Drag Reduction of Transonic Wings with Surrogate-Based Optimization

Abstract

Preliminary design of aircraft wings requires multiple cycles of optimization to compromise the influences from different disciplines and constraints. Surrogate-based optimization is a popular choice in this circumstance because surrogate models can be reused in different cycles once constructed, while gradient-based optimization might be less efficient despite its fast convergence in each cycle. However, surrogate-based optimization suffers from the curse of dimension. The selection of design variables has a big influence on the optimization efficiency and performance. We investigate two different approaches to drag minimization in the design of transonic wings: (1) optimize all the section shapes and their twists independently (2) scale all the wing sections together and optimize them with their twists. We find that surrogate-based optimization by the former approach cannot ensure a better solution in spite of its larger design space. The latter approach is generally more efficient, which is more practical in preliminary design of wings.