Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

SAE TECHNICAL
PAPER SERIES 980031

A Parametric Study on the Ground

Effect of a Simplified Car Model
Antonello Cogotti
Industrie Pininfarina s.p.a.

Reprinted From: Developments in Vehicle Aerodynamics

(SP-1318)

International Congress and Exposition

Detroit, Michigan
February 23-26, 1998

400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

The appearance of this ISSN code at the bottom of this page indicates SAE’s consent that copies of the
paper may be made for personal or internal use of specific clients. This consent is given on the condition,
however, that the copier pay a $7.00 per article copy fee through the Copyright Clearance Center, Inc.
Operations Center, 222 Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sec-
tions 107 or 108 of the U.S. Copyright Law. This consent does not extend to other kinds of copying such as
copying for general distribution, for advertising or promotional purposes, for creating new collective works,
or for resale.

SAE routinely stocks printed papers for a period of three years following date of publication. Direct your
orders to SAE Customer Sales and Satisfaction Department.

Quantity reprint rates can be obtained from the Customer Sales and Satisfaction Department.

To request permission to reprint a technical paper or permission to use copyrighted SAE publications in
other works, contact the SAE Publications Group.

All SAE papers, standards, and selected

books are abstracted and indexed in the
Global Mobility Database

No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written
permission of the publisher.

ISSN 0148-7191
Copyright 1998 Society of Automotive Engineers, Inc.

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. For permission to publish this paper in full or in part, contact the SAE Publications Group.

Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manuscript or a 300
word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE.

