High-lift optimization design using neural networks on a multi-element airfoil

The high-lift performance of a multi-element airfoil was optimized by using neural-net predictions that were trained rising a computational data set. The numerical data was generated using a two-dimensional, incompressible, N avier-Stokes algorithm with the Spalart-Allmaras turbulence model. Because it is difficult to predict maximum lift for high-lift systems, an empirical ly-based maximum lift criteria was used in this study to determine both the maximum lift and the angle of attack at which it occurs. Multiple input, s ingle output networks were trained using the NASA Ames variation of the Lev enberg-Marquardt algorithm for each of the aerodynamic coefficients (lift, drag, and moment). The artificial neural networks were integrated with a gr adient-based optimizer Using independent numerical simulations and experime ntal data for this high-lift configuration, it was shown that this design p rocess successfully optimized flap deflection, gap, overlap, and angle of a ttack to maximize lift. Once the neural networks were trained and integrate d with the optimizer, minimal additional computer resources were required t o perform optimization runs with different initial conditions and parameter s. Applying the neural networks within the high-lift rigging optimization p rocess reduced the amount of computational time and resources by 83% compar ed with traditional gradient-based optimization procedures for multiple opt imization runs.