Simulation of ground vehicle aerodynamics applied to a generic Le Mans prototype
No Thumbnail Available
Date
2013-01-30
Authors
Stevenson, Thomas James
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Optimisation of the aerodynamic design of a vehicle is paramount in achieving peak performance.
In the case of racing cars, it is required that extensive research be conducted into the aerodynamic
response to different vehicle attitudes such that performance and safety envelopes can be established.
Chiefly amongst the numerical techniques typically employed is Computational Fluid Dynamics
(CFD). The objectives of the current research are to simulate the flows around a high performance
ground vehicle, and estimate the performance characteristics and aerodynamic safety envelope; perform
a suitable numerical validation study to ascertain the accuracy and efficacy of the numerical
methods; and to engage design and setup parameters which affect the track-wise performance of
the racing car. In this dissertation, steady-state three-dimensional Reynolds Averaged Navier Stokes
(RANS) equations are employed, where closure is provided by the two-equation Ù−æ SST turbulence
model. Numerical simulation is accomplished with ANSYS ICEM and CFX softwares. Both low
Reynolds number (Re ´ 0.58 × 106) and high Reynolds number (Re ´ 25 × 106) simulations are
conducted. Vehicle ride-heights are modified to establish the relationship and sensitivity of performance
characteristics to physical geometry changes. Downforce generated by the vehicle is shown to
be strongly dependent on the front ride-height hgf , where −0.73 < cl < −1.84 based on the frontal
area of the car. Over the full range of geometries studied, the downforce remains positive, however
the ratio of front / total downforce production is shown to be highly variable. Vehicle drag is found
to vary as 0.36 < cd < 0.45. A strong correlation is found between the production of drag and the
rear ride-height hgr . The overall aerodynamic efficiency Ô is found to vary as 2.2 < Ô < 4.2. A strong
correlation is found between Ô and hgf . The numerical prediction methods are first validated through
a numerical study of the generic Ahmed body, and secondly through an experimental investigation of
a 1/6th scale model of the LMP vehicle in the University’s low speed wind tunnel. Boundary layer
reduction on the ground plane is accomplished through the use of distributed suction on a perforated
raised plane. It is found that insufficient reduction of the boundary layer thickness Ó and displacement
thickness Ó occur before the leading edge of the vehicle. However, through direct comparison of
underbody surface static pressures to numerical predictions, it is found that the suction system does
greatly improve the level of correlation between the experimental and numerical results. As well as
through direct comparison of surface static pressure coefficients, experimental verification is accomplished
using pictorial comparisons of the flow-fields and wall shear stress distributions. Optimisation
of the vehicle’s lap-time is undertaken using quasi-static vehicle dynamics simulations, which employ
numerically-derived aerodynamics data. The resulting optimisation process produces a Le Mans laptime
of 3’ 41” seconds. This prediction is within the performance range of current production racing
cars. Further numerical research is required into the nature of forces which act during vehicle yaw
and extreme pitch scenarios. To facilitate further experimental investigations, a more powerful suction
system is required to remove the boundary layer on the ground plane.