Investigation of pitch damping derivatives for the Standard Dynamic Model at high angles of attack using neural network

Numerical investigations of flow field around the Standard Dynamic Model (SDM) are performed using computational fluid dynamics approach. Initially, the static SDM is studied at various angles of attack up to 70 degrees and an agreement between normal force and pitching moment predictions with experimental data is ensured, thanks to the polyhedral grids. Subsequently, the response of the SDM under single frequency sinusoidal pitching motions is computed and the associated pitching moment coefficient damping is obtained using two methods of classical Fourier coefficients and multilayer perceptron (MLP) artificial neural network. The results are compared to published experimental values. In the final stage, frequency sweep sinusoidal excitation in pitch axis is conducted with 30 degrees amplitude in transonic flow and the MLP is exploited to calculate variable stability derivatives. It is observed that damping derivatives are highly dependent on both amplitude and frequency of oscillation. Also, an increase in the frequency of motion lowers the pitching moment damping. As the motion frequency rises, the pitching moment amplitude increment is seen to be greater than that of normal force. Polyhedral mesh as well as overset grid technique are adopted in flow field computations, leading to high fidelity of numerical simulations. (C) 2019 Elsevier Masson SAS. All rights reserved.