Professional Documents
Culture Documents
0.5 d
tmax
d
t. min
1
Since 10% attening results in very small 0.3% reduction in cross-sectional area; 8% attening in
the present case is expected to have insignicant eect on pressure drop.
3. Results and discussion
Pressure drop and heat transfer results for the layouts tested in water-to-water heat transfer are
shown in Figs. 4 and 5.
3.1. Pressure drop
Fig. 4 represents variation in experimental pressure drop with mass ow rate on hot water side
for all six modules of TTHE. In serpentine layouts, the pressure drop increases with number of
bends. The straight layout TTHE has shown lowest pressure drop amongst all layouts tested,
which is as expected.
2722
2 4 6 8 10 12
Flow Rate of Water, m
h
(kg/min)
0
1
2
3
4
5
6
7
P
r
e
s
s
u
r
e
D
r
o
p
o
n
H
o
t
W
a
t
e
r
S
i
d
e
,
d
p
h
(
b
a
r
)
Straight
Serpentine (3 bends)
Serpentine (6 bends)
Serpentine (7 bends)
Serpentine (8 bends)
Serpentine (9 bends)
Fig. 4. Variations in pressure drop with mass ow rate of water.
0 8 12 16 20 24
Total Pumping Power, W
pp.tot
(W)
1.4
1.6
1.8
2.0
2.2
2.4
2.6
2.8
3.0
3.2
3.4
O
v
e
r
a
l
l
H
e
a
t
T
r
a
n
s
f
e
r
C
o
e
f
f
i
c
i
e
n
t
,
U
i
.
h
(
k
W
/
m
2
K
)
Straight
Serpentine (3 bends)
Serpentine (6 bends)
Serpentine (7 bends)
Serpentine (8 bends)
Serpentine (9 bends)
4
Fig. 5. Variations in experimental overall heat transfer coecient with total pumping power.
2723
3.2. Overall heat transfer coecient
Fig. 5 represents variation of experimentally determined overall heat transfer coecient with
total pumping power for six modules tested. For a xed pumping power, serpentine layout TTHE
with seven bends has shown highest overall heat transfer coecient. Its maximum value is higher
than straight tube TTHE by 17%.
When comparison is done amongst serpentine layouts, the overall heat transfer coecient is
highest in serpentine layout with seven bends and lowest with nine bends. For serpentine layout
with three bends, the overall heat transfer coecient is lower than serpentine layout with seven
bends in this range of ow rate. Thus, the experimental results indicate that in serpentine layout
TTHE, there is a denite optimum for a number of bends for a particular application.
In serpentine layout with three and six bends, with d
bend
/d
ti
of 55.3 and 28.1, respectively, part
of the straight sections does not oer enhancement in heat transfer. Eect of secondary ow
diminishes long before the uid enters the subsequent bend.
In serpentine layouts with eight and nine bends, with d
bend
/d
ti
of 20.9 and 15.8, respectively, the
secondary ow generated by the previous bend probably persists as the uid enters in the next
bend. Since the secondary ow is reversed in subsequent bends in a serpentine layout, pressure
drop increases without commensurate increase in heat transfer.
Benets of the bends are not realized, since number of bends per unit length is large as in the
case of eight and nine bends, the secondary ow is rst neutralized as it passes the bend and re-
versed as it comes out. The consecutive bend interactions are so severe that the performance of the
serpentine layout is worse than that of a straight tube TTHE.
The quantities measured directly include the volume ow rate of water, inlet and outlet temper-
ature of water. The uncertainty in measurement of volume ow rate of water is 2% of reading
(maximum uncertainty: 0.2 lpm). The uncertainty of temperature measurement is 0.3 C.
According to the uncertainty analysis based on Kline and McClintock method illustrated by Mof-
fat [11], the maximum uncertainties of overall heat transfer coecient, heat transfer rate, and log
mean temperature dierence is 3.3%, 3% and 0.76%, respectively.
4. Model development
An eectiveness-NTU method is used in the model developed for performance prediction of
tubetube heat exchanger. The present model developed for straight tubetube heat exchanger
incorporates following attributes:
(a) Layout: straight.
(b) Flow conguration: counter ow or parallel ow.
(c) Geometric parameters of tubes: number, diameter, thickness, material on two sides.
(d) Thermal bonding of tubes: material and its size (/), gap between the adjacent tubes.
Appropriate parameters like length of n, l
n
(part of the tube transferring heat by n eect),
number of tubes on two sides, semi-ll angle / are included in the model to analyse most of the
congurations (various congurations with dierent number of tubes) of tubetube heat exchan-
2724
ger. The performance parameters like overall heat transfer coecient, pressure drop, pumping
power, cost, and size of heat exchanger are calculated for straight layout. Input parameters to
the model are inlet temperatures and mass ow rates of water. Other input parameters necessary
for the analysis are ow arrangement, number of tubes, tube material, tube diameter, tube thick-
ness, thermal bonding material and its size.
Following assumptions were made in the model development.
