Shape, Sizing, and Topology Design of a Wingbox Under Aeroelastic Constraints

This work considers the aeroelastic optimization of transport wingbox structures under three different types of design variables: sizing (for example, skin thickness), shape (planform and airfoil twist/thickness), and topology/layout (rib and stringer placement). The numerical optimization is conducted with a nested bilevel method: with a nongradient-based global optimizer at the outer level (a Bayesian infill method), and a gradient-based optimizer at the inner level. Design variables are divided among the two groups based on the availability of adjoint derivatives. Results are provided in terms of a tradeoff between fuel burn reduction and structural weight reduction, with a focus on demonstrating the importance of including topology/layout variables in the optimization process. These layout variables have historically presented substantial numerical difficulties owing to their nongradient-based nature, and the results shown here are able to quantify the performance degradation (fuel burn, weight) when these layout variables are frozen.