Printed in USA
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018
A Parametric Study on the Ground
Effect of a Simplified Car Model
Copyright © 1998 Society of Automotive Engineers, Inc.
ABSTRACT
Aerodynamic drag of a modern car is generated mainly
by underbody flows.
A better understanding of these flows and of their interac-
tions with the car underbody, may contribute to the future
improvement of the car drag characteristics.
This paper reports the results of a parametric study car-
ried out in the Pininfarina wind tunnel, on a full scale sim-
plified car model, by using the Ground Effect Simulation
System built in 1995.
The main aim of this study was to investigate the effect
on the aerodynamic coefficients produced by important
geometric changes which affect the flows under the car,
in proximity of the ground, and are often difficult or impos-
sible to be modified when tests are made on real cars.
The model chosen for this research program is that
defined by the SAE “Open Jet Interference Committee”
as a reference model to be used for investigating wind
tunnel interference and for comparison between wind
tunnels. In particular it has no wheels.
The model built in full scale by Pininfarina is, in reality, a
super-set of the SAE model, as it includes a number of
additional parts, which are necessary to investigate the
behavior of this model in various conditions of ground
effect.
These add-on parts include:
• Front and rear wheel-housings of three different
sizes.
• Wheels of 2 different sizes / covers
• Some different wheel tracks
• Three different front and rear overhangs.
• Three different rear diffusers.
These parts have been interchanged, to cover a great
number (not all) of the possible different configurations.
1
980031
Antonello Cogotti
Industrie Pininfarina s.p.a.
In addition, each of these configurations has been tested:
• At various ground clearances, from 200 to 50 mm
• In the four classical GESS conditions, namely
• Moving Ground and Rotating Wheels
• Moving Ground Only
• Rotating Wheels Only
• All Static
In addition, pressures have been measured along the
underbody centerline, including the diffuser and the
model base.
Furthermore the flow field under the model has been
measured for some selected configurations, to achieve a
better understanding of the data obtained by the balance.
The analysis of all this large number of measurements
has been rather engaging.
A multiple regression analysis has been used to estimate
the change which can be expected for each coefficient,
CD, CL, CLF, CLR, CMY, when the model configuration
and /or its ground clearance is changed.
These results are reported in the paper, together with
some selected surface pressure diagrams .
It is worth noting that this is the first time that the results
of a systematic study into ground effect at full scale, have
been published.
Although these results are specific for this model shape,
they can contribute to a better understanding of the
effect of the underbody layout changes as well as of the
differences which can be expected for the various GESS
testing conditions.
It is well known that the largest component of car aero-
dynamic drag arises from the energy losses in the flow
along the car underbody, the flow within the wheel-hous-
ings, and the flow around or through the wheels. Evi-
dence of that was shown by flow maps published several
years ago (1) (2). One of these maps, reprinted in fig.1,
shows clearly that most of the momentum loss at the
back of the car, i.e. the most important contribution to the
car CD, is located at a height over the ground between 0
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018
and 500 mm. The car shown in the map, although
designed about ten years ago, had a remarkably low
drag coefficient ( CD ~ 0.29) , not too different from the
CD values of the best cars which are sold nowadays.
Therefore the flow conditions shown behind this car are
still representative of those of many present passenger
cars.
This also means that any realistic possibility of further
significant improvement to the aerodynamic drag of a
modern car is strongly dependent on reducing the losses
from the underbody flows.
However, working on improving underbody flows has
been historically difficult, for a number of reasons.
Car manufacturers are reluctant to invest money (or too
much money ) for streamlining the car underbody. The
Opel Calibra (3) launched in 1986 is probably the first
example of a mass production car where underbody
aerodynamics has been taken into account (fig. 2), and
that has certainly contributed to the excellent CD value of
this car (0.26 - 0.28, depending on the wind tunnel) .
The reluctance of car manufacturers to accept aerody-
namic development of the car underbody is due to sev-
eral factors, namely the increase of cost and weight
associated with fitting the shields to the underbody, the
cost of redesigning some parts of the underbody, the
possibility of thermal problems, and eventually a more
difficult maintenance of the drive line.
Beside these reasons, which are certainly partly true, an
additional important reason which discouraged aerody-
namic work on car underbodies was the lack of experi-
mental facilities able to reproduce flow conditions under
the car which are sufficiently similar to the conditions
existing on the road.
In reality the possibility of reproducing in full scale the car
to ground relative motion and, at the same time, the spin-
ning of the car wheels, is relatively recent. DNW
reported tests of this type on passenger cars in 1989 (4),
and FIAT in 1990 (5) (6). Both used (and still use) a
large belt, and that is certainly an excellent way to pro-
vide a good simulation of the road condition. However the
system to support the car over the belt is relatively com-
plex in both wind tunnels, and require lengthy prepara-
tion of the car and /or the wind tunnel. This last point has
somehow discouraged the use of these large moving
grounds for every day development work.
More recently, in 1995, Pininfarina has built in its Wind
Tunnel a Ground Effect Simulation System (GESS) of
new design. It includes various sub-systems, namely a
narrow moving belt (MB) to cover the area under the car
between the wheels, separate rotating rollers (RR) to
spin the wheels, a boundary layer suction system
(BLSS) followed by a tangential blowing system (TBS),
both upstream of the belt. The latter feeds the belt with a
correct boundary layer and reduces as much as possi-
ble the boundary layer displacement thickness on that
2
part of the test section floor not covered by the moving
belt.
Details on the GESS and on its latest upgrades are
reported in (7) (8) (9). Fig. 3 shows the various sub-sys-
tems and in particular the new rollers.
The main advantage offered by this new type of GESS is
its high level of practicality. Almost any passenger car
can be tested in conditions of moving ground and rotat-
ing wheels, with a very short preparation. And that has
made possible and practical the aerodynamic develop-
ment of the car underbody.
The research program which is reported in this paper is
the logical follow-up to the availability of this Ground
Effect Simulation System, and demonstrates the possibil-
ity of performing a large number of tests, easily and with
a good simulation of the underbody flows.
Beside that, the following points have been taken in
account.
1. In three years of operation the GESS has been used
to test and/or develop more than 100 cars having dif-
ferent body shapes , underbody layouts (from very
rough to very clean), ground clearance ( from the 200
mm typical of passenger cars to the 40 - 45 mm of
some racing cars).While these tests have certainly
provided important information for the specific devel-
opment of each one of these cars, the possibility of
extracting from these results information of more
general value is quite limited. In reality, underbody
flows are in general quite complex due to the many
different details existing on the underbody. Even
using the best techniques which are available to map
the flow under the car (pressure probes, hot wire, ldv)
any attempt to make a complete reconstruction of
these flows gives often only modest results. Conse-
quently it is rather difficult to propose significant
underbody configuration changes or to establish
clear trends of any type, starting from sets of results
measured on real cars.
2. On the other hand, as everyone knows, important
configuration parameters, like wheel-base, track,
front and rear overhangs, wheel-housing size, etc are
virtually impossible to be modified during the wind
tunnel tests on real cars.
Therefore, the potential of aerodynamic improvement
related to this type of configuration changes is practi-
cally unknown, at least for tests on full size cars and
when an appropriate simulation of the ground motion
and wheel rotation are considered.
The research program started in 1996 at Pininfarina,
and partly funded by the Italian National Council for
Research, in the “Progetto Finalizzato Trasporti 2” pro-
gram, intends to investigate this specific field of aerody-
namic effects which are associated with important
geometric changes of a car underbody. And this, by car-
rying out tests on a full scale simplified model and using
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

