Aerodynamics For Formula SAE A Numerical Wind Tunn
Aerodynamics For Formula SAE A Numerical Wind Tunn
Aerodynamics For Formula SAE A Numerical Wind Tunn
net/publication/228351365
Aerodynamics for Formula SAE: A Numerical, Wind Tunnel and On-Track Study
CITATIONS READS
28 2,272
2 authors, including:
Scott Wordley
Monash University (Australia)
7 PUBLICATIONS 192 CITATIONS
SEE PROFILE
All content following this page was uploaded by Scott Wordley on 01 August 2022.
400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760 Web: www.sae.org
The Engineering Meetings Board has approved this paper for publication. It has successfully completed
SAE's peer review process under the supervision of the session organizer. This process requires a
minimum of three (3) reviews by industry experts.
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise,
without the prior written permission of SAE.
SAE Permissions
400 Commonwealth Drive
Warrendale, PA 15096-0001-USA
Email: permissions@sae.org
Tel: 724-772-4028
Fax: 724-776-3036
ISSN 0148-7191
Copyright © 2006 SAE International
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE.
The author is solely responsible for the content of the paper. A process is available by which discussions
will be printed with the paper if it is published in SAE Transactions.
Persons wishing to submit papers to be considered for presentation or publication by SAE should send the
manuscript or a 300 word abstract to Secretary, Engineering Meetings Board, SAE.
Printed in USA
2006-01-0808
Figure 1: Wing profiles (after Zhang and Zerihan [7]) and example
mesh for moving ground 2D CFD validation study.
Figure 4: Front wing profiles and mesh for ground effect 2D CFD,
2D CFD FRONT WING ANALYSIS Compared to the profiles tested by Zhang (which do not
comply with FSAE rules), this design incurs a much
Front Wing in ‘Free-stream’ Results higher drag penalty but is able to achieve the higher
target lift coefficients specified in the initial design.
The performance of the front wing in free-stream air flow
was predicted using two dimensional CFD. The force
coefficients versus angle of attack determined from
numerical modelling are shown below (Fig. 5).
Using two-dimensional CFD, the same wing was then The CFD results predicted that this wing would achieve
numerically modeled with a simulated moving ground the target lift coefficient of 3.5 at an angle of attack of 31
plane for a range of different ground clearance heights degrees. At higher angles of attack a gradual trailing
and angles of attack. A graph of the force coefficients edge separation on the rear most flap was predicted,
obtained from these tests is shown in Figure 6. with downforce steadily decreasing and drag increasing
beyond this angle.
These results show how the lift coefficient achieved by
this profile is dependant on a complicated interaction
between ground clearance and angle of attack. Larger
wing angles of attack appear more sensitive to ground
clearance. Drag is seen to stay reasonably constant with
change in ground clearance, but increases significantly
with higher angles of attack. While it would have been
interesting to model larger ground clearances, they were
neglected in this study because of packaging
considerations.
Figure 8: Full car 3D CFD example results for the 2002 Monash
vehicle, underside view, shaded for pressure (Best grey scale picture
available).
FACILITY AND TESTING METHOD The wind tunnel tests achieved a similar maximum lift
coefficient (2.7) to that predicted by CFD but at a higher
The Monash Full Scale Wind Tunnel is closed return, angle of attack. This was attributed to the effect of the
open jet wind tunnel, located in Melbourne, Australia. small aspect ratio (3.33) compared to the 2D CFD which
The flow properties of this tunnel are well described by assumes an infinite aspect ratio. In the wind tunnel, the
Gilhome [24], and a schematic diagram of the tunnel is front wing was found to begin stalling at 29 degrees
provided in the Appendix. The nozzle of the automotive angle of attack. Smoke visualisation at this setting
working section is 2.6m by 4.0m and is capable of indicated that the flow had separated on the underside
speeds up to 180 km/h. of the rear-most flap. Further increasing this angle of
attack resulted in a leading edge separation at 32
A specialized rig was constructed to allow the small degrees, and a corresponding large decrease in
wheelbase and track Formula SAE car to be mounted to downforce. The measured drag was substantially higher
the tunnel balance. This rig was also designed to allow than that predicted by 2D CFD, most likely due to
wings to be held and tested (with endplates) in ‘free- induced drag which is not accounted for in the CFD
stream’ flow with no car in place. These tests were used results.
