G. E.

Weber Development of a Method to

Commonwealth Edison Co.,
Fossil Technical Services Dept.,
125 South Clark St.,
Evaluate the Design Performance
Chicago, IL 60603
of a Feedwater Heater With a
W. M. Worek
Department of Mechanical Engineering
Short Drain Cooler
(M/C 251),
University of Illinois at Chicago, A method is developed to determine the shell and tube side heat transfer performance
842 West Taylor St., of a feedwater heater with a short drain cooler. The desuperheating, condensing,
Chicago, IL 60607-7022 and drain cooling zones are discussed and analyzed by deriving a modified version
of the Delaware Method of Shell-Side Design.

1 Introduction
Electric utilities purchase feedwater heaters for new plants, designs courtesy of Marley Heat Transfer Corporation (Biar,
where they purchase several at a time, and older plants, where 1991).
the individual heaters are replaced as tube pluggage becomes
excessive. A high-pressure feedwater heater costs between one
and five hundred thousand dollars depending on its size and 2 Shell-Side Heat Transfer Analysis
design. High-pressure feedwater heaters are essential to the A short drain cooler feedwater heater, shown in Fig. 1, has
performance of a unit. If a unit is designed to have feedwater three zones: a desuperheating, a condensing, and a drain cool-
heaters in service and they are removed, the furnace heat input ing zone. Each zone is treated as a separate heat exchanger in
must increase. This will increase unit heat rate and may cause evaluating the heat transfer coefficient and pressure drop. The
furnace tube overheating due to excessive furnace exit gas heater shown in Fig. 1 has a short drain cooler that has all the
temperature. feedwater tubes passing through it and does not span the entire
Since feedwater heaters are so important to unit perform- length of the heater. This is in contrast to a long drain cooler
ance, utilities need a method to evaluate the heater manufac- (not shown) that does not have all the feedwater tubes passing
turer performance specifications and understand the theory through it and spans the entire length of the heater. This paper
behind a heater's design. Currently there is no method available analyzes a heater with a short drain cooler. Another classifi-
to the utility that uses the heater manufacturer's design data cation of feedwater heaters is whether they are high or low
and accurately calculates the zone heat transfer coefficients pressure. A high-pressure heater obtains its extraction steam
and pressure drops when given only the heater design data. supply from either the high-pressure or intermediate-pressure
There are computer programs that the heater manufacturers turbine and obtains its water supply after the boiler feed pumps.
use to design the heater, but these programs are proprietary A low-pressure heater obtains its steam supply from the low-
and would be costly for a utility to use for the small number pressure turbine and its water supply from a point before the
of heaters analyzed on a yearly basis. boiler feed pumps. The method developed in this paper may
This paper uses the Delaware Method for Shell-Side Heat be used for calculating the heat transfer coefficients and shell-
Exchanger Design (Bell, 1986) and modifies the method for side pressure drops for high-pressure heaters but is currently
feedwater heaters currently in use. The Delaware Method, valid only for calculating the heat transfer coefficients for low-
which was designed in 1960, is an excellent heat exchanger pressure heaters. The design parameters needed to perform the
design tool. However, modern feedwater heaters have baffle shell-side heat transfer analysis are Dot\, S, £>,, D0, ms, ixs, Ks, ,
designs that the method cannot accurately model. Therefore,
the Delaware method needs to be modified to predict heat
transfer coefficients and pressure drops in modern feedwater
heaters accurately.
This paper employs the same concepts as the Delaware
Method in calculating an ideal heat transfer coefficient and
extends them to be valid for higher Reynolds number flows.
The heat transfer coefficient and pressure drop correction fac-
tor curves were calibrated using six actual feedwater heater
" DRAIN "RAIN
OUTLET COOLER
INLET MANWAV
Contributed by the Power Division for publication in the JOURNAL OF EN-
GINEERING FOR GAS TURBINES AND POWER. Manuscript received by the Power Fig. 1 Three-zone feedwater heater with short drain cooler, with a de-
Division March 26, 1993. Associate Technical Editor: R. W. Porter. superheating, condensing, and drain cooling zone

4 3 4 / V o l . 116, APRIL 1994 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use
Copyright © 1994 by ASME
 ZONE CENTER

Single Segmental
Flow Distance (Cr)
CH « 1/2

j^ .866 Dot) J
End VIe»

