| Summary: | In the aerospace industry, Multidisciplinary Design Optimisation (MDO) is increasingly being used in the early phase design of aircraft structures.
However, the computational complexity associated to this tool limits its application to coarse FE-models, which are not suffciently accurate to capture the internal deformation of components, such as manholes, cut-outs, bulkheads, and stringer run-outs.
Because of this, aircraft designers are currently unable to evaluate the influence of these components on the sizing of primary structures. This causes the structural designs obtained through MDO to be suboptimal at best and possibly unfeasible, which limits the benefits and thwarts a broader application of this methodology.
In this thesis a novel methodology for the preliminary sizing of aircraft structures is introduced. The proposed global-local MDO procedure relies on the use of a coarse FE- model combined with multiple finer models for the accurate evaluation of components with a complex geometry. Thanks to the introduction of an ad-hoc sensitivity analysis, to consider the coupling between different models, the novel methodology ensures the optimality and feasibility of the computed design.
The impact on computational cost of adopting the proposed global-local strategy is limited, provided that the total number of constraints and design variables is not greatly increased.
Where the reference procedure would fail to find a locally feasible design, the proposed global-local approach successfully finds a multidisciplinary optimal design, which does not violate local constraints. Thus, the methodology enables designers to account for the effect of components with a complex geometry earlier in the design process and reduces the risk of major delays in the product development cycle.
|
|---|