Aeroelastic optimization of adaptive bumps for yaw control

This paper captures the aeroelastic behavior of adaptable bumps on morphing wings using trained neural networks. Parameters of the bump such as the height, size, location, and shape play a major aerodynamic and structural role. Though other issues such as wave drag minimization, boundary layer control are discussed, the primary problem addressed here is the generation of the lateral directional moment. The shape of the bump is optimized so that it produces maximum yaw moment with a minimum actuation energy spent in creating the adaptive bump. The analysis of fluid separation on the airfoil surface due to various types of bumps and its effects on the aerodynamic forces is performed using the computational fluid dynamics (CFD) software FLUENT(TM). This analysis is performed at a low speed of Mach 0.3 and at a transonic speed of Mach 0.7. Structural analysis is performed using the finite element software ANSYS(TM) using a nonlinear beam model of the bump. In order to perform an aeroelastic analysis, the softwares FLUENT and ANSYS have to be interconnected so that the changing aerodynamic pressure with bump deformation is reflected in the structural analysis. Direct coupling of two such numerical codes, one based on CFD and the other based on finite element modeling (FEM), is computationally expensive. Hence, artificial neural networks are trained from these aerodynamic and structural analysis. Two neural networks are trained, one for the aerodynamic pressure and the other for the structural loads and strain energy. These two neural networks serve as an efficient decoupler, that facilitates an aeroelastic optimization procedure to evaluate the best bump shape for maximum drag for providing micro-yaw control while using the minimum actuation energy and minimizing the loss in the lift.