A prior analysis of the effect of steady cross wind on full size cars or models must be conducted when dealing with transient cross wind gust effects on automobiles. The experimental and numerical tests presented in this paper are performed on the Willy square-back test model. This model is realistic compared with a van-type vehicle; its plane underbody surface is parallel to the ground, and separations are limited to the base for moderated yaw angles. Experiments were carried out in the semi-open test section at the Conservatoire National des Arts et Métiers, and computations were performed at the Ecole Centrale de Nantes (ECN). The ISIS-CFD flow solver, developed by the CFD Department of the Fluid Mechanics Laboratory of ECN, used the incompressible unsteady Reynolds-averaged Navier–Stokes equations. In this paper, the results of experiments obtained at a Reynolds number of 0.9×106 are compared with numerical data at the same Reynolds number for steady flows. In both the experiments and numerical results, the yaw angle varies from 0 deg to 30 deg. The comparison between experimental and numerical results obtained for aerodynamic forces, wall pressures, and total pressure maps shows that the unsteady ISIS-CFD solver correctly reflects the physics of steady three-dimensional separated flows around bluff bodies. This encouraging result allows us to move to a second step dealing with the analysis of unsteady separated flows around the Willy model.

Abstract:When analysing the topology of turbulent flow systems for alternative vehicle configurations, conventional wind-tunnel tests may not be sufficient. Vehicle development teams investigating this kind of physical phenomenon are increasingly interested in numerical simulation, but reliable implementation of computational techniques in automotive aerodynamics requires validation to confirm that such methods are capable of accurately reproducing experimentally proven physical behaviour. We address this issue by comparing the results of experimental wind-tunnel tests with simulated values for aerodynamic drag coefficient and vortex wake flows at different angles of car rear window inclination, using an Ahmed [1] reference model. -INTRODUCTIONTo meet targets for shorter lead-times and lower costs in modern automobile design processes, aerodynamics specialists are constantly seeking new solutions capable of providing prompt and accurate guidance in design practice. One such solution is numerical simulation, which provides a valuable complement to results obtained from experimental wind tunnel tests. However, if computational techniques are to become a viable proposition in automotive aerodynamics, they must demonstrate their ability to accurately reproduce the elementary phenomena observed on simple geometric forms in the wind tunnel. With this objective in mind, we discuss computational work on an Ahmed reference model (Ahmed, Ramm and Faltin, 1984 [1]), following the work of Baxendale, Graysmith, Howell and Haynes (1994 [2]). The Ahmed model was chosen for its geometrical simplicity, and because of the body of documented experimental results available for this model in automotive aerodynamics applications. Specifically, we compare computed results against experimental findings for different angles of rear window inclination, and offer qualitative analysis, chiefly concerning aerodynamic drag coefficient and the behaviour of vortex wake flows. -EXPERIMENTAL RESULTSThe Ahmed reference model is shown in figure 1. The angle α represents the angle of inclination of the rear window with respect to the horizontal.

No abstract

The objective of this research is to improve the understanding of unsteady three-dimensional separated flows on bluff bodies. This research focuses on the analysis of the instability and safe maneuverability of passenger vehicles associated with unsteady atmospheric crosswind conditions. Experimental and numerical analysis were performed on the squareback Willy car model animated by an oscillating yaw angle in a steady wind. Experiments were carried out at a Reynolds number of 0.9 10 6 , a frequency of 2 Hz, with an amplitude of Δβ = 10° at the Conservatoire National des Arts et Métiers (CNAM). Unsteady wall pressures were obtained with sensors located inside the model. Computations were performed at the Ecole Centrale de Nantes (ECN). The ISIS-CFD flow solver, developed by the CFD Department of ECN, used the incompressible unsteady Reynolds-Averaged Navier-Stokes equations. Experimental and numerical comparisons are completed with numerical simulations at a larger yaw angle. The analysis of the physics of the unsteady flow around the model is done through the calculation of aerodynamic forces and moments, and cross flow velocities.

13 pagesInternational audienceThe aerodynamic characteristics of automobile vehicles are greatly influenced by the unsteady change in the direction of relative airflow. The aim of this paper is to analyse how such a change influences vehicle wake flow patterns. An analysis was conducted on a simplified model capable of reproducing the typical structures encountered under the aerodynamic conditions of an automobile vehicle. The results were then processed by mapping the steady and unsteady total pressure losses around the model. The findings should enable automobile development engineers inter alia to identify and analyse the physical phenomena that occur when a vehicle is subjected to a sudden gust of side wind