ARTICLE IN PRESS

Journal of Biomechanics 39 (2006) 2535–2541

www.elsevier.com/locate/jbiomech
www.JBiomech.com
Short communication

Skin-friction drag analysis from the forced convection modeling in

simplified underwater swimming
G. Polidoria, R. Taı̈arb, S. Fohannoa,, T.H. Maia, A. Lodinib
a
Laboratoire de Thermomécanique, Moulin de la Housse, BP 1039, 51687 Reims cedex 2, France
b
Laboratoire d’Analyse des Contraintes Mécaniques, Moulin de la Housse, BP 1039, 51687 Reims cedex 2, France
Accepted 18 July 2005

Abstract

This study deals with skin-friction drag analysis in underwater swimming. Although lower than profile drag, skin-friction drag
remains significant and is the second and only other contribution to total drag in the case of underwater swimming. The question
arises whether varying the thermal gradient between the underwater swimmer and the pool water may modify the surface shear
stress distribution and the resulting skin-friction drag acting on a swimmer’s body. As far as the authors are aware, such a question
has not previously been addressed. Therefore, the purpose of this study was to quantify the effect of this thermal gradient by using
the integral formalism applied to the forced convection theory. From a simplified model in a range of pool temperatures (20–30 1C)
it was demonstrated that, whatever the swimming speeds, a 5.3% reduction in the skin-friction drag would occur with increasing
average boundary-layer temperature provided that the flow remained laminar. However, as the majority of the flow is actually
turbulent, a turbulent flow analysis leads to the major conclusion that friction drag is a function of underwater speed, leading to a
possible 1.5% reduction for fast swimming speeds above 1 m/s. Furthermore, simple correlations between the surface shear stress
and resulting skin-friction drag are derived in terms of the boundary-layer temperature, which may be readily used in underwater
swimming situations.
r 2005 Elsevier Ltd. All rights reserved.

Keywords: Skin-friction drag; Forced convection; Underwater swimming

1. Introduction research strategies have been designed to reduce drag.

In human swimming under normal circumstances, in
which a water/air interface occurs, the total drag (Fd)
which is the resistance force on the swimmer’s body, in
motion within a viscous medium, is respectively
composed of the skin-friction drag (Ff), the profile drag
(Fp) and the wave drag (Fw). The contribution of skin-
friction drag is generally assumed to be up to 5% of the
total drag (Toussaint et al., 2002) whereas the contribu-
tion of the wave drag may reach a maximum of 60% of
the total drag (Vennell et al., 2005). In the quest for
higher levels of performance in swimming, many recent
Corresponding author. Tel./fax: +33 326 91 83 10.
E-mail address: stephane.fohanno@univ-reims.fr (S. Fohanno).
As well as improved understanding of the fundamental
mechanisms of swimming, such as optimisation and
coordination of the different movements, technological
developments may offer potential benefits.
The concept of using specifically designed swimming
suits, modeled on shark skin, to achieve drag reduction
by controlling the near-wall turbulence and skin-friction
forces has received much attention. In idealized labora-
tory experimental conditions, Bechert et al. (2000) found
as much as 7.3% decrease in turbulent shear stress, as
compared to a smooth reference plate. Koeltzsch et al.
(2002) confirmed this conclusion from fluid dynamic
experiments showing that small riblet surfaces induced
drag reductions of up to about 10% compared to
smooth surfaces. In contrast, Toussaint et al. (2002)

