Abstract :
[en] Modern aircraft usually have light and flexible, high aspect ratio composite wings, which are
subjected to significant deformations. The aeroelastic behavior of such wings is of paramount
importance as it affects both the structural design and the aircraft performance. Moreover this
behavior needs to be studied in order to take full advantage of the composite material prop-
erties. The present work aims at developing and deploying an efficient framework allowing to
perform aerostructural analysis and optimization for preliminary aircraft design within a reason-
able computational cost.
First, the optimization problem is formulated mathematically, and several methods for solving it
are listed and compared. As there are many more design variables than objectives and con-
straints involved in preliminary aerostructural aircraft design, the adjoint approach is chosen.
To be efficient, the adjoint method requires the different coupled physics to provide their sen-
sitivities. The full potential formulation and the mesh deformation procedure implemented in a
efficient aerodynamic solver, DART, are consequently differentiated analytically to provide these
gradients. MPHYS, a framework designed to perform aeroelastic computations and built on top
of openMDAO, is selected as a basis tool to couple DART to the structural solver TACS, using
MELD to transfer the data and pyGeo to parametrize the geometry. Finally, optimization calcu-
lations are carried out to verify the implementation of the framework. Overall, the results show
that DART provides gradients that can be used to perform optimization tasks through MPHYS.
However, the implementation of the Kutta condition is not yet suited for three-dimensional shape
optimization and should be improved. The next steps also consist in extending the calculations
to include more design variables and to treat full aircraft configurations.