(a) Steady state operation.
(b) No heat loss to the surroundings.
(c) No heat transfer in the direction of ow.
(d) Uniform uid distribution in all tubes in multiple tube tubetube heat exchanger.
(e) One dimensional ow of heat through part of the tube and thermal bonding material.
(f) No phase change.
4.1. Heat transfer equations
The net heat transfer depends on the waterside heat transfer coecients, fouling factors, ther-
mal resistance by tube walls, and thermal bonding material. The correlations for heat transfer
coecients and friction factors reported in the literature have been used in the model. Few impor-
tant equations used in the model are given below.
NTU is calculated by
NTU
UA
C
min
2
Theoretical relations for eectiveness as a function of NTU and heat capacity ratio are available
for dierent ow arrangements like counter ow or parallel ow. For counter ow, equation for
eectiveness is
e
1 expNTU 1 C
ratio
1 C
ratio
expNTU 1 C
ratio
3
where,
C
ratio
C
min
C
max
The maximum possible heat transfer rate is calculated by
Q
max
C
min
t
hi
t
ci
4
The actual heat transfer rate in heat exchanger is calculated by equation
Q e C
min
t
hi
t
ci
5
2725
4.2. Heat transfer coecient
The net heat transfer depends on heat transfer coecients, fouling factors, thermal resistance
by tube walls, and thermal bonding material.
Gnielinskis correlation [9] is used for heat transfer coecient calculation for transition region
as well as fully developed turbulent ow for a straight smooth circular tube
Nu
f
2
Re 1000 Pr
1 12.7
f
2
_
0.5
Pr
2=3
1
6
For 2300 6 Re 6 5 10
6
and 0.5 6 Pr 6 2000.
It is modied version of second Petukhov correlation, which agrees with most reliable experi-
mental data to an accuracy of 5%.
4.3. Friction factor
For friction factor in fully developed turbulent ow for a straight smooth circular tube, Prandtl,
Karman, Nikuradse (PKN) correlation is classical correlation valid for wide range of Re
(4 10
3
6 Re 6 10
7
). The correlation is also used for comparison of recent correlations.
1
f
p 1.7372 ln Re
f
_
_ _
0.3946 7
Predictions of this correlation agree with the extensive experimental measurements within 2 [4].
However, the PKN correlation is not explicit form; Techo, Tickner, James correlation [10] is
used for friction factor calculations.
1
f
p 1.7372 ln
Re
1.964 ln Re 3.8215
_ _
8
This explicit form of PKN correlation agrees within 0.1% of PKN correlation for 10
4
6
Re 6 2.5 10
8
.
5. Overall heat transfer coecient
The overall heat transfer coecient is obtained using following expression
1
UA
1
h
h
r
fh
_ _
1
A
h
g
fsh
R
thh
R
thtbm
R
thc
1
h
c
r
fc
_ _
1
A
c
g
fsc
9
R
thtbm
is thermal resistance of thermal bonding material. Thickness of thermal bonding material
(measured in the direction of heat transfer) varies from minimum at / = 0 to maximum at the
limiting value /
max
. R
thh
and R
thc
are thermal resistance of portion of tubes in contact with ther-
mal bonding material on hot water and cooling water side, respectively. Similarly, r
fh
and r
fc
are
fouling resistances on hot water and cooling water side, respectively. Thermal resistance of ther-
mal bonding material is calculated considering dierential element as shown in Fig. 2(a).
2726
Thermal resistance of dierential element is given by
R
d
to
1 cos / w
g
k
tbm
d
to
2
d/ cos / l
pt
_ _ 10
where, R is thermal resistance of the dierential element in terms of dierential angle d/. The ther-
mal resistances of all elements of TBM are in parallel. Therefore, the overall thermal resistance of
TBM is obtained by integrating 1/R term for the dierential element between the limits /
max
to
/
max
. The derived equation for overall thermal resistance of TBM is
R
thtbm
2
_
/
max
0
k
tbm
d
to
2
cos / l
pt
d
to
1 cos / w
g
d/
_ _1
11
Another approach to obtain thermal resistance of thermal bonding material is by calculating
equivalent/average thickness considering the area equivalence. Following equation gives equi-
valent/average thickness of thermal bonding material.
w
tbmav
w
g
sin /
max
1 cos /
max
1
2
/
max
p
180
sin /
max
cos /
max
_
sin /
max
d
to
12
The thermal resistance of thermal bonding material calculated using Eq. (11) and other calculated
by considering the approach of average thickness of thermal bonding material deviates within
10%.
0.0 0.1 0.2 0.3 0.4 0.5 0.6
Predicted Pressure Drop, dp
h
(bar)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
E
x
p
e
r
i
m
e
n
t
a
l
P
r
e
s
s
u
r
e
D
r
o
p
,
d
p
h
(
b
a
r
)
45 deg line
Fig. 6. Experimental vs. simulated pressure drop.
