International Communications in Heat and Mass Transfer 57 (2014) 5364

Contents lists available at ScienceDirect

International Communications in Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ichmt

Numerical study of an exhaust heat recovery system using corrugated

tube heat exchanger with twisted tape inserts
Vamsi Mokkapati, Chuen-Sen Lin
Mechanical Engineering Department, University of Alaska Fairbanks, P.O. Box 755905, Fairbanks, AK, 99775-5905, USA

a r t i c l e i n f o a b s t r a c t

Available online 15 July 2014 The purpose of this work is to investigate gas to liquid heat transfer performance of concentric tube heat exchang-
er with twisted tape inserted corrugated tube and to evaluate its impact on engine performance and economics
Keywords: through heat recovery from the exhaust of a heavy duty diesel generator (120 ekW rated load). This type of heat
Corrugated tube exchanger is expected to be inexpensive to install and effective in heat transfer and to have minimal effect on ex-
Twisted tapes haust emissions of diesel engines. This type of heat exchanger has been investigated for liquid to liquid heat
Exhaust gas
transfer at low Reynolds number by few investigators, but not for gas to liquid heat transfer. In this paper, a detail
Heavy diesel generator
Waste heat recovery
of heat transfer performance is investigated through simulations using computer software. The software is rst
Concentric tube heat exchanger justied by comparing the simulation results with the developed renowned correlations. Simulations are then
conducted for concentric tube heat exchanger with different twisted tape conguration for optimal design. The
results show that the enhancement in the rate of heat transfer in annularly corrugated tube heat exchanger
with twisted tape is about 235.3% and 67.26% when compared with the plain tube and annularly corrugated
tube heat exchangers without twisted tapes respectively. Based on optimal results, for a 120 ekW diesel genera-
tor, the application of corrugated tube with twisted tape concentric tube heat exchanger can save 2250 gal of fuel,
$11,330 of fuel cost annually and expected payback of 1 month. In addition, saving in heating fuel also reduces in
CO2 emission by 23 metric tons a year.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
In Alaska, there are about 180 villages using independent off grid
diesel generators for many decades and are not equipped with exhaust
heat recovery systems to take advantage of the energy in form of heat
contained in the exhaust which is about 1/3 of fuel energy. Even a frac-
tion of heat energy recovered may have a signicant effect on heating
fuel costs of cold region villages. From village power industry point of
view, the major reasons for not installing exhaust heat recovery system
is the effect on exhaust emissions and may cause maintenance difculty
resulting from soot accumulation and corrosion. In most of the Alaskan
rural villages except in large power plants, maintenance technicians and
engineers are not readily available when a maintenance problem rises
(especially during winters), the shipping of equipment and traveling
are also very expensive for isolated villages in the cold regions.
One of the heat recovery systems which can match the needs is the
vertically installed concentric tube heat exchanger with simple struc-
ture, which has been tested to capture the exhaust heat on a 120 kW
diesel generator and no evidence has been found in emissions and
maintenance. The heat exchanger is installed vertically in order to
Communicated by W.J. Minkowycz.
Corresponding author.
E-mail address: clin@alaska.edu (C.-S. Lin).
reduce the soot accumulation. The water exiting the engine jacket is fur-
ther heated in the current heat recovery system and supplied for space
heating. The advantages of using jacket water for heat recovery are eco-
nomical in installation, no additional pumping power required and no
extra loop cost.
The purpose of this work is to further improve the effectiveness of
heat recovery by using heat augmentation techniques. A literature re-
view has been conducted on concentric tube heat exchangers to deter-
mine the most suitable type of heat exchanger for cold region villages
with higher effectiveness. Heat transfer augmentation techniques are
classied into active and passive methods. Active methods require the
direct external power, whereas passive methods do not require any di-
rect external power. There are various techniques to reduce the thermal
boundary layer thickness by improving good mixing of the uids near
walls and the center of the tube. Swirl ow generating devices provide
chaotic mixing of the uid and also good passive method of heat trans-
fer augmentation.
Many researchers conducted experiments and numerical simula-
tions by inserting a wide range of twisted tapes, twisted coils and
conical-rings into the inner tube of the concentric tube heat exchangers
and proved that the thermodynamic efciency of the system is in-
creased. Heat transfer applications like waste heat recovery methods;
refrigeration and air conditioning systems are using this type of swirl
generators to enhance the heat recovery.