Double Segmental

CM • 1/3

Triple Segmental

CM • l/<

Cross Flo* A m , S. • Cr«S

Front View

Fig. 3 Crossflow area illustration for the desuperheating and drain

Multi Segmental
cooling zones
CM - 1/8

segmental baffles have one half of the tubes available for

crossflow. The double segmental baffles have only 1/3 of the
tubes available for crossflow. The multiplier keeps decreasing
Fig. 2 Baffle types and crossflow multipliers for the desuperheating as the amount of crossflow decreases. The multipliers are ap-
and drain cooling zones proximate, but are fine tuned later in this paper when a cross-
flow correction factor is discussed. The crossflow area can
then be calculated using the formula previously developed (Bell,
Cs, and the baffle type. All of this information is available 1986) and modified (Weber, 1992):
from the heater manufacturer.
Sm = (CAf)(yl)(S)/(0.866 A,it) = (Cr)S (1)
2.1 Desuperheating Zone. The turbine extraction steam where S,„ represents the crossflow area calculated at the center
enters the heater at the desuperheating zone where it is cooled of the desuperheating zone, and Cr represents the crossflow
to within 5-35°C (10-60°F) of its saturation point. Over the distance. The term A is the semicircular zone cross-sectional
last 20 years this has been the area where heater manufacturers area calculated using:
have tried to maximize heater efficiency by minimizing the
steam-side pressure drop. They have done this by modifying A = x(2>oll)V8 (2)
the baffle design to reduce the amount of crossflow. However, Equation (1) divides by the quantity (0.866 Dolt) because the
in minimizing the crossflow, the heat transfer coefficient is crossflow distance Cr is obtained by dividing the corrected
reduced because there is less steam flowing across the tubes. cross-sectional area of the zone by the horizontal distance at
The manufacturers account for this effect by designing a longer its center. The crossflow distance is then multiplied by the
zone with more area. baffle spacing to obtain the crossflow area. Figure 3 illustrates
To reflect the actual zone crossflow accurately, a crossflow the crossflow calculation in the desuperheating zone. The same
area multiplier must be assumed. The multipliers, which are analysis is valid for the drain cooling zone except that the
shown in Fig. 2, correspond to each baffle type. The single semicircle is inverted.

Nomenclature
Jb = bypass correction factor
a -= baffle window vertical f, = friction factor Jc = crossflow correction factor
dimension h = corrected shell-side convec- Jf = friction correction factor
A == area tion heat transfer coeffi- J, = leakage correction factor
A, ~-= inner area cient Js = baffle spacing correction
Am = = mean area h/g = latent heat of condensation factor
An = = outer area hfWj = feedwater enthalpy at heater k = thermal conductivity
b --= baffle window horizontal inlet */ = feedwater thermal conduc-
dimension hjvio = feedwater enthalpy at heater tivity
c -= baffle window vertical outlet ks = steam thermal conductivity
dimension ht = tube-side heat transfer kw = wall material thermal
CM --= crossflow multiplier coefficient conductivity
Cf == tube-side specific heat h0 = shell-side heat transfer L = length
C 5 == shell-side specific heat coefficient Ls = baffle spacing
D, ~-= inner diameter hsi = condensing-zone steam inlet l^si == baffle spacing at zone inlet
Dm '-= mean diameter . enthalpy ^so = baffle spacing at zone outlet
Do == outer diameter hso = condensing-zone enthalpy at mdr = drain inlet mass flow rate
Doti --= outer tube limit saturation ms = steam mass flow rate

Journal of Engineering for Gas Turbines and Power APRIL 1994, Vol. 116/435
Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 Jf( X 10 4 )
!;'','[' KIRS 1 ill ill PI! 11
liflittttttlllllllliil 111; 11 J 1 n iifjiffinct^t -H*±^±iiTri till i EfTiiTEflfltl'fff^^ FfM
T il '7 |'
' 1 II IMlllittlttltlllltt
J - ^

''Pllllllll'llfltfl" ijii iljj iiii |j!ftfMii^Mi]f] 11 j| Ifi 1jljn^wlfrn P'I,

:nBl Uilli 1 1 filflliiilililllll"' jiii ill:|!!HJ!}i[HsS^ffmTiTmmllfttjitfiJ*HliH
: ; • • • • ; •

<„ nil 1 IrffliiiSm hi

;;; ' ' * ' " i l l Hffll 1 Irflliin^ 11 liflllp jjijliijiyjll^
iliiL ti;i;ti 'In llllBlPllliill^BBiflllllimBB _._
" M MiliiiliEffllililili|PIP issiiiifit
If :
ttffl- # ''lit HffiflfflS^p \M\\w lil^fflfiriBPnBHM!tii 111 E l
rjii —
1 • 1] ;:}! 1*1; j | j | j p tpr-
Ttffi ' "ii'-'1wllliOTiii
PiilfTnmfflil P 1 •-
ii.i.ilillllflilflliffltil pilp 1 tl|ji! l i p p i l l ^lllHTOMflffi I I •""*

