Optimization of Blowing Jet Performance on Wind Turbine Airfoil Under Dynamic Stall Conditions Using Active Machine Learning and Computational Intelligence

The dynamic stall is a common phenomenon in horizontal and vertical axis wind, reducing system efficiency. In order to enhance the aerodynamic performance (L/D) of a NACA0012 airfoil under the deep dynamic stall at Reynolds number of 1.35 x 10(5), computational intelligence algorithms were utilized to find the best operational parameters of a continuous blowing jet. The airfoil undergoes a sinusoidal motion between - 5 and 25, and the rotation center is around a quarter of its chord. Unsteady Navier-Stokes equation (URANS) was used with k - omega SST turbulence model. Two types of computational intelligence algorithms, including neural networks and genetic algorithms, were coupled for this optimization. The average lift to drag ratio (L/D) in an oscillation period was considered as the objective function. The blowing jet parameters, which included location, opening length, velocity magnitude and angle of jet, were selected as design variables. Two neural networks have been utilized to find a relation between design variables and the mean lift and drag coefficients over a period to reduce the computational cost of the optimization. The optimization algorithm converged after almost 115 simulations. The ANNs in the last simulation were able to predict the input data with 92% and 93% regression coefficients for average values of drag and lift coefficient in terms of the operational parameters of the jet, respectively. The optimized jet enhanced the mean aerodynamic performance by reducing the drag coefficient and increasing the lift coefficient during a period of oscillation. For the optimal case, this parameter reached the value of 11.727 or 4.717 times the uncontrolled case. The most impact of the jet is in the downward movement. Significant improvement in aerodynamic performance was observed for the optimal blowing jet, which is due to the lack of formation leading edge vortex (LEV), dynamic stall vortex (DSV) and trailing edge vortex (TEV). The results indicated that about 2-5% of the chord is the best location for jet. This location is near the place where the leading edge vortex is formed. Aerodynamic performance improved better when the jet angle was in the range of 55 degrees-70 degrees. Although the jet momentum coefficient was not maximized, jet-opening length and blowing velocity magnitude converged to their maximum values quickly.