the Ground Effect Simulation System mentioned before,
to provide a correct simulation of the underbody flows.
It is a two-year program; this paper reports the results of
the first year of work.
In order to keep the program reasonable in terms of wind
tunnel cost and occupancy time, the following limitations
have been set, at least for this first year of research :
Figure 1. Micro-drag map in the wake of a modern
notchback car, CD = .29 (from ref. 2). Tested
with fixed ground, static wheels.
A second important aim of the research program is to
define, in a systematic way, the differences which can be
expected for the various GESS testing conditions.
The details of this research program and the results
obtained are reported in the following paragraphs.
THE RESEARCH PROGRAM
The aim of the research program reported in this paper,
is, in a broad sense, to increase the understanding of the
underbody flows and of the interactions between these
flows and the main parameters which define the under-
body layout.
To reach this aim, the research program includes the fol-
lowing steps :
• To build a simplified full scale model , having a car-
like shape and a number of interchangeable under-
body parts.
• These parts include :
• front and rear overhangs
• front and rear wheel-housings of different size/diam-
eter
• rear diffuser
• adjustable front and rear tracks
• wheels with different tire width / cover
Furthermore the model ground clearance has to be
adjustable, to investigate the behavior of these various
configuration changes at different ground clearance.
• To carry out a test program , which includes system-
atic force and pressure measurements on the car
model, as well as selected flow maps, mainly along
the model underbody or behind the model.
The test program includes tests with and without
moving ground, and with/without rotating wheels, in
order to quantify the differences which can be
expected in the various conditions of ground simula-
tion.
• A final analysis of the results of this systematic work ,
to define the trend of each aerodynamic coefficient,
versus the various configuration parameters, the
ground clearance and the ground effect simulation
technique.
• the model underbody is smooth and its lower front
end is rounded, so as to prevent flow separations
along the underbody. This means that the effect of
the parameter “underbody roughness” is not
included in this first year program. It is certainly a
very important parameter, however , for its complex-
ity, it may be the subject of a subsequent specific
research program. In any case, tests on a car having
a rough underbody have been recently carried out by
Rover on the road and in the wind tunnel (with and
without moving belt). Results are reported in (12).
• tests are restricted to a yaw angle b = 0 °. Also tests
at b ¼ 0 ° can be regarded as a possible topic for
future research. Furthermore, the same standing
heights and the same track values have been used at
front and at rear.
• tests are also restricted to one rear end shape, i.e.
the notch-back shape, as this shape produces a
wake of small size, without strong longitudinal vorti-
ces. Therefore it should have less interference with
the flows coming from the underbody and which are
the aim of this research program.

3
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(b) development of rocker moulding

(a) development of front spoiler and spoiler corner

shape

(c) development of the tank spoiler to reduce CD

and rear end contamination

Figure 2. Details of the underbody aerodynamic

development work on the Opel Calibra (ref. 3)

4
Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

Figure 3.

5
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

Figure 4. Reference model as defined by the SAE Committee "Open Jet Wind Tunnel Boundary Interference" (from ref. 10)

6
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

Figure 5. The model as built by Pininfarina, with wheels, wheelhousings and various interchangeable parts (a, b, c), and
during the tests in the wind tunnel with moving ground and rotating wheels (d).

