An experimental study of bicycle aerodynamics

Abstract

The primary objective of this research is to measure aerodynamic drag of cyclist using a full scale experimental testing methodology. A full scale testing can provide detailed information about the cyclist along with all gears (bicycle, helmet, cycling suit, shoes, goggle and so on). This paper describes full scale measurements of aerodynamic forces of different bicycles and cyclist's body positions along with various gears under a range of wind speeds. The experimental findings indicate that the time trial bicycle has lower drag than road racing and mountain bicycles. Time trial bicycle has about 36% and road racing bicycle has about 21% less drag than the mountain bicycle. The results also show that the time trial position has the lowest C D value. About 30% and 45% drag reduction are possible in road racing and time trial positions respectively compared to the recreational position.

Full text

269
International Journal of Mechanical and Materials Engineering (IJMME), Vol.6 (2011), No.2, 269-274
AN EXPERIMENTAL STUDY OF BICYCLE AERODYNAMICS
H. Chowdhury, F. Alam and I. Khan
School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University, Melbourne, Australia
Email: harun.chowdhury@rmit.edu.au
Received 14 November 2010, Accepted 18 July 2011
ABSTRACT
The primary objective of this research is to measure
aerodynamic drag of cyclist using a full scale experimental
testing methodology. A full scale testing can provide
detailed information about the cyclist along with all gears
(bicycle, helmet, cycling suit, shoes, goggle and so on). This
paper describes full scale measurements of aerodynamic
forces of different bicycles and cyclist’s body positions
along with various gears under a range of wind speeds. The
experimental findings indicate that the time trial bicycle has
lower drag than road racing and mountain bicycles. Time
trial bicycle has about 36% and road racing bicycle has
about 21% less drag than the mountain bicycle. The results
also show that the time trial position has the lowest CD
value. About 30% and 45% drag reduction are possible in
road racing and time trial positions respectively compared to
the recreational position.
Keywords: Bicycle, Aerodynamic drag, Wind tunnel, Full
scale test, Cycling, Experimental measurement
NOMENCLATURE
A
Projected frontal area of the cyclist, bicycle
and associated gears
CD
Drag coefficient
CR
Coefficient of rolling resistance
F
Total resistance
FD
Aerodynamic drag
FR
Rolling resistance between wheels and road
surface
g
Gravitational acceleration
m
Total mass
V
Cyclist speed
ρ
Air density
1. INTRODUCTION
In cycling, aerodynamics plays a crucial role in the overall
performance of the cyclist. The aerodynamic gains in
bicycle racing are of great significance. It is clearly evident
from the images of the successful cyclists of Tour de France
and Olympic cycling events (regardless of race type, be it
road racing, time trial or track events), aerodynamics plays
an important role in all disciplines when the speed is over 20
km/h. Tour de France 2011 data indicates that the average
speed of the top cyclist is approximately 40 km/h, however,
the speed in the time trial stage is over 55 km/h. Although
the average speed in mountain stages is slightly below 40
km/h, the maximum speed in downhill stages can easily
exceed to 100 km/h. On a flat road, the total resistance
encountered by a cyclist is shown in Equation 1 and
illustrated in Figure 1. The total resistance (F) consists of
the sum of aerodynamic drag (FD) and rolling resistance
between wheels and road surface (FR).
RD FFF 
(1)
The aerodynamic drag increases with the square of a
cyclist’s speed. Aerodynamic drag depends on the projected
frontal area of the cyclist and the bicycle, the drag
coefficient (CD) which is a measure of the flow quality
around the cyclist and the bicycle, density of air, and the
square of the road speed. Aerodynamic drag (FD) is
expressed as:
ACVF DD 2
2
1


