See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/229395394

Analytical modeling of water condensation in condensing heat exchanger

Article  in  International Journal of Heat and Mass Transfer · May 2010

DOI: 10.1016/j.ijheatmasstransfer.2010.02.004

CITATIONS READS

105 2,024

4 authors, including:

Kwangkook Jeong
Arkansas State University - Jonesboro
21 PUBLICATIONS   196 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Predictive Modeling for Fly Ash Deposition in Post-Boiler Heat Exchangers for Coal-Fired Power Plant Applications View project

Numerical Simulation for Water and Heat Recovery for Advanced Organic Rankine Cycle Applications View project

All content following this page was uploaded by Kwangkook Jeong on 15 March 2019.

The user has requested enhancement of the downloaded file.

 International Journal of Heat and Mass Transfer 53 (2010) 2361–2368

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Analytical modeling of water condensation in condensing heat exchanger

Kwangkook Jeong a,*, Michael J. Kessen b, Harun Bilirgen b, Edward K. Levy b
a
Korea Electric Power Corporation, 65 Munji-Ro, Yuseong-Gu, Daejeon 305-380, Republic of Korea
b
Energy Research Center, Lehigh University, 117 ATLSS Drive, Bethlehem, PA 18015, USA

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

Article history: An analytical model of heat and mass transfer processes in a flue gas condensing heat exchanger system
Received 24 August 2009 was developed to predict the heat transferred from flue gas to cooling water and the condensation rate of
Received in revised form 1 February 2010 water vapor in the flue gas. Flue gas exit temperature, cooling water outlet temperature, water vapor
Available online 22 February 2010
mole fraction, and condensation rate of water vapor were computed using the analytic model. A pilot-
scale heat exchanger was used to validate the analytical model. The experimental results show a very
Keywords: good agreement with analytical model results. The performance of the heat exchanger system was eval-
Analytical condensation modeling
uated as functions of the ratio of the mass flow rate of cooling water to the mass flow rate of inlet flue gas,
Boundary value problem
Condensing heat exchanger design
the inlet cooling water temperature and the ratio of flue gas flow rate to total heat transfer surface area.
Heat and mass transfer coefficient Ó 2010 Elsevier Ltd. All rights reserved.
Power plant boiler flue gas
Non-condensable gas
Water recovery

1. Introduction
Thermoelectric power plants utilize significant quantities of
water. For example, a 500 MW power plant that employs once-
through cooling uses 4.5104 m3/h (approximately 45106 kg/h) of
water for cooling and other process requirements. Water supply is-
sues are increasing in importance for new and existing power
plants because the freshwater supply is limited. For companies
considering the development of new thermoelectric power plants,
water is a first-order concern. The impacts of water supply depend
on the regional environment in which the power plant is to be
built. Power plants located in some parts of the United States will
find it increasingly difficult to obtain the large quantities of water
needed to maintain operations. In response to these concerns, DOE
is funding research and development to reduce the amount of
freshwater used by power plants [1].
The focus of this study is water recovery from exhaust gases of
fossil-fired power plants. Power plant exhaust gases release large
amounts of water vapor into the atmosphere. The flue gas is a po-
tential source for obtaining much needed cooling water for a
power plant. There is almost 40% (wet coal mass basis) moisture
in lignite coal, which is translated to 16% moisture by volume
(wet basis) in the flue gas. For example, a 600 MW power plant fir-
ing lignite exhausts a flue gas flow rate of 2.7106 kg/h, which in-
cludes a moisture flow rate of 0.43106 kg/h, or about 16 wt% of
* Corresponding author.
E-mail addresses: kkj206@lehigh.edu, kjeong@kepco.co.kr (K. Jeong).
the flue gas. In contrast, typical cooling tower evaporation rates
for a 600 MW unit are 0.7106 kg/h [2].
If a power plant could recover and reuse a portion of this mois-
ture, it could reduce its total cooling water intake requirement. The
most practical way to recover water from flue gas is to use a con-
densing heat exchanger. The power plant could also recover latent
heat due to condensation as well as sensible heat due to lowering
the flue gas exit temperature. In addition, harmful acid gases such
as H2SO4, HCl, and HNO3 can be condensed by the heat exchanger
preventing them from entering the atmosphere.
Condensation in flue gas is a complicated phenomenon since
heat and mass transfer of water vapor and various acid vapors
simultaneously occur in the presence of non-condensable gases.
To design a full scale heat exchanger for a power plant, the conden-
sation process must be modeled correctly.
Several investigators have proposed analytical solutions. In
1980, Webb and Wanniarachchi [3] developed a one dimensional
numerical model to predict the effect of non-condensable gases
in a 10-row by 10-column finned tube heat exchanger by solving
the Colburn–Hougen equation for refrigerant R-11 and air mixture.
An iterative solution procedure was applied to solve the equation.
The modeling results were not verified with measured data.
From 1999 to 2003, Osakabe et al. [4–7] carried out one-dimen-
sional heat and mass balance calculations for the condensation of
flue gas in bare and finned tube heat exchangers. Experimental
studies using actual flue gas from propane, natural gas and oil com-
bustion were conducted to investigate the effect of parameters
such as flue gas flow rate, cooling water temperature and cooling
water flow rate. The results of both modeling and experiments