http://dx.doi.org/10.1016/j.icheatmasstransfer.2014.07.002
0735-1933/ 2014 Elsevier Ltd. All rights reserved.
54 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364

tape insert tubes and concluded that the heat transfer is being increased
Nomenclature with the increase in the twist ratio. Saha et al. [2] investigated the effect
of regularly spaced twisted tape elements, experimentally in laminar
CP specic heat, J/kg K and turbulent ow regime and reported that the heat transfer increases
d tube diameter, m by 2040%.
dt twisted tape diameter, m Ray and Date [3] conducted numerical analysis inserting a twisted
e enhancement tape in the square duct and developed correlations for friction factor
f Darcy friction factor and Nusselt number and reported that there is a fair agreement between
h heat transfer coefcient, W/m2 K the simulation and the experimental results. Garcia et al. [4] conducted an
H total enthalpy, J experimental study on heat transfer in a plain tube with wire coil inserts
k turbulent kinetic energy, m2/s2, thermal conductivity, in laminar-transition-turbulent regimes and stated that the heat transfer
W/m K rate can be increased by up to 200% keeping pumping power constant.
m mass ow rate, kg/s Eiamsa-ard et al. [5] conducted experiments and reported that
Nu Nusselt number Nusselt number is improved by 160% using full length twisted tape
Pt twisted tape pitch, m and 179% by inserting helical tape with and without rods [6] inside
p pressure, kg/m s2 the tube when compared with the plain tube. Naphon and Sriromruln
Pr Prandtl number [7] conducted experiments by inserting the coiled wire in the plain
Q heat transfer rate, W tube and found that there is a signicant effect of swirl in enhancing
q heat ux, W/m2 the rate of heat transfer.
Re Reynolds number Chang et al. [8] investigated heat transfer in tube with broken twist-
T temperature, K, constant time interval, s ed tape inserts experimentally and reported that the thermal perfor-
t time, s mance is improved by up to 0.91.8 times of those tubes tted with
U quantity of parameters (e.g. velocity) plain twisted tape. Heat transfer behavior with conical ring and twisted
v,u uid velocity, m/s tape insertions was investigated experimentally by Promvonge and
xi, xj cartesian co-ordinates, m Eiamsa-ard [9] and reported that 367% enhancement in the heat trans-
fer over the plain tube.
Promvonge [10] in another report, mentioned that enhancement in
Greek symbols the heat transfer is about 1.21.3 times when square wire coil is inserted
effectiveness in a circular tube. Promvonge [11] conducted experiments to investigate
dissipation, m2/s3 air ow friction and heat transfer characteristics in a round tube tted
density, kg/m3 with both coiled wire and twisted tape and obtained 200350% en-
b viscosity, kg/m s hancement in the heat transfer.
t eddy viscosity, kg/m s Bharadwaj et al. [12] conducted experiments on spirally grooved tube
Kronecker delta with twisted tape and reported 140% of heat transfer enhancement over
viscous stress tensor the plain tube. Rahimi et al. [13] also conducted numerical simulations
Reynolds stress tensor with various twisted tapes and compared with the classic twisted tape
and observed around 31% of enhancement in the heat transfer coefcient.
Eiamsa-ard et al. [14] performed numerical simulations of swirling ow
Subscripts in circular tube by means of twisted tapes and reported that the mean
c cold uid heat transfer rates are about 73.6% higher than that of the plain tube.
h hot uid Instead of changing the type of swirl generating device, Thianpong
l laminar et al. [15] changed the tube type and conducted experiments on a dim-
t turbulent pled tube with classic twisted tape and found the enhancement up to
1.66 to 3.03 times that of the plain tube. Similarly, inserting the twisted
tapes in the corrugated tube is expected to promote the generated swirl
Superscripts than that of plain tube and also can improve the rate of heat transfer. All
() Reynolds average, average the above studies along with the present study are summarized and
() Favre average presented in the Table 1.
() Reynolds uctuation SolidWorks ow simulation software has been used to conduct the
() Favre uctuation numerical study on the effect of swirl generated by twisted tapes of var-
ious congurations in the corrugated tube heat exchangers. This pro-
gram is rst veried with the renowned correlations for heat transfer
coefcient in highly turbulent ow regime and a fair agreement of
Abbreviations
about 10% between the simulation results and correlation prediction
ACT annularly corrugated tube
was observed. The simulation is then applied to diesel engine exhaust
CW corrugation width
gas to liquid heat exchanger. Based on the literature search, this is the
HE heat exchanger
rst study of this type of gas to liquid heat exchanger.
ID inner diameter
OD outer diameter
2. Diesel exhaust heat recovery system
PT plain tube
TT twisted tape
2.1. Current system

The idea has been tested by a power plant in Ruby (Alaska) with ver-
Some of the works using different types of swirl generators listed in tically installed concentric tube heat exchanger with annularly corru-
the literature have been discussed. Al-Fahed et al. [1] conducted the ex- gated tube as inner tube for exhaust and outside tube is plain tube
periments using oil as the working uid for plain, micro n and twisted- with jacket water which is exiting the diesel engine jacket to capture
 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364 55

Table 1
Related studies from the literature.

