Data-Driven Spatiotemporal Modal Analysis of Flow Fields Around Generic and Idealized Automotive Geometries

The complex, unsteady flow dynamics around road vehicles significantly influence their aerodynamic performance. This dissertation investigates the coherent flow structures surrounding the generic, but realistic automotive shape, fastback DrivAer geometry and the idealized squareback Ahmed body at high Reynolds numbers, specifically 2.4E+06 and 7.7E+05, respectively, based on vehicle height; both of these automotive representations are very popular in the study of ground vehicle aerodynamics. Dynamic Mode Decomposition (DMD) is employed to analyze and classify these coherent structures. Initially, DMD was applied to numerical simulation data of the DrivAer model to identify dominant flow modes. This analysis successfully reproduced established modes reported in the literature. However, challenges related to DMD convergence, energy content stability, and computational expense highlighted the need for a more systematic approach when applying DMD to ground vehicle aerodynamics. To establish robust guidelines for DMD parameter selection, a detailed analysis was conducted on the Ahmed body. This study revealed critical parameters for achieving accurate and reliable DMD results. Specifically, a minimum of 3000 snapshots and a sampling frequency ratio between 5 and 10 were found to ensure reconstruction errors below 1%. Furthermore, a normalized sampling period of T*=250 corresponds to 20 cycles of the lowest frequency coherent structures, stabilized DMD mode shapes and energy distributions. With a validated DMD framework, the dominant wake structures of the Ahmed body were extracted and classified. Three primary flow regimes were identified within the Strouhal number range St<0.2: symmetry breaking, bubble pumping, and large-scale vortex shedding. The energy contributions of these regimes were quantified across shear layers, the near-wake, and the far-wake for both pressure and three velocity components. Additionally, Pearson correlation coefficients of the DMD spectral amplitudes were calculated to analyze the interaction between pressure and velocity fluctuations, providing insights into the underlying wake dynamics. This research contributes to a deeper understanding of the aerodynamic forces acting on road vehicles and establishes a systematic methodology for applying DMD to high-Reynolds-number vehicle flows. The developed framework provides critical insights for developing real-time flow control strategies and constructing reduced-order models (ROMs) for efficient aerodynamic predictions.