0017-9310/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2010.02.004
2362 K. Jeong et al. / International Journal of Heat and Mass Transfer 53 (2010) 2361–2368

Nomenclature

A area Greek symbols

a, b, c constants a thermal diffusivity
cp specific heat g efficiency
D binary mass diffusivity k thermal conductivity
d tube diameter q density
F fouling factor
f friction factor Subscripts
h convective heat transfer coefficient c cooling water
kD convective mass transfer coefficient cd water condensate
jH Colburn j factor for heat transfer dew dew point
jm Colburn j factor for mass transfer eff effective
km mass transfer coefficient f liquid film
L tube length g wet flue gas
Le Lewis number i interfacial
l latent heat in inlet
M molecular weight is inner surface
NuD Nusselt number lm log mean difference
m _ mass flow rate m mass transfer
P pressure nb non-condensable at bulk
Pr Prandtl number ni non-condensable at interface
R thermal resistance o outer surface
ReD Reynolds number out outlet
Sc Schmidt number s tube surface
St Stanton number tot total
T temperature w tube wall
U overall heat transfer coefficient
V velocity
Y mole fraction

were compared for variables such as flue gas temperature, cooling measurements. Analyses were also carried out to estimate how
water temperature and condensation rate. For tests using flue gas much flue gas moisture it would be practical to recover from boiler
obtained from oil combustion, cooling water flow rate was varied flue gas and the magnitude of the heat rate improvements which
from 1000 to 3700 kg/h while flue gas flow rate was varied from could be made by recovering sensible and latent heat from flue gas.
60 to 176 kg/h. The results showed that the general temperature This paper describes an analytical model of the heat and mass
profiles and total amount of condensate could be predicted well transfer process in a countercurrent cross flow tubular heat ex-
with one-dimensional mass and heat balance calculations. The pre- changer with cooling water inside and flue gas outside the tubes.
diction for total amount of condensate was more accurate than The analytical results are compared to experiments made using a
that for amount of condensate from each heat exchanger. The con- pilot-scale heat exchanger and the performance of the heat ex-
densation started from the flue gas inlet and the main condensa- changer system is shown as a function of various parameters.
tion region was generated at the heat exchanger inlet due to a
high flow rate of cooling water. It was found that cooling water 2. Analytical modeling
flow rate was an important factor affecting condensation rate.
In 2004, Valencia [8] performed a CFD simulation for the con- 2.1. Control volume and main variables
densation of water vapor and acids on the plate using a commercial
code, FLUENT and a user defined subroutine. A numerical simula- The control volume used for deriving the governing equations is
tion using the commercial code and a simulation based on empir- defined as a fixed region in space that encompasses the flue gas
ical correlations using the Engineering Equation Solver (EES) were and the cooling water tubes. A countercurrent cross flow heat ex-
carried out for a two dimensional (2D) vertical water-cooled plate. changer with smooth-wall tubes is presumed. Fig. 1 shows the
Experiments were conducted for the condensation of nitric acid, variables in the control volume which affect water vapor conden-
sulfuric acid and water vapor in the presence of air on a vertical
water-cooled plate. The discrepancies between experiments and
simulation are in a range of 7–25% depending on the combustion
conditions and the average surface temperature of the plate. The
numerical model was applied to real 3D geometries including an
annular fin heat exchanger and a pin fin heat exchanger.
In 2008, Levy et al. [9,10] investigated use of condensing heat
exchangers to recover water vapor from flue gas at coal-fired
power plants. Pilot scale heat transfer tests were performed to
determine the relationship between flue gas moisture concentra-
tion, heat exchanger design and operating conditions, and water
vapor condensation rate. A theoretical heat and mass transfer mod-
el was developed for predicting rates of heat transfer and water va-
por condensation and comparisons were made with pilot scale
Fig. 1. Control volume and variables for analytical modeling.
 K. Jeong et al. / International Journal of Heat and Mass Transfer 53 (2010) 2361–2368
sation. Heat and mass transfer for water condensation are consid-
ered with this control volume. The flue gas temperatures at the in-
let and outlet of this control volume are expressed as Tg,in and Tg,out,
respectively. The parameters Tc,in and Tc,out express cooling water
temperatures at the inlet and outlet, respectively. Tube wall tem-
perature is expressed by Tw. The parameters Tg and Tc are average
values of flue gas and cooling water temperatures between the in-
let and outlet, respectively. Heat transfer in this control volume is
controlled by heat transfer coefficients on the flue gas side (hg) and
cooling water side (hc).
Mass transfer of water vapor towards the wall occurs at loca-
tions where the tube wall temperature is below the local water
dew point temperature. In this paper, the mole fraction of water
vapor at the inlet and outlet are expressed as yH2O,in and yH2O,out,
respectively. The parameter yH2O is an average mole fraction of
yH2O,in and yH2O,out. The dew point temperature of water vapor Tdew
is the saturation temperature corresponding to the partial pressure
of water vapor in the flue gas. The parameter Ti is temperature of
water vapor at the interface between the gas and liquid film. Mass
transfer coefficient is expressed as km.
In the analytical model, both mass transfer of water vapor and
heat transfer from flue gas to cooling water are considered. Enter-
ing the heat exchanger, the flue gas is first cooled by sensible heat
transfer to cooling water. Then, water vapor in the flue gas con-
denses on the cooling water tube surface if the wall temperature
is lowered below the dew point of the water vapor, simultaneously
releasing latent heat that transfers to the cooling water. Two-phase
flow, which consists of the gas phase (uncondensed flue gas) and
liquid phase (water condensate), is assumed on the flue gas side
(Jeong [11]).
2.2. Governing equations and assumptions
Colburn and Hougen developed a fundamental transport equa-
tion for condensation in the presence of a non-condensable gas
[12]. When the wall temperature is lower than the dew point tem-
perature, water condensation occurs on the tube surface as a result
of diffusion of water vapor through the flue gas to the liquid–vapor
interface. Therefore, water vapor exists in the flue gas as a super-
heated vapor at Tg. Sensible heat transfers from the flue gases to
the liquid–vapor interface in addition to the latent heat transferred
from the condensing vapors. The heat transfer to the cooling water
is the sum of sensible and latent heat. The Colburn–Hougen equa-
tion is defined as follows:
hg ðT g  T i Þ þ km  l  ðyH2 O  yi Þ ¼ U o ðT i  T c Þ
Eq. (1) includes the interfacial temperature, Ti, for the cases
when water is condensing. The parameter l is the latent heat of
water vapor. As shown Fig. 2, Uo is the overall heat transfer coeffi-
cient, taking into account resistances from the condensate liquid
film to the cooling water. These resistances expressed in terms of
the associated heat transfer coefficients as shown below:
Fig. 2. Thermal resistances from condensate to cooling water side.
2363
 
