Flow Control Around a SUV Simplified Vehicle

Abstract

The research on the external aerodynamics of ground vehicles can nowadays be related to sustainable development strategies, confirmed by the worldwide CO2 regulation target. Automotive manufacturers estimate that a drag reduction of 30% contributes to 10g/km of CO2 reduction. However, this drag reduction should be obtained without constraints on the design, the safety, comfort and habitability of the passengers. Thus, it is interesting to find flow control solutions, which will remove or remote recirculation zones due to separation edges with appropriate control techniques. In automotive sales, the SUV, 4x4 and compact cars represent a large part of the market share and the study of control approaches for this geometry is practically useful. In this work, appropriate control techniques are designed to decrease the drag forces around a reduced scale SUV car benchmark called POSUV. The control techniques are based on the DMD (Dynamic Mode Decomposition) algorithms generating an optimized drag reduction procedure. It involves independent transient inflow boundary conditions for flow control actuation in the vicinity of the separation zones and time resolved pressure sensor output signals on the rear end surface of the mockup. This study, that exploits dominant flow features behind the tailgate and the rear bumper, is performed using Large Eddy Simulations on a sufficient run time scale, in order to minimize a cost function dealing with the time and space average pressure coefficient. The resulting dynamic modal decomposition obtained by these LES simulations and by wind tunnel measurements has been compared for the reference case, in order to select the most appropriate run time scale. Analysis of the numerical results shows a significant pressure increase on the tailgate, for independent flow control frequencies. Similar decomposition performed in the wake with and without numerical flow control help understanding the flow modifications in the detachment zones.

Aerodynamical characteristics of a reduced scale ground vehicle according to yaw angle variations

“…Besides wind tunnel measurements, fluid dynamics computations have been performed on a reduced scale model representing a sport utility vehicle (SUV). It should be outlined that experimental data dealing with standard deviation and average rear pressure maps have been previously presented in Edwige et al (2018), where the pressure over the tailgate was optimised with flow control actuations, and Edwige et al (2020) deal with the numerical simulation over the same geometrical model with a large eddy simulation (LES) solver. Main characteristics of the geometrical and the numerical model (Edwige et al , 2018; Edwige, 2019) can be summarized as follows: the reduced SUV model is placed in a wind tunnel of 1.62 m 2 with a ground clearance that has been selected after a DOE leading to the best experimental Cd value of 0.36.…”

“…It should be outlined that experimental data dealing with standard deviation and average rear pressure maps have been previously presented in Edwige et al (2018), where the pressure over the tailgate was optimised with flow control actuations, and Edwige et al (2020) deal with the numerical simulation over the same geometrical model with a large eddy simulation (LES) solver. Main characteristics of the geometrical and the numerical model (Edwige et al , 2018; Edwige, 2019) can be summarized as follows: the reduced SUV model is placed in a wind tunnel of 1.62 m 2 with a ground clearance that has been selected after a DOE leading to the best experimental Cd value of 0.36. Special care has been taken meshing the geometry, including refinements in the underbody, the near wake and in the wall boundary layers of the mock-up that leads to a numerical model of 300 million elements.…”

“…According to the Cd value, the pressure coefficient Cp over the tailgate (Edwige et al , 2018), the velocity and turbulent intensity profiles in the boundary layer and the two cut planes presented in Figure 1 in the near wake, it seems reasonable to use the 3 D average velocity field for streamlines computations. Once validated, the location of the pressure sensors is selected according to the streamline direction, starting in the front of the vehicle mock-up and ending in the wake recirculation zone, as presented in Figure 3.…”

“…According to the streamline direction highlighted in Section 2, side pressure sensors have been added on the two different paths described above to the rear pressure sensors already used on this mock-up for a previous study dealing with flow control (Edwige et al , 2018). These additional sensors enable to look at phase correlations between side and rear pressure frequencies, along these assume convective paths (Figure 4).…”

“…The experimental setup described in Edwige et al (2018) is used to look at synchronized pressure measurements sampled at 2,000 Hz on the side and rear sensors (Figures 4 and 5) of the mock-up. Reference pressure is taken at the Pitot tube located at X = −1.3 m and Z = 1.08 m (Figure 5) from the coordinate system placed on the ground at the maximum X -value of the rear bumper to compute the pressure coefficient Cp, its average and standard deviation, according to the following definition: …”