To CFD or not to CFD? Comparing RANS and viscous panel methods for airfoil shape optimization
E. J. Adler, A. C. Gray, and J. R. R. A. Martins
33rd Congress of the International Council of the Aeronautical Sciences, 2022
As the computational resources available to aerodynamicists increase, so does the number of tools with which they can analyze their designs. The key skill of a good engineer is understanding when each tool produces realistic results. Even if no available tool can capture all relevant physical phenomena, this knowledge allows engineers to alter their designs to account for the missing physics. However, with the increasing use of numerical optimization for aerodynamic shape design, the opportunity for this intuition to shape the design is reduced as the optimizer makes design decisions based purely on numerical results. This makes the choice of analysis method critical. Previous studies have compared panel-based aerodynamic tools and RANS CFD to predict and analyze the key aerodynamic phenomena in different flight conditions. But few, if any, have compared the impact of proper tool selection on the optimal aerodynamic shapes. In this work we show the effect of aerodynamic analysis method selection on the optimal aerodynamic shape and the consequences of choosing one that does not capture the relevant flow physics. In particular, we show that modelling boundary layer transition can produce significantly lower drag designs across almost the entire range of Reynolds numbers at which airfoils are used in practice. Additionally, we find that the simple compressibility correction used in XFOIL is sufficient to produce well performing airfoils for flight Mach numbers up to 0.65, after which point the ability for the flow solver to model shock formation is critical. Our results will provide aerodynamicists with the data and intuition to make informed decisions when selecting analysis methods for optimization.