1 1 1 1
¼ þ F is þ Rw þ ð2Þ
U o Aeff hc Ais hf Aeff
where Aeff and Ais are the heat transfer areas based on the tube outer
diameter including film thickness and the inner diameter of tube,
respectively. hc and hf are the heat transfer coefficients on the cool-
ing water side and liquid film side, respectively. Fis is the fouling
factor on the inside of the tubes. Rw is the conductive resistance
of the tube wall and is expressed as ln(do/dis)/(2pkwL).
The fouling factor and the tube wall resistance are assumed to
be negligible in this analysis. The thermal resistance due to the
condensate film is negligible since it contributes only about 1–3%
of the total thermal resistance [13]. Thicknesses of water film
and tube are neglected in this research, and subsequently, Aeff
and Ais are replaced by Ao (the heat transfer area based on the tube
outer diameter). Then Eq. (2) is simplified to Eq. (3) with the
assumptions.
1 1
 ð3Þ
U o Ao hc Ao
Applying these approximations, the first governing equation is
established as shown in Eq. (4). The heat transfer coefficients on
the flue gas side (hg) and cooling water side (hc) are calculated
using empirical correlations of forced convective heat transfer for
tube banks and inside a tube, respectively.
hg ðT g  T i Þ þ km  l  ðyH2 O  yi Þ ¼ hc ðT i  T c Þ ð4Þ
The second governing equation, Eq. (5) is the energy balance be-
tween the enthalpy change of the flue gas and the heat transfer
rate from the flue gas to the liquid film. This ordinary differential
equation is integrated over the control volume, which is presum-
ably small enough to assume that all other variables are constant.
_ g  cpg  dT ¼ hg  ðT g  T i Þ  dA
m ð5Þ
For the case of no water vapor condensation, heat transfer from
the flue gas is directly transferred to the tube wall, which has a
temperature, Tw, instead of the liquid–film interface. The third gov-
erning equation is for the case when no condensation is occurring
and is shown in Eq. (6). Eq. (6) reflects energy conservation be-
tween the enthalpy change of the flue gas and the heat transfer
rate from the flue gas to the tube wall surface.
_ g  cpg  dT ¼ hg  ðT g  T w Þ  dA
m ð6Þ
The fourth governing equation, the overall energy balance equa-
ð1Þ tion is derived by modifying the Colburn–Hougen equation (4).
Since the total heat transferred to the cooling water, represented
by the right hand side of Eq. (4), is equal to the enthalpy change
of the cooling water, Eq. (4) can be rewritten as Eq. (7).
_ c  C pc  dT c
½hg  ðT g  T i Þ þ km lðyH2 O  yi Þ  dA ¼ m ð7Þ
For the case of no condensation of water vapor, the interfacial
temperature, Ti in Eq. (7), is replaced by tube wall temperature
Tw, and the mass transfer term is eliminated on the left hand side
of Eq. (7). The fifth governing equation is derived from Eq. (7) as
shown hereunder.
_ c  cpc  dT c
hg  ðT g  T w Þ  dA ¼ m ð8Þ
The sixth governing equation balances the enthalpy change of
cooling water with the convective heat transferred from tube wall
to the cooling water. This is written in Eq. (9).
_ c  cpc  dT c
hc  ðT w  T c Þ  dA ¼ m ð9Þ
The ordinary differential equations, from Eqs. (5)–(9), are inte-
grated over a discretized cell, which is presumably small enough
to assume that the all other variables are constant.
2364 K. Jeong et al. / International Journal of Heat and Mass Transfer 53 (2010) 2361–2368

