Nahrain University, College of Engineering Journal (NUCEJ) Vol.10, No.1, 2007 pp.

53-64

Basim O. Hasan
Chemistry. Engineering Dept.- Nahrain
University

Turbulent Prandtl Number and its Use in Prediction of Heat Transfer

Coefficient for Liquids
Basim O. Hasan, Ph.D

Abstract: eddy conductivity is unspecified in the case of heat

transfer. The classical approach for obtaining the
A theoretical study is performed to determine the transport mechanism for the heat transfer problem
turbulent Prandtl number (Prt ) for liquids of wide follows the laminar approach, namely, the momentum
range of molecular Prandtl number (Pr=1 to 600) and thermal transport mechanisms are related by a
under turbulent flow conditions of Reynolds number factor, the Prandtl number, hence combining the
range 10000- 100000 by analysis of experimental molecular and eddy viscosities one obtain the
momentum and heat transfer data of other authors. A Boussinesq relation for shear stress:
semi empirical correlation for Prt is obtained and
employed to predict the heat transfer coefficient for
the investigated range of Re and molecular Prandtl du
( m ) 1
number (Pr). Also an expression for momentum eddy dy
diffusivity is developed. The results revealed that the
Prt is less than 0.7 and is function of both Re and Pr
according to the following relation: and the analogous relation for heat flux:
Prt=6.374Re-0.238 Pr-0.161
q dT
The capability of many previously proposed models of ( h ) 2
Prt in predicting the heat transfer coefficient is cp dy
examined. Cebeci [1973] model is found to give good The turbulent Prandtl number is the ratio between the
accuracy when used with the momentum eddy momentum and thermal eddy diffusivities, i.e., Prt= m/
diffusivity developed in the present analysis. The h. Thus Eq.(2) can be written as:
thickness of thermal sublayer decreases with
Reynolds number and molecular Prandtl number.
q dT
( m /Prt ) 3
Keywords: Turbulent Flow, Heat Transfer cp dy
Coefficient.

1. Introduction Thus if one knows the eddy diffusivity and turbulent

Prandtl number, Prt, the heat transfer problem can be
Despite many years of intensive research into solved [Simpson et. al. 1970]. A number of
turbulent diffusion, it is still poorly understood and can experimental and theoretical investigations have been
only be rather crudely predicted in many cases [Philip devoted to obtain the eddy diffusivity but these for
and Webester 2003]. Because of highly complex turbulent Prandtl number are less
turbulent flow mechanism, the prediction of the As the complex nature of the turbulent transport
transport rates necessarily involves the formulation of process is not yet well understood, the development of
conceptual models which embody many simplifying models for the prediction of Prt requires the
assumptions [Gutfinger 1975]. In the momentum introduction of many tenuous assumptions regarding
problem the eddy viscosity remains unknown while the the behavior of turbulence. The validity of these