Blifflflllliwlipl'' III lii

1
h:!!]:H 'i ' I 1 ill;
i " - ^
i [ 4I'llfili'iii III111'^' Illlil
1 llBIM 1
Hi 1 ffililllilife l l l i p •N^llliliflTl ^ W M f i l 1 mil' illliiiH'n
f l i" ! !
;; -
•:M
if | | lijiil[: ff lliiPti.
:
•

jj ! ! L :i'i
If;i;i|;:
jiiiMi j!:i i l l !
: :
;i |!|!
-*i£j
F
1 III
iiiilxiH;
'''111 H i "IKilllHI|:!ip iii'i^
ji|ji|!i'|ijiiii!li |i-:|j:ii||i"j:j!=j:!;
• •!; : : ::;|!:;iji;B
Ililj! 1 ii'i If t t 1 1
iiiiliiiiliiiiiisillilflfnftlflijlliillllilll ii !,' I'll [
j •'•'

i': ;: H :
:::!!!!!: i!i ;!:: jji; |!jii|!ij:|jl;;|i;;:{:;: !i f i i U u - K -c^
• M T; I =: J • : ( ; : : ; • :
;•;•;:!::

iigjiii: r lilt •I'I f llnllllillwlHlllilllii'i'lH

lil'liijl^
lljiljl iiiiisil jljiiiiipiiik 1111111 H i
*!n!!! =ii|l|j!nHfffflfiPiifflh^i ' • - : : :

Ijjj j |j | jj;P!
Hilt1*!! Hill
':: .:::
13 SffflPMB f mffflffl mm ifiw i"r-tiiilp-i-
w t lMPTFE! ^ B l M l i i•§lIF rwlr- z—
i i ^iiiiHmfliptiiiiiiii'n -r-S
ii";!:;;!:
r
- i? *Pt I i I i niin P'-il-'i i! l | i 1
•;;; f!! huliiiTrf-ffjifffnHfff i ifiii M
%-Li'f[: l •iiiill;
!iiii :ijiipijlfflffj
lli! ill
11
.. 1
' -
i'f!' It! P ' !:!!!'•sislaf n4m ' tl
. [ ; -: i

IIIflfllflP
1111 BT-
" if
I|| 1 ||||i|i;ij
IjliJ fj}}| t
:
:¥- uTi 'Hi
III ;
iiiiii iillll illi PI llf 1

II
I::;:;;;:;: ii ii:; iiii if:: ii
- - ...
:•: i:i!:: :E Niljl IP |''liira-jr- r-
1 • - • • • • :
~;« ipi(ii!|i|ii il
:

W
: ;
j

!!!!lllli iji,ii::.l 1..

10 100 1000
4
Re, ( x 10" )
Fig. 4 Correlation of J factor for ideal tube banks

This allows the calculation of a Reynolds number, which is porating the crossflow area calculated in Eq. (1), we obtain
proportional to the crossflow area given by Eq. (1) and the (Bell, 1986):
steam mass flow rate (Bell, 1986). The Reynolds number is
defined by: jjc,m, r ks n °- 6 6 7
K (4)
Res = D0ms/(ixsS,„) (3)
where Res is proportional to a friction factor Jf, which is The base coefficient is calculated assuming that the entire
obtained using Fig. 4. The friction factor decreases as Res shell-side stream flows across the ideal tube bank formed by
increases because the losses will increase due to higher veloc- the tube array at the centerline of the zone (Bell, 1986). This
ities. is the greatest amount of heat transfer attainable. The value
The base heat transfer coefficient is calculated by utilizing obtained is then multiplied by a series of correction factors:
the Colburn J factor for heat transfer. The / factor is given Jc,Ji,Jb, and Js. These correction factors adjust the base heat
by (Bell, 1986; Incropera and Dewitt, 1981): transfer coefficient (Bell, 1986). They are defined next.
h
0.667 Jc: The Crossflow Correction Factor. This factor is ob-
//=St(Pr)°- 667 = tained using Figs. 5 and 7 where it is plotted versus S„, (Weber,
pvCs 1992). The crossflow correction factor decreases with decreas-
ing crossflow area. As the crossflow area decreases, the amount
where St and Pr are the Stanton and Prandtl numbers, re- of crossflow decreases, which causes a reduction in the heat
spectively. Solving for the heat transfer coefficient and incor- transfer.