Paper Tube Tube diameter Swirl generator Type of HE Increase

Al-Fahed et al. [1] Micron tube 14 mm Twisted tape Steam to oile 225%a
Saha et al. [2] Plain tube 13 mm Spaced twisted tapes Oild 2040%a
Ray and Date [3]b Square duct 9.5 mm 9.5 mm Twisted tapes Waterf
Garcia et al. [4] Plain tube 18 mm Wire coils PG waterc 200%
Eiamsa-ard et al. [5] Plain tube 19 mm Helical tape Water to aire 160%
Eiamsa-ard et al. [6] Plain tube 50 mm Spaced twisted tapes Waterc 179%
Naphon et al. [7] Micron tube 8.92 mm Coiled wire inserts Waterc ~20%
Chang et al. [8] Plain tube 20 mm Broken twisted tape Aird 0.91.8 times
Promvonge et al. [9] Plain tube 48 mm Conical ring and twisted tape Aird 367%
Promvonge [10] Plain tube 47.5 mm Coiled square wires Aird 2.4 times
Promvonge [11] Plain tube 47 mm Wire coil & twisted tape Aird 200350%
Bharadwaj et al. [12] Grooved tube 14.8 mm Twisted tape Waterd 400%a, 140%
Rahimi et al. [13]b Plain tube 17 mm Modied twisted tapes Waterc 2.491.96 times
Eiamsa-ard et al. [14]b Plain tube (2.55.0) P Loose t twisted tapes Waterf 73.6%
Thianpong et al. [15] Dimpled tube 22 mm Twisted tape Waterc 1.663.03times
Present study Corrugated tube 101.5 mm Twisted tape N2 to watere 235.3%

Note: the citations [115] reported are arranged in chronological order.

a
Laminar ow, all other ows are under turbulent region.
b
Numerical (CFD) study, all other reports are experimental investigations.
c
Liquid to Liquid HE.
d
Electric heat source to Liquid/Gas HE.
e
Gas to Liquid HE.
f
Constant heat ux boundary condition.

exhaust heat for space heating. Heat recovery system is attached to John
Deere G4045-T300-HF458 Tier 3, 2010 model year, 1800 RPM, 120 ekW
diesel generator (Fig. 1) [16]. The installation is proved to be retrot and
inexpensive.
The annularly corrugated tube of the heat recovery unit installed in
the Ruby power plant was manufactured by Penex Corporation, and
the following drawings and dimensions are provided on request and di-
mensional details are presented in Table 2.
The outside tube (Shell) and inside tube (ACT) of the heat recovery
system dimensions are presented in Fig. 2 and Table 2. The corrugated
tube has 2.9 times the heat transfer surface area than that of plain
tube and hence can improve the rate of heat transfer. Hence, this is
one of the reasons behind replacing the corrugated tube in the place
of plain tube.
2.2. Proposed system to improve heat transfer performance
The goal of the present work is to maximize the heat recovery rate
by optimizing the design of twisted tape. The controlled variables in-
clude the Reynolds number ranging from 40,000 to 77,000 of exhaust.
The physical size of the heat exchanger, twist ratio and exhaust back
pressure change are the constraints.
3. Annularly corrugated tube heat exchanger with twisted
tape inserts
To improve the heat transfer rate of the heat recovery system, a
modication of inserting a twisted tape into a corrugated tube is inves-
tigated. The modications need to be retrottable and be cost effective
and also have negligible effect on emissions, maintenance and pumping
power. Inserting twisted tape in the present heat recovery system has
almost negligible effect on all the aspects mentioned earlier. The com-
bined effect of the twisted tape and corrugations are expected to better
improve the swirls of exhaust and also the rate of heat transfer.
In order to select the optimal twisted tape insert, four different
twisted tapes with different pitches are modeled and tested by inserting
into the ACT HE. The pitch of the twisted tape is inversely proportional
to the number of twists (Fig. 3).
In general, it is believed that the more the number of twists are, the
more swirl of the uid can be achieved. Too many twists may increase

Fig. 1. Schematic diagram of the system.

56 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364

Table 2 6. The inlet volume ow rate and temperature of water is 63 gpm

Tube details. (0.00397 m3/s) and 195 F (90.55 C) respectively.
Tube Pitch(m) CW(m) OD(m) ID(m) Thickness(m) 7. The inlet volume ow rates and temperature of N2 are 0.3933 m3/s,
Plain 0.102 0.1016 0.61 103
0.3343 m3/s, 0.2753 m3/s, 0.1966 m3/s and 510 C (temperature of
Corrugated 0.0105 0.0075 0.122 0.1016 0.61 103 the engine exhaust).
8. The heat exchanger is installed vertically.

the manufacturing difculty as well as back pressure drop. The 3D
model is presented in Fig. 4 for visualization.
4. Mathematical modeling
Prediction of relationship between the heat transfer and swirling of
the uid inside the tubes is the main goal of the simulations. Simulations
performed for annularly corrugated tube (ACT) heat exchanger with
and without inserting various twisted tapes are mentioned in Table 3
and depicted in Fig. 5. Along with these simulations for plain tube heat
exchanger for reference. In solving the numerical model, the following
assumptions and boundary conditions are considered.
4.1. Validation
The numerical simulations are performed to validate the heat trans-
fer coefcient obtained from simulation with the existing correlations
for Nusselt number. The agreement with the numerical analysis results
and correlation (obtained from the experiment) results of Nusselt num-
ber values are within the 10% difference on the tube side. Fig. 6 presents
the comparison between Numerical analysis, DittusBoelter's correla-
tion, Gnielinski's correlation and Petukhov's correlation [17] in terms
of dimensionless parameter Nusselt number (Nu), Reynolds number
(Re) and Prandtl number (Pr).
Nusselt number

Nu hdh =k: 1
1. Steady operating conditions exist.
2. Radiation effects are negligible. Reynolds number
3. Nitrogen gas is used in the simulations in the place of exhaust gas for
their similar thermal properties and treated as ideal gas. Re vdh =: 2
4. The ambient pressure and the pressure at exit are both assumed to be
at 1-atm. Prandtl number
5. Ambient temperature is assumed to be at 20 C, with heat transfer
coefcient of 15 W/m2K. Pr C p =k: 3

Fig. 2. Exhaust heat recovery unit.