NUCEJ Vol.10, NO.1 Hasan 53

assumptions may be verified only indirectly by have concluded, on the basis of their own results, that
comparing the predicated Prt values with the measured Prt is varyingly affected by the molecular Prandtl
ones. There are basically two routes by which Prt may number, the flow Reynolds number, and the distance
be evaluated experimentally [Gutfinger 1975]: from the wall. The suggested trends of Prt with these
1. Utilizing experimental time-average velocity and parameters for the case of liquid metals (Pr<0.1) differ
temperature (or concentration) profiles together with from those for ordinary fluids (Pr>0.7). For liquid
the integrated Reynolds equations. From equations (1) metals Prt is greater than unity and decreases with
and (2) increasing Reynolds number and the distance away
from the wall (Carr and Balzhiser, 1967). This effect is
du opposite to that for air (Sleicher, 1958), where Prt is
/ 1
less than one, and increases towards unity with
dy
Pr t increase in Re and distance from the wall. Although the
q dT 4 above trends have provided a basis for the formulation
/ 1
C p dy of models for the prediction of Prt to be considered in
predicting heat transfer coefficient, the picture is by no
means conclusive, and many studies have produced
Thus, if the momentum and heat flux variation with the results which exhibit a different behavior. The value of
distance from the wall are known, m, h , and hence Pr=0.7 to 1 is of practical importance because it
Prt can be evaluated. represents most gases. Also the value of Pr=7
2. Direct measurement of the Reynolds transport terms represents water and light liquids near room
temperature [Kays 1993]. Quarmby and Quirk (1974)
( u' ' , T' ' ) and use of the definition of the eddy
have concluded, on the basis of an extensive study of
diffusivity: turbulent flow in a plain circular tube with Prandtl
t du numbers varying from 0.7 to 1200 and Reynolds
u' ' m numbers ranging from 5230 to 23550, that Prt is a
dy simple function of the nondimensional distance from
qt dT the wall, independent of both the Prandtl number and
T' '
h dy Reynolds number. They found that Prt varies smoothly
from 0.5 at the wall to unity at the tube center.
Hence,
Brienkworth and Smith (1969) (Pr=6) and Eckelman
u' ' T' ' and Hanratty (1972) (Pr=0.7) suggest on the other
m and h hand, that Prt is constant and equal to approximately
du / dy dT/dy unity over the whole of the boundary layer. Yet a
different trend was reported by Simpson et al. (1970)
who found that for boundary layer flows of air
or Prt u' ' dT dy (Pr=0.77), the local value of Prt near the wall is
T' ' dy du greater than unity, displaying a maximum value of
Thus, Prt can be evaluated from the experimentally approximately 1.4 at the wall and decreasing to
determined velocity and temperature fields, and the approximately unity in the outer region.
Shlanchyauskas et al. (1974) for Pr>1 found that Prt is
turbulent fluctuating terms u' '
and T' ' . constant and equal to 0.75. In view of the contradictory
The characteristic feature of all the reported experimental data it is significant to note the analysis
experimentally determined Prt is the great scatter of the of Deissler (1963) based on the statistical nature of the
data. The results of different investigations lack in turbulent process. Deissler suggests that Prt approaches
agreement and are often conflicting. This situation unity as the velocity gradient increases, regardless of
arises directly from the necessity to differentiate the the molecular Prandtl number of the fluid. Deissler also
two measured profiles and determine the shear stress suggests that the effect of Pr is much greater at low
and heat flux profiles for the calculation of Prt. The values of Pr (i.e., liquid metals) than at higher ones.
differentiation procedure greatly amplifies the This is consistent with the results of both Sliecher
uncertainties associated with the experimentally (1958) and Quarmby and Quirk (1974). Miyak [1992]
measured velocity and temperature profile. These and Gurniki et. al [2000] stated that the Prt is between
difficulties arise from the inaccessibility of probes to 0.33 and 0.5. Aravinth [2000], developed a resistance-
the thin region adjacent to the wall in which the in-series model to quantitatively predict the heat and
velocity and temperature gradients concentrate, mass transfer processes for turbulent fluid flow through
particularly at high molecular Pr [Gutfinger 1975]. tubes and circular conduits under uniform wall
Experimentally determined turbulent prandtl numbers temperature condition. Kays [1994] examined the
have been surveyed by Kestin and Richardson (1963) available experimental data on Prt for the two
and then by Blom and de Veres (1968). The results of dimension boundary layer in circular tube and flat
surveys together with the more recent data of Bolm plate. Churchill [2002] analyzed the viscosity and eddy
(1970), Simpson et al. (1970), Sleicher et al. (1973), conductivity in fully developed turbulent pipe flow and
and Quarmby and Quirk (1969,1972,1974) and redefined the Prt directly in terms of the time-averaged
Gutfinger (1975) are summarized in Table 1. Although fluctuations and stated that it remains an essential
the results are greatly different, various investigators parametric variable. Toorman (2003) showed that Prt

NUCEJ Vol.10, NO.1 Turbulent Prandtl Number 54

vary from 1 to 10 in the near wall region and fluctuates satisfactory empirical theory. This in turn has
around 1 in the fully turbulent region. Crimaldi et al. necessitated the formulation of semi-empirical theories
[2006] measured the distribution of Prt in the in an attempt to both rationalize the existing data and to
laboratory boundary layer and developed an analytical provide a starting point for heat and mass transfer
model for Prt and found that it is significantly larger calculations. In the present work it is aimed to obtain a
than unity even at large distance from the wall. model for average Prt from available experimental data
The above comments are sufficient to indicate the for wide range of Reynolds and molecular Prandtl
general inconsistencies and lack of agreement which numbers. Also it is aimed to examine the capability of
are characteristic of much of the published work. These many previously proposed models for Prt to predict
inconsistencies, and low accuracy of the available heat transfer coefficient by comparing them with the
measurement, have hindered the development of a experimentally determined heat transfer coefficient.

Table (2) Models for Turbulent PRANDTL Number for Various Authors.

Equation for Prt Author

1 Diessler (1952)
n Re Pr 1 exp n Re Pr
Prt
n=experimental constant=0.000153
1 1 135 Re 0.45 exp 0 .25 Azer and Chao (1961)
0.6<Pr<15
Prt 2 57 Re 0.46 Pr 0.58 exp 0 .25

y
, R=pipe radius
R
1 =0.014Re0.45Pr0.2[1-exp(-1/{0.014 Re0.45Pr0.2}] Aoki(1963)
Prt
Prt =1.0 low turbulence intensity Marchello and Toor (1963)
Prt= Pr high turbulence intensity
2 6 Pr Tyldesley and Silver
Pr t low turbulence intensity (1968)
9 Pr
2 9 Pr
Prt high turbulence intensity
3 9 Pr
1 Graber (1970)
=0.91+0.13Pr0.545
Prt
0.7<Pr<100
f Thomas and Rajagopal
exp y 1 (1973)
2
Pr t Pr
f
exp Pr y 1
2
1 exp( y / A ) Cebeci (1973)
Prt
1 exp( y / B )
A+ is constant =26
1 5
B Ci (log(Pr))i 1
Pr 1
C1=34.96, C2=28.97,C3=33.95, C4=6.33, C5=-1.186
Rosen and Tragrdh (1995)
0.0014{1 exp( Re1 / 2 / 2)}
Prt
0.00124 Pr 0.112

