Abstract: This paper starts with the engineering application of a trade-off between the aerodynamic characteristics and aircraft design requirements in transonic wing design, and proposes a two-step design strategy which combines the wing planform optimization and wing section optimization based on the constraints of constant structure weight and volume. Using the neural network and genetic algorithm, a corresponding design method is set up. For the transonic wing design using the two-step optimization method, improvement of the wing aerodynamic characteristics in the first design step mainly depends on the optimization of the wing planform, which can increase the wing aspect ratio and reduce induced drag. However, the improvement is very limited because the structure weight constraint just allows very little change in the wing planform. In the second design step, based on the wing planform optimized in the first design step and wing volume constraint, improvement of the wing aerodynamic characteristics mainly depends on the optimization of the wing sections which can weaken the shock wave and reduce wave drag at high Mach numbers. By using the two-step optimization method, the lift-drag ratio increments of the two design points can reach 3.02% and 9.96% respectively, and the drag divergence Mach number can be postponed from 0.854 to 0.8653. Moreover, the two-step optimization design method is suitable for engineering application. The pure aerodynamic optimization methods are also applicable to transonic wing design without the constraints of structure weight and volume. The results show that the improvement in aerodynamic characteristics is significant though the method will not then satisfy the design requirements of constant structure weight and volume.

Key words: wing, transonic flow, optimization, engineering application, computational fluid dynamics