Robust design optimization of benchmark aerodynamic case based on polynomial chaos expansion

Transonic airfoil often shows sensitive aerodynamic performance to uncertainties of flight conditions. To build an effective design optimization method for robust transonic airfoil, The AIAA RAE2822 airfoil design optimization test case is studied by robust design optimization (RDO) method. The non-intrusive polynomial chaos expansion (NIPCE) method is applied to provide efficient and accurate uncertainty quantification (UQ). The global evolution optimization algorithm combined with surrogate model is used to build the RDO framework. As a comparison with RDO, deterministic design optimization is firstly performed and researched. These results show that deterministic optimization airfoil is often ill posed and suffering from the drastic increase of drag coefficient at off-design points, even though the drag coefficient of it at design point has been reduced by 48.42% from that of RAE 2822 airfoil. While the RDO airfoil achieves robust drag over a range of Mach numbers and improves the drag-divergence Mach number by 0.05 from RAE2822 airfoil. The case demonstrates that the proposed RDO method provide an effective approach to transonic aerodynamic optimization.