2. Theoretical Aspects The shear stress due to molecular and turbulent

transport of momentum is given by Eq.(1)

NUCEJ Vol.10, NO.1 Hasan 55

 + *
since u =u/u , y =yu / , and
+ * w
=u* y
y (1 )
2 1 R dy 10
* f
Eq.(1) becomes, (1 m ) d(u u ) 0 1 m
u* yu
*
d( ) Since;
qw= Cp( dT
) 11
h dy
or (1 m ) du ; hence:
u *2 dy
h) dT
m ) du qw= Cp( 12
(1 5 d(y / )
w dy
Hence:
r dT
Since where R=radius and
qw= Cpu*(1/Pr+ h
) 13
w R dy
r=distance from the center=R-y; hence
R y R y y
w R R
(1
R
) 6 y dy u* Cp T
dT 14
Eq.(5) becomes: 0 1/Pr h /v q w Tw
y m ) du hence:
(1 ) (1 7
R y
dy dy
*
(T-TW )=(qw/u Cp) 15
0
1/ Pr h / v
f
where R+=(Ru*/ )= Re Eq.(15) gives the temperature at any point (y+). If the
8 temperature profile is written in dimensionless form,
thus: i.e.:
y T+ =(T-Tw)/(Tb-Tw)
(1 )
du R dy 8
If Eq.(14) is integrated from the wall(y+=0 and T=Tw)
1 m to the center of turbulent core (y+ =R+ and T=Tb), one
obtain
Equation (8) can be integrated to the limits that at R
y+=0, u+=0 and at y+=y1+, u+=ub. Where u+b is the *
dy
dimensionless average bulk fluid velocity(ub/u*) and (Tb-Tw )=(qw/u Cp) 16
y1+ is the distance from the wall beyond which the 0
1/ Pr h / v
velocity becomes equal to ub. Therefore Eq. (8)
becomes dividing Eq. (15) by Eq. (16) gives:
y
y (1 )
1 R y
u dy 9 dy
b
0 1 m T TW 1 / Pr h /v
+ 0
T= R 17
u Tb TW dy
since u
ub b 2 , 1 / Pr h /v
b u* f f 0
u
b 2
Since Q=hA(Tb TW )
substitution in Eq. (9) yields h=q/(Tb-Tw)
Sub. in Eq.(16)
R
dy
u* Cp/h=
0
1/ Pr h / v
Since Nu=hd/k

NUCEJ Vol.10, NO.1 Turbulent Prandtl Number 56

 R c- Near turbulent core region:
dy The large eddies in the core region and the small
u* Cpd/Nu k= 18 variation in turbulent intensity in the central region
0
1/ Pr h / v makes the eddy viscosity constant. Therefore in the
present analysis the momentum eddy diffusivity for
u*=ub f / 2 =(Re / d) f /2 central region will be considered constant as did by
hence from Eq.(18): Mizushina et. al. [1971] and Hinze [1975 ]:
Nu= f/2RePr
R dy 19 m / =0.07 R+ y+>y2+ 21c
0 1/Pr h
/
Since Prt= m/ h, thus Eq.(19) becomes: Now, for predicting momentum eddy diffusivity
expression near the wall, Eq. (21a) is inserted in
Eq.(10) and the experimental value of friction factor is
f/2RePr substituted in Eq.(10). The well known Blasius
Nu correlation for friction factor will be adopted, i.e.:
R dy
20
0
1/Pr ( m ) / Pr f=0.079Re-0.25 22
t
Hence Eq.(10) becomes
Accordingly, if the variation of m/ with y+ and Prt
are known, the value of Nu can be estimated from y
Eq.(20). y (1 )
1 R dy 23
5.032 Re0.5
3
3. New Momentum Eddy Diffusivity Model 0 1 Ay
The eddy diffusivity behavior in the viscous sublayer, The upper limit of integration in Eq.(23) is obtained by
damped turbulence layer, and turbulent core affect equating equations (21a) and (21b), i.e.:
greatly the rate of heat (or mass) transport between the A y1+3= 0.45 y1+
wall and bulk. Previous studies [Von Karman 1939, 0.45
Lin et. al. 1953, Deissler 1952, Wasan and Wilk1964, or y1
Rosen and Tragradh 1995, Meignen and Berthoud
A
1998, Lam 1988, Gurniki et. al 2000, and Wang and Thus Eq. (23) becomes
Nesic 2003] showed that the eddy diffusivity is y
function of many variables such as the distance from (0.45/A)
0.5 (1 )
the wall (y+), Re, and Pr. Most studies [Wasan and 5.032 Re0.5 R dy 24
Wilke 1964, Townsend 1961, Hughmark 1969, Shaw 0 1 Ay 3
and Hanraty 1964, Escobedo et. al. 1995, Papavassiliou
Substitution of Re in Eq.(24) and performing the
1997] showed that in the near wall region (the region
integration numerically the value of A can be obtained
of major importance in the property transport) m /
by trial and error for each value of Re. Fig. (1) shows
proportional to y+3, i.e, m / =A y+3.
a plot of A vs Re as obtained from Eq.(24). From best
In the present analysis the region of interest is divided
fit method:
into three zones for momentum eddy diffusivity
variation:
a- Near wall region:
A=0.0064Re-0.322 25a
m / =A y+3 0<y+<y1+ 21a
Substitution of A in Eq.(21a) gives the expression of
where y1+ is the distance beyond which the eddy
eddy diffusivity near the wall.
diffusivity becomes linear with y+ rather than y+3
Therefore,
variation.
b- Region with linear variation of eddy diffusivity: 0.45
Many studies [Sleicher et al.1958, Levich 1962, y1+= =8.4Re0.161 25b
A
Mizushina et. al 1971] showed that in the transition
region the eddy diffusivity vary linearly with y+, i.e.,
y2+ is obtained from equating Eqs. (21b) and Eq.(21c),
i.e.,
m / =B y+ y1+< y+ <y2+ 21b
Best fitting of experimental results for m / obtained
by Sleicher et. al. [1958] indicated that outside the y2+=0.156R+ 25c
viscous sublayer B=0.45. Levich [1962] showed that
B=0.4.