 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364 57

Fig. 3. Twisted tape dimensions.

DittusBoelter's correlation
Density weighted time averaging for velocity and energy i.e., Favre av-
0:8 0:3
Nu 0:024Re Pr : 4 eraging
Z tT
Gnielinski's correlation e 1 lim 1
U e U ; T Constant Time interval:
t U t dtU U

T T t
f =8Re1000 Pr 8
Nu
 5
1 12:7 f =80:5 Pr0:67 1
Favre Averaged Navier Stokes (FANS) [19] equations, i.e., conserva-
where friction factor f in the above equation can be dened using tion equations can be written as follows.
Petukhov's correlation
Continuity equation
2
f 0:79 ln Re1:64 :
uei
0: 9
t xi
Petukhov's correlation

f =8 Re Pr
Nu
 6
C 12:7 f =80:5 Pr0:67 1 Momentum equation
 
where constant C in the above equation can be dened as
uei uei uej p h i
ji 10
C 1:07 900=Re0:63=1 10Pr: t x j xi x j ji

4.2. Governing equations where viscous stress tensor

!
SolidWorks CFD package is employed to perform the numerical uei uej 2 u
analysis. The Favre Averaged Navier Stokes (FANS) equations are ij k ij 11
x j xi 3 xk
used in this CFD package to model the velocity, pressure elds, etc.,
and are combination of law of conservations of mass, momentum
and Reynolds stress tensor
and energy for three dimensional models, also suitable for highly
turbulent ow cases [18], which is density weighted time averaging
ij ui uj : 12
Navier Stokes equation.
Classical time averaging for density and pressure, well known as
Reynolds averaging Boussinesq equation expressing the stress tensor

Z !
1 tT
0 uei uej 2 u
ek 2
U lim U t dtU U U : 7 ij t ij kij : 13
T T t x j xi 3 xk 3

Fig. 4. 3D model of twisted tape swirl generator.

58 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364

(A) (B)

Fig. 5. Cross sectioned view of (A) ACT heat exchanger and (B) ACT heat exchanger with twisted tape.

Total enthalpy Turbulent viscosity factor (f)

" #
  1 p h  i
t
H
x j
e jH
u
x j
qlj qtj ij ui uj ui ui
2
e ji
u
t x j i ji
h  i2  20:5

f 1 exp 0:025Ry : 1
e 1u 1
p RT;e Hh eu
2 i i
e k; k ui ui :
2 2
pRT 17
k k:y
14 RT ;R :
y

where H = total enthalpy, h = Static enthalpy, k = turbulence kinetic
energy, qlj = laminar heat ux vector and qtj = turbulent heat ux vec- The constants C = 0.09, C1 = 1.44, C2 = 1.92, = 1.3, k = 1
tor. The k- turbulence model is used to model the turbulent regime for are dened empirically.
the present study.
For turbulent kinetic energy (k)
  
k k R ui
ui k t ij t P B : 15
t xi xi k xi x j

For dissipation ()
   !
t R u 2
ui C 1 f 1 ij i t C B P B C 2 f 2
t xi xi xi k x j k
!3
 
16
0:05 2
f1 1 ; f 2 1exp RT :
f

Table 3
Dimensional details of various twisted tapes.

Twisted tape(TT) Pitch(Pt) Pt/dt Number of twists

TT-1 0.15 m 1.5 4

TT-2 0.2 m 2 3
TT-3 0.25 m 2.5 2.4
TT-4 0.3 m 3 2 Fig. 6. Comparison of simulation results with existing correlation predictions.
 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364 59

Table 4
Heat transfer comparison for PTHE with twisted tapes at a Reynolds number of 67,000.

PT_TT1 PT_TT2 PT_TT3 PT_TT4

Heat transfer(W) 8524.524 7989.685 7614.559 7516.491

Enhancement (%) 35.53 27.03 21.06 19.5

4.3. Simulation procedure

The basic ow equations are derived in nite volume method using

integral approach. The conservative equations are discretized and
solved throughout the model. The rate of increase in the quantity is
equal to the sum of difference in the ux quantity and generation.
Z Z
! !
UdV s F :d S Qdv 18
t v v