7
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

THE CAR MODEL hang, and the other moved backward by 100 mm , to
increase the front overhang.
The full scale model built for this research program was • n. 3 rear underbody parts, designed in a similar man-
chosen and built following these considerations : ner , so as to have rear wheel-housings at 3 different
• the model has to have a car-like shape, and , at the longitudinal positions.
same time, a shape simple , easy to build and easy • n. 3 add-on inserts to fill the front and rear wheel-
to check for deformations in the future, in case the housings in 3 progressive steps.
same model is used for later research. The wheel-housing size can be either increased by
• the model body has to be made of a central part, 75 mm in height or reduced in diameter by 50 and
and a set of front and rear interchangeable parts, to 100 mm (fig. 5)
reproduce the underbody configurations selected for • n. 2 add-on inserts to fill the rear diffuser, so as to
this research program. produce diffuser angles of 0°, 6° and 12°.
It was decided to use the shape of the reference model Furthermore :
proposed by the SAE Committee “Open Jet Wind Tunnel
Boundary Interference” to check the interference effects • the model is equipped with n. 2 sets of wheels , hav-
which exist in wind tunnels (10), see fig. 4 . ing the same diameter, same type of tires (same
brand and thread), but different tire width (155 and
This model shape has been defined within the Commit- 175 mm).
tee, based on the previous experience made by testing a
Wheel covers can be removed.
similar family of models, namely the “Mira Models”( 11 ).
• front and rear tracks can be easily varied around a
This new model shape has a number of small but signifi- reference value of 1300 mm , which corresponds to a
cant improvements. It is expected to be Reynolds number centered position of the wheels within the wheel-
independent in yaw conditions; the split line of the four housings.
interchangeable rear ends does not affect, by their
• the model is supported on the balance through the
upstream position, the flow over the backlight, it is simple
rockers, by 4 slender cylindrical struts, which are the
to build and to check in the time.
same as those normally used to support the cars
It is therefore suitable for investigating interference tested on the moving ground.
effects of wind tunnels, and to make correlation tests This type of support makes it possible to vary the
between wind tunnels. model standing heights, step by step, in a controlled
In addition to that , this shape , being simple to model, and repeatable way, from a minimum ground clear-
may be used to check the results of CFD codes and/or to ance of 50 mm up to the “reference” value of 200
compare these results with experimental data. mm.
• each model configuration can be tested in the 4 con-
Some models with this shape have already been built or
ditions of GESS, given by the various combinations
are under construction, in various scales, by some auto-
of MB “on”/”off” & Rotating Wheels “on”/”off”.
motive Companies and / or Universities, and this will help
to share experimental and numerical results. • the model and the various interchangeable parts are
equipped with 40 pressure taps , mainly along the
However this SAE reference model has no wheels; (it is underbody and the rear end (fig. 8).
supported by four slender cylindrical legs, so as to be
more effective in assessing wind tunnel interference with-
out any side effect caused by the floor boundary layer),
and no interchangeable parts in the underbody.
Therefore, to carry out this research program, Pininfarina
has designed and built a super-set of additional parts,
specific for investigating the underbody flow field and the
ground effect behavior (fig. 5).
These parts include :
• n. 3 front underbody parts, each one with the relevant
pair of wheel-housings.
Wheel-housings have 3 longitudinal positions, a “ref-
erence” condition (which corresponds to a reference
wheelbase of 2450 mm, front overhang of 925 mm
and rear overhang of 825 mm ) and two more, one
with the front axle moved forward by 100 mm which
increases the wheel-base and reduces the front over-

8
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(a) Example of diagrams CD versus the model ground clearance for various tracks and the four GESS conditions.

9
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(b) Example of diagrams of CLF versus the model ground clearance for various tracks and the four GESS conditions.

10
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(c) Example of diagrams of CLR versus the model ground clearance for various tracks and the four GESS conditions.

11
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(d) Example of diagrams of CL versus the model ground clearance for various tracks and the four GESS conditions

12
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(e) Example of diagrams of CMY versus the model ground clearance for various tracks and the four GESS conditions.

Figure 6. Examples of diagrams versus model ground clearance for various tracks and the four GESS conditions.

13
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

Figure 7. Diagrams of CD, CL, CLF, CLR, CMY coefficients versus the model ground clearance and some underbody
configuration parameters.