0021-9290/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jbiomech.2005.07.013
 ARTICLE IN PRESS
2536 G. Polidori et al. / Journal of Biomechanics 39 (2006) 2535–2541
showed that a statistically non-significant 2% reduction
in drag was found in real pool conditions when wearing
fast-skin suits compared to conventional ones.
In contrast with previous studies, the present analysis
deals especially with the skin-friction drag, which is
approached theoretically from non-isothermal fluid
dynamic theory.
To simplify the analysis, the underwater swimming
situation is considered. This swimming condition is of
interest because it leads to negligible wave drag (Vennell
et al., 2005) and thereby the greatest potential speed
increase for elite swimmers, estimated to be up to 40%
(Laughlin, 2003). This underwater swimming situation
also results in a greater contribution of skin-friction
drag to the total drag. The question arises whether
varying the thermal gradient between the underwater
swimmer and the pool water may modify the surface
shear stress distribution and the resulting skin-friction
drag acting on a swimmer’s body. As far as the authors
are aware, such a question has not yet been addressed.
Therefore, the aim of this study was to quantify the
effect of this thermal gradient on the surface shear stress
distribution and the resulting skin-friction drag acting
on a swimmer’s body. For this purpose, the integral
formalism was applied to forced convection theory.
Furthermore, first-order regression equations for the
surface shear stress and deduced skin-friction drag were
derived in terms of the boundary-layer temperature for
use in underwater swimming situations.
2. Background of the problem
Swimmer’s body geometry can be modeled as a
linkage of complex convex, concave, truncated, ellipsoi-
dal non-rigid surfaces (Yanai, 2001). Both swimmer’s
body shape and the specific swimming propelling
motion induce very complex separated shear layers
combining turbulent fluctuations and large-scale vor-
tices. Such a physical problem seems theoretically too
complex to be analysed and a simplified approach is
required.
In order to reduce profile drag underwater swimmer
should adopt a streamlined shape, as flat as possible, to
prevent the boundary layer from separating. This is the
reason why, using a rough assumption, the physical
model used in the present analysis and presented in
Fig. 1 consists of a flat surface at the human body
surface temperature Tw flowing at a steady constant
velocity UN in a quiescent water pool whose tempera-
ture TN varies in the range 20 1CpTNp30 1C. It should
be noted that this model is also applicable when the
swimmer is wearing a full-body competitive swimsuit. In
such a case, Tw is defined as the external (water-
side) swimsuit surface temperature. Even if suits’
materials are good thermal insulators (l
0:05
Y, V
LAMINAR TURBULENT
U∞ T∞ TRANSITION
δ
DYNAMICAL LAYER
δΤ
TS
THERMAL LAYER x, U, T
ISOTHERMAL BODY SURFACE
Fig. 1. Summary of the simplified boundary-layer forced convection
model.
0:15 W m1 K1), the competitive suits are so thin
(100 mm) that the swimsuit surface temperature will
remain close to that of the real body surface.
The simplified model assumes a uniform temperature
over the whole body surface corresponding to the
average body surface temperature. This choice is
difficult due to the real non-uniformity of the skin
temperature in swimming (higher temperature where
muscles are active, lower elsewhere) and the lack of
measurements in the literature. For example, Brandt
and Pichowsky (1995) measured a 33 1C local skin
temperature over the deltoid muscle after strenuous
exercise. More recently Jansky et al. (2003) investigated
skin temperature changes induced by local cooling. For
example, they measured temperatures between 32.5 and
34.5 1C at the start of their experiments over various
body areas (trunk, forehead, thigh, forearm, fingers, and
palms). These temperatures correspond closely to those
of swimmers at the start of swimming races. Therefore,
an estimated average skin temperature Tw ¼ 33 1C was
chosen for this analysis.
The flat surface is supposed to model a 2D-surface
placed parallel to the mean uniform stream and acting
like the smooth skin surface of a rigid body. Because a
thermal gradient occurs between the flat surface and the
ambient fluid, both dynamic and thermal boundary
layers develop at the body surface.
A common approach in forced convection consists of
evaluating the fluid properties at an average boundary-
layer temperature, called the film temperature, and
defined as
T̄ ¼ ðT w þ T 1 Þ=2. (1)
Table 1 summarizes the various physical properties
of water in the realistic pool temperature range
20 1CpTNp30 1C and used for further analysis.
3. Theoretical modeling
Initially the boundary-layer development is laminar
but at some critical distance xc from the leading edge
small disturbances in the flow begin to be amplified,
 ARTICLE IN PRESS
G. Polidori et al. / Journal of Biomechanics 39 (2006) 2535–2541 2537

Table 1 The continuity and momentum integral equations are

Thermo physical properties of water at different film temperatures
expressed as
Swimmer Tw Water TN Film Density r Kinematic Z