(2)
On the other hand, rolling resistance (FR) depends on the
total mass of the bicycle including the cyclist and the other
accessories. It can be expressed as:
mgCF RR 
(3)
On a flat road, the aerodynamic accounting for almost 50%
to 90% of the total resistance depending on the cyclist speed
and body position during cycling. Kyle and Burke (1984)
reported that the contributions of aerodynamics to the total
drag were 50% (at 3.6 m/s) and 90% (at 8.9 m/s). Luke,
Chin and Haake (2005) stated that the aerodynamic
contribution was 50% at 8.75 m/s for the mountain bicycle.
Apart from the cyclist, a bicycle is comprised of the frame,
forks, wheels, drive train, brakes, handlebars, water bottles,
etc. that interact with the oncoming airflow generating a
complicated and as yet still not fully understood series of

270
perturbations. There are mainly three types of bicycle: road
racing and time trial, mountain bicycle. The bicycle
accounts around 31% to 39% of the total aerodynamic
resistance depending on the bicycle type (Kyle and Burke,
1984; Luke at el., 2005). Out of total aerodynamic drag, the
rider position counts approximately 65 to 80% depending on
body position, helmet and clothing.
Wind
FD
FR
FR
Figure 1 Forces acting on cycling
The remaining drag is coming from bicycle frames, wheels
(mainly front wheels) and other components and adds on.
Prior studies by Faria, Parker and Faria (2005), Jekendup
and Martin (2001) and Lucia, Earnest and Arribas (2003)
reported that the cyclist body position along with a helmet
and suit can significantly minimize the aerodynamic drag
experienced by the cyclist at all stages of racing be it road
racing or time trial. Studies by Brownlie et al (2004) and
more recently by Chowdhury et al. (2008, 2009, 2010)
indicate that the sports apparel can make significant impact
on the aerodynamic drag reduction thus influence on the
outcomes of the events. Studies by Alam et al. (2010), Reid
and Wang (2000) looked at the aerodynamics and thermal
comfort of different bicycle helmets. The reports by Alam et
al. showed that the helmet can produce up to 8% of the total
aerodynamic drag depending on the shape and venting
features of the helmet.
The different body positions are commonly used by
professional cyclists depending on the type of racing and
profile of the terrain. Upright position characterized by the
hands on the upper part of the handlebars, is mainly used in
road racing event. Another cycling position is time trial
position, when the elbows are placed on the pads of the
aero-handlebars, is believed to be the best aerodynamic
position to overcome the aerodynamic drag. However, the
body position remains at almost upright position for the
recreational cyclist and believed to be the most non
aerodynamic cycling position.
In cycling, scant information on full scale testing
methodology and the aerodynamic drag on cyclist are
readily available in the public domain. The primary
objective of this research is to experimentally measure
aerodynamic drag of cyclist using RMIT developed full
scale testing methodology. Details about the methodology
can be found in Chowdhury, Alam and Mainwaring (2011).
2. EXPERIMENTAL PROCEDURE
2.1 Experimental Setup
Figure 2 shows a schematic of the RMIT developed test
setup. The setup consists of a flat wooden platform (1800
mm  850 mm  30 mm) and a stand to hold the bicycle and
rider with the wooden platform firmly. The gap between the
wooden platform and the tunnel floor is 20 mm in order to
avoid any interference between the floor and the wooden
platform. A plastic fairing (as shown in Figure 3) is used at
the front of the platform to minimize the flow separation
from the leading edge of the platform. The whole platform
is connected with a 6-component force sensor (type JR3) via
a 100 mm diameter strut (shown in Figure 2) to measure the
drag, lift and side forces and their corresponding moments
simultaneously. All three categories of bicycles
(recreational, road racing, time trial) along with the rider can
be experimentally evaluated using this setup. The crosswind
effects can also be evaluated using the arrangement. The
developed system minimises error in data recording due to
extraneous cyclist movement or variations in weight
distribution.
The setup is robust enough for both static (cyclist with no
pedalling) and dynamic (cyclist with pedalling) testing. In
order to adjust the cyclist body positions, a video
positioning system has been developed. It consists of two
HD (high definition) digital video cameras detailed
description of this video positioning system can be found on
in Chowdhury et al. (2011). It can minimise any error
occurred due to the change of positions of the cyclist and
equipment as minor position variation can significantly
affect the aerodynamic data. This positioning system is
intended to ensure the reproducibility of the cyclist position
during the experimental procedure. The experimental setup
can be used in any wind tunnel that has appropriate test
section with a solid blockage ratio less than 10%. The
developed setup is well suited to the RMIT Industrial Wind
Tunnel.
The tunnel is a closed return circuit wind tunnel with a
turntable to simulate the cross wind effects. The rectangular
test section dimensions are 3 meters wide, 2 meters high and
9 meters long, and the tunnel’s cross sectional area is 6
square meters. The tunnel is suitable for the full scale
bicycle (along with the cyclist) testing as the solid blockage
ratio of the tunnel is less than 10%. The maximum speed of
the tunnel is approximately 145 km/h. A plan view of the
tunnel is shown in Figure 4.