14
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018
THE TEST PROGRAM
The number of configurations available using the various
interchangeable parts, together with the various possible
standing heights, and the 4 GESS conditions, give a total
number of tests which is far too high to be practically
acceptable, in spite of the limitations already set at the
beginning of the program (flat and smooth underbody,
yaw angle β = 0° and a notch-back rear-end shape,
only).
Therefore to achieve the aim of the research program, i.e.
to find the trend of each aerodynamic coefficient versus
the various configuration parameters and the model
ground clearance, it was decided to perform a shorter
test program, which includes the following steps :
• a first test session on the model having reference
wheelbase, reference front and rear track, basic
wheel-housing, 175 mm tires. For this configuration,
all the various combinations of ground clearance,
rear diffuser angle and GESS condition have been
tested.
• following this initial set of tests, a number of addi-
tional test sessions have been carried out, by chang-
ing a parameter at a time, and repeating for each
session, the various combinations of ground clear-
ance, rear diffuser angle and GESS condition.
• the range of ground clearance, rear diffuser angle
and GESS condition, has been repeated for each
configuration. These parameters are considered
more important from a practical standpoint, because
the first two can be more easily modified on real cars,
while the third, the GESS condition, gives an insight
into the differences which can be expected, for each
parameter, for the different testing conditions of
ground (fixed/moving) and wheels (static/rotating).
• In some cases two configuration parameters have
been changed at the same time, for instance an
increase of track together with a change of front over-
hang.
To complete a total of about 2000 force measurements
and a similar number of pressure scans have been taken.
Fig. 5d shows the model in the wind tunnel during some
of these tests.
ANALYSIS OF RESULTS
A first analysis of this large amount of data has been car-
ried out by presenting the five main aerodynamic coeffi-
cients CD, CL, CLF, CLR, CMY versus the model ground
clearance and by showing in each diagram, a range of
values regarding one of the other parameters.
The aerodynamic coefficients have been presented ver-
sus ground clearance because this is the most important
parameter for the ground effect.
15
Furthermore, as every configuration has been tested in
four different GESS conditions, four diagrams have been
printed on the same page, one for each GESS condition.
The range of values in these diagrams usually refers to:
• different tracks
• different wheel base, i.e. different front and rear over-
hangs
• different wheel-housing sizes
An example of these diagrams can be seen in figs. 6
where the five aerodynamic coeff. CD, CL, CLF, CLR, CMY
are shown versus the ground clearance H (from 200 mm
down to a minimum of 50 mm when feasible) for each of
the four GESS conditions and for the range of track val-
ues.
From the examination of these diagrams, it is certainly
possible to find specific answers to a number of ques-
tions regarding the influence of the GESS condition and /
or of the various car parameters involved in this exercise.
It is for instance easy to recognize that :
• by reducing the ground clearance H , CD values
always become lower, CL often has a minimum at
some height H and CMY in general become lower.
• GESS “on” tests, compared to GESS “off” tests, usu-
ally show lower CD values at high ground clearance
H and so on.
It is however much more difficult, if not impossible, by
looking at these diagrams, to reach conclusions of gen-
eral value
Therefore, a second analysis has been performed by
doing a multiple linear regression on the experimental
data.
Starting from the model base configuration, increments
of CD, CL, CLF, CLR, CMY caused by changes to the fol-
lowing 7 parameters have been analyzed.
The 7 parameters are:
• Ground Clearance
• Diffuser Angle
• Tire Width
• Track
• Wheel-Housing Diameters
• Front Overhangs
• Rear Overhangs
The reference values, and increments are computed for
each parameter and are summarized in table 1.By calcu-
lating, for each aerodynamic coefficient, the multiple lin-
ear regression, corresponding to these seven
parameters, the % change in each aerodynamic coeffi-
cient, corresponding to 1% change for each one of the
seven parameters, is known.
The regression coefficients are reported in table 2. Final
expressions are reported in table 3.
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

