The Relationship Between The Colburn and Silver Methods of Condenser Design
The Relationship Between The Colburn and Silver Methods of Condenser Design
The Relationship Between The Colburn and Silver Methods of Condenser Design
Heat Mass
Tramfer
Pergamon
PI1 : soo17-!4310(%)oooo1-4
Universitlt Stuttgart,
INTRODUCTION
3147
D. R. WEBB et al
3148
NOMENCLATURE
area
molar concentration
molar heat capacity
molar latent heat
molar diffusive flux
molar mass
molar flowrate
number of components
molar flux of condensation
pressure
heat transfer rate
heat flux
dirt resistance
ratio of fluxes, ii,/&
coordinate direction
thickness of mass and heat transfer
films
temperature
overall heat transfer coefficient based
on outside area
liquid mole fraction
vapour mole fraction
overall mole fraction
r&/& by equilibrium and film methods.
Greek symbols
heat transfer coefficient
mass transfer coefficient for the ij pair
;
finite change of quantity
i
binary diffusivity
4
factor Cri&,/a,, ratio of sensible to
8
conductive heat flux
molar vapour fraction
e
thermal conductivity
mass and thermal rate factors
e;:,
factor for binary mixture.
4
THEORY
Dimensionless
groups
Le
Lewis number
Pr
Prandtl number
SC
Schmidt number.
Subscripts
bub
bubble point
C
relating to condensate
dew
Dew point
eff
effective
relating to equilibrium method
cq
f
denoting film average (flux in &r and
f
g
i,i
0
sat
S
t
t
W
wi
wm
wo
A&)
relating to film method
relating to gas phase
indices giving component number
relating to coolant
saturated
relating to interface
total, as in ri,
thermal layer as in s, and 5,
relating to wall
relating to inside wall
relating to mean wall
relating to outside wall as in U,,.
Superscripts
0
modified variables in equations
(63)
*
saturation quantity.
(60)
Overstrikes
molar quantity as in 2
flow rate as in IQ.
The
which
design,
vective
i=l
to
condensation
rates, ri, the number of moles
cross unit surface in unit time, are needed in
and equation (1) shows a diffusive and a concontribution
to these fluxes. The particular
3149
. .
. .
. .
.
.
.
.
1
I
4
:
I
1
I
90
* .
. .,
.
.
.
. .
.
.
. .
*.
:/
To
,
Coolant
.;
!
1
Caate
Coolant
film
Wall
film
Iii;
Gas
Bulk
film
l%=
conditions of the film model, steady-state and onedimensional, ensure that the fii are constant across the
film. Ficks law of diffusion allows the evaluation of
the Ji, the diffusive fluxes, and therefore the prediction
of mass transfer effects in condensation. The term
y#, is the bulk flow or convective contribution to the
condensation rate, induced by the removal of heat to
the coolant.
With two components, which may be two vapours
or a vapour and a noncondensing gas, (n = 2, y,< l),
Ficks Law gives J, as,
s2
4 = h/B12 where/?,, =~f.
(5)
Equation (7) is readily integrated to give the conductive flux at the interface, 4.
L((T,-G)+tL&-T,)
4s = a(ee_l)
4s = % --+r,-T,)
(,p 1)
where
(8)
D. R. WEBB et al.
3150
The heat transfer coefficient, agr is defined by equation (9) and incorporates,
s,, the hypothetical thermal
film thickness. It is obtained empirically. Equation (8)
shows, through the Ackermann [6] correction factor,
5, = a/(e- l), an effect of mass transfer rate on the
heat transfer coefficient. The following definition of
the modified heat transfer coefficient, a,@, is used :
a,. = ag
i-:
(e.
1)
aPi,
The heat transfer coefficient. ag, is that for conductive transfer alone for, in the limit of a approaching
zero, & approaches
1 and LX,= ago. The similar roles
of $J in equation (6) and E in equation (10) are clear.
Equation
(8) leads also to the rate of gas phase
cooling, &, as heat fluxes, 4, and & in Fig. 1, differ by
the sensible heat change over the gas film. The rate of
gas cooling is less than the heat transfer rate at the
condensate surface.
where
ri, = a,,(T,-
7)/
A&+
&A Tg- T)
{I-exp(-c)j
(14)
where
(15)
c
=
(17)
Price and Bell [ 1l] proposed a two-phase enhancement factor to account for the surface wave effects,
but Owen et al. [ 121 found that this was not beneficial
in predicting their data. In the absence of a reliable
study resolving this contradiction,
such a modification
is not recommended.
The method is approximate,
because mass transfer
effects are not rigorously
included by the above
approach, nor as will be shown later is it always safe
in design relative to the Colburn approach. The above
equations are a sufficient set for the determination
of the local condensation
rates at a point within a
condenser by both the film and equilibrium methods.
The following equation, which appears in Schwaab
[ 131, is important because it allows Z,, = 4,/q to be
obtained locally, i.e. at a single plane by the equilibrium method
as vapour.
3151
-TAQ
csate
Fig. 2. Heat fluxes in an interval.