271
Bicycle Trainer
Bicycle
Wooden Platform
6-component load cell
Front Wheel Locker
Front Camera
Side Camera
Figure 2 Schematic of the experimental arrangement
Figure 3 Experimental setup in the wind tunnel
Car
Entrance
Heat Bench
System
Anechoic
Turning
Vanes
Anechoic
Turning
Vanes
Turning
Control
Turntable
Test
Section
Retractable
Turning Vanes
Motor Room
Fan
Diffuser Contraction
Flow
Vanes
Panel
Flow
Heat Bench Pipes
KEY
Figure 4 A plan view of RMIT Industrial Wind Tunnel
The tunnel air speed is measured with a modified National
Physical Laboratory (NPL) ellipsoidal head Pitot-static
tube (located at the entry of the test section) which is
connected through flexible tubing with the Baratron®
pressure sensor made by MKS Instruments, USA. Details
description of the tunnel can be found out in Alam at al.
(2009). The developed test setup was connected through a
mounting stud with the sensor (type JR3) as mentioned
previously. The sensor was used to measure all three
forces (drag, lift and side forces) and three moments (yaw,
pitch and roll moments) simultaneously. Each set of data
point was recorded for 30 seconds with a frequency of 20
Hz ensuring electrical interference is minimized. Multiple
data sets were collected at each speed tested and the
results were averaged for minimizing the further possible
errors in the experimentally acquired data.
In order to evaluate the experimental methodology and test
setup, a series of tests were undertaken using three
different types of bicycles with or without live cyclists
under a range of wind speeds (20 to 70 km/h with an
increment of 10 km/h). The selected speed range is
representative for all major cycling starting from
recreational to time trial.
2.2 Bicycles
Three different types of bicycle were tested at zero degree
yaw angle for the measurement of aerodynamic properties.
The bicycles were selected based on their applications in
cycling. These are: a) mountain bicycle (manufactured by
Kent, USA) used for recreational and mountain cycling, b)
road racing bicycle (manufactured by Orbea, Spain) used

272
specially in professional road racing, and c) time trial
bicycle (manufactured by Louis Garneau, Canada) used in
professional time trial racing. The aerodynamic forces
were measured and compared. Figure 5 shows the
experimental setup with all 3 bicycles inside the wind
tunnel test section.
(a) Mountain bicycle (Kent)
(b) Road racing bicycle (Orbea)
(c) Time trial bicycle (Louis Garneau)
Figure 5 Test bicycles inside the tunnel test section
2.2 Cycling Positions
The aerodynamic forces were measured by using a
recreational and a professional cyclist for three widely
used cycling positions. In wind tunnel testing, a mountain
bicycle (see Figure 6a) and two other racing bicycles
along with appropriate helmets, bicycle and other
accessories were used to replicate the real cycling as
possible.
(a) Recreational position
(b) Road racing position
(c) Time trial position
Figure 6 Cyclist’s body positions and projected frontal
areas