NUCEJ Vol.10, NO.1 Hasan 57

Now to investigate whether Prt is unity over the whole substituted in Eq.(19), i.e., h= m (or Prt=1) to
range of Re and Pr the expressions of momentum eddy estimate Nu and compare with experimental Nu. Thus
diffusivity developed in Eqs. (21a, 21b, 21c) is Eq. (19) becomes:

f/2RePr
Nu
y1 y2 R
dy dy dy 26
0 1/Pr 0.0064 Re 0.322 y 3 y 1 1/Pr 0.45 y y2 1/Pr 0.07 R

Insertion of Re and Pr with f from Eq.(22) and , i.e., h and m are not equal. They also indicate that
performing the integration numerically, Nu can be at a particular Pr, the higher the Re is the higher the
obtained. The results are plotted in Fig. 2 as compared difference between Nu from Eqs.(26) and (27). Also at
with the experimental Nu of Friend and Mitzner a particular Re the higher the Pr is the higher the
[1958]. Friend and Mitzner performed an experimental difference.
heat transfer study for wide range of Pr and obtain the
following relation for 0.5 > Pr> 800: 4. Examination of the capability of
proposed Prt models to predict nu value
f
Re Pr Figure 2 indicates that Prt is not unity. To
Nu 2 27
1/ 3
investigate which models of Prt presented in Table 1
1.2 11.8 f / 2 (Pr 1) Pr gives accurate results, these models are substituted into
equation (20) with m/ taken from Eqs. (21). Thus
Fig. 2 reveals that the Nu obtained from Eq.(26) is Eq.(20) becomes:
different from that obtained from experimental results
indicating that the turbulent Prandtl number is not unity

f/2RePr
Nu
y1 y2 R
dy dy dy 28
0 1/Pr (0.0064Re 0.322 y 3 / Pr ) y 1 1/Pr (0.45y / Pr ) y2 1/Pr (0.07R / Pr )
t t t

Now inserting of various models from Table 1 in Eq. prevent confusion) with experimental Nu of Friend and
(28) and performing the integration numerically, with Mitzner (Eq. (27)). The figure shows that most models
y1+ and y2+ from Eqs.(25b) and (25c) and f from Eq. exhibit considerable deviation. The model of Cebeci
(22), the Nu value can be obtained for each value of Re [1973] exhibits good agreement for high Pr values.
and Pr. Fig. 3 shows comparison between Nu predicted
from various models (some models are avoided to

NUCEJ Vol.10, NO.1 Turbulent Prandtl Number 58

 3

Pr=7
2
Azer and Chao
Deissler
Cebeci
1000
9
Friend and Mitzner (Exp)
8
7 Tyldesley and Silver
6
5

Nu 3

100
9
8
7
6
5

8 9 2 3 4 5 6 7 8 9
10000 100000
Re
Figure (3) Comparison of Nu Obtained from Various Prt Models and Experimental Nu for Pr=7.

