[en] This thesis examines Delayed Detached-Eddy Simulation (DDES) for predicting massively separated flows around an airfoil and a flat plate at high angles of attack, comparing results with experimental data. DDES combines Large-Eddy Simulation (LES) and Reynolds-Averaged Navier-Stokes (RANS) approaches, with RANS used in boundary layers to reduce computational cost and LES in separated flow regions to capture large-scale turbulence. DDES based on the Spalart-Allmaras model accurately predicted aerodynamic force coefficients, closely matching experimental data, whereas unsteady RANS simulations significantly overestimated these coefficients in flows with extensive separation. Fine grid resolution in DDES unveiled the three-dimensional nature of these separated flows, and extending the spanwise domain length reduced force fluctuations and slightly lowered time-averaged lift and drag coefficients. Dynamic Mode Decomposition (DMD) was used to analyze the flow fields, extracting dominant dynamic modes and facilitating comparisons between computational and experimental results. For instance, DMD highlighted asymmetrical vortex shedding in the flow around a flat plate at a high angle of attack. This thesis ultimately demonstrates the superior capability of DDES in accurately predicting complex three-dimensional separated flows, making it a valuable tool in aerodynamic applications.
