HEAT EXCHANGER

DESIGN
Bahan Perkuliahan Perpindahan Panas
By

Dr. H.M. Djoni BUSTAN

BASIC DESIGN PROCEDURE AND THEORY
The general equations for heat transfer across a surface is:
Q = U A AT
m
(1)
where Q = heat transferred per unit time, W,
U = the overall heat transfer coefficient, W/m
2 0
C,
A = heat-transfer area, m
2
,
AT
m
= the mean temperature difference, the temperature
driving force,
0
C.
The prime objective in the design of an exchanger is to determine the
surface area required for the specified duty (rate of heat transfer) using
the temperature differences available.
OVERALL HEAT-TRANSFER COEFFICIENT
The overall coefficient is the reciprocal of the overall resistance to
heat transfer, which is the sum of several individual resistances.
Which are the reciprocals of the individual resistances is given by:
i i
o
id i
o
w
i o o
od o o
h d
d
h d
d
k
d d d
h h U
1 1
2
) / ln( 1 1 1
+ + + + =
where U
o
= the overall coefficient based on the outside area of the
tube, W/m
2 0
C,
h
o
= outside fluid film coefficient, W/m
2

0
C,
h
i
= inside fluid film coefficient, W/m
2

0
C,
h
od
= outside dirt coefficient (fouling factor), W/m
2

0
C,
h
id
= inside dirt coefficient, W/m
2

0
C,
k
w
= thermal conductivity of the tube wall material, W/m
2

0
C,
d
i
= tube inside diameter, m,
d
o
= tube outside diameter, m.
(2)
The steps in a typical design procedure
are given below:
1. Define the duty: heattransfer rate, fluid flowrates, temperatures.
2. Collect together the fluid physical properties required: density,
viscosity, thermal conductivity.
3. Decide on the type of exchanger to be used.
4. Select a trial value for the overall coefficient, U.
5. Calculate the mean temperature difference, AT
m
.
6. Calculate the area required from equation (1)
7. Decide the exchanger layout.
8. Calculate the individual coefficients.
9. Calculate the overall coefficient and compare with the trial value. If
the calculated value differs significantly from the estimated value,
substitute the calculated for the estimated value and return to step
6.
10. Calculate the exchanger pressure drop; if unsatisfactory return to
steps 7 or 4 or 3, in that order of preference.
11. Optimize the design: repeat steps 4 to 10, as necessary, to
determine the cheapest exchanger that will satisfy the duty. Usually
this will be the one with the smallest area.
OVERALL HEAT-TRANSFER COEFFICIENT
Typical values of the overall heat-transfer coefficient for various types of
heat exchanger are given in Table below. More extensive data can be found
in the books by Perry and Chilton (1973), TEMA (1978), and Ludwig (1965).
Shell and Tube Exchangers
Hot fluid Cold fluid U (W/m
2 0
C)
Heat exchangers
Water
Organic solvents
Light oils
Heavy oils
Gases

Coolers
Organic Solvents
Light oils
Heavy oils
Gases
Organic solvents
Water
Gases

Water
Organic solvents
Light oils
Heavy oils
Gases


Water
Water
Water
Water
Brine
Brine
Brine

800-1500
100-300
100-400
50-300
10-50


250-750
350-900
60-300
20-300
150-500
600-1200
15-250
OVERALL HEAT-TRANSFER COEFFICIENT
Shell and Tube Exchangers
Hot fluid Cold fluid U (W/m
2 0
C)
Heaters
Steam
Steam
Steam
Steam
Steam
Dowtherm
Dowtherm
Flue gases
Flue

Condensers
Aqueous vapours
Organic vapours
Organics (some non-condensibles)
Vacuum condensers

Vaporisers
Steam
Steam
Steam

Water
Organic solvents
Light oils
Heavy oils
Gases
Heavy oils
Gases
Steam
Hydrocarbon vapours


Water
Water
Water
Water


Aqueous solutions
Light organics
Heavy organics

1500-400
500-1000
300-900
60-450
30-300
50-300
20-200
30-100
30-100


1000-1500
700-1000
500-700
200-500


1000-1500
900-1200
600-900
(Cont.)
OVERALL HEAT-TRANSFER COEFFICIENT
Aircooled exchangers
Process fluid U (W/m
2 0
C)
Water
Light organics
Heavy organics
Gases, 5-10 bar
10-30 bar
Condensing hydrocarbons
300-450
300-700
50-150
50-100
100-300
300-600
Immersed coils
Coil Pool
Natural circulation
Steam
Steam
Steam
Aqueous solutions
Light oils