where U represents the quantity, V = Volume, S = Surface, F = Flux

and Q = Source.
The second order approximations of uxes (F) are based on modi-
ed Leonard's Quadratic Upstream Interpolation for Convective Kinetics
(QUICK) approximations and the Total Variation Diminishing (TVD)
method. The pressure velocity coupling numerical algorithm i.e., Semi
Implicit Method for Pressure Linked Equations (SIMPLE) is used to esti-
mate the pressure eld [17].
4.4. Mesh independent study
Three different meshes with total number of cells 387,705, 538,467
and 924,491 are initially considered and mesh independent study was
performed to verify the results. The results of heat transfer rate are com-
pared and the discrepancies between the results are less than 0.5%. The
intermediate mesh level is chosen for all the simulations which saves
about 41.76% of simulation time and gives similar results (0.5% devia-
tion) when compared with the ne mesh.
5. Numerical simulations and results
To design the optimal twisted tape for exhaust heat recovery appli-
cation, simulations have been conducted over a range of Reynolds num-
bers (Re) for the plain tube and corrugated tube with and without
twisted tapes and compared to investigate the enhancement in the
heat transfer coefcient (Nu) and rate of heat transfer (Q) as well. The
Reynolds number range is from 33,000 to 68,000 for the gas in the
inner tube.
Fig. 7. The Nu versus Re for the tubes with different twisted tape inserts.
Fig. 8. Twisted tape pitch versus the enhancement percentage.
5.1. Simulations of the PT HE with and without twisted tapes
The simulation results show that the enhancement in the heat trans-
fer of about 35.5%, 27.02%, 21.06% and 19.5% in the PT HE's with twisted
tapes TT1, TT2, TT3 and TT4 respectively when compared with the PT HE
without twisted tape. As expected the tape with maximum number of
twists (TT1) provides more swirl and also yield to the highest rate of
heat transfer (8.5 kW) among all the other tapes.
The PT HE provides a heat transfer of 6.3 kW at a Reynolds number of
67,000. The rate of heat transfer for different twisted tape inserts in a
PTHE is presented and compared with the PTHE in Table 4 and Fig. 7.
It can be observed that the Nusselt number is increasing with the in-
crease in the number of twists. The maximum Nusselt number at the
maximum Reynolds number in the PTHE with TT1 is 200.
5.2. Simulations of ACT heat exchangers with twisted tapes (TT1, TT2, TT3
and TT4)
The simulations for ACT HE are conducted with similar boundary
and inlet conditions which are used for the simulations with PT. The
total heat recovered by ACT heat exchanger without twisted tape is
around 11.48 kW. The further simulations with twisted tapes are com-
pared with the heat transfer of the PT and ACT heat exchangers. The
heat transfer behavior in annularly corrugated tube with different twist-
ed tapes at four different Reynolds numbers of gas are investigated and
numerical simulation results for rate of heat transfer are presented in
Tables 58 and comparison between them is presented in Fig. 8.
The heat transfer enhancement in ACT HE with TT1 is about 235%
and 64.2% at a Reynolds number of 40,000 and about 205.43% and
67.25% at the Reynolds number of 77,700 when compared with the PT
and ACT heat exchanger without TTs respectively. The enhancement
percentage of heat transfer rate in ACT heat exchangers with TTs is ob-
served to be higher at low Reynolds number.
The ow trajectories of the gas inside PT and ACT HE without (Figs. 9,
10) and with TT1 (Figs. 11, 12) respectively are generated by the CFD
program. The gures also represent the temperature variation of the
gas from entrance to exit of HE with respect to the temperature scale be-
side the ow trajectories. Fig. 14 presents the direction of velocity
vectors inside the ACT and ACT with TT1, the dots represent that there
is no tangential velocity inside the tube.
Fig. 13 presents the comparison between the heat transfer rate of
ACT and PT HEs with various twisted tape pitches ranging from 0.15
m to 0.3 m at a nitrogen volume ow rate of 0.3933 m3/s.
60 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364

Fig. 9. Flow trajectories with temperature distribution of the gas inside the plain tube.

Fig. 10. Flow trajectories with temperature distribution of the gas inside the ACT.

Fig. 15 presents the variation in the Nusselt number in the heat
exchangers at different Reynolds numbers for ACT with different
twisted tape inserts. The ACT with twisted tape (TT1) has the maximum
Nusselt number of 201.75 at Reynolds number 77,700 when compared
with the other ACT heat exchangers with and without twisted tape
inserts. TT1 has more number of twists and can generate more swirls
in the ACT than that of other TTs and hence the heat transfer coefcient
is comparatively high.
Figs. 16 and 17 present the contour plots of tangential velocity in-
side the plain and ACT without and with different twisted tapes

Fig. 11. Flow trajectories and temperature distribution of the gas inside the PT with TT-1.