The condensation rate of water vapor is proportional to both the Subsequently, the mass diffusivity of water vapor in flue gas is
mass transfer coefficient, km [kg/s m2 mol] and the concentration estimated from the known properties of water vapor in air and the
differences in the vapor phase between the interface and bulk flow. calculated thermal diffusivity of flue gas,
This can be calculated by integration of the following ordinary dif-  
ag
ferential equation. DH2 O—gas ¼  DH2 O—air ð18Þ
aair
_ cd ¼ km  ðyH2 O  yi Þ  dA
dm ð10Þ
where ag and aair are the thermal diffusivities of flue gas and air,
To calculate km, the relationship between mass diffusion and respectively. Calculated values of thermal diffusivity and mass dif-
heat transfer obtained from the Lewis relation is used. The Colburn fusivity for water vapor in flue gas are listed in Table 1. The mea-
j factors for heat and mass transfer with their applicable ranges are sured mass diffusivity of water vapor in air at 15 oC and one
defined as, [14,15]: atmosphere is 2.61105 m2/s [18]. Al-mutawa [19] used 0.845 as
a Lewis number for water vapor.
h The interfacial mole fraction of water vapor used in Eq. (10) can
jH ¼ St  Pr2=3 ¼  Pr 2=3 0:6 < Pr < 60 ð11Þ
q  cp  V be calculated by Eq. (19), the Antoine equation [20], which is a va-
kD por pressure equation and describes the relation of saturated vapor
jm ¼ St m  Sc2=3 ¼ ðScÞ2=3 0:6 < Sc < 3000 ð12Þ pressure and temperature for pure components:
V
expða  T ibþcÞ
where h is convective heat transfer coefficient [W/m2 K] and kD is yi ¼
convective mass transfer coefficient [m/s]. The Lewis analogy re- Ptot
quires equating Eq. (11) and (12), that is, jH = jm. Then an expression a ¼ 16:262 ð19Þ
is derived for the mass transfer coefficient as: b ¼ 3799:89
hg  M H2 O c ¼ 226:35
km ¼ ð13Þ
cp;g  M g  ylm  Le2=3
H2 O—gas where Ti is interfacial temperature in oC and Ptot is in kPa.
The heat transfer coefficient on the cooling water side was cal-
The mass transfer coefficient is shown to be a function of the
culated by using Eq. (20) provided by Gnielinski [21].
gas side heat transfer coefficient, the logarithmic mole fraction dif-
ference of the non-condensable gas and the properties of flue gas ðf =8ÞðReD  1000ÞPr
NuD ¼
side including Lewis number of water vapor in flue gas. To predict 1 þ 12:7ðf =8Þ1=2 ðPr2=3  1Þ
the convective heat transfer coefficient on the gas side of a bare ð20Þ
0:5 6 Pr 6 2000
tube bank, an empirical correlation, Eq. (14), proposed by Zhukaus-
kas [16] was used. 3000 6 ReD 6 5  106
 1=4
Pr Eq. (20) can be applied for both uniform surface heat flux and
NuD ¼ 0:27  Re0:63
D;max  Pr 0:36
 constant tube wall temperature. The friction factor f can be ob-
Prs
  ð14Þ tained from the Moody diagram. The Reynolds number in Eq.
0:7 6 Pr 6 500
(20) is based on the average velocity Vc of the cooling water in
1000 6 ReD;max 6 2  105
the tubes which ranged from 0.1 to 0.2 m/s in this study.
The Reynolds number appearing in Eq. (14) is based on the Analytical modeling in this study was developed with the fol-
maximum velocity Vg,max which is in the minimum flow area be- lowing assumptions and simplifications.
tween the tubes. The maximum velocity Vg,max ranged from 0.8
to 4.0 m/s, while the upstream flue gas velocity at the duct inlet  Steady state one dimensional flow.
ranged from 0.3 to 1.5 m/s in this study. All properties are calcu-  Countercurrent cross flow condensing heat exchanger.
lated based as mixtures, and ylm, the logarithmic mole fraction dif-  Two phase flow (gas and liquid) for the flue gas side and one
ference of the non-condensable gas between the free stream and phase flow for the cooling water side.
the wall, is calculated as [5,17]:  Film condensation only occurs on the tube wall surface.
 The thermal resistance due to the film is negligible since it con-
yni  ynb tributes about 1–3% of the total thermal resistance.
ylm ¼ ð15Þ
lnðyni =ynb Þ  Negligible thermal resistance from the tube wall.
 No evaporation of water vapor and no chemical reactions.
where yni and ynb are the mole fractions of non-condensable gases  No heat loss to the environment.
at the gas/film interface and in the bulk flow, respectively.  Flue gas contains CO2(g), O2(g), and N2(g) and H2O(g).
The parameter LeH2 O—gas , the Lewis number of water vapor in
flue gas, is defined as Eq. (16), in which DH2 O—gas is the binary mass
diffusion coefficient of water vapor in flue gas.
Table 1
Sc ag Calculated and reference properties.
LeH2 Ogas ¼ ¼ ð16Þ
Pr DH2 O-gas
Properties Values
The Lewis number is typically of order unity for gases. This im- Calculated thermal diffusivity of flue 1.96105 [m2/s]
plies that changes in the thermal and species distributions progress gas at 15 °C in this study
at approximately the same rates in gases that undergo simulta- Reference of thermal diffusivity 2.22105 [m2/s]
of air at 15 °C [18]
neous heat transfer and mass diffusion processes [18,19]. There-
Calculated mass diffusivity of water 2.56105 [m2/s]
fore, unknown mass diffusivity (DAB;unknown ) can be estimated with vapor in flue gas at 15 °C in this study
known other properties from Eq. (17). Reference of mass diffusivity of water 2.61105 [m2/s]
vapor in air at 15 °C [18]
aknown a
known Calculated Le at 15 °C in this study 0.77
Legas  1 ¼ ¼  ð17Þ Reference of Le [19] 0.845
DAB;known DAB;unknown
 K. Jeong et al. / International Journal of Heat and Mass Transfer 53 (2010) 2361–2368
 The chemical composition of non-condensable dry flue gas is
assumed as 15 vol% of CO2, 3.78 vol% of O2 and 81.3 vol% N2 at
dry basis.
 The U-bends in the tube bank are simplified and modeled as a
straight tube having the same length as the multiple bend tube.
2.3. Numerical scheme
The governing equations for the boundary value problem were
solved using an iterative numerical scheme and a one dimensional
finite difference method with forward differencing. The known
variables were: inlet flue gas temperature, inlet flue gas flow rate,