Agitated
Steam
Steam
Steam
Aqueous solutions
Light oils

Dilute aqueous solutions
Light oils
Heavy oils
Water
Water


Dilute aqueous solutions
Light oils
Heavy oils
Water
Water

500-1000
200-300
70-150
200-500
100-150


800-1500
300-500
200-400
400-700
200-300
(Cont.)
OVERALL HEAT-TRANSFER COEFFICIENT
Jacketed vessels
Jacket Vessel
Steam
Steam
Water
Water
500-700
250-500
200-500
200-300
(Cont.)
FOULING FACTORS (DIRT FACTOR)
The deposit material will normally have a relatively low thermal conductivity
and will reduce the overall coefficient.
The effect of fouling is allowed for in design by including the inside and
outside fouling coefficients in equation (2).
Fouling factors are often wrongly used as factors of safety in exchanger
design.
Typical values for the fouling coefficients of common process and service
fluids are given in the next Table. These values are for shell and tube
exchanger with plain (not finned) tubes.
More extensive data on fouling factors are given in the TEMA standards
(1978) and by Ludwig (1965).
FOULING FACTORS (DIRT FACTOR)
Fluid Coefficient (W/m
2

0
C)
River water
Sea water
Cooling water (towers)
Towns water (soft)
Towns water (hard)
Steam condensate
Steam (oil free)
Steam (oil traces)
Refrigerated brine
Air and industrial gases
Flue gases
Organic vapours
Organic liquids
Light hydrocarbons
Heavy hydrocarbons
Boiling organics
Condensing organics
Heat transfer fluids
Aqueous salt solutions
3000-12000
1000-3000
3000-6000
3000-5000
1000-2000
1500-5000
4000-10000
2000-5000
3000-5000
5000-10000
2000-5000
5000
5000
5000
2000
2500
5000
5000
3000-5000
STANDARD DIMENSIONS FOR STEEL TUBES
Outside diameter (mm) Wall thickness (mm)
16
20
25
30
38
50
1.2
-
-
-
-
-
1.6
1.6
1.6
1.6
-
-
2.0
2.0
2.0
2.0
2.0
2.0
-
2.6
2.6
2.6
2.6
2.6
-
-
3.2
3.2
3.2
3.2
TUBESHEET LAYOUT (TUBE COUNT)
The bundle diameter will depend not only on the number of tubes but also
on the number of tube passes, as spaces must be left in the pattern of
tubes on the tube sheet to accommodate the pass partition plates.
1
1
/ 1
1
1
) / (
, ) / (
n
t o b
n
o b t
K N d D
d D K N
=
= (3a)
(3b)
where N
t
= Number of tubes,
D
b
= bundle diameter, mm,
d
o
= tube outside diameter, mm.
Triangular pitch, p
t
= 1.25 d
o
No. passes 1 2 4 6 8
K
1
n1

0.319
0.142
0.249
2.207
0.175
2.285
0.0743
2.499
0.0365
2.675
Square pitch, p
t
= 1.25 d
o

No. passes 1 2 4 6 8
K
1
n1

0.215
2.207
0.156
2.291
0.158
2.263
0.0402
2.617
0.0331
2.643
( )
( )
) (
ln
) (
1 2
2 1
1 2 2 1
t T
t T
t T t T
T
lm

= A (4)
where AT
lm
= log mean temperature difference,
T
1
= inlet shell-side fluid temperature,
T
2
= outlet shell-side fluid temperature,
t
1
= inlet tube-side fluid temperature,
t
2
= outlet tube-side fluid temperature.
For counter-current flow, the logarithmic temperature is given by:
Mean Temperature
Difference (Temperature
Driving Force)
The usual practice in the design of shell and tube exchangers is
to estimate the true temperature difference from the logarithmic
mean temperature by applying a correction factor to allow for the
departure from true counter-current flow:
lm t m
T F T A = A
where AT
m
= true temperature difference, the mean temperature difference
F
t
= the temperature correction factor.
for use in the design equation (1)
(5)
( ) ( ) ( ) | |
( )
( )
( ) (
(

