Aerodynamic design of transonic natural-laminar-flow (NLF) wing via surrogate-based optimization

This paper aims to develop an efficient and robust design method for transonic natural-laminar-flow (NLF) airfoils and wings, based on high-fidelity computational fluid dynamics (CFD) solver. The CFD solver features functionality of automatic transition prediction, by coupling Reynolds-averaged Navier-Stokes (RANS) equations with the linear-stability-theory-based eN method. A laminar supercritical airfoil called LSC 72613 has been designed for cruise Mach number of 0.72 and design lift coefficient of 0.6, attaining about 50-60% laminar flow region on both sides of the airfoil surface. Then an A320-sized transonic NLF wing is designed with this airfoil for cruise condi tion at Mach=0.74, Re=20 million, CL=0.515. In order to further improve the cruise efficiency, this NLF wing is optimized at higher Mach number of 0.75 via an in-house surrogate-based optimizer. The optimization is formulated as a drag minimization problem with constraints on lift, pitching moment and geometric thickness. Through only 130 CFD evaluations, 12.1 counts drag reduction is obtained, while all constraints are strictly satisfied. Further study shows that the drag reduction is contributed by both of shock-wave weakening and laminar-flow extension. On the upper surface, a better trade-off between favorable pressure gradient and shock wave strength is obtained; on the lower surface the cross-flow (CF) instability is effectively suppressed and the laminar flow region is dramatically extended. The improvement of aerodynamic performance is observed not only at design point but also over a certain range of off-design lift coefficients.