273
The measured three positions were: a) recreational
position (generally upright body position), b) road racing
position, and c) Time trial position. For the time trial
position, a professional time trial bicycle (Louis Garneau
as shown in Figure 6c) and a Giro Advantage time trial
helmet were used. For other two positions, a professional
road racing bicycle (Orbea as shown in Figure 6b) and a
Giro Atmos road racing helmet were used. Figure 6 shows
all the configurations for the full scale wind tunnel testing
in 3 different cycling positions. In the figures, the shaded
area (in black colour) represents the projected frontal area
measured with the frontal area measurement system
described in Chowdhury at al. (2011).
3. RESULTS AND DISCUSSION
As mentioned previously, aerodynamic tests were
conducted at speeds (20 km/h to 70 km/h with an
increment of 10 km/h. The net aerodynamic forces acting
on the bicycle or bicycle with cyclist were calculated by
subtracting the force measured with the experimental
setup from the total forces measured with the bicycle or
cyclist with the bicycle including the experimental setup.
Figure 7 shows the drag variation with speeds for a time
trial, a road racing and a mountain bicycle. The figure
indicates that the mountain bicycle has higher drag and the
time trial bicycle has lower drag. It is clearly seen from
Figure 6 that the mountain bicycle is less aerodynamic
than other two. Mountain bicycle has wide and rough tyre,
straight handlebar. On the other hand, other two bicycles
have aerodynamic handlebars, wheels and fork. Because
of the more streamlined features (e.g., fork, handlebar,
wheels) are integrated in the time trial bicycle, it has the
lowest aerodynamic drag among these bicycles tested.
Time trial bicycle has about 36% and road racing bicycle
has about 21% less drag than the mountain bicycle.
However, time trial bicycle has 18% less drag compared to
the road racing bicycle.
0
5
10
15
20
25
30
35
020 40 60 80
Drag Force (N)
Speed (km/h)
Time Trial Bicycle
Road Racing Bicycle
Mountain Bicycle
Figure 7. Drag force variation with speed for three
different types bicycles
The non dimensional drag coefficient (CD) was computed
using the following formula:
AV
F
CD
D2
2
1


(4)
The projected frontal areas (A) were measured with the
method described in Chowdhury at al. (2011). The areas
were obtained as 0.54, 0.41 and 0.38 m² for recreational,
road racing and time trial positions respectively. Projected
frontal area is reduced by about 24% for road racing and
30% for time trial positions compared to recreation
cycling position. Drag coefficients were calculated using
these frontal areas. Figure 8 shows the CD variation with
speed for recreational, road racing and time trial positions.
Results indicate that at recreational position, the CD value
is more than that at other two positions (i.e., road racing
and time trial positions). On the other hand, the time trial
position exhibits the lowest CD value in the speed range
tested. It is clearly evident from Figure 6 that at time trial
position, the projected frontal area is smaller, the bicycle
drag is lower, and also the body position is more inclined
than other two positions. Hence, this position has the
lowest CD value. Results show that about 30% and 45%
drag reductions are possible in road racing and time trial
positions respectively compared to the recreational
position. However, about 21% variation in CD values
between time trial and road racing positions was noted
(see Figure 8).
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
010 20 30 40 50 60 70 80
CD
Speed (km/h)
Recreational
Road Racing
Time Trial
Figure 8. Drag coefficient as a function of speeds for
three different cycling positions
The experimental data was closely monitored for the
repeatability and accuracy. As mentioned earlier, the
position of the cyclist and the experimental setup were
monitored during the data accusation with a video
monitoring system. A small position change of the
experimental setup can cause a large variation in the data