inlet cooling water temperature, cooling water flow rate, and inlet
mole fraction of water vapor.
The inlet cooling water temperature was used as a target value
and served as the criterion for convergence. In order to calculate
inlet cooling water temperature, a value for the exit cooling water
temperature was initially assumed, and the calculations were car-
ried out. The scheme iterated until the correct inlet cooling water
temperature was calculated. The entire system was discretized
with control volumes in each cell. For this study, the total heat
transfer area, 6.74 m2, was discretized into 1000 cells.
3. Experimental
An experimental study was carried out to validate the analytical
model. Fig. 3 shows the overall experimental set up. The experi-
mental setup consisted of six stages of heat exchangers, an inlet
and outlet duct, an induced draft fan, cooling water lines, and an
electric hot water heater. Flue gas was channeled from the exit
duct of the boiler to the inlet of the heat exchanger system. All heat
exchangers were countercurrent cross flow, with flue gas flowing
around the cooling water tube bundles. Cooling water was distrib-
uted to an inline tube bundle arrangement with eight parallel rows
of multiple U-bend tubes (0.0127 m O.D.). Manifolds were at the
inlet and exit of each counter-cross-flow heat exchanger to distrib-
ute the water. Heat exchangers one through six had 3, 5, 9, 13, 13,
and 13 U-bends. The tube banks had a transverse pitch ST of 0.7 and
longitudinal pitch SL of 2.0. The cooling water was taken from the
local municipal water line. The duct housing the heat exchangers
was thermally insulated using flexible fiber glass and rigid foam
insulation. The first heat exchanger (HX1) was not needed in the
experiments described in this paper because of a relatively low in-
let flue gas temperature.
For data acquisition during the test, thermocouples, rotameters,
and a manometer with a pitot tube were used to measure the flue
Flue Gas Flue Gas
Inlet Outlet
Exhaust
Duct
Cooling
Water Outlet Cooling
Water Inlet
HX 1 HX 2 HX 3 HX 4 HX 5 HX 6
Support Frame
Fig. 3. Schematic of experimental setup for bare tube heat exchanger test.
2365
gas and cooling water temperatures, cooling water flow rate, and
flue gas flow rate, respectively. Flue gas temperature was mea-
sured using sheath-type K thermocouples which were inserted into
the center of the rectangular duct at the inlet of each heat exchan-
ger. There was a rotameter on each heat exchanger measuring
cooling water flow rate. Downstream of the heat exchangers, an
eight foot straight section of PVC pipe was used to create a fully
developed flow region for an accurate flow rate measurement. In
this section, a pitot tube and manometer were used to measure
flue gas flow velocity.
Rigid plastic containers were used to collect condensate from
the bottom of each heat exchanger. The mass of condensate and
elapsed time for the test were measured to determine condensa-
tion rates. Moisture content of the flue gas at the inlet of the heat
exchanger was calculated by adding the measured condensation
rate in each heat exchanger to the vapor flow rate in the flue gas,
on a molar basis. Wet bulb and dry bulb temperature measure-
ments made in the first few tests showed the flue gas at the exit
of the apparatus was saturated. As a result, the exit vapor flow rate
in the flue gas was calculated by assuming the flue gas to be fully
saturated at the cooling water tube-wall temperature at the exit of
the heat exchanger system.
Tests were carried out with flue gas from three different fuels:
#6 oil, natural gas and coal. Tests with the flue gas from #6 oil
and natural gas were performed in the Boiler House at Lehigh Uni-
versity [9,10]. Tests using coal combustion gas were conducted at a
coal-fired power plant. In this paper, test results related only to
flue gas derived from coal firing are presented. All of the data were
obtained by using bare tube bank heat exchangers. The parameters
which could have an impact on the performance of condensing
heat exchangers were selected for experimental study: (1) inlet
cooling temperature, (2) cooling water flow rate, and (3) flue gas
flow rate.
A term used in this study, condensation efficiency, is the
weight% ratio of total condensation rate to inlet water vapor flow
rate. This is used to evaluate the performance of the condensing
heat exchangers. Condensation efficiency is expressed as follows:
P5
_
m
gcd ½wt% ¼ _i¼1 cd;i  100 ð21Þ
mH2 O;in
The numerator represents total condensation rate, which is cal-
culated by summing individual condensation rates from the vari-
ous heat exchanger stages. The denominator represents the mass
flow rate of water vapor at the system inlet.
It was found that the mass flow rate of cooling water is an
important factor affecting condensation efficiency since it is re-
lated to the heat absorption capacity of the cooling water. How-
ever, the mass flow rate of cooling water (m _ c ) is more
meaningful when known relative to the mass flow rate of wet flue
_ g;in ). Moreover, better condensation efficiencies
gas at the inlet (m
were expected with higher ratios of cooling water to flue gas flow
rates.
The ratio of the mass flow rate of wet flue gas to the total heat
transfer surface area ðm _ g;in =Ao ½kg=h m2 Þ is another important
Fan parameter. It is expected that higher ratios of flue gas flow rate
to heat transfer surface area result in poor performance since it
could mean relatively insufficient heat transfer surface area.
4. Results and discussion
Analytical modeling for water vapor condensation in pilot scale
condensing heat exchangers was performed using the experimen-
tal test conditions. Five input variables are needed to simulate the
pilot scale test: inlet flue gas flow rate and temperature; inlet cool-
ing water flow rate and temperature; and inlet moisture fraction.
2366 K. Jeong et al. / International Journal of Heat and Mass Transfer 53 (2010) 2361–2368

