Airfoil Aerodynamic Optimization Based on an Improved Genetic Algorithm

In order to boost the convergence speed of Genetic algorithms (GAs), some amelioration was made on the standard Genetic algorithm, a new crossover and a new mutation were designed in this paper. According to the numerical tests, the convergence speed of ameliorated Genetic algorithms accelerated apparently. In the Rae2822 supercritical airfoil drag reduction optimization, we obtained a satisfying result by only evolving only 20 steps, the drag descended 32.09 percent totally. In the optimization, 16 variables were used, with constrains of lift, maximal thick, maximal area no descent, a Bezier-Bernstein function was used to parameterize the airfoil configuration and using N-S field solver to obtain the objective function. Because of the natural parallel characteristic of Genetic algorithms, the optimization was run on the Linux clusters for reducing the time cost.