Results of this analysis can now be shown in a few dia- is “off”, while the same decrease of front lift caused
grams only (fig. 7). by the increase of the wheel-housing size and of the
front overhang is under-estimated when GESS is
For example fig. 7a, which presents the % drag incre-
“off”.
ment per 1% change of each parameter, shows that a 1%
increase of ground clearance, increases the drag coeffi- Again, the wheel rotation alone or the ground motion
cient by about 0.28% in condition of GESS “on”, 0.34% in alone, sometime over-corrects, sometime under-cor-
condition of GESS “off “, and respectively 0.33% with rects the GESS “off” values.
moving ground only and 0.30% with rotating wheels only. • Rear Lift is in this case mainly dependent on the rear
diffuser angle.
From fig. 7c, it can be seen that a change of 1% of the
rear diffuser angle produces a change in the rear lift of - Again the decrease of rear lift, caused by the
4.5% and -3.9%, depending on the GESS conditions. increase of the rear diffuser angle, is a bit under-esti-
mated in case of GESS “off” or rotating wheels only,
These results are certainly specific for this simplified car while it is over-estimated when the moving ground
model and therefore cannot be considered valid for every only is operating.
car, mainly because of the characteristics of this model,
• Total Lifts can be reduced by increasing the diffuser
as mentioned at the beginning of the paper (completely
angle and the (front) track. GESS “off” under-esti-
smooth floor, rounded front end, etc.).
mated the diffuser effect and over-estimated the
However, they clearly show show the trends which are Track effect.
specific for each parameter and they also give an indica- • Pitching Moment can be reduced by reducing the
tion about how the trend of each parameter can change ground clearance, or the diffuser angle, or the tire
when the condition of flow simulation along the under- size.
body changes due to the ground motion or wheel rota-
tion. A further similar analysis was then performed by looking
at the pressure coefficients measured on the model sur-
As an example, some of these trends are mentioned here face, along the underbody and at the rear, in the various
under : configuration conditions.
• CD values increase when ground clearance , or diff. The position of the pressure taps along the model center
angle, tire size, wheel-housing size increases. How- line is shown in fig. 8 .
ever the increase of drag caused by the increase of
the ground clearance or the Wheel-Housing size is An example of pressure values measured for the four
over-estimated when GESS is “off”, while the GESS conditions is shown in fig. 9.
increase of drag caused by the increase of the diff. Some additional diagrams are reported in fig. 10, where
angle or the tire size is under-estimated when GESS the pressure coefficients are shown for the 4 GESS con-
is “off”. ditions and respectively for :
• CD values decrease when the track or the front and
rear overhangs are increased. However the Figure Ground Diffuser Wheel Hous-
decrease of drag caused by the increase of the Clearance Angle ing Diameter
Track is over-estimated when GESS is “off”, while mm deg mm
the decrease of drag caused by the increase of the 10a 200 0-6-12 +75
front and rear overhangs show a very small depen- 10b 100 0-6-12 +75
dency on the GESS condition. 10c 50 0-6-12 +75

• The wheel rotation alone usually improves the simu-

lation, compared to the GESS “off” condition, how- These diagrams clearly show some additional effects,
ever it often “over-corrects” the over or under- such as the important change of pressure, upstream of
estimated values which are measured when GESS the diffuser, when the diffuser angle changes from 0 to
is “off”. 12°.
Similar considerations can be easily made for the lift and Other effects are more difficult to define, so, as with the
the pitching moment coefficients (fig. 7). force coefficients, it was decided to perform a multiple lin-
ear regression for the values of selected pressure taps
• Front Lift can be reduced (at least in the case of this
versus the seven parameters (ground clearance, etc.)
model) by increasing the front track, or the wheel-
already mentioned.
housing size or the front overhang (fig.7). From a
physical standpoint this can be seen as a conse-
quence of an accelerated flow under the model front
end. This is confirmed by the pressure values mea-
sured under the model , see tap n.40 in fig. 11a.
However the decrease of front lift caused by the
increase of the track is over-estimated when GESS

16
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

Figure 8. Pressure taps along the underbody centerline and the model rear end.

Figure 9. Example of pressures measured along the venter line, on the underbody and at the rear end of the model.

17
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(a) Example of pressures measured along the center line, on the underbody and at the rear end of the model at H=200 mm
and diffuser angle = 0°, 6°, 12°.

18
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(b) Example of pressures measured along the center line, on the underbody and at the rear end of the model at H=100 mm
and diffuser angle = 0°, 6°, 12°.

19
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

(c) Example of pressures measured along the center line, on the underbody and at the rear end of the model at H=50 mm
and diffuser angle = 0°, 6°, 12°.

Figure 10. Examples of measured pressures.