CL-m
cn,,_ylsj
= exp
-- 1
4
8,2 = ew[-41.
(19)
Tdew
T,
T,
%,a
T.
D. R. WEBB et al
3152
3-
Le=l
Lewis Number
1.2
stam-Helium
E
0.8
0.4
0.0
ERRORS INCURRED
IN USING
THE SILVER
METHOD
- 0.4
- 1
(20)
Lewis Number
rs
Y, -
18
fouling factor may be allowed. The overall heat transfer coefficient may not, as a result, be strongly affected
by the gas side. Nevertheless, the discrepancies are
very significant, and the Colburn method is always
preferable in design.
THE RELATIONSHIP
BETWEEN
AND COLBURN
THE SILVER
METHODS
3153
=-
(21)
5,
-=
5
E (e@-1)
-__
$J(e-1)
(24)
Equations (23) and (24) explain the discrepancy
between the film and equilibrium methods for binary
mixtures. The effective gas side heat transfer
coefficients are in the ratio shown by the slopes of the
condensation paths in Fig. 7. This allows the relationship between the methods to be quantified.
When (dT,/dY& > (dTJdJ&
= (dT,ld.Y&, the
Silver method is unsafe compared with the film
method, as used in industrial design, showing a higher
gas film heat transfer coefficient. This is the usual case,
because, as seen in Fig. 7, a comparison is being made
between the chord and the tangent at B. When Le > 1,
the value of dT,/dj& is even greater and the Silver
method still less safe in design. Conversely at Le < 1,
the methods are in better agreement. If Le is small
enough the Silver method will become safe.
3154
Equilibrium
Method
??
Unmodified
l..aviac0rrccti00 onv
Film Method
A
Unmodified
The features seen in Fig. 5 are now readily understood. It is the difference between the slopes of the
tangent and the chord in Fig. 7 and their appearance
in equation (24) that causes the approximate
method
to be unsafe at Lc = I. The Lewis number, through
the factor Lr. is responsible for the genera1 shape of
Fig. 5. The general variability is the effect of the flux
ratio 5,/f and liquid subcooling.
The role of an increased gas side temperature driving force is also clear, because it serves to increase the
slope of the chord BF in Fig. 7. It is also important
to recognize that the treatment
of the equilibrium
method here is differential. When the cooling curve is
calculated at discrete intervals the condensation
path
follows a chord in Fig. 7, improving the agreement
between film and equilibrium approaches.
The Silver approach is more commonly used in the
industrial design ofmulticomponent
condensers. With
unspecified mixtures. where only a cooling curve is
available, it must be used. Its advantage is that no
further vapour-liquid
equilibrium
calculations
are
needed, irrespective of geometry. It has been corrected
and improved by use of the more physically realistic
film model. but it is shown above that substantial
discrepancies
between the methods can occur. even
when all published correction
factors are included.
The success of the Silver method in industrial practice
must be ascribed to the fact that industrial design is
extremely conservative.
A NEW CORRECTION
METHOD
The experimental
verification of the present work
has been carried out by Fahrner [ 151, using the experimental data of Lehr [5]. Figs. 8 and 9. which are for
binary condensation
of benzene and toluene, respectively, from nitrogen. The comparisons
allow the
assessment of the relative predictions of the film and
equilibrium methods as described in this paper, and
the improvements
that are possible based on appli-
cation of equation (24). which provides a way of correcting the gas side heat transfer resistance to bring
the equilibrium method into closer agreement with the
film model.
(25)
3155
1.0
0.0
Unmdikd
2.0
Equilibrium Method
??
3.0
UlUlNdfficd
4.0
5.0
3156
D. R. WEBB
et al.
8. D. R. Webb, Multicomponent
condensation.
Plenary
Lecture. 9th Int. Heat Trans. Conf., Jerusalem. August
(1990).
9. J. McNaught.
Mass transfer correction terms in design
18th
methods for multi-component-partial
condensers,
Natl. Heat Transfer Co& San Diego, ASME-AIChE
(1979).
10. J. McNaught, An assessment of design methods for condensation
of vapours
from a non-condensing
gas,
ICHMT Seminar on Advances in Heat E.xchangers, Dubrovnik (1981).
Il. B. C. Price and K. J. Bell, Design of binary vapour
condensers using the Colburn-Drew
equations. AIChEJ
Symp. Ser. 8,267-272 (1974).
12. R. G. Owen, R. G. Sardesai and D. J. Pulling. Transfer
coefficients for condensation
in a vertical tube, Paper 80HT-55, 19th Natl. Heat Transfer Corzf. ASME,lAIChE.
Orlando (1980).
of condenser design methods.
13. R. Schwaab, Comparison
Diplomarbeit
University
of Stuttgart/UMIST,
Manchester ( 1993).
14. A. P. Colburn and A. G. Edison, Prevention of fog in
Ind. Engng Chem. 33,457458
( 194 1).
cooler-condensers,
Development
and experimental
vali15. M. E. Fahrner,
dation of condenser design methods, Diplomarbeit,
University of Stuttgart/UMIST,
Manchester (1994).