Combining an RBF-based morpher with continuous adjoint for low-speed aeronautical optimization applications

Abstract

In this paper, the continuous adjoint method, developed by NTUA in the Open-FOAM® environment, is coupled with an RBF-based morpher developed by UTV to tackle optimization problems in low-speed aeronautics. The adjoint method provides a fast and accurate way for computing the sensitivity derivatives of the objective functions (here, drag, lift and losses) with respect to the design variables. The latter are defined as a set of variables controlling a group of RBF control points used to deform both the surface and volume mesh of the computational domain. The use of the RBF-based morpher provides a fast and robust way of handling mesh and geometry deformations, facing two challenging tasks related to shape optimization with the same tool. The coupling of the above-mentioned tools is used to tackle (a) the minimization of the cooling losses for an electric motor installed on a lightweight aircraft, by controlling the cooling air intake shape and (b) the shape optimization of a glider geometry targeting maximum lift-to-drag ratio by mainly optimizing the wing-fuselage junction. Regarding problem (a), a porous media is utilized to simulate the pressure drop caused by the radiator; the adjoint to this porosity model is developed and presented. This work was carried out in the framework of the EU-funded RBF4AERO project and the presented methods are available through the RBF4AERO platform (www.rbf4aero.eu).