Experimental data selected to verify the analytical model were

conducted with inlet cooling water temperatures ranging from
23.9 oC to 32.3 oC.
Fig. 4 shows axial variations of predicted and measured flue gas
and cooling water temperatures. The x-axis represents the normal-
ized cumulative surface area of the heat exchangers starting from
the HX2 in the flue gas flow direction. The operating conditions
for the test are listed under the graph. For this particular test con-
_ c =m
dition in Fig. 4, the ratio of m _ g;in was 2.92. Given this high ratio,
it was expected to have high condensation efficiency even though
the inlet cooling water temperature was relatively high. The ratio
of m_ g;in =Ao was 27.6 [kg/hm2] for this test point.
Predicted and measured data are expressed with lines and sym-
bols, respectively. It is shown with the measured data that the flue
gas was cooled from 149.5 °C to 32 °C while the cooling water tem-
perature was increased from 31 °C to 51.8 °C. As can be seen from Fig. 5. Individual and total condensation rates from each heat exchanger for Test
this figure, there is good agreement between the measured and the 0806 T90c (Inlet wet flue gas flowrate: 185.7 kg/h, inlet flue gas temperature:
predicted values of both flue gas and cooling water temperatures. 149.5 °C, inlet moisture fraction: 14.4 vol%, cooling water flowrate: 542.9 kg/h and
These results show that the theoretical model used in this study inlet cooling water temperature: 31.0 °C).
can predict the heat transfer phenomenon in the condensing heat
exchangers with good accuracy.
Fig. 5 shows a comparison between calculated and measured
condensation rates for the same test point (0816 T90c), in terms
of individual and total condensation rates. Predicted individual
condensation rates agree with measured rates to within 16% and
total condensation rates agree with measured values to within
0.1% according to the errors obtained by the least squares method.
Individual condensation rates have relatively greater differences
than the total condensation rate does. The average uncertainties
of the measured individual and total condensation rates for all
tests were 3.4% and 1.7%, respectively. The analytical model cor-
rectly predicted the distribution of condensate among the five heat
exchangers; the maximum condensation occurring in HX4, and the
minimum occurring in HX2. The condensation efficiency predicted
by the analytical model was the same as what was measured in
experiment, being 67.6%. The results showed that the analytical
model can accurately predict the mass transfer phenomenon for
water condensation among the heat exchangers.
Fig. 6 shows the effect of inlet cooling water temperature on the
Fig. 6. Effect of inlet cooling water temperature on condensation efficiency for coal
condensation efficiency. The results show strong dependence of flue gas condensation modeling (inlet wet flue gas flowrate: 152–192 kg/h, inlet
condensation efficiency on inlet cooling water temperature since flue gas temperature: 140.9–150.9 °C, inlet moisture fraction: 11.9–14.5 vol%,
condensation efficiency linearly decreased from 74 to 46 wt% as in- cooling water flowrate: 281–404 kg/h and inlet cooling water temperature: 24.4–
let cooling water temperature increased from 24.4 °C to 38.1 °C. 38.1 °C).

