Aerostructural topology optimization using high fidelity modeling

Abstract

We investigate the use of density-based topology optimization for the aeroelastic design of very flexible wings. This is achieved with a Reynolds-averaged Navier–Stokes finite volume solver, coupled to a geometrically nonlinear finite element structural solver, to simulate the large-displacement fluid-structure interaction. A gradient-based approach is used with derivatives obtained via a coupled adjoint solver based on algorithmic differentiation. In the example problem, the optimization uses strong coupling effects and the internal topology of the wing to allow mass reduction while maintaining the lift. We also propose a method to accelerate the convergence of the optimization to discrete topologies, which partially mitigates the computational expense of high-fidelity modeling approaches.