Fig. 12. Flow trajectories and temperature distribution of the gas inside ACT with TT-1.
 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364
Fig. 13. Heat transfer rate versus pitch of the twisted tape inserts in the ACTHE and PTHE.
respectively. High tangential velocity is observed in the ACT with the
similar twisted tape and uid ow conditions, which in other words,
also conrms more swirling.
5.3. Effectiveness () of PT and ACT heat exchangers
Effectiveness is a measure of heat transfer performance of a heat ex-
changer and is denoted by . Effectiveness of a heat exchanger can be
dened as the ratio of the total amount of heat that has been exchanged
to the maximum possible heat transfer rate [20].
Q tangential velocity inside the HE (Figs. 16 and 17).
   Q max    19
where C c T c;o T c;i C h T h;i T h;o ; Q max C min T h;i T c;i
C c mc Cpc ; C h mh Cph ; C min Minimum of C c and C h
Using the above equations the Effectiveness of the PT and ACT (with
and without twisted tapes) heat exchangers under similar conditions
are calculated from the simulation results. The effectiveness of ACT HE
with and without TT1 is about 23.67% and 14.35%, whereas the effec-
tiveness of PT HE with and without TT1 is about 10.67% and 8.17%
respectively.
6. Discussion of results
Simulation results show that, for the same heat exchanger dimen-
sion (i.e., height and diameter), the heat transfer rate (11.4 kW) of an
annularly corrugated heat exchanger (ACTHE) is higher than that
Fig. 14. Velocity vectors inside ACT and ACT with TT1 at tube exit.
61
(6.3 kW) of plain tube (PTHE). This may have resulted from the more
surface area of the ACTHE.
The results also show that the twisted tape (TT) inserts can improve
heat transfer related performance parameters, such as Nusselt number
(Fig. 7), heat transfer enhancement (Fig. 8), and heat transfer rates
(Tables 4 to 8) against both plain tube and corrugated tube. This can at-
tribute to the swirls caused by the twist tape inserts, which can be found
in Figs. 9 to 12.
Figs. 9 and 11 show that the ow swirl (represented by stream lines)
of plain tube with TT-1 insert is much higher than that of a plain tube
without a TT. A similar observation is found in Figs. 10 and 12 for corru-
gated tube.
From Tables 4 and 6, it can be found that, for about the same level of
Reynolds number, the enhancement percentages resulted from TT in-
sertion for corrugated tube (Table 5. Enhancement from 23.3% to
67.26%) is signicantly higher than that for plain tube (Table 4. En-
hancement from 19.5% to 35.53%). The cause may be that the corrugated
tube help in promoting the swirl and as a result heat transfer rate is
largely increased.
According to Figs. 8 and 9, the ow trajectory inside the plain tube
and annularly corrugated tube is very similar to each other; the
advantages of corrugations are not completely utilized. Inserting the
twisted tape in the plain tube and corrugated tubes, the swirling of
the uid can be achieved and a signicantly more amount of the
temperature is being dropped in the corrugated tube (with TT1)
than that in the plain tube with TT1, this can be observed in
Figs. 11 and 12.
From observation, the heat transfer rate increases as the pitch of the
TT decreases (Fig. 13). This may be a result that the shorter the pitch
means the more the twists for the same length HE. The more twists
may guide the uid to travel more turns along the HE and cause higher
From Fig. 8, the enhancement percentage is about 235.32% and
205.43% at a Reynolds number of 40,000 and 77,000 respectively i.e.
the twisted tapes are performing more effectively at lower ow rates
in annularly corrugated tube. From Table 9, the effectiveness of PTHE
is about 8.17% and when the PT is replaced by ACT the effectiveness is
increased to 14.34%. But inserting twisted tape in the ACT HE, the effec-
tiveness is drastically increased to 23.6% which is about 9.26% more than
that of the PTHE.
According to simulation results, the maximum (~ 67.25%) rate of
heat transfer was observed when TT-1 is inserted in the current trial
system of Ruby power plant, Alaska.
7. Economic analysis
For the Ruby village power plant diesel generator equipped with ACT
vertical HE is able to recover 11.48 kW of heat from the diesel engine ex-
haust. The present ACT HE unit cost is $2,215; approximated cost for the
62 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364
Fig. 15. Nusselt number versus Reynolds number for ACT.
twisted tape (TT1) is $500 (estimated by a local machine shop). Simula-
tion results show that using twisted tape with a pitch of 0.15 m can en-
hance the heat recovery rate by up to 67.25%.
The cost of the twisted tape = $500.
The total system cost = $2,715 (i.e., unit cost + tape cost = $2,215
+ $500).
Heating value for the fuel at 75% boiler efciency = 101,250 BTU/
gal.
Total heat recovery rate of ACT HE = 11.48 kW = 39,171.36 BTU/h.
Heat recovered per year = 39,171.36 BTU/h 24 h/day 363days/
year = 341.22 MBTU/year.
Based on $5.00 per gallon of fuel cost, which is common for rural Alas-
kan villages, the estimated payback period for ACT heat exchanger with
TT1 is about 1 month and practically has no effect in the maintenance. A
summary of the economic analysis is listed in the rst row of Table 10.
Fig. 16. Tangential Velocity Contour plot at exit of PTHE without and with TT (14).
Fig. 17. Tangential Velocity Contour plot at exit of ACTHE without and with twisted tapes (14).
Similarly, economic analysis results for ACT with TT1, which show
the most improvement among all the TTs are also listed in row 2 of
Table 10. Saving fuel not only saves money but also reduces the amount
of CO2 released into the atmosphere. About 10.15 kg of CO2 is released
by burning 1 gal of heating oil [21]. Using ACT HE with twisted tape
can save up to 5,636.5 gal per year which is 2,266 gal more than that
of ACT HE (presently using HE), while also reducing emissions of CO2
by up to 23 metric tons a year.
8. Conclusions
A study to improve the effectiveness of a diesel exhaust heat recov-
ery system (concentric tube heat exchanger) of Ruby, Alaska has been
conducted. The improved system is the original heat exchanger inserted
with a twisted tape. The goal of the present work is to maximize the
heat recovery rate by optimizing the design of twisted tape insert. The
physical size of the heat exchanger, twist ratio and exhaust back pres-
sure change are the constraints. In this study, the outer tube is a plain
tube and the inner tube is an annularly corrugated tube with and with-
out twisted tape inserts. The following paragraphs summarize the nd-
ings obtained from this study.
1. The effect of twisted tape insert was found to cause signicant im-
provement in heat transfer performance of the plain tube heat ex-
changer. Among all the twisted tapes tested (TT1, TT2, TT3 & TT4)
the one with lower pitch i.e. TT1 (~0.15 m) has more effect on the
heat transfer enhancement which is about 35.5%.
2. The corrugated tube heat exchanger improves the rate of heat trans-
fer up to 82% and 35% when compared with the plain tube without
and with TT1 respectively.
3. The corrugated tube heat exchanger with twisted tape insert has
more heat transfer rate when compared with the corrugated tube
alone. The effect of twisted tape insert was found to cause signicant
improvement in the heat transfer performance. The heat transfer rate
increased by inserting the twisted tapes (TT1, TT2, TT3 and TT4) in
ACT heat exchanger is about 67.25%, 49.89%, 33.83% and 23.3% respec-
tively when compared with the ACT HE without twisted tape.
 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364 63

