Towards real-time vehicle aerodynamic design via multi-fidelity data-driven reduced order modeling

Designing a vehicle exterior shape takes place at the intersection between styling and aerodynamics-two disciplines with often competing notions. This calls for an interactive shape design process that delivers aerodynamic responses to design modifications in realtime. Reduced Order Modeling (ROM) is a well-known concept to accelerate aerodynamic computations. The main objective of this investigation is therefore to assess the performance of ROM for industrial-sized problems in terms of accuracy and computational cost for real-time vehicle aerodynamics. We focus specifically on an interpolation-based approach, combining Proper Orthogonal Decomposition (POD) and the statistical Response Surface Modeling (RSM) technique Kriging. A new extension to this known approach is presented, which is able to handle data of different levels of accuracy. Both the conventional single-and the newly developed multi-fidelity approach are applied to time-averaged Detached-Eddy Simulations (DES) in combination with Reynolds-Averaged Navier-Stokes simulations (RANS) of a production passenger car.