(1C) (1C) temperature (kg/m3) viscosity n
q d qU


ðU 1  UÞU dy ¼ n
, (2)
T̄ð CÞ (m2/s)  106 qx 0 qy y¼0

33 20 26.5 996.68 0.865 Using the integral method inevitably leads to a choice of
33 22 27.5 996.41 0.846 a polynomial order characterizing both velocity and
33 24 28.5 996.12 0.829 temperature profiles. From mathematical criteria a
33 26 29.5 995.83 0.811
33 28 30.5 995.53 0.794
fourth order shape has been advocated by Polidori
33 30 31.5 995.21 0.777 et al. (1999). Thus the velocity profile across the
dynamical boundary-layer is expressed as a fourth-order
polynomial using the Pohlhausen method (Polidori and
Padet, 2002)
 3  4
Uðx; y; T̄Þ y y y
¼2 2 þ .
U1 dðx; T̄Þ dðx; T̄Þ dðx; T̄Þ
(3)
Resolution of momentum integral Eq. (2) leads to the
velocity boundary layer thickness
rffiffiffiffiffi
35 x
dðxÞ ¼ 6 pffiffiffiffiffiffiffiffiffiffiffiffiffi . (4)
Fig. 2. Anthropometric parameters of subject Yann in a prone 37 ReðT̄Þ
position with the arms extended above the head and the head facing
the bottom of the pool. In laminar forced convection, the local shear stress at
the surface tw is defined as
 
qUðT̄Þ
characterizing a transition process towards turbulence tw ¼ rðT̄ÞnðT̄Þ . (5)
qy y¼0
(Fig. 1). In submerged swimming, the governing para-
meters reduce to the Reynolds number ReL ¼ U 1 L=n Incorporating Eqs. (3) and (4) in the shear stress
(UN is the swimmer’s speed, L is the swimmer’s body expression (5) leads to the analytical surface shear stress
length with stretched arms defined in Fig. 2 and v is the expression
water kinematic viscosity). It is usually assumed that the rffiffiffiffiffi sffiffiffiffiffiffiffiffiffi
swimmer’s body acts like a rigid streamlined body and 1 37 nðT̄Þ
tw jLAM ¼ rðT̄Þ ½U 1 3=2 . (6)
the critical location xc of the transition is estimated to 3 35 x
correspond to a Reynolds number Rexc ¼ 5  105 . For
example, for a swimmer’s speed U 1 ¼ 2 m=s and for
L ¼ 2 m, the transition from laminar to turbulent occurs 3.2. Shear stress modeling in the turbulent case
for xc
0:2 m. About 10% of the skin surface is
subjected to a laminar flow regime so that the major The transitional region is very complex and not yet
part of the body surface will experience a turbulent understood, so that the turbulent regime is usually
boundary layer. Because both dynamic and thermal considered as idealized and the corresponding turbulent
viewpoints are of interest, the simplified model is based theory is applied beyond the critical Reynolds number
on external boundary-layer theory applied to steady value. In turbulent flow the governing parameters are
forced convection flow conditions. For this purpose the made up of both mean and time-fluctuating compo-
Karman-Pohlhausen integral approach (Kakac- and nents, so that it is convenient to consider their time-
Yener, 1995; Padet, 1997; Polidori et al., 2003) has been averaged expressions to solve corresponding problems.
used and developed in the range of swimming Reynolds It has been shown (Kakac- and Yener, 1995) that in the
numbers inducing both laminar and turbulent effects. fully turbulent region (outside the laminar sub-layer) the
average velocity increases approximately as a one-
seventh-power law resulting in the corresponding profile
3.1. Shear stress modeling in the laminar case  
U y 1=7
¼ . (7)
U1 dðT̄Þ
This section deals with the laminar boundary-layer
theory of the forced convection developed in the range Incorporating this velocity profile in the integral
Rexc o5  105 . momentum Eq. (2) gives the following analytical form
 ARTICLE IN PRESS
2538 G. Polidori et al. / Journal of Biomechanics 39 (2006) 2535–2541

