Aerodynamic optimization design of the aerofoil based on genetic algorithms and neural network

Genetic algorithm has a primary disadvantage that computational cost increases greatly for overmuch evaluation of objective functions and their fitness. To improve efficiency of optimization by means of genetic algorithm, an improved method in aerodynamic optimization design of aerofoil is constructed by combining artificial neural network with genetic algorithm. B-Spline method was adopted to parameterize the airfoil, then, followell the uniform experimental design method,with the help of computational program of two-dimensional cascade profile flow field, the distribution of the artificial neural network sample points were founded. Optimize an initial aerofoil by choosing the power coefficient of the curve reference points as optimize variables, and using the lift-drags ratio and change, rate of the aerofoil area as optimization objectives. The examples indicate that the hybrid algorithm is effective and trustiness. It is proved that the improved method is valuable on engineering application.