20
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018
The five selected pressure taps are respectively:
• pressure tap n. 40 (underbody front end)
• pressure tap n. 8 (mid underbody)
• pressure tap n. 31 (rear underbody upstream the rear
diffuser)
• pressure tap n. 35 (on the rear diffuser)
• pressure tap n. 27 (on the model rear end)
Results of this multiple linear regression are shown in
figs. 11a, b, c, d, e.
From these diagrams, it can be seen that, for instance:
• the static pressure on the front tap 40 is mainly
dependent on the front overhang. By increasing the
front overhang, it is easy to reduce the pressure
under the model front end, on the centerline. How-
ever this effect is over-estimated when GESS is “off”
as well as when only the wheels are rotating or only
the moving ground is in operation. This can be
explained by considering that , in condition of GESS
“off”, the flow is slowed down by the fixed wheels and
by the static ground, and therefore speeds up along
the centerline, where the pressure locally goes down.
Only when the wheels are rotating and the ground is
moving, the flow can accelerate along the under-
body sides and slow down along the centerline.
• pressure values on the tap n. 31 (upstream of the dif-
fuser) depends mainly on the diffuser angle. Fig 11c
shows that the presence of a diffuser is felt well
upstream of the diffuser itself. In this case the GESS
“off” test condition , under-estimates the pressure
decrease.
• pressures on the car rear end (tap n. 27, fig. 11e)
seem to depend mainly on the tire size and front
overhang.
However, the correlation coefficient R2 of these regres-
sions is lower than the corresponding regressions
already determined for the force coefficients. This sug-
gests that in general, the linear regressions made on the
pressure values are less reliable, than those already
seen for the force coefficients, in defining the trends
associated with each parameter.
Furthermore, in the case of tap 27 , on the model rear
end, R2 is very low . This suggests that, compared to the
taps on the underbody, the base pressure is less depen-
dent on the various underbody parameters previously
mentioned.
ADDITIONAL COMMENTS
• A multiple linear regression might not be the best
technique to process this data, considering that
some trends ( see for instance CLF versus ground
clearance in fig. 6b ) are certainly not linear. How-
ever, some attempts made to introduce quadratic
terms in the regression, didn’t show any clear
improvement of the linear regression results and did
21
not change the conclusions which can be drawn
from these results.
• Aerodynamic coefficients are always referred to a
frontal area S = 1.9 m2. No change is made to
account for the different ground clearances. This
means that all the coefficient values at, for instance,
H = 50 mm should be lowered by 2.6%, which is
unimportant in the context of this work.
• In order to correctly perceive the effects which can be
produced by each one of the 7 parameters, i.e. their
importance, it is helpful to remember that the range
of possible variations of each parameter, referred to
its reference value, can be (and is) quite different.
For instance, ground clearance H can easily be reduced
by 50%, from 200 mm (reference) to 100 mm. And that
means that, by using the coefficients reported in tables 2
and 3 (see for instance the case of GESS “on”, the CD
value reduces by 0.280 * (-50%) = -14% (or even more in
case of H = 50 mm).
The parameter wheel-housing diameter (which shows a
similar regression coefficient = 0.283) can be reduced at
maximum by -50 mm in the radius or, in percentage of
the base diameter = -10.8%.
Correspondingly, the CD value reduces, at maximum, by
only 0.283 * (-10.8%) = -3%. All that means that the
importance of each parameter depends on the corre-
sponding regression coefficients and, at the same time,
on the extent of the parameter possible variations.
CONCLUSIONS
As a result of this first year of research work carried out
on a full-scale simplified car model, with a number of
interchangeable underbody parts, different wheels, vari-
ous ground clearances and 4 GESS conditions, these
main conclusions can be drawn.
The effect of important changes of the underbody layout
has been defined through a systematic set of tests.
These results are certainly valid in the case of car models
having a rounded lower front end , and a flat underbody
shape, which may be seen as the trend for the underbody
shape of future cars.
Any extrapolation of these results to conventional under-
body shapes (i.e. front dam and rough surface) has to be
taken with caution .
The following trends are established :
• Aerodynamic Drag is mainly improved by reducing
the ground clearance and the wheel-housing size, or
by increasing the track (within the limit of the body
width).
• Front Lift is improved (i.e. reduced) by reducing the
ground clearance, and the tire width, or by increasing
the track, the wheel-housing size and the front over-
hang.
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018
• Rear Lift is improved (i.e. reduced) by increasing the