feedwater outlet tempera-

mx = fraction of drain inlet that Qs heat transferred from steam ture from zone
flashes Re condensing heat transfer feedwater temperature leav-
K = number of tubes in cross- resistance ing drain cooler
flow Re Reynolds number T • 1
SI
steam inlet temperature
P = pitch Rfo shell-side fouling resistance T steam outlet temperature
Pr = Prandtl number Rh tube-side heat transfer 1
so total zone heat transfer
St = Stanton number resistance coefficient
Vcond = heat transferred in condens- RT total resistance u baffle window vertical
ing zone RIS total side fouling resistance dimension
Qdc = heat transferred in drain Rw tube wall resistance w = quality of drains inlet satu-
X
cooling zone
Qds = heat transferred in desuper-
s spacing
crossflow area AP 6
rated steam
Sm baffle pressure drop
heating zone t thickness window pressure drop
Qdsc = desuperheating steam heat T
1
feedwater inlet temperature A7},„ log mean temperature
CI
release in condensing zone zone difference
Qfw = heat transferred to feed- T • feedwater inlet temperature fis = steam viscosity
' ci.cond
water to condensing zone (if = feedwater viscosity

4 3 6 / V o l . 116, APRIL 1994 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 Sm m 2 X 10"2 (ft 2 ) 2260 rn/i kg/m288c (lb m /ft8 88o)
(460)
(0.23]
2.10
2 0 0 0 - (400)
(0.22! £.

2.00 >'*'
~? .
5
L_ ^
^ 1760 — (360) ixx
7
j *

(0.21
7 ^_
1.90- .^ 1800 (300)
yy
«/
(0.201

1.80 1260 (260)

(0.19!
^
s'
i^^\- 1000 (200)
1.70 y
(0.18:
/
£ ,'
(160)
tx
760
1.60
(0.17
600 (100) ir
1.50
(0.18:
250 (60)

1 40
- to.i6:
0.14 0.16 0.18 0.17 0.18 0.19 0.20
0.46 0.60 0.66 0.60 0.66 0.70 0.76
Jo \
Fig. 5 Crossflow correction factor for the desuperheating zone Fig. 6 Leakage correction factor for the desuperheating and drain cool-
ing zones

//: The Leakage Correction Factor. This factor includes conduction heat transfer. Condensing heat transfer, Rc, is ob-
both shell-to-baffle and tube-to-baffle leakage. It is obtained tained as (Clemmer and Lemezis, 1965):
using Fig. 6 where it is plotted versus the ratio of ms to S,„ Rc = 0.0704592 m 2 o C/kW (0.0004 hrft 2 °F/Btu)
(Weber, 1992). As ms/S„, increases, more steam or water passes
through the zone for a given crossflow area. This causes more if r s a l >160°C(320°F)
leakage to occur. Rc = 12.03769 (1.8 r sat + 32)-°- 8912 m 2 °C/kW
Jt,: The Zone Bypass Correction Factor. This factor is
defined as the steam that passes through the zone along the (0.06834 (r s a t r 0 - 8 9 1 2 hrft 2 °F/Btu)
outer circumference and doesn't come in contact with any tubes if 7; a t<160 o C(320°F) (6)
or baffles. Manufacturers try to minimize leakage by using The only heaters tested in the derivation of this paper had
sealing strips or other methods but still leakage cannot be steam saturation temperatures above 160°C (320°F).
completely eliminated. Design data to calculate this factor are
The tube-side convective resistance R/, is calculated by mul-
very difficult or impossible to obtain from the heater manu-
tiplying hc by the ratio of the tube inlet and outlet diameters
facturer. Therefore this factor is assumed as 0.9, which is
and taking the inverse (Clemmer and Lemezis, 1965). This is
representative of a "well-designed" heater (Mueller, 1986).
expressed as:
Js: The Inlet and Outlet Baffle Spacing Correction Fac-
tor. Depending on the heater design, this factor could be \/Rh = hiDi/D0 (7)
greater than or less than 1 if the inlet and outlet spacing is The tube-side conduction resistance Rw is calculated by using:
closer than or greater than the remaining baffles. This factor
is obtained using Fig. 8 where it is plotted versus NB and ls R„ = tD0/(kDm) (8)
(Bell, 1986). The tube-side fouling resistance is calculated using:
The corrected shell-side zone heat transfer coefficient is de- Rts = 3.522xlO-5D0/Di m 2 o C/kW
termined using:
(0.0002 D0/Di hrft 2 °F/Btu) (9)
h0 = hbzxJcJtJbJs (5)
The total resistance is the sum of the resistances, given by:
The correction factors JC,JL, and Jb are not those calculated
using the conventional Delaware Method. New equations were R, = RC + Rh + R„ + Rls. (10)
formulated for these factors to reflect the behavior of modern The condensing zone heat transfer is the inverse of R,:
feedwater heaters accurately. The factor Js is obtained by using
the same graph as the Delaware Method for turbulent flow. U=l/R, (11)