of the surface shear stress Table 2

Transition critical location xc between the laminar and turbulent
7 qdðT̄Þ regimes
tw ¼ rðT̄ÞU 21 . (8)
72 qx Underwater swimming speed
A common way to express the surface shear stress
developing at the surface of a flat wall is as an empirical 1 m/s 1.5 m/s 2 m/s 2.8 m/s
T̄ ¼ 26:5  C 0.43 m 0.29 m 0.22 m 0.16 m
form from measurements of turbulent flows in circular
T̄ ¼ 31:5  C 0.39 m 0.26 m 0.19 m 0.14 m
ducts (Kakac- and Yener, 1995)
 1=4
45 nðT̄Þ with outstretched arms L are of interest in the modeling,
tw ¼  103 rðT̄ÞU 21 (9)
2 U 1 dðT̄Þ as shown in Fig. 2.
Thus combining Eqs. (8) and (9) leads to the expression The global skin-friction drag may be obtained by
of the dynamical boundary-layer thickness integrating the local shear stress over the simplified
anthropometric dimensions of the swimmer with respect
dðT̄Þ 0:371 to the flow regime conditions.
¼ . (10)
x Rex ðT̄Þ1=5  Z xc
Incorporating (10) in Eq. (9) provides an expression for F f ¼ P̄H tw jLAM ðx; T̄; U 1 Þ dx
0
the turbulent surface shear stress Z L 
1 þ tw jTURB ðx; T̄; U 1 Þ dx . ð13Þ
tw jTURB ¼ 2:88  102 rðT̄ÞU 21 . (11) xc
Rex ðT̄Þ1=5
Incorporating Eq. (12) into expression (13) leads to the
First-order regression equations for both laminar and skin-friction drag acting on the swimmer’s body model
turbulent regimes have been deduced from the previous expressed as a function of the average thermal
parametric analyses in the complete range of considered boundary-layer temperature:
parameters and can be expressed in terms of film pffiffiffiffiffi
temperature and swimming speed. It should be noted F f ¼ P̄H fð6:80  103 T̄ þ 0:816ÞU3=2
1 xc
that these correlations are only suitable in the range of þ ð9:94  103 T̄ þ 2:462ÞU9=5
1 ½L
4=5
 x4=5
c g.
film temperature 26:5pT̄ð CÞp31:5 corresponding to ð14Þ
realistic indoor pool temperatures
rffiffiffiffiffiffiffiffi
U1 4. Results and discussion
3
tw jLAM ¼ ð3:40  10 T̄ þ 0:408ÞU 1 ;
x
The question that has motivated the present analysis
Rex o5  105 ,
was to theoretically quantify the potential influence of
 4 1=5
3 U1 varying the pool temperature on the laminar and
tw jTURB ¼ ð7:95  10 T̄ þ 1:970ÞU1 ; turbulent surface shear stress distributions and the
x
deduced skin-friction drag calculations. To make that
Rex 45  105 . ð12Þ
quantification easier a drag performance percentage was
It may be recalled that there is no satisfactory theory also established from a reference thermal situation.
dealing with the transitional region between laminar and
turbulent regimes but the analysis suggests that the 4.1. Shear stresses analysis
transitional region corresponds to the beginning of the
turbulent regime. Table 2 shows the various transition First, the influence of the temperature gradient on the
critical locations for the chosen parameter range. surface shear stresses was evaluated at given streamwise
body abscissae for both laminar and turbulent regimes.
3.3. Overall skin-friction drag For elite swimmers, underwater turn techniques lead
competitors to have forward speed of gliding after turn
The skin-friction drag force in the present analysis and push up to 2.8 m/s.
reduces to the skin friction Ff exerted by the external In Fig. 3 the surface shear stress is plotted against the
ambient water on the body surface area of the swimmer streamwise body abscissa for the range of standardized
in motion. To avoid complex difficulties in the body and real maximum average underwater swimming
geometry simulation the simplified theoretical model speeds (1.5–2.8 m/s). It is shown that the surface shear
considers the body swimmer to act like a fictitious stresses are strongly dependent on the location of the
parallelepipedal shape made of planar surfaces. To laminar-turbulent transition, i.e. the flow regime.
suggest anthropometric dimensions both the average The laminar to turbulent mathematical transition
body hydraulic perimeter P̄H and the swimmer’s height results in a drastic increase in the surface shear stress

