The reduction of energy consumption of cars is always a significant issue in automotive design. Turbulent flow around a car body is very difficult to simulate accurately due to the complexity of the flow conditions around the body, such as complex flow separation and laminar to turbulent flow transition. In particular, flow over the Ahmed body with a rear angle of 25o is considered a challenging problem for the RANS approach with two-equation turbulence models. In this study, we aim to analyze the Kelvin-Helmholtz instability associated with this flow with a URANS approach. Methodology for utilizing the URANS method is fully discussed. The predicted velocity profiles and drag coefficient are com-pared with experimental results. Three turbulence models, such as the k-ε, k-ω and SST models, are assessed and validated with experimental data. The aim of the study is to evaluate the performance of these models for the study of the Kel-vin-Helmholtz instability over the Ahmed body and for car bodies generally us-ing experimental data for their validations. It is found that the URANS approach with the turbulence models with proper numerical treatment can perform as well as or even better than the LES. And the SST model shows the best performance compared with other turbulence models.

This paper presents two novel automated optimization approaches. The first one proposes a framework to optimize wind turbine blades by integrating multidisciplinary 3D parametric modeling, a physics-based optimization scheme, the Inverse Blade Element Momentum (IBEM) method, and 3D Reynolds-averaged Navier–Stokes (RANS) simulation; the second method introduces a framework combining 3D parametric modeling and an integrated goal-driven optimization together with a 4D Unsteady Reynolds-averaged Navier–Stokes (URANS) solver. In the first approach, the optimization toolbox operates concurrently with the other software packages through scripts. The automated optimization process modifies the parametric model of the blade by decreasing the twist angle and increasing the local angle of attack (AoA) across the blade at locations with lower than maximum 3D lift/drag ratio until a maximum mean lift/drag ratio for the whole blade is found. This process exploits the 3D stall delay, which is often ignored in the regular 2D BEM approach. The second approach focuses on the shape optimization of individual cross-sections where the shape near the trailing edge is adjusted to achieve high power output, using a goal-driven optimization toolbox verified by 4D URANS Computational Fluid Dynamics (CFD) simulation for the whole rotor. The results obtained from the case study indicate that (1) the 4D URANS whole rotor simulation in the second approach generates more accurate results than the 3D RANS single blade simulation with periodic boundary conditions; (2) the second approach of the framework can automatically produce the blade geometry that satisfies the optimization objective, while the first approach is less desirable as the 3D stall delay is not prominent enough to be fruitfully exploited for this particular case study.