2.3 Drain Cooling Zone. The drain cooling zone heat

2.2 Condensing Zone. The condensing-zone shell-side transfer coefficient is calculated in a manner similar to the way
heat transfer coefficient is the easiest to obtain. The method it is calculated in the desuperheating zone with some exceptions
used is explained in detail by Clemmer and Lemezis (1965). noted below.
The shell-side heat transfer coefficient is obtained by inverting The same steam mass flow rate used in the desuperheating
the sum of the heat transfer resistances. There are four re- zone is used in the drain cooling zone if there is no drain inlet
sistances to calculate: condensing heat transfer; shell-side foul- to the heater. If a unit is designed with multiple feedwater
ing; tube-side convective heat transfer; and tube-wall heaters, the condensate exiting a heater is usually sent to the

Journal of Engineering for Gas Turbines and Power APRIL 1994, Vol. 116/437
Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 Sm m2 x 10'1 (ft2)
1.7S)
; ; ; | ] i ' •

1.60-
'' ! \ 1\ \
1.36) ;

i
i

,
1
1.60 I
\'s :
1.88) 1 . ^.
1.40 1
s '.

1.4S) '
J_j_ \ \
1.80-- \S \
• v

1.3S) ; ! V
X-
•
1.20- : ,
; •

• ~t T
1.2S)
/
1
, ! ; | 1 /
'
1.10 1
1
i
j_-t__^f' '
1.18) y
ill ;. ' s
10 14 18 22
I ! j j

1 . 0 0 - - 1.08) •
• '

!
•
.1 NB. NUMBER OF BAFFLES
/ I Fig. 8 Baffle spacing correction factor for the drain cooling and de-
. S^
y ' superheating zone (courtesy of Taylor and Francis, Washington, DC)
0 . 9 0 - - :0.9B)
A ;
y, ;
;1 •
1

y i ' 1
1 !
0.88) y i ;1 ;'
0.80 1
s/
• |

1 f0 (13)
/ 1\ 1 h hD k„A,„ hi A;
/
• |

0.78) \ : : ;l !1 ! 1 1 where Rf„ is the outside fouling resistance. This is typically

0.70 - 1 - / .
\ I

o.3o o.to 0.70 Q.m o.m 1.00 1.10 i.ao taken as 5.283 X 1(T 5 m 2 o C/kW (0.0003 hrft 2o F/Btu).
Jo
Fig. 7 Crossflow correction factor for the desuperheating and drain 5 Pressure Drop Analysis
cooling zones
5.1 Desuperheating Zone. As mentioned previously, the
design of the desuperheating zone has been improved signif-
heater upstream in the feedwater flow path. On a horizontal icantly over the last several years. Heater manufacturers have
heater, the drain inlet from an upper heater flows into the minimized the steam-side pressure drop in this zone by using
condensing zone as in Fig. 1 and mixes with the condensate triple and multisegmental baffles. Common multisegmental
already in the heater as it flows to the drain cooling zone. zone pressure drops range from 3.5 to7.0kPa(0.5 to l.Opsia).
Because of the increased crossflow in the drain cooling zone, Modifying the original Delaware Method to calculate the
its heat transfer coefficient could be as much as ten times pressure drop accurately in this zone is not a straightforward
greater than it is in the desuperheating zone. Figure 7 gives task. This is an area of considerable uncertainty even for the
the correction factor Jc as a function of Sm. The curve has a heater manufacturer. Because of the low-zone pressure drop
parabolic shape because as the amount of crossflow increases it is very difficult to calculate accurately. The modified Del-
there exists a point where further increasing the amount of aware Method given in this paper can be as much as 30 percent
crossflow will result in channeling and poor flow distribution, higher or lower than the value quoted by the manufacturer.
which will reduce heat transfer. For example, if the method calculates a 4.8 kPa (0.7 psia)
The leakage correction factor is calculated the same way as pressure drop, the actual pressure drop could be 3.36 to 6.24
it is in the desuperheating zone using Fig. 6. The leakage will kPa (0.47 to 0.93 psia). Therefore, the method is accurate to
usually be less in this zone because the crossflow area S„, is within 1.4 kPa (0.23 psi) for this case. This is a relatively small
greater. However, if the drain inlet flow from the upper heater amount of error in the overall pressure drop. When analyzed
is sufficiently large, it could result in more leakage for the in this manner, the method's accuracy seems acceptable, es-
drain cooling zone. Therefore, one should never assume that pecially in light of the fact that the manufacturer is not certain
the drain cooler leakage is the same as the desuperheater leak- of his published pressure drop.
age. The method calculates two kinds of pressure drops and sums
them over the entire zone length. The pressure drop APb is the
pressure drop in one crossflow section if there is no leakage
3 Tube Side Heat Transfer Analysis
or bypass flow. The pressure drop APW is the pressure drop
The tube-side heat transfer coefficient is calculated using in one baffle window section if there is no leakage or bypass
the Dittus-Boelter relation given in Eq. (12) (Clemmer and flow. The crossflow pressure drop is given by (Bell, 1986):
Lemezis, 1965). The information needed to complete this por-
tion of the analysis is the feedwater properties kf, nf, pf, Cj, 2fiih*Nc
APfi = (13)
and V, PsgcSm
/•
where/- is the friction factor plotted versus Res in Fig. 9 and
h,= (*//A)(0.023)Re°-8Pr0-4 (12) Nc is the number of tubes in crossflow given by:
where Pr = C^ij/kj and Re = pf VfDj/jXf. (CM)(A) Sm
Nc (14)
Dot[(0.866)P S(P)-
4 Overall Zone Transfer Coefficient The baffle window pressure drop is given by:
The overall heat transfer coefficient is calculated by taking
the inverse of the sum of the heat transfer resistances given in -AP„ = (15)
Eq. (7) (Bell, 1986): gcS wPs