G. Polidori et al. / Journal of Biomechanics 39 (2006) 2535–2541
SURFACE SHEAR STRESSτw (Pa)
xc (T, U∞)
17
LAMINAR
MODELING
13
2.8 m/s
9 TURBULENT
2 m/s
5 matic viscosity of water decreases as its temperature
1.5 m/s
1 leads to a strong increase in the local surface shear stress
0.01 0.10 1.00
STREAMWISE BODY ABSCISSA (m)
Fig. 3. Surface shear stress evolution versus the streamwise body
abscissa.
level. To minimize the friction drag it would be ideal
to keep a laminar boundary layer along the whole
body surface. However, it seems difficult to maintain a
laminar boundary layer with such complex flow
behavior where adverse pressure fields and inherent
layer separations occur. Now, especially focusing on the
turbulent regime, which is the state of the majority of
the flow around the swimmer, one can observe that
increasing the streamwise body abscissa leads to a
decrease in the local surface shear stress acting on the
swimmer’s body model. Moreover, whatever the swim-
ming speed may be, the increase of the film temperature
results in a decrease of the local surface shear stress.
This conclusion is important. Indeed, the lack of
maintenance of a laminar regime over the whole body
surface may result, a priori, in the inconvenience of an
earlier transition to the turbulent regime when increas-
ing the thermal gradient. However, this is rapidly
balanced by the reduction of the local surface shear
stress due to the increase of the film temperature.
To quantify the surface shear stress reduction for a
given temperature, T̄, by comparison with that at a
reference temperature, T̄ ref , and at a given swimming
speed, the following formula is deduced from Eq. (12)
and written as
   
tw ðT̄ ref Þ  tw ðT̄Þ T̄  T̄ ref
E t ð%Þ ¼ 100  ¼ 100 
tw ðT̄ ref Þ b  T̄ ref
(15)
( majority of the flow regime around the swimmer is
b ¼ 120 in the laminar region Rex o5  105 ;
with
b ¼ 247 in the turbulent region Rex 45  105 :
One will notice that b is a parameter that is deduced
from the present calculation and depends only on the
flow regime. In Eq. (15), b, T̄ and T̄ ref are given in
degrees celsius (1C).
ARTICLE IN PRESS
2539
For example, consider the increase of the average
T = 26.5°C
boundary layer temperature from T̄ ref ¼ 26:5  C to
T = 31.5°C
T̄ ¼ 31:5  C. Applying the expression (15) leads to an
estimated 5.3% local shear stress enhancement in the
laminar regime and a 1.5% in the turbulent regime.
4.2. Deduced skin-friction drag
MODELING In the considered film temperature range, the kine-
increases (Table 1). Therefore, the mathematical transi-
tion to turbulence occurs earlier for warmer water. This
10.00
(Fig. 3) for the warmer water that will only occur a little
further downstream for the colder water. As a con-
sequence, the benefits of a higher temperature gradient
are locally overcome by the disadvantages of the
turbulent regime, compared to the laminar regime, with
respect to the local surface shear stress. When the
transition to the turbulent regime has occurred for both
film temperatures, the benefits of the increase of the film
temperature are recovered for a sufficiently large
streamwise body abscissa.
To clarify and quantify the previous conclusion the
skin-friction drag performance has been expressed as a
function of the swimmer’s body length with stretched
arms. For this purpose, the low film temperature T̄ ref ¼
26:5  C was taken as the reference case. The skin-friction
drag performance was defined as:
 
F f ðT̄ ref Þ  F f ðT̄Þ
Performanceð%Þ ¼ 100  ,
F f ðT̄ ref Þ
(
Performance40 drag reduction;
) ð16Þ
Performanceo0 drag increase:
The skin-friction drag performance parameter, Perfor-
mance (%), is plotted against the streamwise body
abscissa in Fig. 4 for different swimming speeds varying
from 1 to 2.8 m/s.
It can be seen that the magnitude of the skin-friction
drag performance depends on the regime of the flow.
Indeed it appears that according to the flow regime, the
skin-friction drag performance is either swimmer speed
dependent in the case of the turbulent flow regime, or
not in the laminar case. Increasing the average boundary
layer temperature in the laminar flow regime results in a
better situation for skin-friction drag enhancement up to
5.3% whatever speed the swimmer has. However, the
actually turbulent. Therefore, turbulent regime analysis
is of great importance. By contrast to the laminar case,
the turbulent skin-friction drag performance is highly
Reynolds number or swimmer’s speed dependent.
The major conclusion was that at the beginning of the
turbulent regime, one observes a degradation in the
skin-friction drag performance due to the sudden and
 ARTICLE IN PRESS
2540 G. Polidori et al. / Journal of Biomechanics 39 (2006) 2535–2541

10 Table 3
Example of skin-friction drag evaluation for U 1 ¼2.8 m/s;
FRICTION DRAG PERFORMANCE (%)

ENHANCEMENT
(FRICTION DRAG REDUCTION) P̄H ¼0.81 m
5
2.8 m/s Water Film Friction drag Friction drag
0 temperature temperature (N) swimmer (N) swimmer
(1C) (1C) female male