Table 5 Table 9
Simulation results of ACT heat exchanger with twisted tapes at Re = 77,700. Heat exchanger effectiveness.

ACT_TT1 ACT_TT2 ACT_TT3 ACT_TT4 HE Without TT TT1 TT2 TT3 TT4

Heat transfer(W) 19211.17 17217.42 15372.72 14162.57 PT 8.17% 10.68% 10% 9.52% 9.4%
e (%)-PT 205.52 173.73 144.4 125.17 ACT 14.34% 23.6% 21.29% 19.05% 17.57%
e (%)-ACT 67.26 49.9 33.83 23.3
The e(%) -PT refers the enhancement percentage when compared with PT HE without
TTs.
The e(%) -ACT refers the enhancement percentage when compared with ACT HE Table 10
without TTs. Economic analysis.

Heat exchanger ACT (current ACT with TT1(proposed

installation) installation)
Table 6
Simulation results of ACT heat exchanger with twisted tapes at Re = 66,200. Initial unit cost $2,150 $2,650
Heat recovered 341.22 MBtu/year 570.69 MBtu/year
ACT_TT1 ACT_TT2 ACT_TT3 ACT_TT4 annually
Annual fuel savings 3,370.1 gal/year 5,636.5 gal/year
Heat transfer(W) 16828.29 15391.3 13952.96 12887.44
Annual savings $16,850/year $28,181/year
e (%)-PT 196.63 171.3 145.95 127.17
Estimated payback 1.5 months 1 months
e (%)-ACT 63.41 50.98 38.59 27.13
period
The e(%) -PT refers the enhancement percentage when compared with PT HE without
TTs.
The e(%) -ACT refers the enhancement percentage when compared with ACT HE
without TTs.