274
measurement. Therefore, to minimize the error in the
measurement, several measurements (at least 3 times)
were taken in each position. Data were checked for the
consistency with the video feedback. Re-tests were
performed for any large variations in the measurements.
Table 1 shows the average CD with standard deviation.
The results in the table indicate that the data variation is
minimal. Therefore, the measurement is repeatable and
reliable. Table 1 Error analysis
Speed (km/h)
CD
20
1.1395 ± 0.0172
30
1.1411 ± 0.0077
40
1.1393 ± 0.0040
50
1.1340 ± 0.0026
60
1.1231 ± 0.0034
70
1.1045 ± 0.0013
4. CONCLUSION
Time trial bicycle has lower drag than road racing and
mountain bicycles. Time trial bicycle has about 36% and
road racing bicycle has about 21% less drag than the
mountain bicycle. Time trial position has the lowest CD
value. About 30% and 45% drag reduction are possible in
road racing and time trial positions respectively compared
to recreational position. The RMIT developed
experimental arrangement allows aerodynamic evaluation
not only for the cyclist but also the bicycle, helmet and
other accessories (wheels, handle bars, cycling suits, bells
and horns, lights, water bottle, etc) with high level
accuracy. The experimental measurement method is
simple and reliable. The repositioning system developed
here is simple, user friendly and accurate.
ACKNOWLEDGEMENTS
The authors express their sincere thanks to Mr Jordi
Beneyto-Ferre, School of Applied Sciences, RMIT
University for his assistance with the full scale testing in
the wind tunnel.
REFERENCES
Alam, F. Chowdhury, H. Theppadungporn, C. Subic, A.
and Khan, M. 2009. Aerodynamic properties of
badminton shuttlecock. International Journal of
Mechanical and Materials Engineering 4 (3): 266-
272.
Alam, F., Chowdhury, H., Elmira, Z., Sayogoa, A., Love,
J. and Subic, A. 2010. An experimental study of
thermal comfort and aerodynamic efficiency of
recreational and racing bicycle helmets. Procedia
Engineering 2:2413-2418.
Brownlie, L.W., Kyle, C.R., Harber, E., MacDonald, R.,
Shorten, M.R., 2004. Reducing the aerodynamic
drag of sports apparel: Development of the NIKE
Swift sprint running and SwiftSkin speed skating
suits. In: Hubbard M, Mehta, R.D, Pallis J.M (ed)
The Engineering of Sport 5, International Sports
Engineering Association, UK: 90-96.
Chowdhury, H, Beneyto-Ferre, J., Tate, M., Alam, F.,
Mainwaring, D., Forster, D. and Subic, A. 2009.
Effects of Textile and Garment Design on
Aerodynamic Characteristics Applied to Cycling
Apparel. In: The Impact of Technology on Sports
III, RMIT University, Australia: 131-136.
Chowdhury, H., Alam, F. and Subic, A. 2010.
Aerodynamic Performance Evaluation of Sports
Textile. Procedia Engineering 2: 2517-2522.
Chowdhury, H., Alam, F., Mainwaring, D. 2011. A full
scale bicycle aerodynamics testing methodology.
Procedia Engineering 13: 94-99.
Chowdhury, H., Alam, F., Mainwaring, D.E., Subic, A.,
Tate, M. and Forster, D. 2008. Methodology for
aerodynamic testing of sports garments. In: The
Proceedings of the 4th BSME-ASME International
Conference for Thermal Engineering, BSME,
Dhaka: 409-414.
Faria, E., Parker, D. and Faria, I. 2005. The Science of
Cycling Factors Affecting Performance – Part 2,
Sports Medicine 35: 313-337.
Jekendup, A.E. and Martin, J. 2001. Improving cycling
performance: how should we spend our time and
money, Sports Med 31: 559-569
Kyle, C.R. and Burke, E.R. 1984. Improving the racing
bicycle, Mechanical Engineering 106 (9): 34–35.
Lucia, A., Earnest, C. and Arribas, C. 2003. The Tour de
France: a physiological review, Scand J Med Sci
Sports 13: 275-283.
Lukes, R.A., Chin, S.B., Haake, S.J. 2005. The
understanding and development of cycling
aerodynamics, Sports Engineering 8: 59-74.
Reid, J. and Wang, E.L. 2000. A system for quantifying the
cooling effectiveness of bicycle helmets, J Biomech
Eng. 122: 475-460.