4 3 8 / V o l . 116, APRIL 1994 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 Re 8 (x10- 5 )
Fig. 9 Pressure drop correction factor for the desuperheating zone

These pressure drops are then corrected for leakage, bypass,

and baffle spacing using Jh Jb, and Js, respectively. These are
1 10 100
the same three factors used in the heat transfer analysis. This
is in contrast to the Delaware method, which derives new baffle Re ( x 10"4 )
leakage, bypass, and spacing correction factors in calculating
Fig. 10 Pressure drop correction factor for the drain cooling zone
the pressure drop.
The total zone shell-side pressure drop is given by (Bell,
1986):
APs = ((Nb-l)APbJb + NbAPw)J,+ 2APbJbJs (16) APW= y : (is)
where Nb is the number of baffles given by:
Equation (18) may be used in the drain cooling zone because
Nb = L-L^Lso+L (1?) the baffles in this zone are more conventional (i.e., single and
s double segmental), which is the type of baffles that are better
suited for the Delaware method.
Equation (12) for APb is the same equation used in the
original Delaware Method except that a different curve is used 6 Calculation of Feedwater Outlet Temperature
for/; and the viscosity ratio is not used. Equation (15) for
APW0 is different from the original Delaware Method in that After a heat transfer coefficient and pressure drop for each
it assumes that there is no crossflow in the baffle window and zone has been obtained, the last task in evaluating the feedwater
it divides by the quantity S2W instead of S,„SW (Weber, 1992). heater design performance is to calculate the feedwater outlet
These modifications were necessary because of the reduced temperature and compare it to what is guaranteed by the heater
amount of crossflow and increased amount of baffle window manufacturer. The heater manufacturer guarantees a terminal
area in modern feedwater heaters. If the baffle window area temperature difference (TTD) and drain cooler approach
were used to predict the pressure drop in this zone the calculated (DCA) for the heater.
values would be as much as ten times greater than what is
quoted by the manufacturer. 6.1 Drain Cooling Zone Heat Loss (Qdc). The drain cool-
ing zone heat loss Qdc is calculated first because this is where
5.2 Condensing Zone. The pressure drop in the con- the feedwater flows initially as it enters the heater. Therefore
densing zone is negligible and assumed equal to zero. This is the feedwater inlet conditions are unknown. The shell-side
common practice when analyzing feedwater heaters. As the condensate outlet conditions are also given by the manufac-
steam exits the desuperheating zone it expands to fill the entire turer in this zone and is used as a first guess in calculating the
condensing zone and then falls to the bottom of the heater as heat transfer. This guess may need revision after the total zone
it condenses. Therefore there is negligible crossflow or window heat transfer is calculated. The shell-side inlet condensate is
pressure drop. assumed to be saturated liquid (quality = 0.0) at heater op-
erating pressure.
5.3 Drain Cooling Zone. The drain cooling zone pressure The shell-side heat transfer may now be calculated and is
drop is calculated using the same equations as the desuper- given by:
heating zone except for the baffle window pressure drop, APW.
The window pressure drop is calculated with the same equation Qdc = ms(hsi-hso) (19)
used in the Delaware method (Bell, 1986): Using this value and the Law of Conservation of Energy, the

Journal of Engineering for Gas Turbines and Power APRIL 1994, Vol. 116/439
Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 Semicircle
of Radius
R —.

