Surrogate-based aerodynamics for composite wing box sizing

The main objective of this paper is to propose an aeroelastic optimization approach capable of performing structural sizing optimization considering structural and aerodynamic constraints. The proposed approach uses a detailed FE model of a composite wing with shell elements in order to obtain realistic weight estimation and structural responses. Moreover, a surrogate model based on rigid RANS computations provides a high-fidelity lift and drag coefficient estimation during the optimization as constraints. The use of RANS computations allows the surrogate model to consider all drag components and not the induced drag only. An approximation of the structural displacement is proposed based on modal projection and principle component analysis. Results shows that a polynomial regression of order four is accurate enough to be used as surrogate model of the drag coefficient and of the lift-to-drag ratio. Moreover, it is possible to conclude that improvements in aerodynamic performance comes at the price of a heavier and stiffer structure.