5. More accurate model for turbulent where y+ varies from 0 to y1+. By performing the
PRANDTL number integral u+ is obtained for each y+. Fig. 4 shows the
velocity profile at various Re as compared with other
To obtain more accurate expression for Prt from authors. The figure shows that the velocity profile
experimental results of Friend and Mitzner, Nu is obtained from present analysis is in good agreement
estimated from Eq. (27) of Friend and Mitzner and with previous experimental and theoretical studies.
substituted in Eq. (28). Performing the integration of Also the figure reveals that very close to the wall the
the denominator numerically, Prt is obtained by trial velocity profile is linear and Re has no effect on u+.
and error for each value of Re and Pr. In present
analysis the variation of Prt with the radial distance is 30
ignored, i.e., the average turbulent Prandtl number is
determined. Using statistical method the following Present Analysis [Re=10000]
relation is obtained: Wasan and Wilke
Sleicher et al (exp)

20 Von Karman
Prt=6.374Re-0.238 Pr-0.161
29
C.C=0.98
u+

The results shows that for liquids of Pr > 1 the Prt is

less than one indicating that the thermal eddy diffusion 10

is larger than momentum eddy diffusion. Also Prt

decreases with increasing Re and Pr indicating that the
increase in thermal eddy diffusion is larger than that in
momentum eddy diffusion. Many studies [Deissler
0
1952, Jenkis 1951, Azer and Chao 1961] showed that
for liquid metals (Pr< 0.1) the Prt is higher than one.
Also it is to be mentioned that Miyak [1992] and
Gurniki et. al [2000] found that the Prt is between 0.33 0 10 20 30 40 50
and 0.5. y+

Figure (4) Velocity Profile as Compared with other

6. Velocity Profile in the Near Wall Region Workers
The velocity profile in the near wall region is obtained
by substituting Eq. (21a) into Eq.(8) and integrating 7. Temperature Profile
both sides, i.e.:
To obtain temperature profile Eq.(17) is used with h=
m/Prt and Prt from Eq. (29) and the expression of
y m/ from Eqs.(21) for each region of y+. Thus Eq.
y (1 )
u R dy 30 (17) becomes:
0.322 y 3
0 1 0.0064 Re

NUCEJ Vol.10, NO.1 Hasan 59

 y
dy
0 .238 0 .161
0 1 / Pr ( m / v ) / 6 .374 Re Pr
T
y1 y2 R 31
dy dy dy
0 0 .0064 Re 0 .322 y 3 y 1 0 .45 y y2 0 .07 R
1/Pr 1/Pr 1/Pr
6 .374 Re 0 .238 Pr 0 .161 6 .374 Re 0 .238 Pr 0 .161 6 .374 Re 0 .238 Pr 0 .161

The denominator of Eq.(31) is constant because the

limits of integration are constants while the integral in 1.0

the numerator varies with y+, i.e., for each value of y+

there is a value of numerator and consequently a value
of T+. The expression of m/ in the numerator is taken 0.8

from Eqs.(21a) according to y+, the limit of integral. Re=40000

Pr=600
The values of T+ that are obtained from Eq. (31) are Pr=100
0.6
shown graphically in Fig. 5 for Re=40000. The same

T+
Pr=40

trend in Fig. 5 is for other values of Re (10000 and Pr=20

Pr=7
10000). The figure shows the effect of Pr on 0.4
temperature profile. In a turbulent boundary layer, the
gradient is very steep near the wall and weaker farther
from the wall where the eddies are larger and turbulent 0.2
mixing is more efficient [Lienhard 2001]. As the Pr
increases the temperature profile close to the wall
becomes more flat (the slope increases) indicating that 0.0

the dimensionless thickness of the of thermal sublayer 0 5 10 15 20

y+
( +T) decreases. Fig. 6 reveals that increasing Re
leads to slight increase in +T. Using statistical Figure (5) Temperature Profile Obtained from the
methods the following relation is obtained for +T: Models of Eddy Diffusivity and Prt of Present
Analysis at Various Pr and Re=40000.

+ T=8.635Re-0.067Pr-0.245; C.C=0.99 32
1.0

or
0.8

2
T/d=8.635 Re-1.067Pr-0.245 33
f
Pr=7
0.6
Re=10000

Hence the thickness of thermal sublayer decreases with

T+

Re=40000
Re=100000
Re and Pr. Levich [1962] proposed the following
0.4
relation for the ratio of viscous sublayer to thermal
sublayer:
0.2

b Pr1/3 34
T 0.0

0 5 10 15 20
Hence the thickness of viscous sublayer is y+
Figure (6) Temperature Profile Obtained from the
Models of Eddy Diffusivity and Prt of Present
2
b/d =8.635 Re-1.067Pr0.088 35 Analysis at Various Re and Pr=7.
f

Predicting Heat Transfer Coefficient

The expressions of momentum eddy
diffusivity and turbulent Prandtl number developed in,
Eqs. (21) and Eq. (29) respectively are substituted in
Eq.(20) to obtain Nu. Figs. 7 to 9 show a comparison
between Nu predicted from present analysis with other
experimental works. It is evident that Nu predicted