rear diffuser angle.
• Total Lift is improved (i.e. reduced) by increasing the
rear diffuser angle and the (front) track.
The magnitude of these improvements may be somewhat
dependent on the type of the regression which has been
used to process the experimental data, however the cor-
relation coefficient R2 is high enough to ensure the reli-
ability of these results.
The consequences of using different GESS conditions on
the results of the various configuration changes have
been also clearly defined through these tests. In reality,
these different GESS conditions correspond to the test
conditions which can be found in various wind tunnels.
The difference of results obtained for each underbody
parameter, when tested in the 4 different GESS condi-
tion, may seem , at first, very small and therefore negligi-
ble. This impression is misleading. In fact, the sum of a
number of small differences caused by different test con-
ditions, can lead to overall results (i.e. a final car configu-
ration) which are quite different.
The attempt at improving the simulation, by spinning the
wheels without simulating the ground motion, often over-
corrects the CD values (i.e. gives CD values which are
often too low).
Furthermore there is no clear trend for the Lift values,
which sometime are increased, sometime decreased by
the wheel rotation, depending on the underbody parame-
ter which is modified.
The alternative attempt at improving the simulation, by
providing a ground motion, without spinning the wheels,
usually increases the CD values in a unrealistic manner ,
as the moving belt increases the flow hitting onto the
underbody and the static wheels.
The decrease in lift values, are usually over-corrected by
the ground motion alone, mainly when the ground clear-
ance is reduced.
Therefore, from a practical standpoint, it may be con-
cluded that, if a complete dynamic simulation (moving
ground and rotating wheels) is not possible, the second
best choice is probably a complete conventional static
simulation (i.e. GESS off ).The effects of wheel rotation
and moving ground partly compensate each other,
although in an unpredictable way. The wheel rotation
alone or the ground motion alone can be used, when
available, to analyze their separate effects, rather than to
improve the overall flow-field simulation.
Future work within this research program might include a
sistematic investigation of some of the following vari-
ables:
1. Front Air Dam and Underbody Roughness
2. Model Angle of Attack, i.e. different standing heights
at front and rear
3. Different Track Values between front and rear
22
4. Behaviour at Yaw
5. Different Upper Rear End Shapes
REFERENCES
1. A. Cogotti, “A Strategy for Optimum Surveys of Passenger
Car Flow Fields”, SAE paper 890374, Intl’ SAE Congress,
Detroit, February 1989
2. A. Cogotti, “Prospects for Aerodynamic Research in the
Pininfarina Wind Tunnel”, FISITA paper 905149, XXIII
FISITA Congress, Torino, May 1990
3. H.J. Hemmelmann, H. Berneburg, J. Schulze, “The Aero-
dynamic Development of the Opel Calibra”, SAE paper
900327, Intl’ SAE Congress, Detroit, February 1990
4. Mercker E., Knape H.W., “Ground Simulation with Moving
Belt and Tangential Blowing for Full-Scale Automotive Test-
ing in a Wind Tunnel”, SAE paper 890367, Detroit 1989
5. B. Bonis, F.B. uagliottiQuQuagliotti, G.P. Dallara, “Is the
Moving Belt Ground Plane really necessary for the Auto-
mobile Testing in the Wind Tunnel?”, SAE paper 870719,
Intl’ SAE Congress, Detroit, February 1987
6. C. Berta, “Full-Scale Moving Belt in Fiat Aerodynamic Wind
Tunnel”, FISITA paper 905147, XXIII FISITA Congress,
Torino, May 1990
7. A. Cogotti, “Ground Effect Simulation for Full-Scale Cars in
the Pininfarina Wind Tunnel”, CNR-Pininfarina Workshop
on Wind Tunnel Simulation on Ground Effect, Turin, May
1994
8. A. Cogotti, “Ground Effect Simulation for Full-Scale Cars in
the Pininfarina Wind Tunnel”, SAE paper 950996, Intl’ SAE
Congress, Detroit, February 1995
9. A. Cogotti, “Ground Effect Simulation in the Pininfarina
Wind Tunnel”, Status Report, EADE Meeting, Turin, Octo-
ber 1996
10. “Aerodynamic Testing of Road Vehicles in Open Jet Wind
Tunnel”, SAE Report J2071 (Open Throat Wind Tunnel
Adjustments), Revision of October 97
11. G.W. Carr, W.R. Stapleford, “Blockage Effects in Automo-
tive Wind Tunnel Testing”, SAE paper 860093, Intl’ SAE
Congress, Detroit, February 1986
12. G.M.Le Good, J.P. Howell, M.A. Passmore, A. Cogotti, “A
Comparison of On-Road Aerodynamic Drag Measure-
ments with Wind Tunnel Data from Pininfarina and Mira”,
SAE paper 98......, Intl’ SAE Congress, Detroit, February
1998
Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

23
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

Table2. Coefficients of the multiple linear regression for each of the 5 aerodynamic coefficients, the 4 Gess conditions and
the 7 pararmeters.

24
 Downloaded from SAE International by University of Liverpool, Sunday, August 12, 2018

Table3. Final expressions given by the regressions for the 4 Gess conditions, versus the 7 parameters.

25