+ + +
+ +

+
=
] 1 1 [ 2
] 1 1 [ 2
ln 1
1 / 1 ln 1
2
2
2
R R S
R R S
R
RS S R
F
t (6)
) /( ) (
) /( ) (
1 1 1 2
1 2 2 1
t T t t S
t t T T R
=
=
Conductivity of Metal
Metal Temperature (
0
C) Kw(W/m
2 0
C)
Alumunium

Brass
(70Cu, 30Zn)

Copper

Nickel

Cupro-nickel (10 per cent Ni)
Monel
Stainless steel (18/8)
Steel


Titanium
0
100
0
100
400
0
100
0
212
0-100
0-100
0-100
0
100
600
0-100
202
206
97
104
116
388
378
62
59
45
30
16
45
45
36
16
Tubeside heattransfer coefficient
(Single Phase)
c b a
w C Nu ) / ( Pr Re =
Turbulent flow
where Nu

= Nusselt number =
Re = Reynolds number =
Pr = Prandtl number =
and h
i
= inside coefficient, W/m
2

0
C
d
e
= equivalent (or hydraulic) diameter, m
The index for Reynolds number is generally taken as 0.8. The Prandtl number
can range from 0.3 for cooling to 0.4 for heating. The index for viscosity
factor is normally taken as 0.14 for flow in tubes, from the work of Sieder and
tate (1936).
where C

= 0.021 for gases,
0.023 for non-viscous liquids,
0.027 for viscous liquids.
f
e i
k
d h

e t e t
d G d u
=
f
p
k
C
(7)
Tubeside heattransfer coefficient
(Single Phase)
Laminar flow
where L is the length of the tube, metres
If the Nusselt number given by equation (8) is less than 3.5, it should be
taken as 3.5. In laminar flow the length of the tube can have a marked effect
on the heat-transfer rate for length to diameter ratios less than 500.
14 . 0 33 . 0 33 . 0
) / ( ) / ( Pr) (Re 86 . 1
w e
L d Nu =
(8)
Heattransfer factor, j
h
The use of the j
h
factor enables data for laminar and turbulent flow to be
represented on the same graph.
(9)
( )
14 . 0
67 . 0
/ Pr

=
w h
St j
Tubeside heattransfer coefficient
(Single Phase)
Tube-side heat-transfer factor
Tubeside heattransfer coefficient
(Single Phase)
The equation below has been adapted from data given by Eagle and Ferguson
(1930):
Coefficients for water
2 . 0 8 . 0
/ ) 02 . 0 35 . 1 ( 4200
i t i
d u t h + =
where h
i
= inside coefficient, for water, W/m
2

0
C
t = water temperature,
0
C
u
t
= water velocity, m/s
d
i
= tube inside diameter, mm
Tubeside pressure drop (Single Phase)
The basic equation for isothermal flow in pipes (constant temperature) is:
where j
f
is the dimensionless friction factor and L is the effective pipe length.

The flow in a heat exchanger will clearly not be isothermal, normally only the
change in viscosity is considered:


m = 0.25 for laminar flow, Re < 2100,
= 0.14 for turbulent flow, Re < 2100.
Values of j
f
for heat exchanger tubes can be obtained from the next graphic.

2
) / ' ( 8
2
t
i f
u
d L j P

= A
( )
m
w
t
i f
u
d L j P

= A /
2
) ' ( 8
2
Tubeside pressure drop (Single Phase)
Tube-side friction factor
Correlations were based on the total stream flow, and empirical methods
were used to account for the performance of real exchangers compared
with that for cross flow over ideal tube banks.
This typical methods is known as bulkflow method.
DESIGN METHODS
The complex pattern on the shell-side, and the great number
of variables involved, make it difficult to predict the shell-
side coefficient and pressure drop with complete assurance.
Kern (1950) and Donohue (1955)
He developed a semi-analytical method based on work done in the
cooperative research program on Shell and Tube Exchanger.
His method accounts for the major bypass and leakage streams and is
suitable for a manual calculation.
Bell (1960, 1963)
The heattransfer coefficient and pressure drop are estimated
from correlations for flow over ideal tube-banks, and the effects
of leakage, bypassing and flow in the window zone are allowed
for by applying correction factors.
Heattransfer Coefficients
L b w n oc s
F F F F h h =
The shell-side heat transfer coefficient is given by:
where h
oc
= heat transfer coefficient calculated for cross-flow over an ideal tube
bank, no leakage or bypassing,
Fn = correction factor to allow for the effect of the number of vertical tube rows,
F
w
= window effect correction factor
F
b
= bypass stream correction factor
F
L
= leakage correction factor
The total correction will vary from 0-6 for a poorly designed exchanger with
large clearances to 0.9 for a well-designed exchanger.
h
oc
, ideal cross-flow coefficient
( )
14 . 0
/ Pr Re
3
1
w h
f
o oc
j
k
d h
=
F
n
, tube row correction factor
N
cv
is number of constrictions crossed = number of tube rows between the baffle
tips.
1. Re > 2000, turbulent; take F
n
from fig. above
2. Re > 100 to 2000, transition region, take F
n
= 1.0
3. Re < 100; laminar region, F
n
(N