NUCEJ Vol.10, NO.1 Turbulent Prandtl Number 60

from present analysis is in good agreement with other 5

experimental works for the whole range of Re and Pr. 4

Pr=600

3 Present Analysis
7 Colburn

6 Petkhov
Pr=7 2
Notter and Sleicher
5
Present Analysis

4 Allen and Eckert

Slaiman and Hasan

Nu
3 Petkhov 1000
9
8
7
Nu

2 6

3
100.00
9

8
2
7 8 9 2 3 4 5 6 7 8 9
10000 100000
6

8 9 2 3 4 5 6 7 8 9
Re
10000 100000
Re
Figure (9) Comparison between Nu Obtained from
Present Analysis with Other Works at Pr=600..
Figure (7) Comparison between Nu Obtained from
Present Analysis with Other Works at Pr=7.
CONCLUSIONS

3
The following points are concluded for the
investigated range of Re and Pr:
2
Pr=100
1- Turbulent Prandtl number plays an important role in
present analysis
determining the value of heat transfer coefficient and
Colburn the assumption of unity turbulent Prandtl number is
Petkhov very poor for fluids of Pr>1, i.e., the momentum eddy
1000 Notter and Sleicher
9 diffusivity is not equal to thermal eddy diffusivity.
8
7
2- The Turbulent Prandtl number depends on both
6 Reynolds and molecular Prandtl numbers and its value
Nu

5
is generally less than 0.7.
4 3- The models of turbulent Prandtl number and
3
momentum eddy diffusivity developed in present
analysis give more accurate results in predicting the
2
heat transfer coefficient than previously proposed
models.
4- The thickness of thermal sublayer is function of Re
and Pr while the thickness of viscous sublayer is
100
8 9 2 3 4 5 6 7 8 9
strongly affected by Re and slightly by Pr.
10000 100000
Re
Figure (8) Comparison between Nu Obtained from
Present Analysis with Other Works at Pr=100. REFERENCES

1. Allen R. W. and E. R. G. Echert,( 1969), "Friction

and heat transfer measurements to turbulent pipe
flow of water" J. Heat Transfer, August, P.301.
2. Aoki S. 1963 "Consideration on the heat transfer
in liquid metals", Bull, Tokyo, Inst. Tech., 54, 63-
73.
3. Aravinth, S., 2000, Prediction of heat and mass
transfer for fully developed turbulent Fluid Flow
through tubes, International Journal of Heat and
Mass Transfer, (43), pp.1399-1408
4. Azer N. Z., and B. T. Chao, 1961, Turbulent Heat
Transfer in Liquid Metals- Fully Developed Pipe
Flow with Constant Wall Temperature, Int. J.
Heat Mass Transfer 3,77-83.
5. Blom, J., (1970) An Experimental Determination
of the Turbulent Prandtl Number in a Developing

