其中涉及增强模型。特别是，在全局和局部尺度上评估结构响应，避免使用近似分析方法。为此，完全参数化的全局和局部有限元模型与内部遗传算法相连接。仅针对结构的最关键区域创建精细模型，并通过专用的子建模方法链接到全局区域。