Viscous and Non-Viscous Approaches for a Delta Wing Benchmark

This study presents an investigation of the sharp-edged delta wing from the Advisory Group of Aerospace Research and Development number 247 (AGARD-247), performing a quantitative comparison of different Computational Fluid Dynamics (CFD) modeling strategies. The analysis encompasses inviscid (Euler) and viscous Reynolds-Averaged Navier-Stokes (RANS) approaches, utilizing the one-equation Spalart-Allmaras with the Quadratic Constitutive Relation from 2000 and the two-equation k-omega Shear Stress Transport turbulence models. A cross-validation of these methods is performed on two distinct platforms, the open-source suite SU2 and the commercial solver FLUENT, with all numerical results being benchmarked against established experimental data. The findings confirm that viscous simulations are superior in predicting the flow, especially at high angles of attack, whereas the non-viscous approach overpredicts lift and fails to capture stall. Notable differences were observed between the turbulence models, particularly in the post-stall regime, and between the solvers, with FLUENT demonstrating greater mesh sensitivity. Although the viscous models showed closer agreement with the experimental data, none could accurately predict the pitching moment coefficient or capture the formation of secondary vortices, indicating the limitations of steady-state RANS models for flows dominated by vortex breakdown. The work concludes that, despite their limitations, viscous modeling is crucial for the analysis and highlights the need for scale-resolving simulations to more accurately capture the complex flow dynamics.