NUCEJ Vol.10, NO.1 Hasan 61

 Temperature Boundary Layer, presented at The Technology, Israel, Hemisphere Publishing
Fourth International Heat Transfer Conf., Paris, Corporation, John Wiley and Sons, Washington,.
Versailles. 22. Hinze J.O., 1975 Turbulance Phenomena, 2nd
6. Blom, J., and deVries, D. A., (1968) On the Edition., McGraw-Hill, USA.
Values of the Turbulent Prandtl Number, Third 23. Hasan, 2003, « Heat, Mass, and Momentum
All Union Heat and Mass Transfer Conference, Analogies to Estimate Corrosion Rate under
Paper No. 1.8, Minsk. Turbulent Flow Conditions », Ph.D. Thesis,
7. Brinkworth, B. T., and Smith, P. C., (1969) Chem.Eng. Dept, Nahrain University, Baghdad,
Velocity Distribution in the Core of Turbulent 2003.
Pipe Flow, Chem. Eng. Sci., 24, 787. 24. Jenkis R., (1951) Variation of Eddy Conductivity
8. Carr, A. D., and Balzhiser, R. E., (1967) with Prandtl Modulus and it is Use in Prediction
Temperature Profiles and Eddy Diffusivities in of turbulent heat Transfer Coefficient, Heat
Liquid Metals, Birt. Chem. Eng. 12, 53. Transfer and Fluid Mechanics Institute , Stanford
9. Cebeci, T., (1973) A Model for Eddy Univ. Press., pp. 147-158.
Conductivity and Turbulent Prandtl Number, 25. Kays M., Convective Heat and Mass Transfer,
ASME J. Heat Mass Transfer, 95, 227. 3nd Edition, McGraw-Hill, United State, 1993
10. Churchill S. W., (2002), "A Reinterpretation of 26. Kestin, J., and Rechardson, P. D., (1967) Heat
the Turbulent Prandtl Number", Ind. Eng. Chem. Transfer Across Turbulent, Incompressible
Res., 41 (25), 6393-6401, 2002. Boundary Layers, Int. J. Heat Mass Transfer, 6,
11. Crimaldi J. P., J.R. Koseff, and S. G. Monismith, 147(1967).
(2006)" A mixing-length formulation for the 27. Lam C.Y., (1988) An Eddy Viscosity Turbulence
turbulent Prandtl number in wall-bounded flows Model, Int. Com. Heat Mass Transfer, Vol.15, pp.
with bed roughness and elevated scalar sources"; 355-363.
Physics of Fluids, 18, 095102, September. 28. Levich W.G., (1962) Physicchemical
12. Danckwerts, P.V., (1951) "Significance of Liquid Hydrodynamis, New Jersey, Prentice Hall,
Film Coefficient in Gas Absorption"; Ind. And pp.689,1962.
Eng. Chem., 43, 1460. 29. Lienhard J. H., (2001), "Heat Transfer Text
13. Deissler R. G., 1952 "Analysis of Fully Book", 3rd Ed., Pholgiston, Mech. Eng. Dep.,
Developed Turbulent Heat Transfer at Low Peclet Houston, U.S.A.
Number in Smooth Tubes with Application to 30. Lin, C.S., R.W Moulton, and G. L. Putnam,
Liquid Metals, NACA RM E 52F05. (1953) "Mass Transfer between Solid Wall and
14. Deissler, G., (1955), "Analysis of tubulent heat Fluid Streams", Ind. Eng. Chem. 43,636.
transfer, mass transfer and friction in smooth 31. Marchello J.M. and H. L. Toor, (1963), "A
tubes at high Pranadtl and Schmidt numbers", Mixing Model for Transfer Near a Boundary",
NASA technical report Ind. Eng. Chem. Fund., 2, 1, 8.
15. Deissler, R. G., .(1952). Turbulent Heat Transfer 32. Meignen R. and G. Berthoud, 1998 Int. J. Heat
and Temperature Fluctuations in a Field with Mass Transfer 41, 3373.
Uniform Velocity and Temperature Gradients, 33. Miyake Y., (1992), Computational Fluid
Int. J. Heat Mass Transfer, 6, 257 Dynamics, Ed. M. Yasuhara and H. Daiguji,
16. Eckelman,L. D., and Hanratty , T. J., (1972) Chap. 10, Univ.Tokyo Press, pp.223 .
Interpretation of Measured Variations of the 34. Mizushina T. and F. Ogino, (1970) "Eddy
Eddy Conductivity, Int. J. Heat Mass Transfer, viscosity and universal Velocity Profile in
15, 2231. Turbulent Flow in a Straight Pipe" J. of Chem.
17. Escobedo J. and G.A. Mansoori and U. L. Eng. of Japan, 3, 166.
Chicago, (1995) Solid Particle Deposition 35. Mizushina T., F. Ogino, Y. Oka and H. Fukuda,
During Turbulent Flow Production Operations, (1971) Int. J. Heat Mass Trans.,Vol. 14, P. 1705.
Society of Petroleum Eng., SPE 29488, 36. Notter, R.H. and C. A. Sleicher (1972)" A
Oklahama, April. solution of Turbulent Graetz Problem", Chem
18. Friend, W.L. and A. B. Metzner, , (1958) AICHE. Eng. Sci., 27,2073-2093.
J., Vol. 4, No. 4, P. 393,. 37. Papavassiliou D.V. and J. Thomas Hanratty,
19. Graber H., 1970, Der Warmeubergang in Glatten (1997)Int. J. Heat Mass Transfer 40, pp. 1303,.
Rohren, zwischen parallelen, in Ringaspalten and 38. Petukhov B. S., (1970) Heat Transfer and Friction
langs Rohrbudlen bei exponentiller in Turbulent Pipe Flow with Variable Physical
Warmefluissverteiiung in erzwungener laminarer Properties, Advances in Heat Trans., Vol. 6, P.
oder turbulenter Stromung, Int. J. Heat Mass 503,.
Transfer, 13, 1645-1703. 39. Philip J.W. and D. R. Webster, (2003)" Turbulent
20. Gurniki F,, K. Fukagata, S. Zahrai, and F.H. Bark, Diffusion ", School of Civil and Environmental
(2000)" LES of turbulent Channel Flow of Binary Eng., Georgia Institute of Technology, Atlanta,
Electrolyte", J. Appl. Electrochem., 30, 1335- Georgia,30332-0355..
1343. 40. Quarmby, A. and R. K. Annad, (1969),
21. Gutfinger G., 1975 "Topics in Transport "Axisymmetric Turbulent Mass Transfer in a
Phenomena", Mech. Eng. Dept., Israel Institute of Circular Tube", J. Fluid Mechanics, 38, (3), 433.