When inlet cooling water temperature was increased, both tube

Fig. 4. Predicted and measured temperatures of flue gas and cooling water for Test
0806 T90c (inlet wet flue gas flowrate: 185.7 kg/h, inlet flue gas temperature:
149.5 °C, inlet moisture fraction: 14.4 vol%, cooling water flowrate: 542.9 kg/h and
inlet cooling water temperature: 31.0 °C).
wall temperature and interfacial temperature increased. Then,
interfacial water vapor mole fraction increased due to its depen-
dence on interfacial temperature. So the smaller difference in
moisture fraction (see Eq. (10)) reduced the condensation rate
and efficiency.
Fig. 7 shows the effect of the ratio of the mass flow rate of cool-
ing water to the mass flow rate of inlet flue gas ðm _ g;in Þ on con-
_ c =m
densation efficiency. It is also shown that predicted condensation
efficiencies are in good agreement with the corresponding mea-
sured efficiencies. As shown with the measured data, condensation
efficiencies improved from 56 to 67 wt% as m _ c =m_ g;in increased from
1.5 to 3.5. Condensation efficiency is strongly affected by changing
m_ c =m _ g;in . The higher values of m _ c =m
_ g;in result in higher condensa-
tion efficiencies owing to higher thermal mass ðm _ c  cp;c Þ of cooling
water side.
Fig. 8 shows the condensation efficiency as a function of the ra-
tio of inlet wet flue gas flow rate to total heat transfer surface area,
m_ g;in =Ao . It is seen that condensation efficiency linearly decreased
from 77 to 70 wt% as m _ g;in =Ao decreased from 22.4 to 28.3 kg/
h m2, which implies a weak dependence on m _ g;in =Ao .
 K. Jeong et al. / International Journal of Heat and Mass Transfer 53 (2010) 2361–2368
_ c =m
Fig. 7. Effect of m _ g;in on condensation efficiency for coal flue gas condensation
modeling (inlet wet flue gas flowrate: 177–190 kg/h, inlet flue gas temperature:
122.9–152.9 °C, inlet moisture fraction: 12.4–14.5 vol%, cooling water flowrate:
281–662 kg/h and inlet cooling water temperature: 30.9–32.9 °C).
_ g;in =Ao on condensation efficiency for coal flue gas condensation
Fig. 8. Effect of m
modeling (inlet wet flue gas flowrate: 150–192 kg/h, inlet flue gas temperature:
145.9 –150.9 °C, inlet moisture fraction: 11.9–13.5 vol%, cooling water flowrate:
363–401 kg/h and inlet cooling water temperature: 23.9–25.9 °C).
Fig. 9 illustrates predicted condensation efficiency versus cor-
responding measured condensation efficiency for all the coal-
fired tests. The dotted line in Fig. shows the ±10% standard devi-
ation band. All predicted data are within ±10% with an average
discrepancy of 2.5% between experimental data and predicted re-
sults. Therefore, the analytical model was able to predict heat
and mass transfer in the condensing heat exchangers with good
accuracy.
A case study was performed to predict condensation effi-
ciency for 0:5 6 m _ g;in 6 3:5 after the accuracy of the analyti-
_ c =m
cal model was verified. This broad range of flow conditions was
not tested in experiments due to limitations in the experimental
setup. In these cases, cooling water flow rate was fixed at
453.6 kg/h and 31.9 °C to assure fully turbulent flow and the in-
let wet flue gas flow rate was varied from 907 to 129 kg/h.
Fig. 10 illustrates the modeling results and several measured
points for which experiments were performed. The results show
that condensation efficiency decreases steadily with a decrease
in m _ g;in , and condensation efficiencies from 10 to 30 wt% will
_ c =m
occur for values 0:5 6 m _ g;in 6 1:0 .
_ c =m
2367
Fig. 9. Predicted vs. measured condensation efficiencies for coal flue gas conden-
sation (inlet wet flue gas flowrate: 152–192 kg/h, inlet flue gas temperature: 140.9–
150.9 °C, inlet moisture fraction: 11.9–14.5 vol%, cooling water flowrate: 281–
404 kg/h and inlet cooling water temperature: 23.9–37.9 °C).
Fig. 10. Effect of m_ c =m
_ g;in on condensation efficiency for 0:5 6 m
_ c =m
_ g;in 6 3:5
(measured: inlet wet flue gas flowrate: 181–189 kg/h, inlet flue gas temperature:
136.9–152.9 °C, inlet moisture fraction: 12.3–14.5 vol%, cooling water flowrate:
519–662 kg/h and inlet cooling water temperature: 30.9–32.9 °C, predicted: inlet
wet flue gas flowrate: 129–907 kg/h, inlet flue gas temperature: 148.9 °C, inlet
moisture fraction: 14.5 vol%, cooling water flowrate: 453 kg/h and inlet cooling
water temperature: 31.9 °C).
5. Conclusions
An analytical model of heat and mass transfer processes in a
flue gas condensing heat exchanger system was developed using
fundamental heat and mass transfer relations. The modeling ap-
proach is based on conservation of energy and mass for the flue
gas and cooling water. All governing equations were solved using
an iterative solution technique with appropriate assumptions and
simplifications.
Experiments were carried out to validate the analytical model.
Flue gas exhausted from a boiler was channeled to the pilot scale