Baffle W n d o w
Baffle

b Is given by the heater manufacturer

a Is calculated using:
b is given by the beater manufacturer Baffle
» 2 - R2 - <R-b) 2
a Is calculated using:
t2 - R2 - (R-u) 2
a 2 . R 2 - (R-b) 2 Stf - area of baffle window
S„ . area of baffle window - _2_ (ab + uv)
3
- _2_ <a) ( b )
3

End View

Fig. 12 Baffle window area calculation for double segmental baffles

End View

Fig. 11 Baffle window area calculation for single segmental baffles

heat transfer to the feedwater is equal to the heat lost by the

condensate, Qdc.
Therefore:
Qdc = ms(hsi - hso) = ms„(hfwo - hfwi) (20)
The feedwater side outlet conditions may now be obtained by
solving Eq. (20) for hfwo.
The last step in evaluating the drain cooling zone heat trans-
fer is to set Eq. (20) equal to the overall heat transfer for the
zone:
Qdc = ms(hsi - hso) = mfw{hfwo - hfwi) = U(A )AT /m (F) (21)
where:
U is the calculated overall zone heat transfer coefficient
A is the calculated total zone area
ATi„, is the log mean temperature difference calculated using:
V-* si •* co! ~ \* so ^ ci)
AT,, (22)
T -T
In 1
SI ±
CO
T
1
-T
SO * CI

F is the log mean temperature difference correction factor F

calculated using the method explained by Bowman et al. (1940).
The correction factor is obtained by calculating two temper-
ature ratios, R and P:
S„ • Total Semi-Circle Area - Total Baffle Area
R = (TSi — TS0)/(TC0 — Tci)
P=(Tco-Tci)/(Tsi-Tci) - ( W 2 X R 2 ) - 2RC

These values are utilized to obtain a pure crossflow correction Fig. 13 Baffle window area calculation for multisegmental baffles
factor, CF:
(R2+lf\ 1
log — adjusted with the crossflow multiplier discussed previously.
R-l l-PR The adjusted correction factor F i s given by (Weber, 1992):
CF=- (23)
(2/P)-l-R + (R2+l)W2 F= 1 - [CM(1 - CF)]
log
(2/P)-l-R-{R2+l)W2
The outlet condensate conditions must be adjusted by trial
CF assumes that all of the tubes in the zone are in crossflow. and error until Eq. (21) is satisfied within 0.05 percent error.
However, this is not the case in any of the feedwater heaters The calculated value of the feedwater outlet temperature of
that the author has analyzed. Therefore the log mean tem- the zone, Tcodc is the inlet temperature to the next zone it
perature difference crossflow correction factor CF must be enters, which is the condensing zone. That is: TCOidc= Fc,iCOnd.

440 / V o l . 116, APRIL 1994 Transactions of the ASME

Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use
 The heat loss by the shell-side fluid in the condensing zone Table 1 Heat transfer coefficient and pressure drop comparison
is the addition of three sources: Qdsc, Qcd, and Qdr. HEAT TRANSFER COEFFICIENT AND PRESSURE DROP COMPARISON
HEAT TRANSFER COEFFICIENT PRESSURE DROPS
ZONE
6.2 Condensing (QCOnd)- (1) Qdsc = the heat release by AVG. % ERROR MAX .% ERROR AVG. % ERROR MAX. % ERROR
the 6-33°C (10-60°F) superheated steam as it enters the zone
from the desuperheating zone. This heat release is given by: DESUPERHEATING" 4.6% 10.5% 27.8 72.5%

