The development and accelerated use of optimization frameworks in aircraft design is a testament to their ability to identify optimal and often non-intuitive shapes as a result of multi-disciplinary design objectives. Airfoil design is a continuously revised multi-disciplinary problem, and is pivotal to illustrate the performance of optimization frameworks involving numerical simulation, flexible shape parametrization, and intelligent evolutionary algorithms. An often overlooked component of this classic problem is to consider the dynamic aeroelastic behavior under trim conditions, which can generate explicit boundaries to the flight envelope. Trim introduces a significantly strong coupling with objectives governing static performance, e.g. aerodynamic and/or structural, thereby resulting in a highly nonlinear and discontinuous design space. In this paper, a multi-objective particle swarm optimization framework for multi-disciplinary performance improvement is presented, pertaining to aerodynamic, structural and aeroelastic design criteria at trim conditions. The framework is assisted by the construction of adaptive Kriging surrogates, which is cooperatively used with the numerical solver to identify optimal solutions within a computational constraint. Designer preferences are introduced to reflect the optimal compromise between the objectives. Results of the optimization process indicate a large spread in design variable influence and interaction, and a subtle yet clear distinction between all objectives is illustrated through the catalog of final airfoil candidates obtained.
|ジャーナル||Journal of Intelligent Material Systems and Structures|
|出版ステータス||出版済み - 5 1 2017|
All Science Journal Classification (ASJC) codes