-5 1 m/s 20 26.5 19.7 22.0

DEGRADATION
-10
(FRICTION DRAG INCREASE)
1.5 m/s
-15
2 m/s
-20
0 0.5 1 1.5
STREAMWISE BODY ABSCISSA (m)
Fig. 4. Friction drag performance when increasing the average
thermal boundary-layer from T̄ ref ¼ 26:5  C to T̄ ¼ 31:5  C.
sharp increase in the surface shear stress. Then, from a
given streamwise body abscissa the trend is reversed and
the performance is observed to reach 1.5% for fast
underwater speeds. This enhancement (i.e. reduction) in
the skin-friction drag occurs earlier as the speed is
increased. Comparisons with experimental results are
essential but difficult, since solutions for solving such
problems have not yet been developed.
4.3. Tentative dimensional analysis
A tentative evaluation is made from the male and
female anthropometric parameters given in the USA
swimming data (see for example the web site www.
usswim.org). The difference between male and female
swimmers is statistically significant. The average height
and weight for male swimmers are respectively
187.177.2 cm and 80.477.0 kg while the average height
and weight for female swimmers are 173.075.5 cm and
65.676.6 kg. From the following commonly used
formula (Mosteller, 1987) defining the body surface
area:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 Height ðcmÞ  WeightðkgÞ
Body Surface Area ðm Þ ¼
3600
corresponding body surface areas are respectively
found to be 2.04 m2 for males and 1.78 m2 for females.
To get some idea of the theoretical skin-friction drag,
the average swimmer hydraulic perimeter (P̄H ) is
defined as the ratio between the body surface area and
the L swimmer length with outstretched arms. The
average swimmer’s body length with stretched arms is
L ¼ 2.50 m for male swimmers and L ¼ 2.20 m for
female swimmers. The values for the average hydraulic
perimeter were found to be comparable for both female
30 31.5 19.4 21.6
and male swimmers, namelyP̄H ¼0.81 m. Examples of
friction drag values in steady flowing state are given in
2 2.5 Table 3 for a 2.8 m/s speed. Average friction drags of 20
and 22 N were found respectively for female and male
swimmers. Increase in pool temperature from 20 to
30 1C resulted in a decrease of the friction drag of
about 0.4 N.
Because the theoretical model used is a simplified one,
the present intention is only to provide the order of
magnitude of the skin-friction drag.
4.4. Comparison with Toussaint’s results
A thorough comparison with previous results cannot
be made because, to the authors’ knowledge, no
numerical solution or experimental data concerning
friction drag in underwater swimming are available in
literature.
Nevertheless a rough comparison can be attempted
with the measurements of Toussaint et al. (2004) in the
case of swimming at the air/water interface. In such a
case, they established the following relationship between
the total drag, Fd, and the swimming velocity, UN, for
velocities less than 2 m/s.
F d ¼ 21:33ðU 1 Þ2:34 . (18)
According to Toussaint et al. (2002), the friction drag
contribution is less than 5% of the total drag. At a speed
of 2 m/s, this gives: F f jToussaint
5:4 N.
Now let consider the present analysis and try to arrive
as close as possible to Toussaint’s experimental condi-
tions. Under the assumption that the friction drag
mainly occurs at the water/skin interface and that half of
the body is underwater so that P̄H ¼ ð0:81=2Þ m, the
(17)
present modeling gives, respectively F f jMale ¼ 5:8 N and
F f jFemale ¼ 5:2 N for male and female swimmers corre-
sponding to about 4–5% of the above mentioned
Toussaint’s total drag, which is in good agreement with
Toussaint’s results.
5. Conclusion
The main objective of this study was to evaluate and
theoretically quantify the effect of the temperature
 ARTICLE IN PRESS
G. Polidori et al. / Journal of Biomechanics 39 (2006) 2535–2541
gradient between the swimmer’s body and the pool
ambient water on the surface shear stress distribution
and deduced skin-friction drags, in a pool temperature
range varying from 20 to 301C. This was achieved using
the integral formalism modeling extended to laminar
and turbulent forced convection theory under the
assumption of a streamlined swimmer. The results of
this study were limited to steady flow. For different
swimming speeds, the surface shear stresses and the
