Optimization of active flow control over an airfoil using surrogate management framework

An efficient method based on the "surrogate management framework" is excised to optimize the actuation parameters of the active flow control over an airfoil with a synthetic jet. The sample points in the design space are selected according to Design of Experiments (DoE) theory, and the approximation models are constructed based on the sampled data obtained from the Unsteady Reynolds Averaged Navier-Stokes (URANS) simulations. The accuracy of these approximation models is evaluated at test points by comparing the approximated values with the data obtained from URANS simulations. Three types of approximation models, quadratic Response Surface Model (RSM), Kriging model and Radial Basis Function Neutral Network (RBF-NN), are created from the same dataset. The model with highest accuracy is selected as the surrogate model used to replace the URANS analysis during optimization. The optimization objective is to maximize the lift coefficient of a NACA 0015 airfoil at given angles of attack (12° to 22°), with the jet momentum coefficient, non-dimensional frequency and jet angle (with respect to the wall) as the design variables. The surrogate model is coupled with a Simulated Annealing Genetic Algorithm (SAGA) to efficiently obtain the global optimum. Thanks to the optimization process the lift coefficient at the critical angle of attack is increased by 21.5% and the drag due to flow separation is decreased by 31.8%. It is preliminarily shown that the presented method is efficient and usable for optimizing the active flow control via synthetic jet technology.