Abstract :
[en] In order to reduce aircraft fuel consumption and improve the reliability of the design process, multi-disciplinary optimization is nowadays carried out during the preliminary design stage. These optimization calculations must model the aeroelastic behavior of the aircraft in order to be effective and to take full advantage of composite materials. Since many design variables are involved during the early stages, the optimization problem is usually solved using the adjoint method. Moreover, the fluid model and the associated numerical simulation method must be selected with care, as they are the main contributors to the overall computational cost. The present work aims at developing and deploying an efficient framework allowing to perform aerostructural analysis and optimization for preliminary aircraft design within a reasonable computational cost.
The optimization problem is first formulated mathematically. Then, an optimization package is implemented using MPHYS, a framework designed to perform aeroelastic computations and built on top of OpenMDAO. Several numerical tools are subsequently selected, interfaced and integrated in the framework. More specifically, pyGeois used to parametrize the geometry, and the aerodynamic solver DART is coupled to the structural solver TACS, using MELD to transfer the data. Although the current framework is restricted to static aeroelastic analysis and optimization, unsteady aerodynamic modeling methodologies are also being developed so that dynamic aeroelasticity can be included in the package. Finally, optimization calculations are carried out to demonstrate the implementation of the framework. Overall, the results indicate that the newly developed package is able to perform optimization calculations. In particular, the adjoint nonlinear potential formulation yields optimized wing shapes and structures at very low computational cost. The next steps consist in continuing to improve the framework, improving the robustness of DART, and continuing the development of unsteady aerodynamic modeling and integrating it in the framework.