generators in rural Alaskan villages, the potential effect on economy of

4. The heat transfer rate in corrugated tube heat exchanger with twist- diesel power industry may reach to a savings of $4 million per year
ed tape insert is 67% more than that of corrugated tube alone. When based on 370,000 MW-h power consumption in 2007 [22].
the same twisted tape is inserted in the plain tube heat exchanger,
the heat transfer rate is about 35.3% more than that of tube alone.
Acknowledgments
This clearly states that corrugated tube is promoting the effect of
swirl than that of plain tube.
The authors would like to gratefully acknowledge the support of
5. Using ACT HE with twisted tape can save up to 5,636.5 gal of fuel per
Alaska Center for Energy and Power, Mechanical Engineering Depart-
year which is 2,266 gal more than that of presently installed heat re-
ment, College of Engineering and Mines at University of Alaska Fairbanks.
covery system and also reducing emissions of CO2 by up to 23 metric
tons annually.
6. Payback period for the proposed exhaust heat recovery system is References
about 1 month in comparison with the currently installed setup
[1] S. Al-Fahed, L.M. Chamra, W. Chakroun, Pressure drop and heat transfer comparison
which is 1.5 months. for both micron tube and twisted-tape inserts in laminar ow, Exp. Thermal Fluid
7. The pressure drop of the exhaust gas in this heat recovery system is Sci. 18 (1999) 323333.
about 0.257 psi (1.77 kpa), which conveys that the effect of inserting [2] S.K. Saha, A. Dutta, S.K. Dhal, Friction and heat transfer characteristics of laminar
swirl ow through a circular tube tted with regularly spaced twisted-tape ele-
twisted tape on the engine performance is minimal.
ments, Int. J. Heat Mass Transf. 44 (22) (2001) 42114223.
[3] S. Ray, A.W. Date, Friction and heat transfer characteristics of ow through square
Based on this study, the proposed exhaust heat recovery system
duct with twisted tape insert, Int. J. Heat Mass Transf. 46 (5) (2003) 889902.
with TT1 is simple, relatively inexpensive, easy to maintain and no prac- [4] A. Garca, P.G. Vicente, A. Viedma, Experimental study of heat transfer enhancement
tical effect on exhaust emissions. Also it has a good economic effect on with wire coil inserts in laminar-transition-turbulent regimes at different Prandtl
Ruby diesel power application. If the system is applied to all the diesel numbers, Int. J. Heat Mass Transf. 48 (2122) (2005) 46404651.
[5] S. Eiamsa-ard, P. Promvonge, Enhancement of heat transfer in a tube with regularly-
spaced helical tape swirl generators, Sol. Energy 78 (4) (2005) 483494.
[6] S. Eiamsa-ard, C. Thianpong, P. Promvonge, Experimental investigation of heat
Table 7 transfer and ow friction in a circular tube tted with regularly spaced twisted
Simulation results of ACT heat exchanger with twisted tapes at Re = 53,000. tape elements, Int. Commun. Heat Mass Transfer 33 (10) (2006) 12251233.
[7] P. Naphon, P. Sriromruln, Single-phase heat transfer and pressure drop in the micro-
ACT_TT1 ACT_TT2 ACT_TT3 ACT_TT4 n tubes with coiled wire insert, Int. Commun. Heat Mass Transfer 33 (2) (2006)
Heat transfer(W) 14546.09 13650.8 12530.05 11494.28 176183.
[8] S.W. Chang, T.L. Yang, J.S. Liou, Heat transfer and pressure drop in tube with broken
e (%)-PT 203.44 184.77 161.38 139.78
twisted tape insert, Exp. Thermal Fluid Sci. 32 (2) (2007) 489501.
e (%)-ACT 60.89 50.98 38.59 27.13
[9] P. Promvonge, S. Eiamsa-ard, Heat transfer behaviors in a tube with combined
The e(%) -PT refers the enhancement percentage when compared with PT HE without conical-ring and twisted-tape insert, Int. Commun. Heat Mass Transfer 34 (7)
TTs. (2007) 849859.
The e(%) -ACT refers the enhancement percentage when compared with ACT HE [10] P. Promvonge, Thermal performance in circular tube tted with coiled square wires,
without TTs. Energy Convers. Manag. 49 (5) (2008) 980987.
[11] P. Promvonge, Thermal augmentation in circular tube with twisted tape and wire
coil turbulators, Energy Convers. Manag. 49 (11) (2008) 29492955.
Table 8 [12] P. Bharadwaj, A.D. Khondge, A.W. Date, Heat transfer and pressure drop in a spirally
grooved tube with twisted tape insert, Int. J. Heat Mass Transf. 52 (78) (2009)
Simulation results of ACT heat exchanger with twisted tapes at Re = 40,000.
19381944.
ACT_TT1 ACT_TT2 ACT_TT3 ACT_TT4 [13] M. Rahimi, S.R. Shabanian, A.A. Alsaira, Experimental and CFD studies on heat
transfer and friction factor characteristics of a tube equipped with modied twisted
Heat transfer(W) 11765.7 10821.17 10379.16 9401.002 tape inserts, Chem. Eng. Process. Process Intensif. 48 (3) (2009) 762770.
e (%)-PT 235.32 208.40 195.80 167.93 [14] S. Eiamsa-ard, K. Wongcharee, S. Sripattanapipat, 3-D numerical simulation of
e (%)-ACT 64.294 51.105 44.93 31.27 swirling ow and convective heat transfer in a circular tube induced by means of
loose-t twisted tapes, Int. Commun. Heat Mass Transfer 36 (9) (2009) 947955.
The e(%)-PT refers the enhancement percentage when compared with PT HE without
[15] C. Thianpong, et al., Compound heat transfer enhancement of a dimpled tube with a
TTs. twisted tape swirl generator, Int. Commun. Heat Mass Transfer 36 (7) (2009)
The e(%)-ACT refers the enhancement percentage when compared with ACT HE 698704.
without TTs. [16] Ruby Power Plant Exhaust Heat Recovery Testing, 2011.
64 V. Mokkapati, C.-S. Lin / International Communications in Heat and Mass Transfer 57 (2014) 5364

[17] Introduction to Heat Transfer, by Frank P. Incropera, David P. DeWitt, Third edition
John Wiley & Sons, Inc., 1996
[18] SolidWorks Flow Simulation 2012, Technical Reference, 2012.
[19] Turbulent Flows: Fundamentals, Experiments and Modeling by G. Biswas, V.
Eswaran, Science, CRC Press, 2002. (456 Pages).
[20] Fundamentals of Heat Exchanger Design by Ramesh K. Shah and Dusan P. Sekulic,
John Wiley & Sons, Inc., 2003
[21] U.S. Energy Information Administration, www.eia.gov/oiaf/1605/coefcients.html
2011.
[22] Capture of heat energy from diesel engine exhaust by Lin CS, Final Report Prepared
for National Energy Technology Laboratory, DOE Award # DE-FC26-01NT41248, No-
vember 2008.