c
)
-0.18


Where N

c
is the number of rows crossed in series from end to end of shell, and
depends on the number of baffles.

Heat-transfer factor for cross-flow tube banks
Window Correction Factor (F
w
)
( ) ( )
(

=
3
1
/ 2 1 exp
cv s
s
b
b
N N
A
A
F o
F
b
, by pass correction factor
By-pass correction factor
o = 1.5 for laminar flow, Re < 100,
o= 1.35 for transitional & turbulent flow Re > 100
A
b
= clearance area between bundle & shell
A
s
= maximum area for cross-flow
N
s
= number of sealing strips encountered by the
bypass stream in the cross-flow zone
N
cv
= number of constrictions, tube rows,
encountered in the cross-flow section.
Coefficient for F
L
, heat transfer
Pressure Drop
The pressure drop in the cross-flow and window zones are determined
separately, and summed to give the total shell-side pressure drop.
Cross-flow zones
The pressure drop in the cross-flow zones between the baffle tips is calculated
from:
L b i c
F F P P ' ' A = A
AP
c
= pressure drop in a cross-flow zone
AP
i
= pressure drop calculated for an equivalent ideal tube bank
F
b
= by-pass correction factor
F
L
= leakage correction factor
Friction factor for cross-flow tube banks
By-pass factor for pressure drop F
b
Coefficient for F
L
, pressure drop
Windowzones pressure drop
Bell used a method proposed by Colburn. Corrected for leakage, the window
drop for turbulent flow is given by:
u
z
= the geometric mean velocity
u
z
=
u
w
= the velocity in the window zone
u
w
=
W
s
= shell-side fluid mass flow, kg/s
N
wv
= number of restrictions for cross-flow in window zone, approximately
equal to the number of rows
2
) 6 . 0 2 ( '
2
z
wv L w
u
N F P

+ = A
s w
u u

w
s
A
W
Shell and bundle
geometry
Clearance and flow areas in the shell-side
H
c
= baffle cut heigh = D
s
x B
c
,
B
c
= the baffle cut as a fraction,
H
b
= height from baffle chord to
the top of the tube bundle,
B
b
= bundle cut = H
b
/D
b
,
u
b
= angle subtended by baffle
chord, rads,
D
b
= bundle diameter

The number of tubes in window
zone N
w
is given by:
N
w
= N
t
x R
a


Shell and bundle
geometry
Baffle and Tube Geometry
The number of tubes in a cross-flow zone
N
c
is given by
N
c
= N
t
2N
w
and
R
w
= 2 N
w
/N
t


R
a
is obtained from fig. beside, for the
appropriate baffle cut B
c
|
|
.
|

\
|

|
|
.
|

\
|
=
4 4
2 2
o
w a
s
w
d
N R
D
A
t t
( )
w t
o t
tb
N N
d c
A =
2
t
Baffle Geometrical Factors
Effect of fouling on pressure drop
Bells method gives an estimate of the shell-side pressure drop for the
exchanger in the clean condition. Devore (1961) has estimated the affect of
fouling on pressure drop by calculating the pressure drop in an exchange in
the clean condition and with the clearance reduced by fouling, using Tinkers
method. The ratios given in the Table below, which are adapted from Devores
figures, can be used to make a rough estimate of the effect of fouling on
pressure drop.
Fouling Coefficient

(W/m
2 0
C)
Shell diameter / baffle spacing
1.0 2.0 5.0
Laminar flow
6000
2000
< 1000

Turbulent flow
6000
2000
< 1000

1.06
1.19
1.32


1.12
1.37
1.64

1.20
1.44
1.99


1.38
2.31
3.44

1.28
1.55
2.38


1.55
2.96
4.77