Airfoil Optimization Through Differential Evolution

Abstract: This thesis presents the development of a numerical optimization algorithm for airfoils, and how it can be used in design of wind turbine blades. It was found that the developed algorithm successfully improves the goal parameters under given conditions and constraints. This research was conducted on behalf of Winfoor AB who has developed a conceptually new blade design, in which every single blade is made up of three individual blades, kept together by rods in a truss like manner. Their wish was to develop a new airfoil for their turbine, with higher performance and a more docile stall, and yet remaining a high airfoil thickness in order to not alter structural stability. The task was conducted by describing the airfoils with B-splines and writing an optimization algorithm in MATLAB in which the flow characteristics of the airfoils were determined by the external software XFOIL. This thesis shows how to characterize numerical optimization problems, what differential evolution is and how it can be implemented in a MATLAB-code, how airfoils can be described with B-splines, the usage of XFOIL and how penalty functions can be imposed for constrained optimization problems, to mention some of the wisdoms this work has brought. This research is important as numerical optimization of airfoils is not yet the standard method for airfoil design, and thus it can possibly contribute with valuable insights and results to further development of airfoil optimization.