skin-friction drags were investigated and presented in
both analytical and graphical forms. In order to check
the performance degree in friction drag with varying the
pool temperature both laminar and turbulent regimes
have been considered. The laminar flow regime provides
a better situation for friction drag enhancement up to
5.3% whatever speed the swimmer attains. This conclu-
sion is in close agreement with the common boundary
layer theory. By contrast, the turbulent skin-friction
drag performance is highly Reynolds number or
swimmer’s speed dependent. In turbulent and fast
swimming speed increasing the average boundary layer
temperature in the considered range gives a 1.5%
friction drag reduction for sufficient body lengths.
Additional parametric dimensional analysis gives aver-
age skin-drag values of about 20–22 N. Increasing the
pool temperature from 20 to 30 1C results in a decrease
of the friction drag of about 0.4 N.
It is not our intention and it seems difficult to say if
these skin-friction drag performances are significant or
not. Indeed relative weighting of both skin-friction drag
and profile drag is unknown and varying the thermal
field inevitably leads to altered boundary-layer separa-
tion points along the body and consequently the
turbulent wake size and the profile drag. A compromise
has to be taken between favoring either the laminar or
the turbulent regimes although a laminar boundary
layer is normally considered as ideal. When the flow
regime is laminar, separation at the body surface starts
almost as soon as the pressure gradient becomes adverse
and a larger wake forms while when the flow regime is
turbulent, separation is delayed and the corresponding
wake is smaller. The importance of keeping the
boundary layer attached to the swimmer body surface
2541
is so important that swimwear manufacturers sometimes
purposely cause the boundary layer to become turbu-
lent.
The authors hope that the work presented here be
used as an starting point for further discussion.
References
Bechert, D.W., Bruse, M., Hage, W., 2000. Experiments with three-
dimensional riblets as an idealized model of shark skin. Experi-
ments in Fluids 28, 403–412.
Brandt, R.A., Pichowsky, M.A., 1995. Conservation of energy in
competitive swimming. Journal of Biomechanics 28, 925–933.
Jansky, L., Vavra, V., Jansky, P., Kunc, P., Knizkova, I., Jandova, D.,
Slovacek, K., 2003. Skin temperature changes in human induced by
local peripheral cooling. Journal of Thermal Biology 28, 429–437.
Kakac- , S., Yener, Y., 1995. Convective Heat Transfer. CRC Press,
Boca Raton.
Koeltzsch, K., Dinkelacker, A., Grundmann, R., 2002. Flow over
convergent and divergent riblets. Experiments in Fluids 33,
346–350.
Laughlin, T., 2003. The pushoff—free distance, free speed. Where’s the
best place to be fishlike? Under water. Total swim online magazine,
November, www.totalimmersion.net/2003-archive-links.html
Mosteller, R.D., 1987. Simplified calculation of body surface area.
New England Journal of Medicine 317, 1098.
Padet, J., 1997. Principes des transferts convectifs. Polytechnica, Paris.
Polidori, G., Rebay, M., Padet, J., 1999. Retour sur les résultats de la
théorie de la convection forcée laminaire établie en écoulement de
couche limite externe 2D. International Journal of Thermal
Sciences 38, 398–409.
Polidori, G., Padet, J., 2002. Transient laminar forced convection with
arbitrary variation in the wall heat flux. Heat and Mass Transfer
38, 301–307.
Polidori, G., Popa, C., Mai, T.H., 2003. Transient flow rate behaviour
in an external natural convection boundary layer. Mechanics
Research Communications 30, 615–621.
Toussaint, H.M., Truijen, M., Elzinga, M.-J., van de Ven, A., de Best,
H., Snabel, B., de Groot, G., 2002. Effect of a fast-skin body suit
on drag during front crawl swimming. Sport Biomechanics 1, 1–10.
Toussaint, H.M., Roos, P.E., Kolmogorov, S., 2004. The determina-
tion of drag in front crawl swimming. Journal of Biomechanics 37,
1655–1663.
Vennell R., Pease, D., Wilson B., 2005. Wave drag on human
swimmers, Journal of Biomechanics, in press, doi:10.1016/
j.jbiomech.2005.01.023.
Yanai, T., 2001. Rotational effect of buoyancy in frontcrawl: does it
really cause the legs to sink? Journal of Biomechanics 34, 235–243.