A new approach to high-fidelity aerodynamic design optimization of a wind turbine blade configuration is offered. This method combines Blade Element Momentum (BEM) theory with the high fidelity aerodynamic shape optimization of an airfoil. The chord length and the twist angle of the blade at various radiuses have been calculated by BEM. The Navier Stokes equations are solved to simulate both two and three dimensional flows. The Results which are obtained from 2D Computational Fluid Dynamics (CFD) have been utilized to train a Neural Network (NN). E387 Eppler is used as the base cross section of the blade. In the process of airfoil optimization, Genetic Algorithm (GA) is coupled with trained NNto attain the best airfoil shape at each angle of the attack. The simulation and validation of the base wind turbine with calculated pitch angle, twist angle, chord profile and base airfoil have been performed. The comparison of the results of this turbine with optimized one, illustrates a significant improvement in power factor.

Keywords