Rapid Sizing of a Composite Wing Structure

摘要

This paper presents a method for rapidly optimising an aircraft composite wing-box structure to ascertain the level of meaningful information that can be realised from multi-fidelity wing-box models. The wing-box is optimised for minimum structural weight whilst maintain structural credibility via an aerospace structural design, analysis, and optimisation software called HyperSizer. The optimisation tool is capable of efficiently narrowing the material trade space by selecting optimal laminates from a given set of laminate library and find an optimal geometry. The proposed procedure consists of constructing, and optimising three smeared models (low-, medium-, and high-fidelity) with increasing discretisation, and a more refined discrete model, akin to that employed in real aerospace design practice using a finite element (FE) wing model. These models are discretised into "components" which act as optimisation zones. The ability of the smeared models to provide the optimal design (minimum weight) and identify the critical load cases are benchmarked against the discrete model. The results indicated that the high-fidelity model can accurately estimate the wing-box weight within 1% error. Even with a low-fidelity model the critical load cases contributing to 96% of the wing-box weight can be accurately predicted within a fraction of the time that would be needed by the discrete model. The performance of the low-fidelity model suggests that it can offer a significant opportunity to down-select critical load cases from a pool of load cases, and facilitate in making key structural decisions before investing time and resource in a more detailed FEM. The time savings demonstrated canbeused to provide more design iterations in a given time frame thereby providing a design close to the optimal (minimum weight) than would otherwise be possible using traditional methods.