NUCEJ Vol.10, NO.1 Turbulent Prandtl Number 62

41. Quarmby, A., and Quirk, R. K., (1974) 54. Von Karman T., (1939) "Mechanishe Ahnlichkeit
Aximmetric and Non Aximmetric Turbulent in und Turbulenz" Proc. 3rd. Int. Congress Apll.
a Plain Circular Tube at High Schmidt Mech. Stockholm 1, 85.
Number, Int. J. Heat Mass Transfer, 17,143. 55. Wang S. and S. Nesic (2003) " On Coupling Co2
42. Quarmby, A., and R. K Quirk, (1972), Corrosion and Multiphse Flow Model", Corrosion
"Measurement of Radial and Tangential Eddy J., Paper No. 03631, P.1.
Diffusivity of Heat and Mass Transfer in 56. Wasan, D.T and C. R. Wike, (1964), "Turbulent
Turbulent Flow in a Plan Tube" Int. J. Heat Mass exchange of momentum, mass and heat between
Transfer, 15, 2309. fluid streams and pipe wall", Int. J. Heat Mass
43. Richrdit, H. NACA TM, 1047, 1943 Trans.,Vol. 7, P. 87.
44. Rosen C. and C. Tragradh, (1995), "Prediction of
turbulent high Schmidt number mass transfer NOMENCLATURE
using a low Reynolds number k e turbulence
model, Chem Eng. J., 59, 153-159.
45. Sherwood T. K., K. A. Smith, and P.E. Fowles, A Area, m2
(1968) The Velocity and Eddy Viscosity Cp Specific heat, kJ/kg.K
Distribution in the wall Region of Turbulent Pipe d Diameter
Flow"; Chem. Eng. Sci., Vol.23, P. 1225-1236,. f Fanning friction factor
46. Shlanchyauskas, A. A., P. P.Vaitekunas, and A. h Heat transfer coefficient, W/m2. C
A. Zhukauskas, (1974), Calculations of heat Nu Nusselt number, hd/k
Transfer in Turbulent Boundary Layer of Viscous Pr Prandtl number
liquids and an Evaluation pf the Effect of Prt", Prt Turbulent Prandtl Number
Int. Chem. Eng., 14, (1), 33. qw Heat Flux at the wall (W/m2)
47. Simpson, R. L, D.G. Whitten, R. J. Moffat, qt Turbulent heat Flux at the wall (W/m2)
(1970), "An experimental study of the turbulent R+ Dimensionless radius
Prandtl number of air with injuction and suction", Re Reynolds number, du/
Int. J. of Heat and Mass Transfer, Vol.13, pp. 125. T Temperature
48. Sleicher C. A, A. S. Awad, R. H. Notter, (1973), " T+ Dimensionless temperature
Temperature and Eddy Diffusivity Profile in T' Fluctuating temperature
NaK", Int. J. Heat Mass transfer, 16, 1565. u Local velocity
49. Sleicher C.A., J.R. Emeryville, and Calif, (1958) u* Friction velocity, m/s
Experimental Velocity and Temperature Profile y+ Dimensionless distance from the wall
for Air in Turbulent Pipe Flow, Trans. ASME, pp. R Radius
693, April. y Distance from the wall
50. Thomas L. C., and R. Rajagopal, (1974) u' Fluctuating velocity in axial direction
Adaptation of Stochastic Formulation of the v' Fluctuating velocity in redial direction
Surface Rejuuvenation Model to Turbulent
Convection Heat Transfer, Chem. Eng. Sci. 29, GREEK LETTERS
1639.
51. Toorman, E.A., (2003), "Validation of Molecular thermal diffusivity, (m2/s).
macroscopic modeling of particle-laden turbulent T Thermal layer thickness, m.
flows", Proceedings 6th Belgian National b Viscous sublayer thickness, m.
Congress on Theoretical and Applied m Eddy diffusivity for momentum
Mechanics,Gent, 26-27 May. transfer, m2/s.
52. Townsend A.A ,( 1961) J. Fluid Mech., Vol. 11, h Eddy diffusivity for heat transfer,
P. 97. m2/s.
53. Tyldesley J.R. and R. S. Silver, (1968) The
Prediction of Transport Properties of a Turbulent
Flow, Int. J. Heat Mass Trans., 11,1325..

Kinematic Viscosity m2/s

Shear Stress (N/m2)
Density (kg/m3)
Ratio of the distance from the wall to pipe radious

SUBSCRIPTS
b Bulk
m Momentum
h Heat
w Wall
t Turbulent

NUCEJ Vol.10, NO.1 Hasan 63

 Prt
10000 -100000
Re Prt
Pr
Pr Re 0.7 Prt

Prt 6.374 Re 0.238 Pr 0.161

Prt
Cebeci 1973

NUCEJ Vol.10, NO.1 Turbulent Prandtl Number 64

This document was created with Win2PDF available at http://www.daneprairie.com.
The unregistered version of Win2PDF is for evaluation or non-commercial use only.