2727
6. Experimental validation of model
The present model developed for straight tube TTHE is validated experimentally for water-
to-water heat transfer by testing a typical module of straight tube TTHE. The experimental setup
used for conducting experiments is explained in earlier section.
The experimental and simulated results on pressure drop and overall heat transfer coecient
for straight tube TTHE are represented in Figs. 6 and 7. The results show good agreement bet-
ween the results prediction by model and the experimental results. The average dierence between
the predicted and experimental results of pressure drop on hot water side is 2.3% and the maxi-
mum dierence is 4.2%. The average dierence between the predicted and experimental overall
heat transfer coecient is 2% and maximum dierence is 6.2%.
7. Conclusions
Five dierent serpentine layouts of TTHE have been experimentally evaluated for the eect of
varying straight lengths between bends on thermo-hydraulic performance. Tubes used in all ve
serpentine layout TTHE are 9.525 mm OD and 43 mm bend diameter. Maximum overall heat
transfer coecient is obtained for the serpentine layout TTHE with seven bends when the perfor-
mance is compared at same pumping power. Also, the overall heat transfer coecient in serpen-
tine layout TTHE with seven bends is higher than TTHE with straight layout by 17%.
1.0 1.5 2.0 2.5 3.0
Predicted Overall Heat Transfer Coefficient, U
i.h
(kW/m
2
K)
1.0
1.5
2.0
2.5
3.0
E
x
p
e
r
i
m
e
n
t
a
l
O
v
e
r
a
l
l
H
e
a
t
T
r
a
n
s
f
e
r
C
o
e
f
f
i
c
i
e
n
t
,
U
i
.
h
(
k
W
/
m
2
K
)
45 deg line
Fig. 7. Experimental vs. predicted overall heat transfer coecient.
2728
In waterwater heat transfer, not all serpentine TTHE layouts are better than straight layout.
The serpentine layout TTHE with three and nine bends have shown lower heat transfer perfor-
mance than straight layout TTHE. However, at very low pumping power (below 3 W), serpentine
layout with three bends has shown slightly better heat transfer performance than straight tube
TTHE. The maximum dierence in performance of optimum serpentine layout TTHE (seven
bends) and nine bend serpentine TTHE, which is a non-optimum, is 30%. Thus, due care should
be taken while designing TTHE with serpentine layout to maximize the benets of this design.
In case of serpentine layout TTHE, the experimental results indicate that there is a denite opti-
mum for a number of bends for a particular application. In the present case, the optimum number
of bends in serpentine layout is 7.
An analytical model for simulation of straight tube TTHE in waterwater heat transfer is val-
idated experimentally. Eectiveness-NTU approach is used in the model for performance predic-
tions. The average and maximum deviation between predicted and experimental values of overall
heat transfer coecient is 2% and 6.2%, respectively. Similarly, average and maximum deviation
is pressure drop 2.3% and 4.2%, respectively.
References
[1] M.V. Rane, M.S. Tandale, Tubetube heat exchangers, Filed PCT/IN03/00377, 2003.
[2] M.V. Rane, M.S. Tandale, An experimental study on various layouts of tubetube heat exchanger in steam
condensation, Experimental Thermal and Fluid Science, submitted for publication.
[3] M.V. Rane, M.S. Tandale, Benets of superheat recovery on chillers- case study for a hotel installation, Paper
presented at 21st IIR International Congress of Refrigeration, Washington, DC, August 2003, pp. 1722.
[4] R.K. Shah, S.D. Joshi, Convective heat transfer in curved ducts, in: S. Kakak, R.K. Shah (Eds.), Handbook of
Single-phase Convective Heat Transfer, rst ed., John Wiley and Sons, 1987.
[5] I.Y. Chen, J.C. Huang, C.C. Wang, Singe-phase and two-phase frictional characteristics of small U-tube wavy
tubes, International Communication Heat Mass Transfer 31 (3) (2004) 303314.
[6] I.Y. Chen, S.K. Lai, C.C. Wang, Frictional performance of small diameter U-type wavy tubes, ASME Journal of
Fluids Engineering 47 (2003) 22412249.
[7] M.M. Ohadi, E.M. Sparrow, A. Walawalkar, A.I. Ansari, Pressure drop characteristics for a turbulent ow in a
straight circular tube situated downstream of a bend, International Journal of Heat Mass Transfer 33 (4) (1990)
583591.
[9] V. Gnielinski, New equations for heat and mass transfer in turbulent pipe and channel ow, International
Chemical Engineering 16 (2) (1976) 359368.
[10] R. Techo, R.R. Tickner, R.E. James, An accurate equation for the computation of the friction factor from smooth
pipes from the Reynolds number, ASME Transaction Journal of Applied mechanics 32 (1965) 443.
[11] R.J. Moat, Describing the uncertainties in experimental results, Experimental Thermal and Fluid Science 1 (1988)
317.
[12] D.F. Geary, Return bend pressure drop in refrigeration systems, ASHRAE Transactions 2342 (1980) 250265.
2729