to understand and tune the performance of the wings in
isolation from the car, and will be examined first. The Rear Wing Tests
minor amount of drag and lift generated by the rig itself
was subtracted from all results. The lift and drag coefficients, measured in the wind
tunnel, for the rear wing in free-stream are shown in
WINGS IN ‘FREE-STREAM’ TESTING Figure 8 below. The CFD predictions are provided on
the same graph for comparison.
Front Wing Tests
The wind tunnel results showed a trailing edge
The lift and drag coefficients, measured in the wind separation on the rear most flap beginning at 38
tunnel, for the front wing in free-stream are shown in degrees. By 40 degrees the underside of this flap was
Figure 7 below. The CFD predictions are provided on fully separated, resulting in a plateau in the CL curve.
the same graph for comparison. The flow on the underside of the main plane remained
attached until 48 degrees, beyond which a complete
leading edge separation was observed. The
considerable difference between the stall angles
predicted by CFD and measured in the wind tunnel was
most likely due to the extremely small aspect ratio of the
wind tunnel tested wing (1.72). Again, the wind tunnel
drag is believed to be higher than predicted from CFD
due to induced drag. It is interesting to note that further
testing showed that removal of the leading edge slat
reduced the maximum CL angle of attack of this wing by
8 degrees and decreased the maximum lift coefficient by
0.2 CL.
Figure 13: Example ‘symmetry plane’ test configuration, after Katz [5].
As expected, these results were lower than predicted by A CAD drawing and photograph of the final design is
the two dimensional CFD analysis due to: shown below (Figs. 16 and 17).
• The shortened second flap, the centre third of During testing calibration weights were added at the
which is permanently removed to enable the beginning and end of each run to allow offset and
wing to clear the nose cone (Fig. 14); and scaling errors to be compensated for, in post analysis. A
reference plane was used to set the ride height and
• The blockage and interference effects caused each test run was repeated four times, against and with
by the vehicle and, in this case, the stationary the wind (if any). Each calibration and run was exported
front wheels. to a spreadsheet for analysis and correction of the
individual offsets by linear interpolation. The scaling
Further testing using the symmetry rig examined the from the calibration runs was then applied to correct the
effect of adding a 45 mm foot plate to the outside lower logged data. The corrected data was then plotted
edge of the endplates. Such a flap is designed to help against air speed squared and a linear regression
prevent flow migrating from the outside of the endplate, performed on the data set. The least squares fit
underneath into the low pressure region developed by returned the v2 coefficient for each angle of attack and
the wing. Tests at a range of ground heights and wing ride height setting, allowing lift coefficients to be
angles of attack showed significantly improved calculated.
downforce (8%) from the use of such a flap.
Unfortunately, because such a flap is considered part of
the wing, it must be located within the maximum
allowable envelope for wings defined by the rules [2],
and illustrated in [1]. This means that wing span must
be reduced by a corresponding amount. Tests using
different span wings suggested that a 13% loss in
downforce would result from the required 90 mm
reduction in wing span, and result in a net loss in
downforce for this sized foot plate. Although not tested,
it is thought that a smaller foot plate, in the order of 20
mm, might result in a net downforce gain.
Figure 18: Front wing down force measured on-track, plotted against
Figure 16: CAD design of the parallel linkage system for measuring air speed squared, 27 deg AOA, ground clearance: 0.07 chord.
downforce on the front wing. Tie rod links attach to outboard wishbone 2
Average v coefficient of 0.76 yields a CL of 2.42.
pick-ups.