heat exchanger. Measurements were made and experimental re-
sults were compared to the results from the theoretical model.
The term ‘condensation efficiency’ was defined to make a quan-
titative performance evaluation of the condensing heat exchang-
ers. The ratio of the mass flow rate of cooling water to the mass
flow rate of inlet flue gas was defined as m _ g;in , which was the
_ c =m
most important operating parameter used to evaluate the conden-
sation efficiency.
2368 K. Jeong et al. / International Journal of Heat and Mass Transfer 53 (2010) 2361–2368
The results from analytical modeling, using the same opera-
tional conditions as the experiments, agreed well with experimen-
tal results. The condensation rates in each heat exchanger and
temperature distributions for the flue gas and cooling water were
used to compare the tests. The average discrepancy between the
results of the analytical model and experiments were within a
few percent.
In order to determine how condensation efficiency with rela-
tively low ratios of m _ g;in , the analytical model developed in this
_ c =m
study was used to determine the condensation efficiency. It was
predicted that the condensation efficiency would range from 10
to 30 wt% as the ratio of m _ c =m_ g;in was varied from 0.5 to 1.0.
Acknowledgements
This research and paper were prepared with the support of the
U.S. Department of Energy, under Award No. DE-FC26-06NT42727.
However, any opinions, findings, conclusions, or recommendations
expressed herein are those of the authors and do not necessarily
reflect the views of the DOE. The authors are also grateful to the
Pennsylvania Infrastructure Technology Alliance for providing par-
tial funding.
References
[1] T.J. Feeley, S. Pletcher, B. Carney, A.T. McNemar, Department of Energy/
National Energy Technology Laboratory’s Power Plant – Water R&D Program,
Power-Gen International 2006, 2006.
[2] D.G. Kroger, Air-Cooled Heat Exchangers and Cooling Towers, PennWell
Corporation, USA, 2004.
[3] R.L. Webb, A.S. Wanniarachchi, The effect of non-condensible gases in water
chiller condensers – literature survey and theoretical predictions, ASHRAE
Trans. 80 (1980) 142–159.
View publication stats
[4] M. Osakabe, T. Itoh, K. Yagi, Condensation heat transfer of actual flue gas on
horizontal tubes, in: Proceedings of the 5th ASME/JSME Joint Thermal
Engineering Conference, 1999.
[5] M. Osakabe, Thermal-hydraulic behavior and prediction of heat exchanger for
latent heat recovery of exhaust flue gas, in: Proceedings of the ASME Heat
Transfer Division, vol. 2, 1999.
[6] M. Osakabe, Latent Heat Recovery from Oxygen-Combustion Flue Gas, Energy
Conversion Engineering Conference and Exhibit, vol. 2, 2000, pp. 804–812.
[7] M. Osakabe, K. Yagi, T. Itoh, K. Ohmasa, Condensation heat transfer on tubes in
actual flue gas (parametric study for condensation behavior), Heat Transfer –
Asian Res. 32 (2003).
[8] M.P.P. Valencia, Condensation of Water Vapor and Acid Mixtures from Exhaust
Gases, Dr. Ing Dissertation, Technical University of Berlin, 2004.
[9] E. Levy, H. Bilirgen, C. Samuelson, K. Jeong, M. Kessen, C. Whitcomb, Separation
of water and acid vapors from boiler flue gas in a condensing heat exchanger,
in: Proceedings of the 33rd International Technical Conference on Coal
Utilization and Fuel Systems, 2008. .
[10] E. Levy, H. Bilirgen, K. Jeong, M. Kessen, C. Samuelson, C. Whitcombe, Recovery
of Water from Boiler Flue Gas, DOE Final Technical Report (952467), 2008.
[11] K. Jeong, Condensation of Water and Sulfuric Acid Vapor in Boiler Flue Gas,
Ph.D. Dissertation, Lehigh University, AAT 3354749, 2009.
[12] A.P. Colburn, O.A. Hougen, Design of cooler condensers for mixtures of vapors
with non-condensing gases, Ind. Eng. Chem. 26 (1934) 1178–1182.
[13] M. Goldbrunner, Lokale Phänomene bei der Kondensation von Dampf in
Anwesenheit eines nightkondensierbaren Gases, VDI Verlag. Dissertation
Universität Munchen, 2003.
[14] A.P. Colburn, Trans. Am. Inst. Chem. Eng. 29 (1933) 174.
[15] T.H. Chilton, A.P. Colburn, Ind. Eng. Chem. 26 (1934) 1183.
[16] A. Zhukauskas, Heat transfer from tubes in cross flow, Advances Heat Transfer,
vol. 8, Academic Press, New York, 1972.
[17] K. Stephan, Heat Transfer in Condensation and Boiling, Springer-Verlag, New
York, 1992.
[18] F.P. Incropera, D.P. Dewitt, T.L. Bergman, A.S. Lavine, Fundamentals of Heat and
Mass Transfer, sixth ed., Wiley, 2007.
[19] N. Al-mutawa, Experimental Investigations of Frosting and Defrosting of
Evaporator Coils at Freezer Temperature, Ph.D. dissertation, The University of
Florida, 1997.
[20] C. Antoine, Tensions des vapeurs; nouvelle relation entre les tensions et les
températures, Comptes Rendus des Séances de l’Académie des Science 107
(1888) 681–837.
[21] V. Gnielinski, New equations for heat and mass transfer in turbulent pipe and
channel flow, Int. Chem. Eng. 16 (1976) 359–368.