Qdsc = ms{hsi-hs0) (24) CONDENSING 1.4% 1.8% NA NA

where: hsi is the inlet enthalpy to the zone; hso is the saturated DRAIN COOLING \.\°k 2.4% 2.8 4.1
vapor enthalpy at operating pressure. Using the same logic for
the shell-side condensate: •Four of the six heaters analyzed had desuperheating zones. ThB other two had only a
condensing and drain cooling zone.
—T
si,dc J so.cond
(2) Qcd = the latent heat of condensation of the steam is given 6.4 Thermal Temperature Difference and Drain Cooler
by: Approach. The utility normally purchases a feedwater heater
Qcd = mshfg (25) by specifying a terminal temperature difference (TTD) and
where hje is the latent heat of condensation at heater operating drain cooler approach (DCA). The terminal temperature dif-
pressure. ference is defined as the saturation temperature at the extrac-
(3) Qdr = the condensation of drain inlet. If condensate from tion steam inlet temperature minus the feedwater outlet
an upstream heater flows into the heater being analyzed the temperature:
condensate flashes (vaporizes) due to the pressure differential. TTD = TSM,ext - Tfwo (32)
The amount that flashes must be recondensed. This heat release
is given by: The drain cooler approach is defined as the drain cooler
condensate outlet temperature minus the feedwater inlet tem-
Qdr = mxhfg (26) perature to the drain cooling zone:
where hjg is the latent heat of condensation evaluated at the r>CA= TCOidc-Tfwi (33)
drain inlet temperature and pressure and mx is the fraction of
the drain inlet that flashes given by: 7 Conclusions
mx = Xmdr (27) This paper has derived a method to evaluate the heat transfer
The value X is the quality of the saturated steam evaluated at coefficients and pressure drops in a modern feedwater heater.
the heater operating pressure (ASME, 1967). For the six feedwater heaters utilized to design the method,
the average and maximum percentage of error of the heat
The total condensing zone heat release is the addition of the transfer coefficient and pressure drops are listed in Table 1.
three sources: The heat transfer coefficients exhibit excellent correlation if
Qcond = Qdsc + Qcd + Qdr (28) the calculated Reynolds number is below 106. Above 106 the
The zone feedwater outlet temperature is calculated in a percentage of error will be 10 percent or greater. All heat
manner similar to the drain cooling zone with the use of Eq. transfer coefficients in the drain cooling and condensing zones
(21), except that since there is zero crossflow in this zone, correlate very well.
F=1.0. The percentage of error in the desuperheating zone is high
The zone inlet steam temperature must be adjusted by trial because this is an area of considerable uncertainty, even for
and error until Eq. (21) is satisfied within 0.05 percent error the heater manufacturer. This is due to the very low shell-side
(Weber, 1992). pressure drop (often < 1 psia) and high Reynolds number.
The calculated value of Tco leaving the condensing zone is Note that the method is not valid for drain cooling zones for
the inlet temperature to the desuperheating zone. That is: low-pressure heaters. A new friction factor curve needs to be
calibrated for the drain coolers of low-pressure heaters. How-
-* co,cond — ^ ci,ds \^J ever, since the condensate enthalpy is relatively unaffected by
Similarly, the outlet steam temperature from the desuper- a pressure change of less than 10 psia, the overall effect of
heating zone is equal to the inlet steam temperature to the drain cooler pressure drop on heater performance is minimal.
condensing zone: The heat transfer coefficients and pressure drops can then
be used to calculate the end results, which are the terminal
* si,cond ~ * so,ds \^^)
temperature difference and drain cooler approach. The tem-
perature differences may then be compared to what the heater
6.3 Desuperheating Zone (Qds). The heat supplied in the manufacturer specifies.
desuperheating zone is given by:
Qds = rns{hsiids-hSOtds) (31) References
The zone feedwater outlet temperature is then evaluated as ASME, 1967, Steam Tables, New York.
it was in the drain cooling zone using Eqs. (7), (21), (22), and Bell, K. J., 1986, Delaware Method for Shell Side Design, pp. 129-166, cited
(23), except that the steam conditions in this zone are fixed by in Palen, J. W., Heat Exchanger Sourcebook, Hemisphere Publishing Corp.,
the utility's unit that balance at the zone inlet and by the already ,New York.
Biar, M., 1991, "Feedwater Heater Design Data," Marley Heat Transfer
calculated zone outlet conditions that were used to balance the Corp., Houston, TX.
condensing zone. Therefore if Eq. (21) is not satisfied, the Bowman, R. A., Mueller, A. C , and Nagle, W. M., 1940, "Mean Difference
extraction steam flow to the heater must be adjusted and the in Design," ASME Trans., Vol. 62.
Clemmer, A. B., and Lemezis, S., 1965, "Selection and Design of Closed
feedwater and steam conditions recalculated beginning with Feedwater Heaters," presented at the ASME 1965 Winter Annual Meeting,
the drain cooling zone. This trial and error process proceeds Chicago, 1L.
until Eq. (21) is satisfied for all three zones. Incropera, F. P., and Dewitt, D. P., 1981, Fundamentals of Heat Transfer,
The feedwater outlet temperature calculated in this zone is Wiley, New York, pp. 302-303.
Mueller, A. C , 1986, Shell and Tube Exchanger Design, pp. 167-171, cited
the final feedwater outlet temperature from the heater. This in Palen, J. W., Heat Exchanger Sourcebook, Hemisphere Publishing Corp.,
value is then compared to what is desired or to what the feed- New York.
water heater manufacturer specifies. Weber, G. E., 1992, Master's Thesis, University of Illinois at Chicago.

Journal of Engineering for Gas Turbines and Power APRIL 1994, Vol. 116/441
Downloaded From: https://gasturbinespower.asmedigitalcollection.asme.org on 06/30/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use