Aerodynamic Performance Based Globally Robust Design Optimization Of Wind Turbine Airfoils

The aerodynamic performance of wind turbine blades directly decides the whole wind turbine unit’s wind catching rate,power and load characteristic,meanwhile it is the key factor in wind turbine system optimization design.The blade is a assembled superposition through different airfoils along spanwise distribution,therefore,designing airfoil with high-performance is the basic way and measure to improve the wind utilization rate of wind turbine.Traditional airfoil optimization design is always aimed at the aerodynamic performance of certain blade airfoil based on deterministic analysis,ignoring the randomness of overall blades.Although the aerodynamic performance of airfoil obtained through this method is improved,the overall blade airfoil aerodynamic performance does not be improved as expected.In order to improve the whole blade airfoil aerodynamic performance,wind turbine airfoil robust optimization is researched in the thesis,and the main work includes the following three aspects:1.Discuss the main factors which influence the aerodynamic performance,laying a foundation for the set of constraints in later optimization.Compare the accuracy of CST and improved Hicks-Henne airfoil parametric modeling method for fitting NACA63418 and DU93-W-210 airfoil under different orders.The accuracy of XFOIL software analyzing aerodynamic performance is verified through compared with wind tunnel test.And the automated analysis of airfoil aerodynamic performance is realized by combining XFOIL and MATLAB software.2.Halton quasi-random sequence and the inverse cosine hyperbolic inertia weight function is introduced in standard particle swarm optimization algorithm to improve the performance of the original algorithm.Based on MATLAB toolbox,the implementation process and parameter settings of genetic algorithm,simulated annealing algorithm and the improved particle swarm algorithm are presented.Through optimizing typical test functions,the optimization accuracy,stability and efficiency of the above three algorithms are analyzed to provide reference for the choice of optimization algorithm.3.Based on Latin hypercube sampling and second order polynomial,establish the aerodynamic response surface surrogate to shorten the optimization time.Considering of the randomness of Reynolds and proportion along blade spanwise distribution,robust design objective function and constraint conditions are built.Based on improved Hicks-Henne airfoil parametric modeling approach,the NACA63418 and DU93-W-210 airfoils are selected for second optimization and high-performanced airfoil is obtained.